Emacs Lisp

This is edition 3.0 of the GNU Emacs Lisp Reference Manual,
corresponding to Emacs version 23.3.

The homepage for GNU Emacs is at http://www.gnu.org/software/emacs/.
For information on using Emacs, refer to the Emacs Manual.
To view this manual in other formats, click here.

Copyright © 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011 Free Software Foundation, Inc.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.3 or any later version published by the Free Software Foundation; with the Invariant Sections being “GNU General Public License,” with the Front-Cover texts being “A GNU Manual,” and with the Back-Cover Texts as in (a) below. A copy of the license is included in the section entitled “GNU Free Documentation License.”

(a) The FSF's Back-Cover Text is: “You have the freedom to copy and modify this GNU manual. Buying copies from the FSF supports it in developing GNU and promoting software freedom.”

Introduction	Introduction and conventions used.
Lisp Data Types	Data types of objects in Emacs Lisp.
Numbers	Numbers and arithmetic functions.
Strings and Characters	Strings, and functions that work on them.
Lists	Lists, cons cells, and related functions.
Sequences Arrays Vectors	Lists, strings and vectors are called sequences. Certain functions act on any kind of sequence. The description of vectors is here as well.
Hash Tables	Very fast lookup-tables.
Symbols	Symbols represent names, uniquely.
Evaluation	How Lisp expressions are evaluated.
Control Structures	Conditionals, loops, nonlocal exits.
Variables	Using symbols in programs to stand for values.
Functions	A function is a Lisp program that can be invoked from other functions.
Macros	Macros are a way to extend the Lisp language.
Customization	Writing customization declarations.
Loading	Reading files of Lisp code into Lisp.
Byte Compilation	Compilation makes programs run faster.
Advising Functions	Adding to the definition of a function.
Debugging	Tools and tips for debugging Lisp programs.
Read and Print	Converting Lisp objects to text and back.
Minibuffers	Using the minibuffer to read input.
Command Loop	How the editor command loop works, and how you can call its subroutines.
Keymaps	Defining the bindings from keys to commands.
Modes	Defining major and minor modes.
Documentation	Writing and using documentation strings.
Files	Accessing files.
Backups and Auto-Saving	Controlling how backups and auto-save files are made.
Buffers	Creating and using buffer objects.
Windows	Manipulating windows and displaying buffers.
Frames	Making multiple system-level windows.
Positions	Buffer positions and motion functions.
Markers	Markers represent positions and update automatically when the text is changed.
Text	Examining and changing text in buffers.
Non-ASCII Characters	Non-ASCII text in buffers and strings.
Searching and Matching	Searching buffers for strings or regexps.
Syntax Tables	The syntax table controls word and list parsing.
Abbrevs	How Abbrev mode works, and its data structures.
Processes	Running and communicating with subprocesses.
Display	Features for controlling the screen display.
System Interface	Getting the user id, system type, environment variables, and other such things.
Appendices
Antinews	Info for users downgrading to Emacs 22.
GNU Free Documentation License	The license for this documentation.
GPL	Conditions for copying and changing GNU Emacs.
Tips	Advice and coding conventions for Emacs Lisp.
GNU Emacs Internals	Building and dumping Emacs; internal data structures.
Standard Errors	List of all error symbols.
Standard Buffer-Local Variables	List of variables buffer-local in all buffers.
Standard Keymaps	List of standard keymaps.
Standard Hooks	List of standard hook variables.
Index	Index including concepts, functions, variables, and other terms.

Detailed Node Listing

Here are other nodes that are inferiors of those already listed, mentioned here so you can get to them in one step:

Introduction
Caveats	Flaws and a request for help.
Lisp History	Emacs Lisp is descended from Maclisp.
Conventions	How the manual is formatted.
Version Info	Which Emacs version is running?
Acknowledgements	The authors, editors, and sponsors of this manual.
Conventions
Some Terms	Explanation of terms we use in this manual.
nil and t	How the symbols nil and t are used.
Evaluation Notation	The format we use for examples of evaluation.
Printing Notation	The format we use when examples print text.
Error Messages	The format we use for examples of errors.
Buffer Text Notation	The format we use for buffer contents in examples.
Format of Descriptions	Notation for describing functions, variables, etc.
Format of Descriptions
A Sample Function Description	A description of an imaginary function, foo.
A Sample Variable Description	A description of an imaginary variable, electric-future-map.
Lisp Data Types
Printed Representation	How Lisp objects are represented as text.
Comments	Comments and their formatting conventions.
Programming Types	Types found in all Lisp systems.
Editing Types	Types specific to Emacs.
Circular Objects	Read syntax for circular structure.
Type Predicates	Tests related to types.
Equality Predicates	Tests of equality between any two objects.
Programming Types
Integer Type	Numbers without fractional parts.
Floating Point Type	Numbers with fractional parts and with a large range.
Character Type	The representation of letters, numbers and control characters.
Symbol Type	A multi-use object that refers to a function, variable, or property list, and has a unique identity.
Sequence Type	Both lists and arrays are classified as sequences.
Cons Cell Type	Cons cells, and lists (which are made from cons cells).
Array Type	Arrays include strings and vectors.
String Type	An (efficient) array of characters.
Vector Type	One-dimensional arrays.
Char-Table Type	One-dimensional sparse arrays indexed by characters.
Bool-Vector Type	One-dimensional arrays of t or nil.
Hash Table Type	Super-fast lookup tables.
Function Type	A piece of executable code you can call from elsewhere.
Macro Type	A method of expanding an expression into another expression, more fundamental but less pretty.
Primitive Function Type	A function written in C, callable from Lisp.
Byte-Code Type	A function written in Lisp, then compiled.
Autoload Type	A type used for automatically loading seldom-used functions.
Character Type
Basic Char Syntax	Syntax for regular characters.
General Escape Syntax	How to specify characters by their codes.
Ctl-Char Syntax	Syntax for control characters.
Meta-Char Syntax	Syntax for meta-characters.
Other Char Bits	Syntax for hyper-, super-, and alt-characters.
Cons Cell and List Types
Box Diagrams	Drawing pictures of lists.
Dotted Pair Notation	A general syntax for cons cells.
Association List Type	A specially constructed list.
String Type
Syntax for Strings	How to specify Lisp strings.
Non-ASCII in Strings	International characters in strings.
Nonprinting Characters	Literal unprintable characters in strings.
Text Props and Strings	Strings with text properties.
Editing Types
Buffer Type	The basic object of editing.
Marker Type	A position in a buffer.
Window Type	Buffers are displayed in windows.
Frame Type	Windows subdivide frames.
Terminal Type	A terminal device displays frames.
Window Configuration Type	Recording the way a frame is subdivided.
Frame Configuration Type	Recording the status of all frames.
Process Type	A subprocess of Emacs running on the underlying OS.
Stream Type	Receive or send characters.
Keymap Type	What function a keystroke invokes.
Overlay Type	How an overlay is represented.
Font Type	Fonts for displaying text.
Numbers
Integer Basics	Representation and range of integers.
Float Basics	Representation and range of floating point.
Predicates on Numbers	Testing for numbers.
Comparison of Numbers	Equality and inequality predicates.
Numeric Conversions	Converting float to integer and vice versa.
Arithmetic Operations	How to add, subtract, multiply and divide.
Rounding Operations	Explicitly rounding floating point numbers.
Bitwise Operations	Logical and, or, not, shifting.
Math Functions	Trig, exponential and logarithmic functions.
Random Numbers	Obtaining random integers, predictable or not.
Strings and Characters
String Basics	Basic properties of strings and characters.
Predicates for Strings	Testing whether an object is a string or char.
Creating Strings	Functions to allocate new strings.
Modifying Strings	Altering the contents of an existing string.
Text Comparison	Comparing characters or strings.
String Conversion	Converting to and from characters and strings.
Formatting Strings	format: Emacs's analogue of printf.
Case Conversion	Case conversion functions.
Case Tables	Customizing case conversion.
Lists
Cons Cells	How lists are made out of cons cells.
List-related Predicates	Is this object a list? Comparing two lists.
List Elements	Extracting the pieces of a list.
Building Lists	Creating list structure.
List Variables	Modifying lists stored in variables.
Modifying Lists	Storing new pieces into an existing list.
Sets And Lists	A list can represent a finite mathematical set.
Association Lists	A list can represent a finite relation or mapping.
Rings	Managing a fixed-size ring of objects.
Modifying Existing List Structure
Setcar	Replacing an element in a list.
Setcdr	Replacing part of the list backbone. This can be used to remove or add elements.
Rearrangement	Reordering the elements in a list; combining lists.
Sequences, Arrays, and Vectors
Sequence Functions	Functions that accept any kind of sequence.
Arrays	Characteristics of arrays in Emacs Lisp.
Array Functions	Functions specifically for arrays.
Vectors	Special characteristics of Emacs Lisp vectors.
Vector Functions	Functions specifically for vectors.
Char-Tables	How to work with char-tables.
Bool-Vectors	How to work with bool-vectors.
Hash Tables
Creating Hash	Functions to create hash tables.
Hash Access	Reading and writing the hash table contents.
Defining Hash	Defining new comparison methods.
Other Hash	Miscellaneous.
Symbols
Symbol Components	Symbols have names, values, function definitions and property lists.
Definitions	A definition says how a symbol will be used.
Creating Symbols	How symbols are kept unique.
Property Lists	Each symbol has a property list for recording miscellaneous information.
Property Lists
Plists and Alists	Comparison of the advantages of property lists and association lists.
Symbol Plists	Functions to access symbols' property lists.
Other Plists	Accessing property lists stored elsewhere.
Evaluation
Intro Eval	Evaluation in the scheme of things.
Forms	How various sorts of objects are evaluated.
Quoting	Avoiding evaluation (to put constants in the program).
Eval	How to invoke the Lisp interpreter explicitly.
Kinds of Forms
Self-Evaluating Forms	Forms that evaluate to themselves.
Symbol Forms	Symbols evaluate as variables.
Classifying Lists	How to distinguish various sorts of list forms.
Function Indirection	When a symbol appears as the car of a list, we find the real function via the symbol.
Function Forms	Forms that call functions.
Macro Forms	Forms that call macros.
Special Forms	"Special forms" are idiosyncratic primitives, most of them extremely important.
Autoloading	Functions set up to load files containing their real definitions.
Control Structures
Sequencing	Evaluation in textual order.
Conditionals	if, cond, when, unless.
Combining Conditions	and, or, not.
Iteration	while loops.
Nonlocal Exits	Jumping out of a sequence.
Nonlocal Exits
Catch and Throw	Nonlocal exits for the program's own purposes.
Examples of Catch	Showing how such nonlocal exits can be written.
Errors	How errors are signaled and handled.
Cleanups	Arranging to run a cleanup form if an error happens.
Errors
Signaling Errors	How to report an error.
Processing of Errors	What Emacs does when you report an error.
Handling Errors	How you can trap errors and continue execution.
Error Symbols	How errors are classified for trapping them.
Variables
Global Variables	Variable values that exist permanently, everywhere.
Constant Variables	Certain "variables" have values that never change.
Local Variables	Variable values that exist only temporarily.
Void Variables	Symbols that lack values.
Defining Variables	A definition says a symbol is used as a variable.
Tips for Defining	Things you should think about when you define a variable.
Accessing Variables	Examining values of variables whose names are known only at run time.
Setting Variables	Storing new values in variables.
Variable Scoping	How Lisp chooses among local and global values.
Buffer-Local Variables	Variable values in effect only in one buffer.
File Local Variables	Handling local variable lists in files.
Directory Local Variables	Local variables common to all files in a directory.
Frame-Local Variables	Frame-local bindings for variables.
Variable Aliases	Variables that are aliases for other variables.
Variables with Restricted Values	Non-constant variables whose value can not be an arbitrary Lisp object.
Scoping Rules for Variable Bindings
Scope	Scope means where in the program a value is visible. Comparison with other languages.
Extent	Extent means how long in time a value exists.
Impl of Scope	Two ways to implement dynamic scoping.
Using Scoping	How to use dynamic scoping carefully and avoid problems.
Buffer-Local Variables
Intro to Buffer-Local	Introduction and concepts.
Creating Buffer-Local	Creating and destroying buffer-local bindings.
Default Value	The default value is seen in buffers that don't have their own buffer-local values.
Functions
What Is a Function	Lisp functions vs. primitives; terminology.
Lambda Expressions	How functions are expressed as Lisp objects.
Function Names	A symbol can serve as the name of a function.
Defining Functions	Lisp expressions for defining functions.
Calling Functions	How to use an existing function.
Mapping Functions	Applying a function to each element of a list, etc.
Anonymous Functions	Lambda expressions are functions with no names.
Function Cells	Accessing or setting the function definition of a symbol.
Obsolete Functions	Declaring functions obsolete.
Inline Functions	Defining functions that the compiler will open code.
Declaring Functions	Telling the compiler that a function is defined.
Function Safety	Determining whether a function is safe to call.
Related Topics	Cross-references to specific Lisp primitives that have a special bearing on how functions work.
Lambda Expressions
Lambda Components	The parts of a lambda expression.
Simple Lambda	A simple example.
Argument List	Details and special features of argument lists.
Function Documentation	How to put documentation in a function.
Macros
Simple Macro	A basic example.
Expansion	How, when and why macros are expanded.
Compiling Macros	How macros are expanded by the compiler.
Defining Macros	How to write a macro definition.
Backquote	Easier construction of list structure.
Problems with Macros	Don't evaluate the macro arguments too many times. Don't hide the user's variables.
Indenting Macros	Specifying how to indent macro calls.
Common Problems Using Macros
Wrong Time	Do the work in the expansion, not in the macro.
Argument Evaluation	The expansion should evaluate each macro arg once.
Surprising Local Vars	Local variable bindings in the expansion require special care.
Eval During Expansion	Don't evaluate them; put them in the expansion.
Repeated Expansion	Avoid depending on how many times expansion is done.
Writing Customization Definitions
Common Keywords	Common keyword arguments for all kinds of customization declarations.
Group Definitions	Writing customization group definitions.
Variable Definitions	Declaring user options.
Customization Types	Specifying the type of a user option.
Customization Types
Simple Types	Simple customization types: sexp, integer, number, string, file, directory, alist.
Composite Types	Build new types from other types or data.
Splicing into Lists	Splice elements into list with :inline.
Type Keywords	Keyword-argument pairs in a customization type.
Defining New Types	Give your type a name.
Loading
How Programs Do Loading	The load function and others.
Load Suffixes	Details about the suffixes that load tries.
Library Search	Finding a library to load.
Loading Non-ASCII	Non-ASCII characters in Emacs Lisp files.
Autoload	Setting up a function to autoload.
Repeated Loading	Precautions about loading a file twice.
Named Features	Loading a library if it isn't already loaded.
Where Defined	Finding which file defined a certain symbol.
Unloading	How to "unload" a library that was loaded.
Hooks for Loading	Providing code to be run when particular libraries are loaded.
Byte Compilation
Speed of Byte-Code	An example of speedup from byte compilation.
Compilation Functions	Byte compilation functions.
Docs and Compilation	Dynamic loading of documentation strings.
Dynamic Loading	Dynamic loading of individual functions.
Eval During Compile	Code to be evaluated when you compile.
Compiler Errors	Handling compiler error messages.
Byte-Code Objects	The data type used for byte-compiled functions.
Disassembly	Disassembling byte-code; how to read byte-code.
Advising Emacs Lisp Functions
Simple Advice	A simple example to explain the basics of advice.
Defining Advice	Detailed description of defadvice.
Around-Advice	Wrapping advice around a function's definition.
Computed Advice	...is to defadvice as fset is to defun.
Activation of Advice	Advice doesn't do anything until you activate it.
Enabling Advice	You can enable or disable each piece of advice.
Preactivation	Preactivation is a way of speeding up the loading of compiled advice.
Argument Access in Advice	How advice can access the function's arguments.
Advising Primitives	Accessing arguments when advising a primitive.
Combined Definition	How advice is implemented.
Debugging Lisp Programs
Debugger	How the Emacs Lisp debugger is implemented.
Edebug	A source-level Emacs Lisp debugger.
Syntax Errors	How to find syntax errors.
Test Coverage	Ensuring you have tested all branches in your code.
Compilation Errors	How to find errors that show up in byte compilation.
The Lisp Debugger
Error Debugging	Entering the debugger when an error happens.
Infinite Loops	Stopping and debugging a program that doesn't exit.
Function Debugging	Entering it when a certain function is called.
Explicit Debug	Entering it at a certain point in the program.
Using Debugger	What the debugger does; what you see while in it.
Debugger Commands	Commands used while in the debugger.
Invoking the Debugger	How to call the function debug.
Internals of Debugger	Subroutines of the debugger, and global variables.
Edebug
Using Edebug	Introduction to use of Edebug.
Instrumenting	You must instrument your code in order to debug it with Edebug.
Edebug Execution Modes	Execution modes, stopping more or less often.
Jumping	Commands to jump to a specified place.
Edebug Misc	Miscellaneous commands.
Breaks	Setting breakpoints to make the program stop.
Trapping Errors	Trapping errors with Edebug.
Edebug Views	Views inside and outside of Edebug.
Edebug Eval	Evaluating expressions within Edebug.
Eval List	Expressions whose values are displayed each time you enter Edebug.
Printing in Edebug	Customization of printing.
Trace Buffer	How to produce trace output in a buffer.
Coverage Testing	How to test evaluation coverage.
The Outside Context	Data that Edebug saves and restores.
Edebug and Macros	Specifying how to handle macro calls.
Edebug Options	Option variables for customizing Edebug.
Breaks
Breakpoints	Breakpoints at stop points.
Global Break Condition	Breaking on an event.
Source Breakpoints	Embedding breakpoints in source code.
The Outside Context
Checking Whether to Stop	When Edebug decides what to do.
Edebug Display Update	When Edebug updates the display.
Edebug Recursive Edit	When Edebug stops execution.
Edebug and Macros
Instrumenting Macro Calls	The basic problem.
Specification List	How to specify complex patterns of evaluation.
Backtracking	What Edebug does when matching fails.
Specification Examples	To help understand specifications.
Debugging Invalid Lisp Syntax
Excess Open	How to find a spurious open paren or missing close.
Excess Close	How to find a spurious close paren or missing open.
Reading and Printing Lisp Objects
Streams Intro	Overview of streams, reading and printing.
Input Streams	Various data types that can be used as input streams.
Input Functions	Functions to read Lisp objects from text.
Output Streams	Various data types that can be used as output streams.
Output Functions	Functions to print Lisp objects as text.
Output Variables	Variables that control what the printing functions do.
Minibuffers
Intro to Minibuffers	Basic information about minibuffers.
Text from Minibuffer	How to read a straight text string.
Object from Minibuffer	How to read a Lisp object or expression.
Minibuffer History	Recording previous minibuffer inputs so the user can reuse them.
Initial Input	Specifying initial contents for the minibuffer.
Completion	How to invoke and customize completion.
Yes-or-No Queries	Asking a question with a simple answer.
Multiple Queries	Asking a series of similar questions.
Reading a Password	Reading a password from the terminal.
Minibuffer Commands	Commands used as key bindings in minibuffers.
Minibuffer Contents	How such commands access the minibuffer text.
Minibuffer Windows	Operating on the special minibuffer windows.
Recursive Mini	Whether recursive entry to minibuffer is allowed.
Minibuffer Misc	Various customization hooks and variables.
Completion
Basic Completion	Low-level functions for completing strings.
Minibuffer Completion	Invoking the minibuffer with completion.
Completion Commands	Minibuffer commands that do completion.
High-Level Completion	Convenient special cases of completion (reading buffer name, file name, etc.).
Reading File Names	Using completion to read file names and shell commands.
Completion Styles	Specifying rules for performing completion.
Programmed Completion	Writing your own completion-function.
Command Loop
Command Overview	How the command loop reads commands.
Defining Commands	Specifying how a function should read arguments.
Interactive Call	Calling a command, so that it will read arguments.
Distinguish Interactive	Making a command distinguish interactive calls.
Command Loop Info	Variables set by the command loop for you to examine.
Adjusting Point	Adjustment of point after a command.
Input Events	What input looks like when you read it.
Reading Input	How to read input events from the keyboard or mouse.
Special Events	Events processed immediately and individually.
Waiting	Waiting for user input or elapsed time.
Quitting	How C-g works. How to catch or defer quitting.
Prefix Command Arguments	How the commands to set prefix args work.
Recursive Editing	Entering a recursive edit, and why you usually shouldn't.
Disabling Commands	How the command loop handles disabled commands.
Command History	How the command history is set up, and how accessed.
Keyboard Macros	How keyboard macros are implemented.
Defining Commands
Using Interactive	General rules for interactive.
Interactive Codes	The standard letter-codes for reading arguments in various ways.
Interactive Examples	Examples of how to read interactive arguments.
Input Events
Keyboard Events	Ordinary characters--keys with symbols on them.
Function Keys	Function keys--keys with names, not symbols.
Mouse Events	Overview of mouse events.
Click Events	Pushing and releasing a mouse button.
Drag Events	Moving the mouse before releasing the button.
Button-Down Events	A button was pushed and not yet released.
Repeat Events	Double and triple click (or drag, or down).
Motion Events	Just moving the mouse, not pushing a button.
Focus Events	Moving the mouse between frames.
Misc Events	Other events the system can generate.
Event Examples	Examples of the lists for mouse events.
Classifying Events	Finding the modifier keys in an event symbol. Event types.
Accessing Mouse	Functions to extract info from mouse events.
Accessing Scroll	Functions to get info from scroll bar events.
Strings of Events	Special considerations for putting keyboard character events in a string.
Reading Input
Key Sequence Input	How to read one key sequence.
Reading One Event	How to read just one event.
Event Mod	How Emacs modifies events as they are read.
Invoking the Input Method	How reading an event uses the input method.
Quoted Character Input	Asking the user to specify a character.
Event Input Misc	How to reread or throw away input events.
Keymaps
Key Sequences	Key sequences as Lisp objects.
Keymap Basics	Basic concepts of keymaps.
Format of Keymaps	What a keymap looks like as a Lisp object.
Creating Keymaps	Functions to create and copy keymaps.
Inheritance and Keymaps	How one keymap can inherit the bindings of another keymap.
Prefix Keys	Defining a key with a keymap as its definition.
Active Keymaps	How Emacs searches the active keymaps for a key binding.
Searching Keymaps	A pseudo-Lisp summary of searching active maps.
Controlling Active Maps	Each buffer has a local keymap to override the standard (global) bindings. A minor mode can also override them.
Key Lookup	Finding a key's binding in one keymap.
Functions for Key Lookup	How to request key lookup.
Changing Key Bindings	Redefining a key in a keymap.
Remapping Commands	A keymap can translate one command to another.
Translation Keymaps	Keymaps for translating sequences of events.
Key Binding Commands	Interactive interfaces for redefining keys.
Scanning Keymaps	Looking through all keymaps, for printing help.
Menu Keymaps	Defining a menu as a keymap.
Menu Keymaps
Defining Menus	How to make a keymap that defines a menu.
Mouse Menus	How users actuate the menu with the mouse.
Keyboard Menus	How users actuate the menu with the keyboard.
Menu Example	Making a simple menu.
Menu Bar	How to customize the menu bar.
Tool Bar	A tool bar is a row of images.
Modifying Menus	How to add new items to a menu.
Defining Menus
Simple Menu Items	A simple kind of menu key binding, limited in capabilities.
Extended Menu Items	More powerful menu item definitions let you specify keywords to enable various features.
Menu Separators	Drawing a horizontal line through a menu.
Alias Menu Items	Using command aliases in menu items.
Major and Minor Modes
Hooks	How to use hooks; how to write code that provides hooks.
Major Modes	Defining major modes.
Minor Modes	Defining minor modes.
Mode Line Format	Customizing the text that appears in the mode line.
Imenu	How a mode can provide a menu of definitions in the buffer.
Font Lock Mode	How modes can highlight text according to syntax.
Desktop Save Mode	How modes can have buffer state saved between Emacs sessions.
Hooks
Running Hooks	How to run a hook.
Setting Hooks	How to put functions on a hook, or remove them.
Major Modes
Major Mode Basics
Major Mode Conventions	Coding conventions for keymaps, etc.
Auto Major Mode	How Emacs chooses the major mode automatically.
Mode Help	Finding out how to use a mode.
Derived Modes	Defining a new major mode based on another major mode.
Generic Modes	Defining a simple major mode that supports comment syntax and Font Lock mode.
Mode Hooks	Hooks run at the end of major mode functions.
Example Major Modes	Text mode and Lisp modes.
Minor Modes
Minor Mode Conventions	Tips for writing a minor mode.
Keymaps and Minor Modes	How a minor mode can have its own keymap.
Defining Minor Modes	A convenient facility for defining minor modes.
Mode Line Format
Mode Line Basics	Basic ideas of mode line control.
Mode Line Data	The data structure that controls the mode line.
Mode Line Top	The top level variable, mode-line-format.
Mode Line Variables	Variables used in that data structure.
%-Constructs	Putting information into a mode line.
Properties in Mode	Using text properties in the mode line.
Header Lines	Like a mode line, but at the top.
Emulating Mode Line	Formatting text as the mode line would.
Font Lock Mode
Font Lock Basics	Overview of customizing Font Lock.
Search-based Fontification	Fontification based on regexps.
Customizing Keywords	Customizing search-based fontification.
Other Font Lock Variables	Additional customization facilities.
Levels of Font Lock	Each mode can define alternative levels so that the user can select more or less.
Precalculated Fontification	How Lisp programs that produce the buffer contents can also specify how to fontify it.
Faces for Font Lock	Special faces specifically for Font Lock.
Syntactic Font Lock	Fontification based on syntax tables.
Setting Syntax Properties	Defining character syntax based on context using the Font Lock mechanism.
Multiline Font Lock	How to coerce Font Lock into properly highlighting multiline constructs.
Multiline Font Lock Constructs
Font Lock Multiline	Marking multiline chunks with a text property.
Region to Fontify	Controlling which region gets refontified after a buffer change.
Documentation
Documentation Basics	Good style for doc strings. Where to put them. How Emacs stores them.
Accessing Documentation	How Lisp programs can access doc strings.
Keys in Documentation	Substituting current key bindings.
Describing Characters	Making printable descriptions of non-printing characters and key sequences.
Help Functions	Subroutines used by Emacs help facilities.
Files
Visiting Files	Reading files into Emacs buffers for editing.
Saving Buffers	Writing changed buffers back into files.
Reading from Files	Reading files into buffers without visiting.
Writing to Files	Writing new files from parts of buffers.
File Locks	Locking and unlocking files, to prevent simultaneous editing by two people.
Information about Files	Testing existence, accessibility, size of files.
Changing Files	Renaming files, changing protection, etc.
File Names	Decomposing and expanding file names.
Contents of Directories	Getting a list of the files in a directory.
Create/Delete Dirs	Creating and Deleting Directories.
Magic File Names	Defining "magic" special handling for certain file names.
Format Conversion	Conversion to and from various file formats.
Visiting Files
Visiting Functions	The usual interface functions for visiting.
Subroutines of Visiting	Lower-level subroutines that they use.
Information about Files
Testing Accessibility	Is a given file readable? Writable?
Kinds of Files	Is it a directory? A symbolic link?
Truenames	Eliminating symbolic links from a file name.
File Attributes	How large is it? Any other names? Etc.
Locating Files	How to find a file in standard places.
File Names
File Name Components	The directory part of a file name, and the rest.
Relative File Names	Some file names are relative to a current directory.
Directory Names	A directory's name as a directory is different from its name as a file.
File Name Expansion	Converting relative file names to absolute ones.
Unique File Names	Generating names for temporary files.
File Name Completion	Finding the completions for a given file name.
Standard File Names	If your package uses a fixed file name, how to handle various operating systems simply.
File Format Conversion
Format Conversion Overview	insert-file-contents and write-region.
Format Conversion Round-Trip	Using format-alist.
Format Conversion Piecemeal	Specifying non-paired conversion.
Backups and Auto-Saving
Backup Files	How backup files are made; how their names are chosen.
Auto-Saving	How auto-save files are made; how their names are chosen.
Reverting	revert-buffer, and how to customize what it does.
Backup Files
Making Backups	How Emacs makes backup files, and when.
Rename or Copy	Two alternatives: renaming the old file or copying it.
Numbered Backups	Keeping multiple backups for each source file.
Backup Names	How backup file names are computed; customization.
Buffers
Buffer Basics	What is a buffer?
Current Buffer	Designating a buffer as current so that primitives will access its contents.
Buffer Names	Accessing and changing buffer names.
Buffer File Name	The buffer file name indicates which file is visited.
Buffer Modification	A buffer is modified if it needs to be saved.
Modification Time	Determining whether the visited file was changed ``behind Emacs's back''.
Read Only Buffers	Modifying text is not allowed in a read-only buffer.
The Buffer List	How to look at all the existing buffers.
Creating Buffers	Functions that create buffers.
Killing Buffers	Buffers exist until explicitly killed.
Indirect Buffers	An indirect buffer shares text with some other buffer.
Swapping Text	Swapping text between two buffers.
Buffer Gap	The gap in the buffer.
Windows
Basic Windows	Basic information on using windows.
Splitting Windows	Splitting one window into two windows.
Deleting Windows	Deleting a window gives its space to other windows.
Selecting Windows	The selected window is the one that you edit in.
Cyclic Window Ordering	Moving around the existing windows.
Buffers and Windows	Each window displays the contents of a buffer.
Displaying Buffers	Higher-level functions for displaying a buffer and choosing a window for it.
Choosing Window	How to choose a window for displaying a buffer.
Dedicated Windows	How to avoid displaying another buffer in a specific window.
Window Point	Each window has its own location of point.
Window Start and End	Buffer positions indicating which text is on-screen in a window.
Textual Scrolling	Moving text up and down through the window.
Vertical Scrolling	Moving the contents up and down on the window.
Horizontal Scrolling	Moving the contents sideways on the window.
Size of Window	Accessing the size of a window.
Resizing Windows	Changing the size of a window.
Coordinates and Windows	Converting coordinates to windows.
Window Tree	The layout and sizes of all windows in a frame.
Window Configurations	Saving and restoring the state of the screen.
Window Parameters	Associating additional information with windows.
Window Hooks	Hooks for scrolling, window size changes, redisplay going past a certain point, or window configuration changes.
Frames
Creating Frames	Creating additional frames.
Multiple Terminals	Displaying on several different devices.
Frame Parameters	Controlling frame size, position, font, etc.
Terminal Parameters	Parameters common for all frames on terminal.
Frame Titles	Automatic updating of frame titles.
Deleting Frames	Frames last until explicitly deleted.
Finding All Frames	How to examine all existing frames.
Frames and Windows	A frame contains windows; display of text always works through windows.
Minibuffers and Frames	How a frame finds the minibuffer to use.
Input Focus	Specifying the selected frame.
Visibility of Frames	Frames may be visible or invisible, or icons.
Raising and Lowering	Raising a frame makes it hide other windows; lowering it makes the others hide it.
Frame Configurations	Saving the state of all frames.
Mouse Tracking	Getting events that say when the mouse moves.
Mouse Position	Asking where the mouse is, or moving it.
Pop-Up Menus	Displaying a menu for the user to select from.
Dialog Boxes	Displaying a box to ask yes or no.
Pointer Shape	Specifying the shape of the mouse pointer.
Window System Selections	Transferring text to and from other X clients.
Drag and Drop	Internals of Drag-and-Drop implementation.
Color Names	Getting the definitions of color names.
Text Terminal Colors	Defining colors for text-only terminals.
Resources	Getting resource values from the server.
Display Feature Testing	Determining the features of a terminal.
Frame Parameters
Parameter Access	How to change a frame's parameters.
Initial Parameters	Specifying frame parameters when you make a frame.
Window Frame Parameters	List of frame parameters for window systems.
Size and Position	Changing the size and position of a frame.
Geometry	Parsing geometry specifications.
Window Frame Parameters
Basic Parameters	Parameters that are fundamental.
Position Parameters	The position of the frame on the screen.
Size Parameters	Frame's size.
Layout Parameters	Size of parts of the frame, and enabling or disabling some parts.
Buffer Parameters	Which buffers have been or should be shown.
Management Parameters	Communicating with the window manager.
Cursor Parameters	Controlling the cursor appearance.
Font and Color Parameters	Fonts and colors for the frame text.
Positions
Point	The special position where editing takes place.
Motion	Changing point.
Excursions	Temporary motion and buffer changes.
Narrowing	Restricting editing to a portion of the buffer.
Motion
Character Motion	Moving in terms of characters.
Word Motion	Moving in terms of words.
Buffer End Motion	Moving to the beginning or end of the buffer.
Text Lines	Moving in terms of lines of text.
Screen Lines	Moving in terms of lines as displayed.
List Motion	Moving by parsing lists and sexps.
Skipping Characters	Skipping characters belonging to a certain set.
Markers
Overview of Markers	The components of a marker, and how it relocates.
Predicates on Markers	Testing whether an object is a marker.
Creating Markers	Making empty markers or markers at certain places.
Information from Markers	Finding the marker's buffer or character position.
Marker Insertion Types	Two ways a marker can relocate when you insert where it points.
Moving Markers	Moving the marker to a new buffer or position.
The Mark	How "the mark" is implemented with a marker.
The Region	How to access "the region".
Text
Near Point	Examining text in the vicinity of point.
Buffer Contents	Examining text in a general fashion.
Comparing Text	Comparing substrings of buffers.
Insertion	Adding new text to a buffer.
Commands for Insertion	User-level commands to insert text.
Deletion	Removing text from a buffer.
User-Level Deletion	User-level commands to delete text.
The Kill Ring	Where removed text sometimes is saved for later use.
Undo	Undoing changes to the text of a buffer.
Maintaining Undo	How to enable and disable undo information. How to control how much information is kept.
Filling	Functions for explicit filling.
Margins	How to specify margins for filling commands.
Adaptive Fill	Adaptive Fill mode chooses a fill prefix from context.
Auto Filling	How auto-fill mode is implemented to break lines.
Sorting	Functions for sorting parts of the buffer.
Columns	Computing horizontal positions, and using them.
Indentation	Functions to insert or adjust indentation.
Case Changes	Case conversion of parts of the buffer.
Text Properties	Assigning Lisp property lists to text characters.
Substitution	Replacing a given character wherever it appears.
Transposition	Swapping two portions of a buffer.
Registers	How registers are implemented. Accessing the text or position stored in a register.
Base 64	Conversion to or from base 64 encoding.
MD5 Checksum	Compute the MD5 "message digest"/"checksum".
Atomic Changes	Installing several buffer changes "atomically".
Change Hooks	Supplying functions to be run when text is changed.
The Kill Ring
Kill Ring Concepts	What text looks like in the kill ring.
Kill Functions	Functions that kill text.
Yanking	How yanking is done.
Yank Commands	Commands that access the kill ring.
Low-Level Kill Ring	Functions and variables for kill ring access.
Internals of Kill Ring	Variables that hold kill ring data.
Indentation
Primitive Indent	Functions used to count and insert indentation.
Mode-Specific Indent	Customize indentation for different modes.
Region Indent	Indent all the lines in a region.
Relative Indent	Indent the current line based on previous lines.
Indent Tabs	Adjustable, typewriter-like tab stops.
Motion by Indent	Move to first non-blank character.
Text Properties
Examining Properties	Looking at the properties of one character.
Changing Properties	Setting the properties of a range of text.
Property Search	Searching for where a property changes value.
Special Properties	Particular properties with special meanings.
Format Properties	Properties for representing formatting of text.
Sticky Properties	How inserted text gets properties from neighboring text.
Lazy Properties	Computing text properties in a lazy fashion only when text is examined.
Clickable Text	Using text properties to make regions of text do something when you click on them.
Fields	The field property defines fields within the buffer.
Not Intervals	Why text properties do not use Lisp-visible text intervals.
Non-ASCII Characters
Text Representations	How Emacs represents text.
Converting Representations	Converting unibyte to multibyte and vice versa.
Selecting a Representation	Treating a byte sequence as unibyte or multi.
Character Codes	How unibyte and multibyte relate to codes of individual characters.
Character Properties	Character attributes that define their behavior and handling.
Character Sets	The space of possible character codes is divided into various character sets.
Scanning Charsets	Which character sets are used in a buffer?
Translation of Characters	Translation tables are used for conversion.
Coding Systems	Coding systems are conversions for saving files.
Input Methods	Input methods allow users to enter various non-ASCII characters without special keyboards.
Locales	Interacting with the POSIX locale.
Coding Systems
Coding System Basics	Basic concepts.
Encoding and I/O	How file I/O functions handle coding systems.
Lisp and Coding Systems	Functions to operate on coding system names.
User-Chosen Coding Systems	Asking the user to choose a coding system.
Default Coding Systems	Controlling the default choices.
Specifying Coding Systems	Requesting a particular coding system for a single file operation.
Explicit Encoding	Encoding or decoding text without doing I/O.
Terminal I/O Encoding	Use of encoding for terminal I/O.
MS-DOS File Types	How DOS "text" and "binary" files relate to coding systems.
Searching and Matching
String Search	Search for an exact match.
Searching and Case	Case-independent or case-significant searching.
Regular Expressions	Describing classes of strings.
Regexp Search	Searching for a match for a regexp.
POSIX Regexps	Searching POSIX-style for the longest match.
Match Data	Finding out which part of the text matched, after a string or regexp search.
Search and Replace	Commands that loop, searching and replacing.
Standard Regexps	Useful regexps for finding sentences, pages,...
Regular Expressions
Syntax of Regexps	Rules for writing regular expressions.
Regexp Example	Illustrates regular expression syntax.
Regexp Functions	Functions for operating on regular expressions.
Syntax of Regular Expressions
Regexp Special	Special characters in regular expressions.
Char Classes	Character classes used in regular expressions.
Regexp Backslash	Backslash-sequences in regular expressions.
The Match Data
Replacing Match	Replacing a substring that was matched.
Simple Match Data	Accessing single items of match data, such as where a particular subexpression started.
Entire Match Data	Accessing the entire match data at once, as a list.
Saving Match Data	Saving and restoring the match data.
Syntax Tables
Syntax Basics	Basic concepts of syntax tables.
Syntax Descriptors	How characters are classified.
Syntax Table Functions	How to create, examine and alter syntax tables.
Syntax Properties	Overriding syntax with text properties.
Motion and Syntax	Moving over characters with certain syntaxes.
Parsing Expressions	Parsing balanced expressions using the syntax table.
Standard Syntax Tables	Syntax tables used by various major modes.
Syntax Table Internals	How syntax table information is stored.
Categories	Another way of classifying character syntax.
Syntax Descriptors
Syntax Class Table	Table of syntax classes.
Syntax Flags	Additional flags each character can have.
Parsing Expressions
Motion via Parsing	Motion functions that work by parsing.
Position Parse	Determining the syntactic state of a position.
Parser State	How Emacs represents a syntactic state.
Low-Level Parsing	Parsing across a specified region.
Control Parsing	Parameters that affect parsing.
Abbrevs and Abbrev Expansion
Abbrev Mode	Setting up Emacs for abbreviation.
Abbrev Tables	Creating and working with abbrev tables.
Defining Abbrevs	Specifying abbreviations and their expansions.
Abbrev Files	Saving abbrevs in files.
Abbrev Expansion	Controlling expansion; expansion subroutines.
Standard Abbrev Tables	Abbrev tables used by various major modes.
Abbrev Properties	How to read and set abbrev properties. Which properties have which effect.
Abbrev Table Properties	How to read and set abbrev table properties. Which properties have which effect.
Processes
Subprocess Creation	Functions that start subprocesses.
Shell Arguments	Quoting an argument to pass it to a shell.
Synchronous Processes	Details of using synchronous subprocesses.
Asynchronous Processes	Starting up an asynchronous subprocess.
Deleting Processes	Eliminating an asynchronous subprocess.
Process Information	Accessing run-status and other attributes.
Input to Processes	Sending input to an asynchronous subprocess.
Signals to Processes	Stopping, continuing or interrupting an asynchronous subprocess.
Output from Processes	Collecting output from an asynchronous subprocess.
Sentinels	Sentinels run when process run-status changes.
Query Before Exit	Whether to query if exiting will kill a process.
System Processes	Accessing other processes running on your system.
Transaction Queues	Transaction-based communication with subprocesses.
Network	Opening network connections.
Network Servers	Network servers let Emacs accept net connections.
Datagrams	UDP network connections.
Low-Level Network	Lower-level but more general function to create connections and servers.
Misc Network	Additional relevant functions for network connections.
Serial Ports	Communicating with serial ports.
Byte Packing	Using bindat to pack and unpack binary data.
Receiving Output from Processes
Process Buffers	If no filter, output is put in a buffer.
Filter Functions	Filter functions accept output from the process.
Decoding Output	Filters can get unibyte or multibyte strings.
Accepting Output	How to wait until process output arrives.
Low-Level Network Access
Network Processes	Using make-network-process.
Network Options	Further control over network connections.
Network Feature Testing	Determining which network features work on the machine you are using.
Packing and Unpacking Byte Arrays
Bindat Spec	Describing data layout.
Bindat Functions	Doing the unpacking and packing.
Bindat Examples	Samples of what bindat.el can do for you!
Emacs Display
Refresh Screen	Clearing the screen and redrawing everything on it.
Forcing Redisplay	Forcing redisplay.
Truncation	Folding or wrapping long text lines.
The Echo Area	Displaying messages at the bottom of the screen.
Warnings	Displaying warning messages for the user.
Invisible Text	Hiding part of the buffer text.
Selective Display	Hiding part of the buffer text (the old way).
Temporary Displays	Displays that go away automatically.
Overlays	Use overlays to highlight parts of the buffer.
Width	How wide a character or string is on the screen.
Line Height	Controlling the height of lines.
Faces	A face defines a graphics style for text characters: font, colors, etc.
Fringes	Controlling window fringes.
Scroll Bars	Controlling vertical scroll bars.
Display Property	Enabling special display features.
Images	Displaying images in Emacs buffers.
Buttons	Adding clickable buttons to Emacs buffers.
Abstract Display	Emacs' Widget for Object Collections.
Blinking	How Emacs shows the matching open parenthesis.
Usual Display	The usual conventions for displaying nonprinting chars.
Display Tables	How to specify other conventions.
Beeping	Audible signal to the user.
Window Systems	Which window system is being used.
The Echo Area
Displaying Messages	Explicitly displaying text in the echo area.
Progress	Informing user about progress of a long operation.
Logging Messages	Echo area messages are logged for the user.
Echo Area Customization	Controlling the echo area.
Reporting Warnings
Warning Basics	Warnings concepts and functions to report them.
Warning Variables	Variables programs bind to customize their warnings.
Warning Options	Variables users set to control display of warnings.
Overlays
Managing Overlays	Creating and moving overlays.
Overlay Properties	How to read and set properties. What properties do to the screen display.
Finding Overlays	Searching for overlays.
Faces
Defining Faces	How to define a face with defface.
Face Attributes	What is in a face?
Attribute Functions	Functions to examine and set face attributes.
Displaying Faces	How Emacs combines the faces specified for a character.
Face Remapping	Remapping faces to alternative definitions.
Face Functions	How to define and examine faces.
Auto Faces	Hook for automatic face assignment.
Font Selection	Finding the best available font for a face.
Font Lookup	Looking up the names of available fonts and information about them.
Fontsets	A fontset is a collection of fonts that handle a range of character sets.
Low-Level Font	Lisp representation for character display fonts.
Fringes
Fringe Size/Pos	Specifying where to put the window fringes.
Fringe Indicators	Displaying indicator icons in the window fringes.
Fringe Cursors	Displaying cursors in the right fringe.
Fringe Bitmaps	Specifying bitmaps for fringe indicators.
Customizing Bitmaps	Specifying your own bitmaps to use in the fringes.
Overlay Arrow	Display of an arrow to indicate position.
The display Property
Replacing Specs	Display specs that replace the text.
Specified Space	Displaying one space with a specified width.
Pixel Specification	Specifying space width or height in pixels.
Other Display Specs	Displaying an image; adjusting the height, spacing, and other properties of text.
Display Margins	Displaying text or images to the side of the main text.
Images
Image Formats	Supported image formats.
Image Descriptors	How to specify an image for use in :display.
XBM Images	Special features for XBM format.
XPM Images	Special features for XPM format.
GIF Images	Special features for GIF format.
TIFF Images	Special features for TIFF format.
PostScript Images	Special features for PostScript format.
Other Image Types	Various other formats are supported.
Defining Images	Convenient ways to define an image for later use.
Showing Images	Convenient ways to display an image once it is defined.
Image Cache	Internal mechanisms of image display.
Buttons
Button Properties	Button properties with special meanings.
Button Types	Defining common properties for classes of buttons.
Making Buttons	Adding buttons to Emacs buffers.
Manipulating Buttons	Getting and setting properties of buttons.
Button Buffer Commands	Buffer-wide commands and bindings for buttons.
Abstract Display
Abstract Display Functions	Functions in the Ewoc package.
Abstract Display Example	Example of using Ewoc.
Display Tables
Display Table Format	What a display table consists of.
Active Display Table	How Emacs selects a display table to use.
Glyphs	How to define a glyph, and what glyphs mean.
Operating System Interface
Starting Up	Customizing Emacs startup processing.
Getting Out	How exiting works (permanent or temporary).
System Environment	Distinguish the name and kind of system.
User Identification	Finding the name and user id of the user.
Time of Day	Getting the current time.
Time Conversion	Converting a time from numeric form to calendrical data and vice versa.
Time Parsing	Converting a time from numeric form to text and vice versa.
Processor Run Time	Getting the run time used by Emacs.
Time Calculations	Adding, subtracting, comparing times, etc.
Timers	Setting a timer to call a function at a certain time.
Idle Timers	Setting a timer to call a function when Emacs has been idle for a certain length of time.
Terminal Input	Accessing and recording terminal input.
Terminal Output	Controlling and recording terminal output.
Sound Output	Playing sounds on the computer's speaker.
X11 Keysyms	Operating on key symbols for X Windows.
Batch Mode	Running Emacs without terminal interaction.
Session Management	Saving and restoring state with X Session Management.
Starting Up Emacs
Startup Summary	Sequence of actions Emacs performs at startup.
Init File	Details on reading the init file.
Terminal-Specific	How the terminal-specific Lisp file is read.
Command-Line Arguments	How command-line arguments are processed, and how you can customize them.
Getting Out of Emacs
Killing Emacs	Exiting Emacs irreversibly.
Suspending Emacs	Exiting Emacs reversibly.
Terminal Input
Input Modes	Options for how input is processed.
Recording Input	Saving histories of recent or all input events.
Tips and Conventions
Coding Conventions	Conventions for clean and robust programs.
Key Binding Conventions	Which keys should be bound by which programs.
Programming Tips	Making Emacs code fit smoothly in Emacs.
Compilation Tips	Making compiled code run fast.
Warning Tips	Turning off compiler warnings.
Documentation Tips	Writing readable documentation strings.
Comment Tips	Conventions for writing comments.
Library Headers	Standard headers for library packages.
GNU Emacs Internals
Building Emacs	How the dumped Emacs is made.
Pure Storage	A kludge to make preloaded Lisp functions sharable.
Garbage Collection	Reclaiming space for Lisp objects no longer used.
Memory Usage	Info about total size of Lisp objects made so far.
Writing Emacs Primitives	Writing C code for Emacs.
Object Internals	Data formats of buffers, windows, processes.
Object Internals
Buffer Internals	Components of a buffer structure.
Window Internals	Components of a window structure.
Process Internals	Components of a process structure.

1 Introduction

Most of the GNU Emacs text editor is written in the programming language called Emacs Lisp. You can write new code in Emacs Lisp and install it as an extension to the editor. However, Emacs Lisp is more than a mere “extension language”; it is a full computer programming language in its own right. You can use it as you would any other programming language.

Because Emacs Lisp is designed for use in an editor, it has special features for scanning and parsing text as well as features for handling files, buffers, displays, subprocesses, and so on. Emacs Lisp is closely integrated with the editing facilities; thus, editing commands are functions that can also conveniently be called from Lisp programs, and parameters for customization are ordinary Lisp variables.

This manual attempts to be a full description of Emacs Lisp. For a beginner's introduction to Emacs Lisp, see An Introduction to Emacs Lisp Programming, by Bob Chassell, also published by the Free Software Foundation. This manual presumes considerable familiarity with the use of Emacs for editing; see The GNU Emacs Manual for this basic information.

Generally speaking, the earlier chapters describe features of Emacs Lisp that have counterparts in many programming languages, and later chapters describe features that are peculiar to Emacs Lisp or relate specifically to editing.

This is edition 3.0 of the GNU Emacs Lisp Reference Manual, corresponding to Emacs version 23.3.

• Caveats: Flaws and a request for help.
• Lisp History: Emacs Lisp is descended from Maclisp.
• Conventions: How the manual is formatted.
• Version Info: Which Emacs version is running?
• Acknowledgements: The authors, editors, and sponsors of this manual.

1.1 Caveats

This manual has gone through numerous drafts. It is nearly complete but not flawless. There are a few topics that are not covered, either because we consider them secondary (such as most of the individual modes) or because they are yet to be written. Because we are not able to deal with them completely, we have left out several parts intentionally.

The manual should be fully correct in what it does cover, and it is therefore open to criticism on anything it says—from specific examples and descriptive text, to the ordering of chapters and sections. If something is confusing, or you find that you have to look at the sources or experiment to learn something not covered in the manual, then perhaps the manual should be fixed. Please let us know.

As you use this manual, we ask that you send corrections as soon as you find them. If you think of a simple, real life example for a function or group of functions, please make an effort to write it up and send it in. Please reference any comments to the node name and function or variable name, as appropriate. Also state the number of the edition you are criticizing.

Please send comments and corrections using M-x report-emacs-bug.

1.2 Lisp History

Lisp (LISt Processing language) was first developed in the late 1950s at the Massachusetts Institute of Technology for research in artificial intelligence. The great power of the Lisp language makes it ideal for other purposes as well, such as writing editing commands.

Dozens of Lisp implementations have been built over the years, each with its own idiosyncrasies. Many of them were inspired by Maclisp, which was written in the 1960s at MIT's Project MAC. Eventually the implementors of the descendants of Maclisp came together and developed a standard for Lisp systems, called Common Lisp. In the meantime, Gerry Sussman and Guy Steele at MIT developed a simplified but very powerful dialect of Lisp, called Scheme.

GNU Emacs Lisp is largely inspired by Maclisp, and a little by Common Lisp. If you know Common Lisp, you will notice many similarities. However, many features of Common Lisp have been omitted or simplified in order to reduce the memory requirements of GNU Emacs. Sometimes the simplifications are so drastic that a Common Lisp user might be very confused. We will occasionally point out how GNU Emacs Lisp differs from Common Lisp. If you don't know Common Lisp, don't worry about it; this manual is self-contained.

A certain amount of Common Lisp emulation is available via the cl library. see Overview.

Emacs Lisp is not at all influenced by Scheme; but the GNU project has an implementation of Scheme, called Guile. We use Guile in all new GNU software that calls for extensibility.

1.3 Conventions

This section explains the notational conventions that are used in this manual. You may want to skip this section and refer back to it later.

• Some Terms: Explanation of terms we use in this manual.
• nil and t: How the symbols nil and t are used.
• Evaluation Notation: The format we use for examples of evaluation.
• Printing Notation: The format we use when examples print text.
• Error Messages: The format we use for examples of errors.
• Buffer Text Notation: The format we use for buffer contents in examples.
• Format of Descriptions: Notation for describing functions, variables, etc.

1.3.1 Some Terms

Throughout this manual, the phrases “the Lisp reader” and “the Lisp printer” refer to those routines in Lisp that convert textual representations of Lisp objects into actual Lisp objects, and vice versa. See Printed Representation, for more details. You, the person reading this manual, are thought of as “the programmer” and are addressed as “you.” “The user” is the person who uses Lisp programs, including those you write.

Examples of Lisp code are formatted like this: (list 1 2 3). Names that represent metasyntactic variables, or arguments to a function being described, are formatted like this: first-number.

1.3.2 nil and t

In Lisp, the symbol nil has three separate meanings: it is a symbol with the name ‘nil’; it is the logical truth value false; and it is the empty list—the list of zero elements. When used as a variable, nil always has the value nil.

As far as the Lisp reader is concerned, ‘()’ and ‘nil’ are identical: they stand for the same object, the symbol nil. The different ways of writing the symbol are intended entirely for human readers. After the Lisp reader has read either ‘()’ or ‘nil’, there is no way to determine which representation was actually written by the programmer.

In this manual, we write () when we wish to emphasize that it means the empty list, and we write nil when we wish to emphasize that it means the truth value false. That is a good convention to use in Lisp programs also.

     (cons 'foo ())                ; Emphasize the empty list
     (setq foo-flag nil)           ; Emphasize the truth value false

In contexts where a truth value is expected, any non-nil value is considered to be true. However, t is the preferred way to represent the truth value true. When you need to choose a value which represents true, and there is no other basis for choosing, use t. The symbol t always has the value t.

In Emacs Lisp, nil and t are special symbols that always evaluate to themselves. This is so that you do not need to quote them to use them as constants in a program. An attempt to change their values results in a setting-constant error. See Constant Variables.

1.3.3 Evaluation Notation

A Lisp expression that you can evaluate is called a form. Evaluating a form always produces a result, which is a Lisp object. In the examples in this manual, this is indicated with ‘⇒’:

     (car '(1 2))
          ⇒ 1

You can read this as “(car '(1 2)) evaluates to 1.”

When a form is a macro call, it expands into a new form for Lisp to evaluate. We show the result of the expansion with ‘==>’. We may or may not show the result of the evaluation of the expanded form.

     (third '(a b c))
          ==> (car (cdr (cdr '(a b c))))
          ⇒ c

Sometimes to help describe one form we show another form that produces identical results. The exact equivalence of two forms is indicated with ‘==’.

     (make-sparse-keymap) == (list 'keymap)

1.3.4 Printing Notation

Many of the examples in this manual print text when they are evaluated. If you execute example code in a Lisp Interaction buffer (such as the buffer ‘*scratch*’), the printed text is inserted into the buffer. If you execute the example by other means (such as by evaluating the function eval-region), the printed text is displayed in the echo area.

Examples in this manual indicate printed text with ‘-|’, irrespective of where that text goes. The value returned by evaluating the form (here bar) follows on a separate line with ‘⇒’.

     (progn (prin1 'foo) (princ "\n") (prin1 'bar))
          -| foo
          -| bar
          ⇒ bar

1.3.5 Error Messages

Some examples signal errors. This normally displays an error message in the echo area. We show the error message on a line starting with ‘error-->’. Note that ‘error-->’ itself does not appear in the echo area.

     (+ 23 'x)
     error--> Wrong type argument: number-or-marker-p, x

1.3.6 Buffer Text Notation

Some examples describe modifications to the contents of a buffer, by showing the “before” and “after” versions of the text. These examples show the contents of the buffer in question between two lines of dashes containing the buffer name. In addition, ‘-!-’ indicates the location of point. (The symbol for point, of course, is not part of the text in the buffer; it indicates the place between two characters where point is currently located.)

     ---------- Buffer: foo ----------
     This is the -!-contents of foo.
     ---------- Buffer: foo ----------
     
     (insert "changed ")
          ⇒ nil
     ---------- Buffer: foo ----------
     This is the changed -!-contents of foo.
     ---------- Buffer: foo ----------

1.3.7 Format of Descriptions

Functions, variables, macros, commands, user options, and special forms are described in this manual in a uniform format. The first line of a description contains the name of the item followed by its arguments, if any. The category—function, variable, or whatever—appears at the beginning of the line. The description follows on succeeding lines, sometimes with examples.

• A Sample Function Description: A description of an imaginary function, foo.
• A Sample Variable Description: A description of an imaginary variable, electric-future-map.

1.3.7.1 A Sample Function Description

In a function description, the name of the function being described appears first. It is followed on the same line by a list of argument names. These names are also used in the body of the description, to stand for the values of the arguments.

The appearance of the keyword &optional in the argument list indicates that the subsequent arguments may be omitted (omitted arguments default to nil). Do not write &optional when you call the function.

The keyword &rest (which must be followed by a single argument name) indicates that any number of arguments can follow. The single argument name following &rest will receive, as its value, a list of all the remaining arguments passed to the function. Do not write &rest when you call the function.

Here is a description of an imaginary function foo:

Any argument whose name contains the name of a type (e.g., integer, integer1 or buffer) is expected to be of that type. A plural of a type (such as buffers) often means a list of objects of that type. Arguments named object may be of any type. (See Lisp Data Types, for a list of Emacs object types.) Arguments with other sorts of names (e.g., new-file) are discussed specifically in the description of the function. In some sections, features common to the arguments of several functions are described at the beginning.

See Lambda Expressions, for a more complete description of optional and rest arguments.

Command, macro, and special form descriptions have the same format, but the word `Function' is replaced by `Command', `Macro', or `Special Form', respectively. Commands are simply functions that may be called interactively; macros process their arguments differently from functions (the arguments are not evaluated), but are presented the same way.

Special form descriptions use a more complex notation to specify optional and repeated arguments because they can break the argument list down into separate arguments in more complicated ways. ‘[optional-arg]’ means that optional-arg is optional and ‘repeated-args...’ stands for zero or more arguments. Parentheses are used when several arguments are grouped into additional levels of list structure. Here is an example:

1.3.7.2 A Sample Variable Description

A variable is a name that can hold a value. Although nearly all variables can be set by the user, certain variables exist specifically so that users can change them; these are called user options. Ordinary variables and user options are described using a format like that for functions except that there are no arguments.

Here is a description of the imaginary electric-future-map variable.

User option descriptions have the same format, but `Variable' is replaced by `User Option'.

1.4 Version Information

These facilities provide information about which version of Emacs is in use.

The following two variables have existed since Emacs version 19.23:

1.5 Acknowledgements

This manual was written by Robert Krawitz, Bil Lewis, Dan LaLiberte, Richard M. Stallman and Chris Welty, the volunteers of the GNU manual group, in an effort extending over several years. Robert J. Chassell helped to review and edit the manual, with the support of the Defense Advanced Research Projects Agency, ARPA Order 6082, arranged by Warren A. Hunt, Jr. of Computational Logic, Inc. Additional sections were written by Miles Bader, Lars Brinkhoff, Chong Yidong, Kenichi Handa, Lute Kamstra, Juri Linkov, Glenn Morris, Thien-Thi Nguyen, Dan Nicolaescu, Martin Rudalics, Kim F. Storm, Luc Teirlinck, and Eli Zaretskii.

Corrections were supplied by Drew Adams, Juanma Barranquero, Karl Berry, Jim Blandy, Bard Bloom, Stephane Boucher, David Boyes, Alan Carroll, Richard Davis, Lawrence R. Dodd, Peter Doornbosch, David A. Duff, Chris Eich, Beverly Erlebacher, David Eckelkamp, Ralf Fassel, Eirik Fuller, Stephen Gildea, Bob Glickstein, Eric Hanchrow, Jesper Harder, George Hartzell, Nathan Hess, Masayuki Ida, Dan Jacobson, Jak Kirman, Bob Knighten, Frederick M. Korz, Joe Lammens, Glenn M. Lewis, K. Richard Magill, Brian Marick, Roland McGrath, Stefan Monnier, Skip Montanaro, John Gardiner Myers, Thomas A. Peterson, Francesco Potorti, Friedrich Pukelsheim, Arnold D. Robbins, Raul Rockwell, Jason Rumney, Per Starbäck, Shinichirou Sugou, Kimmo Suominen, Edward Tharp, Bill Trost, Rickard Westman, Jean White, Eduard Wiebe, Matthew Wilding, Carl Witty, Dale Worley, Rusty Wright, and David D. Zuhn.

2 Lisp Data Types

A Lisp object is a piece of data used and manipulated by Lisp programs. For our purposes, a type or data type is a set of possible objects.

Every object belongs to at least one type. Objects of the same type have similar structures and may usually be used in the same contexts. Types can overlap, and objects can belong to two or more types. Consequently, we can ask whether an object belongs to a particular type, but not for “the” type of an object.

A few fundamental object types are built into Emacs. These, from which all other types are constructed, are called primitive types. Each object belongs to one and only one primitive type. These types include integer, float, cons, symbol, string, vector, hash-table, subr, and byte-code function, plus several special types, such as buffer, that are related to editing. (See Editing Types.)

Each primitive type has a corresponding Lisp function that checks whether an object is a member of that type.

Lisp is unlike many other languages in that its objects are self-typing: the primitive type of each object is implicit in the object itself. For example, if an object is a vector, nothing can treat it as a number; Lisp knows it is a vector, not a number.

In most languages, the programmer must declare the data type of each variable, and the type is known by the compiler but not represented in the data. Such type declarations do not exist in Emacs Lisp. A Lisp variable can have any type of value, and it remembers whatever value you store in it, type and all. (Actually, a small number of Emacs Lisp variables can only take on values of a certain type. See Variables with Restricted Values.)

This chapter describes the purpose, printed representation, and read syntax of each of the standard types in GNU Emacs Lisp. Details on how to use these types can be found in later chapters.

• Printed Representation: How Lisp objects are represented as text.
• Comments: Comments and their formatting conventions.
• Programming Types: Types found in all Lisp systems.
• Editing Types: Types specific to Emacs.
• Circular Objects: Read syntax for circular structure.
• Type Predicates: Tests related to types.
• Equality Predicates: Tests of equality between any two objects.

2.1 Printed Representation and Read Syntax

The printed representation of an object is the format of the output generated by the Lisp printer (the function prin1) for that object. Every data type has a unique printed representation. The read syntax of an object is the format of the input accepted by the Lisp reader (the function read) for that object. This is not necessarily unique; many kinds of object have more than one syntax. See Read and Print.

In most cases, an object's printed representation is also a read syntax for the object. However, some types have no read syntax, since it does not make sense to enter objects of these types as constants in a Lisp program. These objects are printed in hash notation, which consists of the characters ‘#<’, a descriptive string (typically the type name followed by the name of the object), and a closing ‘>’. For example:

     (current-buffer)
          ⇒ #<buffer objects.texi>

Hash notation cannot be read at all, so the Lisp reader signals the error invalid-read-syntax whenever it encounters ‘#<’. In other languages, an expression is text; it has no other form. In Lisp, an expression is primarily a Lisp object and only secondarily the text that is the object's read syntax. Often there is no need to emphasize this distinction, but you must keep it in the back of your mind, or you will occasionally be very confused.

When you evaluate an expression interactively, the Lisp interpreter first reads the textual representation of it, producing a Lisp object, and then evaluates that object (see Evaluation). However, evaluation and reading are separate activities. Reading returns the Lisp object represented by the text that is read; the object may or may not be evaluated later. See Input Functions, for a description of read, the basic function for reading objects.

2.2 Comments

A comment is text that is written in a program only for the sake of humans that read the program, and that has no effect on the meaning of the program. In Lisp, a semicolon (‘;’) starts a comment if it is not within a string or character constant. The comment continues to the end of line. The Lisp reader discards comments; they do not become part of the Lisp objects which represent the program within the Lisp system.

The ‘#@count’ construct, which skips the next count characters, is useful for program-generated comments containing binary data. The Emacs Lisp byte compiler uses this in its output files (see Byte Compilation). It isn't meant for source files, however.

See Comment Tips, for conventions for formatting comments.

2.3 Programming Types

There are two general categories of types in Emacs Lisp: those having to do with Lisp programming, and those having to do with editing. The former exist in many Lisp implementations, in one form or another. The latter are unique to Emacs Lisp.

• Integer Type: Numbers without fractional parts.
• Floating Point Type: Numbers with fractional parts and with a large range.
• Character Type: The representation of letters, numbers and control characters.
• Symbol Type: A multi-use object that refers to a function, variable, or property list, and has a unique identity.
• Sequence Type: Both lists and arrays are classified as sequences.
• Cons Cell Type: Cons cells, and lists (which are made from cons cells).
• Array Type: Arrays include strings and vectors.
• String Type: An (efficient) array of characters.
• Vector Type: One-dimensional arrays.
• Char-Table Type: One-dimensional sparse arrays indexed by characters.
• Bool-Vector Type: One-dimensional arrays of t or nil.
• Hash Table Type: Super-fast lookup tables.
• Function Type: A piece of executable code you can call from elsewhere.
• Macro Type: A method of expanding an expression into another expression, more fundamental but less pretty.
• Primitive Function Type: A function written in C, callable from Lisp.
• Byte-Code Type: A function written in Lisp, then compiled.
• Autoload Type: A type used for automatically loading seldom-used functions.

2.3.1 Integer Type

The range of values for integers in Emacs Lisp is −536870912 to 536870911 (30 bits; i.e., -2**29 to 2**29 - 1) on most machines. (Some machines may provide a wider range.) It is important to note that the Emacs Lisp arithmetic functions do not check for overflow. Thus (1+ 536870911) is −536870912 on most machines.

The read syntax for integers is a sequence of (base ten) digits with an optional sign at the beginning and an optional period at the end. The printed representation produced by the Lisp interpreter never has a leading ‘+’ or a final ‘.’.

     -1               ; The integer -1.
     1                ; The integer 1.
     1.               ; Also the integer 1.
     +1               ; Also the integer 1.
     1073741825       ; Also the integer 1 on a 30-bit implementation.

As a special exception, if a sequence of digits specifies an integer too large or too small to be a valid integer object, the Lisp reader reads it as a floating-point number (see Floating Point Type). For instance, on most machines 536870912 is read as the floating-point number 536870912.0.

See Numbers, for more information.

2.3.2 Floating Point Type

Floating point numbers are the computer equivalent of scientific notation; you can think of a floating point number as a fraction together with a power of ten. The precise number of significant figures and the range of possible exponents is machine-specific; Emacs uses the C data type double to store the value, and internally this records a power of 2 rather than a power of 10.

The printed representation for floating point numbers requires either a decimal point (with at least one digit following), an exponent, or both. For example, ‘1500.0’, ‘15e2’, ‘15.0e2’, ‘1.5e3’, and ‘.15e4’ are five ways of writing a floating point number whose value is 1500. They are all equivalent.

See Numbers, for more information.

2.3.3 Character Type

A character in Emacs Lisp is nothing more than an integer. In other words, characters are represented by their character codes. For example, the character A is represented as the integer 65.

Individual characters are used occasionally in programs, but it is more common to work with strings, which are sequences composed of characters. See String Type.

Characters in strings and buffers are currently limited to the range of 0 to 4194303—twenty two bits (see Character Codes). Codes 0 through 127 are ASCII codes; the rest are non-ASCII (see Non-ASCII Characters). Characters that represent keyboard input have a much wider range, to encode modifier keys such as Control, Meta and Shift.

There are special functions for producing a human-readable textual description of a character for the sake of messages. See Describing Characters.

• Basic Char Syntax: Syntax for regular characters.
• General Escape Syntax: How to specify characters by their codes.
• Ctl-Char Syntax: Syntax for control characters.
• Meta-Char Syntax: Syntax for meta-characters.
• Other Char Bits: Syntax for hyper-, super-, and alt-characters.

2.3.3.1 Basic Char Syntax

Since characters are really integers, the printed representation of a character is a decimal number. This is also a possible read syntax for a character, but writing characters that way in Lisp programs is not clear programming. You should always use the special read syntax formats that Emacs Lisp provides for characters. These syntax formats start with a question mark.

The usual read syntax for alphanumeric characters is a question mark followed by the character; thus, ‘?A’ for the character A, ‘?B’ for the character B, and ‘?a’ for the character a.

For example:

     ?Q ⇒ 81     ?q ⇒ 113

You can use the same syntax for punctuation characters, but it is often a good idea to add a ‘\’ so that the Emacs commands for editing Lisp code don't get confused. For example, ‘?\(’ is the way to write the open-paren character. If the character is ‘\’, you must use a second ‘\’ to quote it: ‘?\\’.

You can express the characters control-g, backspace, tab, newline, vertical tab, formfeed, space, return, del, and escape as ‘?\a’, ‘?\b’, ‘?\t’, ‘?\n’, ‘?\v’, ‘?\f’, ‘?\s’, ‘?\r’, ‘?\d’, and ‘?\e’, respectively. (‘?\s’ followed by a dash has a different meaning—it applies the “super” modifier to the following character.) Thus,

     ?\a ⇒ 7                 ; control-g, C-g
     ?\b ⇒ 8                 ; backspace, <BS>, C-h
     ?\t ⇒ 9                 ; tab, <TAB>, C-i
     ?\n ⇒ 10                ; newline, C-j
     ?\v ⇒ 11                ; vertical tab, C-k
     ?\f ⇒ 12                ; formfeed character, C-l
     ?\r ⇒ 13                ; carriage return, <RET>, C-m
     ?\e ⇒ 27                ; escape character, <ESC>, C-[
     ?\s ⇒ 32                ; space character, <SPC>
     ?\\ ⇒ 92                ; backslash character, \
     ?\d ⇒ 127               ; delete character, <DEL>

These sequences which start with backslash are also known as escape sequences, because backslash plays the role of an “escape character”; this terminology has nothing to do with the character <ESC>. ‘\s’ is meant for use in character constants; in string constants, just write the space.

A backslash is allowed, and harmless, preceding any character without a special escape meaning; thus, ‘?\+’ is equivalent to ‘?+’. There is no reason to add a backslash before most characters. However, you should add a backslash before any of the characters ‘()\|;'`"#.,’ to avoid confusing the Emacs commands for editing Lisp code. You can also add a backslash before whitespace characters such as space, tab, newline and formfeed. However, it is cleaner to use one of the easily readable escape sequences, such as ‘\t’ or ‘\s’, instead of an actual whitespace character such as a tab or a space. (If you do write backslash followed by a space, you should write an extra space after the character constant to separate it from the following text.)

2.3.3.2 General Escape Syntax

In addition to the specific escape sequences for special important control characters, Emacs provides several types of escape syntax that you can use to specify non-ASCII text characters.

You can specify characters by their Unicode values. ?\unnnn represents a character that maps to the Unicode code point ‘U+nnnn’ (by convention, Unicode code points are given in hexadecimal). There is a slightly different syntax for specifying characters with code points higher than U+ffff: \U00nnnnnn represents the character whose code point is ‘U+nnnnnn’. The Unicode Standard only defines code points up to ‘U+10ffff’, so if you specify a code point higher than that, Emacs signals an error.

This peculiar and inconvenient syntax was adopted for compatibility with other programming languages. Unlike some other languages, Emacs Lisp supports this syntax only in character literals and strings.

The most general read syntax for a character represents the character code in either octal or hex. To use octal, write a question mark followed by a backslash and the octal character code (up to three octal digits); thus, ‘?\101’ for the character A, ‘?\001’ for the character C-a, and ?\002 for the character C-b. Although this syntax can represent any ASCII character, it is preferred only when the precise octal value is more important than the ASCII representation.

     ?\012 ⇒ 10         ?\n ⇒ 10         ?\C-j ⇒ 10
     ?\101 ⇒ 65         ?A ⇒ 65

To use hex, write a question mark followed by a backslash, ‘x’, and the hexadecimal character code. You can use any number of hex digits, so you can represent any character code in this way. Thus, ‘?\x41’ for the character A, ‘?\x1’ for the character C-a, and ?\x8e0 for the Latin-1 character ‘a’ with grave accent.

2.3.3.3 Control-Character Syntax

Control characters can be represented using yet another read syntax. This consists of a question mark followed by a backslash, caret, and the corresponding non-control character, in either upper or lower case. For example, both ‘?\^I’ and ‘?\^i’ are valid read syntax for the character C-i, the character whose value is 9.

Instead of the ‘^’, you can use ‘C-’; thus, ‘?\C-i’ is equivalent to ‘?\^I’ and to ‘?\^i’:

     ?\^I ⇒ 9     ?\C-I ⇒ 9

In strings and buffers, the only control characters allowed are those that exist in ASCII; but for keyboard input purposes, you can turn any character into a control character with ‘C-’. The character codes for these non-ASCII control characters include the 2**26 bit as well as the code for the corresponding non-control character. Ordinary terminals have no way of generating non-ASCII control characters, but you can generate them straightforwardly using X and other window systems.

For historical reasons, Emacs treats the <DEL> character as the control equivalent of ?:

     ?\^? ⇒ 127     ?\C-? ⇒ 127

As a result, it is currently not possible to represent the character Control-?, which is a meaningful input character under X, using ‘\C-’. It is not easy to change this, as various Lisp files refer to <DEL> in this way.

For representing control characters to be found in files or strings, we recommend the ‘^’ syntax; for control characters in keyboard input, we prefer the ‘C-’ syntax. Which one you use does not affect the meaning of the program, but may guide the understanding of people who read it.

2.3.3.4 Meta-Character Syntax

A meta character is a character typed with the <META> modifier key. The integer that represents such a character has the 2**27 bit set. We use high bits for this and other modifiers to make possible a wide range of basic character codes.

In a string, the 2**7 bit attached to an ASCII character indicates a meta character; thus, the meta characters that can fit in a string have codes in the range from 128 to 255, and are the meta versions of the ordinary ASCII characters. See Strings of Events, for details about <META>-handling in strings.

The read syntax for meta characters uses ‘\M-’. For example, ‘?\M-A’ stands for M-A. You can use ‘\M-’ together with octal character codes (see below), with ‘\C-’, or with any other syntax for a character. Thus, you can write M-A as ‘?\M-A’, or as ‘?\M-\101’. Likewise, you can write C-M-b as ‘?\M-\C-b’, ‘?\C-\M-b’, or ‘?\M-\002’.

2.3.3.5 Other Character Modifier Bits

The case of a graphic character is indicated by its character code; for example, ASCII distinguishes between the characters ‘a’ and ‘A’. But ASCII has no way to represent whether a control character is upper case or lower case. Emacs uses the 2**25 bit to indicate that the shift key was used in typing a control character. This distinction is possible only when you use X terminals or other special terminals; ordinary terminals do not report the distinction to the computer in any way. The Lisp syntax for the shift bit is ‘\S-’; thus, ‘?\C-\S-o’ or ‘?\C-\S-O’ represents the shifted-control-o character.

The X Window System defines three other modifier bits that can be set in a character: hyper, super and alt. The syntaxes for these bits are ‘\H-’, ‘\s-’ and ‘\A-’. (Case is significant in these prefixes.) Thus, ‘?\H-\M-\A-x’ represents Alt-Hyper-Meta-x. (Note that ‘\s’ with no following ‘-’ represents the space character.) Numerically, the bit values are 2**22 for alt, 2**23 for super and 2**24 for hyper.

2.3.4 Symbol Type

A symbol in GNU Emacs Lisp is an object with a name. The symbol name serves as the printed representation of the symbol. In ordinary Lisp use, with one single obarray (see Creating Symbols), a symbol's name is unique—no two symbols have the same name.

A symbol can serve as a variable, as a function name, or to hold a property list. Or it may serve only to be distinct from all other Lisp objects, so that its presence in a data structure may be recognized reliably. In a given context, usually only one of these uses is intended. But you can use one symbol in all of these ways, independently.

A symbol whose name starts with a colon (‘:’) is called a keyword symbol. These symbols automatically act as constants, and are normally used only by comparing an unknown symbol with a few specific alternatives.

A symbol name can contain any characters whatever. Most symbol names are written with letters, digits, and the punctuation characters ‘-+=*/’. Such names require no special punctuation; the characters of the name suffice as long as the name does not look like a number. (If it does, write a ‘\’ at the beginning of the name to force interpretation as a symbol.) The characters ‘_~!@$%^&:<>{}?’ are less often used but also require no special punctuation. Any other characters may be included in a symbol's name by escaping them with a backslash. In contrast to its use in strings, however, a backslash in the name of a symbol simply quotes the single character that follows the backslash. For example, in a string, ‘\t’ represents a tab character; in the name of a symbol, however, ‘\t’ merely quotes the letter ‘t’. To have a symbol with a tab character in its name, you must actually use a tab (preceded with a backslash). But it's rare to do such a thing.

Common Lisp note: In Common Lisp, lower case letters are always “folded” to upper case, unless they are explicitly escaped. In Emacs Lisp, upper case and lower case letters are distinct.

Here are several examples of symbol names. Note that the ‘+’ in the fifth example is escaped to prevent it from being read as a number. This is not necessary in the fourth example because the rest of the name makes it invalid as a number.

     foo                 ; A symbol named ‘foo’.
     FOO                 ; A symbol named ‘FOO’, different from ‘foo’.
     1+                  ; A symbol named ‘1+’
                         ;   (not ‘+1’, which is an integer).
     \+1                 ; A symbol named ‘+1’
                         ;   (not a very readable name).
     \(*\ 1\ 2\)         ; A symbol named ‘(* 1 2)’ (a worse name).
     
     
     +-*/_~!@$%^&=:<>{}  ; A symbol named ‘+-*/_~!@$%^&=:<>{}’.
                         ;   These characters need not be escaped.

Normally the Lisp reader interns all symbols (see Creating Symbols). To prevent interning, you can write ‘#:’ before the name of the symbol.

2.3.5 Sequence Types

A sequence is a Lisp object that represents an ordered set of elements. There are two kinds of sequence in Emacs Lisp, lists and arrays. Thus, an object of type list or of type array is also considered a sequence.

Arrays are further subdivided into strings, vectors, char-tables and bool-vectors. Vectors can hold elements of any type, but string elements must be characters, and bool-vector elements must be t or nil. Char-tables are like vectors except that they are indexed by any valid character code. The characters in a string can have text properties like characters in a buffer (see Text Properties), but vectors do not support text properties, even when their elements happen to be characters.

Lists, strings and the other array types are different, but they have important similarities. For example, all have a length l, and all have elements which can be indexed from zero to l minus one. Several functions, called sequence functions, accept any kind of sequence. For example, the function elt can be used to extract an element of a sequence, given its index. See Sequences Arrays Vectors.

It is generally impossible to read the same sequence twice, since sequences are always created anew upon reading. If you read the read syntax for a sequence twice, you get two sequences with equal contents. There is one exception: the empty list () always stands for the same object, nil.

2.3.6 Cons Cell and List Types

A cons cell is an object that consists of two slots, called the car slot and the cdr slot. Each slot can hold or refer to any Lisp object. We also say that “the car of this cons cell is” whatever object its car slot currently holds, and likewise for the cdr.

A note to C programmers: in Lisp, we do not distinguish between “holding” a value and “pointing to” the value, because pointers in Lisp are implicit.

A list is a series of cons cells, linked together so that the cdr slot of each cons cell holds either the next cons cell or the empty list. The empty list is actually the symbol nil. See Lists, for functions that work on lists. Because most cons cells are used as part of lists, the phrase list structure has come to refer to any structure made out of cons cells.

Because cons cells are so central to Lisp, we also have a word for “an object which is not a cons cell.” These objects are called atoms.

The read syntax and printed representation for lists are identical, and consist of a left parenthesis, an arbitrary number of elements, and a right parenthesis. Here are examples of lists:

     (A 2 "A")            ; A list of three elements.
     ()                   ; A list of no elements (the empty list).
     nil                  ; A list of no elements (the empty list).
     ("A ()")             ; A list of one element: the string "A ()".
     (A ())               ; A list of two elements: A and the empty list.
     (A nil)              ; Equivalent to the previous.
     ((A B C))            ; A list of one element
                          ;   (which is a list of three elements).

Upon reading, each object inside the parentheses becomes an element of the list. That is, a cons cell is made for each element. The car slot of the cons cell holds the element, and its cdr slot refers to the next cons cell of the list, which holds the next element in the list. The cdr slot of the last cons cell is set to hold nil.

The names car and cdr derive from the history of Lisp. The original Lisp implementation ran on an IBM 704 computer which divided words into two parts, called the “address” part and the “decrement”; car was an instruction to extract the contents of the address part of a register, and cdr an instruction to extract the contents of the decrement. By contrast, “cons cells” are named for the function cons that creates them, which in turn was named for its purpose, the construction of cells.

• Box Diagrams: Drawing pictures of lists.
• Dotted Pair Notation: A general syntax for cons cells.
• Association List Type: A specially constructed list.

2.3.6.1 Drawing Lists as Box Diagrams

A list can be illustrated by a diagram in which the cons cells are shown as pairs of boxes, like dominoes. (The Lisp reader cannot read such an illustration; unlike the textual notation, which can be understood by both humans and computers, the box illustrations can be understood only by humans.) This picture represents the three-element list (rose violet buttercup):

         --- ---      --- ---      --- ---
        |   |   |--> |   |   |--> |   |   |--> nil
         --- ---      --- ---      --- ---
          |            |            |
          |            |            |
           --> rose     --> violet   --> buttercup

In this diagram, each box represents a slot that can hold or refer to any Lisp object. Each pair of boxes represents a cons cell. Each arrow represents a reference to a Lisp object, either an atom or another cons cell.

In this example, the first box, which holds the car of the first cons cell, refers to or “holds” rose (a symbol). The second box, holding the cdr of the first cons cell, refers to the next pair of boxes, the second cons cell. The car of the second cons cell is violet, and its cdr is the third cons cell. The cdr of the third (and last) cons cell is nil.

Here is another diagram of the same list, (rose violet buttercup), sketched in a different manner:

      ---------------       ----------------       -------------------
     | car   | cdr   |     | car    | cdr   |     | car       | cdr   |
     | rose  |   o-------->| violet |   o-------->| buttercup |  nil  |
     |       |       |     |        |       |     |           |       |
      ---------------       ----------------       -------------------

A list with no elements in it is the empty list; it is identical to the symbol nil. In other words, nil is both a symbol and a list.

Here is the list (A ()), or equivalently (A nil), depicted with boxes and arrows:

         --- ---      --- ---
        |   |   |--> |   |   |--> nil
         --- ---      --- ---
          |            |
          |            |
           --> A        --> nil

Here is a more complex illustration, showing the three-element list, ((pine needles) oak maple), the first element of which is a two-element list:

         --- ---      --- ---      --- ---
        |   |   |--> |   |   |--> |   |   |--> nil
         --- ---      --- ---      --- ---
          |            |            |
          |            |            |
          |             --> oak      --> maple
          |
          |     --- ---      --- ---
           --> |   |   |--> |   |   |--> nil
                --- ---      --- ---
                 |            |
                 |            |
                  --> pine     --> needles

The same list represented in the second box notation looks like this:

      --------------       --------------       --------------
     | car   | cdr  |     | car   | cdr  |     | car   | cdr  |
     |   o   |   o------->| oak   |   o------->| maple |  nil |
     |   |   |      |     |       |      |     |       |      |
      -- | ---------       --------------       --------------
         |
         |
         |        --------------       ----------------
         |       | car   | cdr  |     | car     | cdr  |
          ------>| pine  |   o------->| needles |  nil |
                 |       |      |     |         |      |
                  --------------       ----------------

2.3.6.2 Dotted Pair Notation

Dotted pair notation is a general syntax for cons cells that represents the car and cdr explicitly. In this syntax, (a . b) stands for a cons cell whose car is the object a and whose cdr is the object b. Dotted pair notation is more general than list syntax because the cdr does not have to be a list. However, it is more cumbersome in cases where list syntax would work. In dotted pair notation, the list ‘(1 2 3)’ is written as ‘(1 . (2 . (3 . nil)))’. For nil-terminated lists, you can use either notation, but list notation is usually clearer and more convenient. When printing a list, the dotted pair notation is only used if the cdr of a cons cell is not a list.

Here's an example using boxes to illustrate dotted pair notation. This example shows the pair (rose . violet):

         --- ---
        |   |   |--> violet
         --- ---
          |
          |
           --> rose

You can combine dotted pair notation with list notation to represent conveniently a chain of cons cells with a non-nil final cdr. You write a dot after the last element of the list, followed by the cdr of the final cons cell. For example, (rose violet . buttercup) is equivalent to (rose . (violet . buttercup)). The object looks like this:

         --- ---      --- ---
        |   |   |--> |   |   |--> buttercup
         --- ---      --- ---
          |            |
          |            |
           --> rose     --> violet

The syntax (rose . violet . buttercup) is invalid because there is nothing that it could mean. If anything, it would say to put buttercup in the cdr of a cons cell whose cdr is already used for violet.

The list (rose violet) is equivalent to (rose . (violet)), and looks like this:

         --- ---      --- ---
        |   |   |--> |   |   |--> nil
         --- ---      --- ---
          |            |
          |            |
           --> rose     --> violet

Similarly, the three-element list (rose violet buttercup) is equivalent to (rose . (violet . (buttercup))). It looks like this:

         --- ---      --- ---      --- ---
        |   |   |--> |   |   |--> |   |   |--> nil
         --- ---      --- ---      --- ---
          |            |            |
          |            |            |
           --> rose     --> violet   --> buttercup

2.3.6.3 Association List Type

An association list or alist is a specially-constructed list whose elements are cons cells. In each element, the car is considered a key, and the cdr is considered an associated value. (In some cases, the associated value is stored in the car of the cdr.) Association lists are often used as stacks, since it is easy to add or remove associations at the front of the list.

For example,

     (setq alist-of-colors
           '((rose . red) (lily . white) (buttercup . yellow)))

sets the variable alist-of-colors to an alist of three elements. In the first element, rose is the key and red is the value.

See Association Lists, for a further explanation of alists and for functions that work on alists. See Hash Tables, for another kind of lookup table, which is much faster for handling a large number of keys.

2.3.7 Array Type

An array is composed of an arbitrary number of slots for holding or referring to other Lisp objects, arranged in a contiguous block of memory. Accessing any element of an array takes approximately the same amount of time. In contrast, accessing an element of a list requires time proportional to the position of the element in the list. (Elements at the end of a list take longer to access than elements at the beginning of a list.)

Emacs defines four types of array: strings, vectors, bool-vectors, and char-tables.

A string is an array of characters and a vector is an array of arbitrary objects. A bool-vector can hold only t or nil. These kinds of array may have any length up to the largest integer. Char-tables are sparse arrays indexed by any valid character code; they can hold arbitrary objects.

The first element of an array has index zero, the second element has index 1, and so on. This is called zero-origin indexing. For example, an array of four elements has indices 0, 1, 2, and 3. The largest possible index value is one less than the length of the array. Once an array is created, its length is fixed.

All Emacs Lisp arrays are one-dimensional. (Most other programming languages support multidimensional arrays, but they are not essential; you can get the same effect with nested one-dimensional arrays.) Each type of array has its own read syntax; see the following sections for details.

The array type is a subset of the sequence type, and contains the string type, the vector type, the bool-vector type, and the char-table type.

2.3.8 String Type

A string is an array of characters. Strings are used for many purposes in Emacs, as can be expected in a text editor; for example, as the names of Lisp symbols, as messages for the user, and to represent text extracted from buffers. Strings in Lisp are constants: evaluation of a string returns the same string.

See Strings and Characters, for functions that operate on strings.

• Syntax for Strings: How to specify Lisp strings.
• Non-ASCII in Strings: International characters in strings.
• Nonprinting Characters: Literal unprintable characters in strings.
• Text Props and Strings: Strings with text properties.

2.3.8.1 Syntax for Strings

The read syntax for a string is a double-quote, an arbitrary number of characters, and another double-quote, "like this". To include a double-quote in a string, precede it with a backslash; thus, "\"" is a string containing just a single double-quote character. Likewise, you can include a backslash by preceding it with another backslash, like this: "this \\ is a single embedded backslash".

The newline character is not special in the read syntax for strings; if you write a new line between the double-quotes, it becomes a character in the string. But an escaped newline—one that is preceded by ‘\’—does not become part of the string; i.e., the Lisp reader ignores an escaped newline while reading a string. An escaped space ‘\ ’ is likewise ignored.

     "It is useful to include newlines
     in documentation strings,
     but the newline is \
     ignored if escaped."
          ⇒ "It is useful to include newlines
     in documentation strings,
     but the newline is ignored if escaped."

2.3.8.2 Non-ASCII Characters in Strings

You can include a non-ASCII international character in a string constant by writing it literally. There are two text representations for non-ASCII characters in Emacs strings (and in buffers): unibyte and multibyte. If the string constant is read from a multibyte source, such as a multibyte buffer or string, or a file that would be visited as multibyte, then the character is read as a multibyte character, and that makes the string multibyte. If the string constant is read from a unibyte source, then the character is read as unibyte and that makes the string unibyte.

You can also represent a multibyte non-ASCII character with its character code: use a hex escape, ‘\xnnnnnnn’, with as many digits as necessary. (Multibyte non-ASCII character codes are all greater than 256.) Any character which is not a valid hex digit terminates this construct. If the next character in the string could be interpreted as a hex digit, write ‘\ ’ (backslash and space) to terminate the hex escape—for example, ‘\x8e0\ ’ represents one character, ‘a’ with grave accent. ‘\ ’ in a string constant is just like backslash-newline; it does not contribute any character to the string, but it does terminate the preceding hex escape.

You can represent a unibyte non-ASCII character with its character code, which must be in the range from 128 (0200 octal) to 255 (0377 octal). If you write all such character codes in octal and the string contains no other characters forcing it to be multibyte, this produces a unibyte string. However, using any hex escape in a string (even for an ASCII character) forces the string to be multibyte.

You can also specify characters in a string by their numeric values in Unicode, using ‘\u’ and ‘\U’ (see Character Type).

See Text Representations, for more information about the two text representations.

2.3.8.3 Nonprinting Characters in Strings

You can use the same backslash escape-sequences in a string constant as in character literals (but do not use the question mark that begins a character constant). For example, you can write a string containing the nonprinting characters tab and C-a, with commas and spaces between them, like this: "\t, \C-a". See Character Type, for a description of the read syntax for characters.

However, not all of the characters you can write with backslash escape-sequences are valid in strings. The only control characters that a string can hold are the ASCII control characters. Strings do not distinguish case in ASCII control characters.

Properly speaking, strings cannot hold meta characters; but when a string is to be used as a key sequence, there is a special convention that provides a way to represent meta versions of ASCII characters in a string. If you use the ‘\M-’ syntax to indicate a meta character in a string constant, this sets the 2**7 bit of the character in the string. If the string is used in define-key or lookup-key, this numeric code is translated into the equivalent meta character. See Character Type.

Strings cannot hold characters that have the hyper, super, or alt modifiers.

2.3.8.4 Text Properties in Strings

A string can hold properties for the characters it contains, in addition to the characters themselves. This enables programs that copy text between strings and buffers to copy the text's properties with no special effort. See Text Properties, for an explanation of what text properties mean. Strings with text properties use a special read and print syntax:

     #("characters" property-data...)

where property-data consists of zero or more elements, in groups of three as follows:

     beg end plist

The elements beg and end are integers, and together specify a range of indices in the string; plist is the property list for that range. For example,

     #("foo bar" 0 3 (face bold) 3 4 nil 4 7 (face italic))

represents a string whose textual contents are ‘foo bar’, in which the first three characters have a face property with value bold, and the last three have a face property with value italic. (The fourth character has no text properties, so its property list is nil. It is not actually necessary to mention ranges with nil as the property list, since any characters not mentioned in any range will default to having no properties.)

2.3.9 Vector Type

A vector is a one-dimensional array of elements of any type. It takes a constant amount of time to access any element of a vector. (In a list, the access time of an element is proportional to the distance of the element from the beginning of the list.)

The printed representation of a vector consists of a left square bracket, the elements, and a right square bracket. This is also the read syntax. Like numbers and strings, vectors are considered constants for evaluation.

     [1 "two" (three)]      ; A vector of three elements.
          ⇒ [1 "two" (three)]

See Vectors, for functions that work with vectors.

2.3.10 Char-Table Type

A char-table is a one-dimensional array of elements of any type, indexed by character codes. Char-tables have certain extra features to make them more useful for many jobs that involve assigning information to character codes—for example, a char-table can have a parent to inherit from, a default value, and a small number of extra slots to use for special purposes. A char-table can also specify a single value for a whole character set.

The printed representation of a char-table is like a vector except that there is an extra ‘#^’ at the beginning.

See Char-Tables, for special functions to operate on char-tables. Uses of char-tables include:

• Case tables (see Case Tables).
• Character category tables (see Categories).
• Display tables (see Display Tables).
• Syntax tables (see Syntax Tables).

2.3.11 Bool-Vector Type

A bool-vector is a one-dimensional array whose elements must be t or nil.

The printed representation of a bool-vector is like a string, except that it begins with ‘#&’ followed by the length. The string constant that follows actually specifies the contents of the bool-vector as a bitmap—each “character” in the string contains 8 bits, which specify the next 8 elements of the bool-vector (1 stands for t, and 0 for nil). The least significant bits of the character correspond to the lowest indices in the bool-vector.

     (make-bool-vector 3 t)
          ⇒ #&3"^G"
     (make-bool-vector 3 nil)
          ⇒ #&3"^@"

These results make sense, because the binary code for ‘C-g’ is 111 and ‘C-@’ is the character with code 0.

If the length is not a multiple of 8, the printed representation shows extra elements, but these extras really make no difference. For instance, in the next example, the two bool-vectors are equal, because only the first 3 bits are used:

     (equal #&3"\377" #&3"\007")
          ⇒ t

2.3.12 Hash Table Type

A hash table is a very fast kind of lookup table, somewhat like an alist in that it maps keys to corresponding values, but much faster. The printed representation of a hash table specifies its properties and contents, like this:

     (make-hash-table)
          ⇒ #s(hash-table size 65 test eql rehash-size 1.5
                                  rehash-threshold 0.8 data ())

See Hash Tables, for more information about hash tables.

2.3.13 Function Type

Lisp functions are executable code, just like functions in other programming languages. In Lisp, unlike most languages, functions are also Lisp objects. A non-compiled function in Lisp is a lambda expression: that is, a list whose first element is the symbol lambda (see Lambda Expressions).

In most programming languages, it is impossible to have a function without a name. In Lisp, a function has no intrinsic name. A lambda expression can be called as a function even though it has no name; to emphasize this, we also call it an anonymous function (see Anonymous Functions). A named function in Lisp is just a symbol with a valid function in its function cell (see Defining Functions).

Most of the time, functions are called when their names are written in Lisp expressions in Lisp programs. However, you can construct or obtain a function object at run time and then call it with the primitive functions funcall and apply. See Calling Functions.

2.3.14 Macro Type

A Lisp macro is a user-defined construct that extends the Lisp language. It is represented as an object much like a function, but with different argument-passing semantics. A Lisp macro has the form of a list whose first element is the symbol macro and whose cdr is a Lisp function object, including the lambda symbol.

Lisp macro objects are usually defined with the built-in defmacro function, but any list that begins with macro is a macro as far as Emacs is concerned. See Macros, for an explanation of how to write a macro.

Warning: Lisp macros and keyboard macros (see Keyboard Macros) are entirely different things. When we use the word “macro” without qualification, we mean a Lisp macro, not a keyboard macro.

2.3.15 Primitive Function Type

A primitive function is a function callable from Lisp but written in the C programming language. Primitive functions are also called subrs or built-in functions. (The word “subr” is derived from “subroutine.”) Most primitive functions evaluate all their arguments when they are called. A primitive function that does not evaluate all its arguments is called a special form (see Special Forms).

It does not matter to the caller of a function whether the function is primitive. However, this does matter if you try to redefine a primitive with a function written in Lisp. The reason is that the primitive function may be called directly from C code. Calls to the redefined function from Lisp will use the new definition, but calls from C code may still use the built-in definition. Therefore, we discourage redefinition of primitive functions.

The term function refers to all Emacs functions, whether written in Lisp or C. See Function Type, for information about the functions written in Lisp.

Primitive functions have no read syntax and print in hash notation with the name of the subroutine.

     (symbol-function 'car)          ; Access the function cell
                                     ;   of the symbol.
          ⇒ #<subr car>
     (subrp (symbol-function 'car))  ; Is this a primitive function?
          ⇒ t                       ; Yes.

2.3.16 Byte-Code Function Type

The byte compiler produces byte-code function objects. Internally, a byte-code function object is much like a vector; however, the evaluator handles this data type specially when it appears as a function to be called. See Byte Compilation, for information about the byte compiler.

The printed representation and read syntax for a byte-code function object is like that for a vector, with an additional ‘#’ before the opening ‘[’.

2.3.17 Autoload Type

An autoload object is a list whose first element is the symbol autoload. It is stored as the function definition of a symbol, where it serves as a placeholder for the real definition. The autoload object says that the real definition is found in a file of Lisp code that should be loaded when necessary. It contains the name of the file, plus some other information about the real definition.

After the file has been loaded, the symbol should have a new function definition that is not an autoload object. The new definition is then called as if it had been there to begin with. From the user's point of view, the function call works as expected, using the function definition in the loaded file.

An autoload object is usually created with the function autoload, which stores the object in the function cell of a symbol. See Autoload, for more details.

2.4 Editing Types

The types in the previous section are used for general programming purposes, and most of them are common to most Lisp dialects. Emacs Lisp provides several additional data types for purposes connected with editing.

• Buffer Type: The basic object of editing.
• Marker Type: A position in a buffer.
• Window Type: Buffers are displayed in windows.
• Frame Type: Windows subdivide frames.
• Terminal Type: A terminal device displays frames.
• Window Configuration Type: Recording the way a frame is subdivided.
• Frame Configuration Type: Recording the status of all frames.
• Process Type: A subprocess of Emacs running on the underlying OS.
• Stream Type: Receive or send characters.
• Keymap Type: What function a keystroke invokes.
• Overlay Type: How an overlay is represented.
• Font Type: Fonts for displaying text.

2.4.1 Buffer Type

A buffer is an object that holds text that can be edited (see Buffers). Most buffers hold the contents of a disk file (see Files) so they can be edited, but some are used for other purposes. Most buffers are also meant to be seen by the user, and therefore displayed, at some time, in a window (see Windows). But a buffer need not be displayed in any window. Each buffer has a designated position called point (see Positions); most editing commands act on the contents of the current buffer in the neighborhood of point. At any time, one buffer is the current buffer.

The contents of a buffer are much like a string, but buffers are not used like strings in Emacs Lisp, and the available operations are different. For example, you can insert text efficiently into an existing buffer, altering the buffer's contents, whereas “inserting” text into a string requires concatenating substrings, and the result is an entirely new string object.

Many of the standard Emacs functions manipulate or test the characters in the current buffer; a whole chapter in this manual is devoted to describing these functions (see Text).

Several other data structures are associated with each buffer:

• a local syntax table (see Syntax Tables);
• a local keymap (see Keymaps); and,
• a list of buffer-local variable bindings (see Buffer-Local Variables).
• overlays (see Overlays).
• text properties for the text in the buffer (see Text Properties).

The local keymap and variable list contain entries that individually override global bindings or values. These are used to customize the behavior of programs in different buffers, without actually changing the programs.

A buffer may be indirect, which means it shares the text of another buffer, but presents it differently. See Indirect Buffers.

Buffers have no read syntax. They print in hash notation, showing the buffer name.

     (current-buffer)
          ⇒ #<buffer objects.texi>

2.4.2 Marker Type

A marker denotes a position in a specific buffer. Markers therefore have two components: one for the buffer, and one for the position. Changes in the buffer's text automatically relocate the position value as necessary to ensure that the marker always points between the same two characters in the buffer.

Markers have no read syntax. They print in hash notation, giving the current character position and the name of the buffer.

     (point-marker)
          ⇒ #<marker at 10779 in objects.texi>

See Markers, for information on how to test, create, copy, and move markers.

2.4.3 Window Type

A window describes the portion of the terminal screen that Emacs uses to display a buffer. Every window has one associated buffer, whose contents appear in the window. By contrast, a given buffer may appear in one window, no window, or several windows.

Though many windows may exist simultaneously, at any time one window is designated the selected window. This is the window where the cursor is (usually) displayed when Emacs is ready for a command. The selected window usually displays the current buffer, but this is not necessarily the case.

Windows are grouped on the screen into frames; each window belongs to one and only one frame. See Frame Type.

Windows have no read syntax. They print in hash notation, giving the window number and the name of the buffer being displayed. The window numbers exist to identify windows uniquely, since the buffer displayed in any given window can change frequently.

     (selected-window)
          ⇒ #<window 1 on objects.texi>

See Windows, for a description of the functions that work on windows.

2.4.4 Frame Type

A frame is a screen area that contains one or more Emacs windows; we also use the term “frame” to refer to the Lisp object that Emacs uses to refer to the screen area.

Frames have no read syntax. They print in hash notation, giving the frame's title, plus its address in core (useful to identify the frame uniquely).

     (selected-frame)
          ⇒ #<frame emacs@psilocin.gnu.org 0xdac80>

See Frames, for a description of the functions that work on frames.

2.4.5 Terminal Type

A terminal is a device capable of displaying one or more Emacs frames (see Frame Type).

Terminals have no read syntax. They print in hash notation giving the terminal's ordinal number and its TTY device file name.

     (get-device-terminal nil)
          ⇒ #<terminal 1 on /dev/tty>

2.4.6 Window Configuration Type

A window configuration stores information about the positions, sizes, and contents of the windows in a frame, so you can recreate the same arrangement of windows later.

Window configurations do not have a read syntax; their print syntax looks like ‘#<window-configuration>’. See Window Configurations, for a description of several functions related to window configurations.

2.4.7 Frame Configuration Type

A frame configuration stores information about the positions, sizes, and contents of the windows in all frames. It is not a primitive type—it is actually a list whose car is frame-configuration and whose cdr is an alist. Each alist element describes one frame, which appears as the car of that element.

See Frame Configurations, for a description of several functions related to frame configurations.

2.4.8 Process Type

The word process usually means a running program. Emacs itself runs in a process of this sort. However, in Emacs Lisp, a process is a Lisp object that designates a subprocess created by the Emacs process. Programs such as shells, GDB, ftp, and compilers, running in subprocesses of Emacs, extend the capabilities of Emacs.

An Emacs subprocess takes textual input from Emacs and returns textual output to Emacs for further manipulation. Emacs can also send signals to the subprocess.

Process objects have no read syntax. They print in hash notation, giving the name of the process:

     (process-list)
          ⇒ (#<process shell>)

See Processes, for information about functions that create, delete, return information about, send input or signals to, and receive output from processes.

2.4.9 Stream Type

A stream is an object that can be used as a source or sink for characters—either to supply characters for input or to accept them as output. Many different types can be used this way: markers, buffers, strings, and functions. Most often, input streams (character sources) obtain characters from the keyboard, a buffer, or a file, and output streams (character sinks) send characters to a buffer, such as a *Help* buffer, or to the echo area.

The object nil, in addition to its other meanings, may be used as a stream. It stands for the value of the variable standard-input or standard-output. Also, the object t as a stream specifies input using the minibuffer (see Minibuffers) or output in the echo area (see The Echo Area).

Streams have no special printed representation or read syntax, and print as whatever primitive type they are.

See Read and Print, for a description of functions related to streams, including parsing and printing functions.

2.4.10 Keymap Type

A keymap maps keys typed by the user to commands. This mapping controls how the user's command input is executed. A keymap is actually a list whose car is the symbol keymap.

See Keymaps, for information about creating keymaps, handling prefix keys, local as well as global keymaps, and changing key bindings.

2.4.11 Overlay Type

An overlay specifies properties that apply to a part of a buffer. Each overlay applies to a specified range of the buffer, and contains a property list (a list whose elements are alternating property names and values). Overlay properties are used to present parts of the buffer temporarily in a different display style. Overlays have no read syntax, and print in hash notation, giving the buffer name and range of positions.

See Overlays, for how to create and use overlays.

2.4.12 Font Type

A font specifies how to display text on a graphical terminal. There are actually three separate font types—font objects, font specs, and font entities—each of which has slightly different properties. None of them have a read syntax; their print syntax looks like ‘#<font-object>’, ‘#<font-spec>’, and ‘#<font-entity>’ respectively. See Low-Level Font, for a description of these Lisp objects.

2.5 Read Syntax for Circular Objects

To represent shared or circular structures within a complex of Lisp objects, you can use the reader constructs ‘#n=’ and ‘#n#’.

Use #n= before an object to label it for later reference; subsequently, you can use #n# to refer the same object in another place. Here, n is some integer. For example, here is how to make a list in which the first element recurs as the third element:

     (#1=(a) b #1#)

This differs from ordinary syntax such as this

     ((a) b (a))

which would result in a list whose first and third elements look alike but are not the same Lisp object. This shows the difference:

     (prog1 nil
       (setq x '(#1=(a) b #1#)))
     (eq (nth 0 x) (nth 2 x))
          ⇒ t
     (setq x '((a) b (a)))
     (eq (nth 0 x) (nth 2 x))
          ⇒ nil

You can also use the same syntax to make a circular structure, which appears as an “element” within itself. Here is an example:

     #1=(a #1#)

This makes a list whose second element is the list itself. Here's how you can see that it really works:

     (prog1 nil
       (setq x '#1=(a #1#)))
     (eq x (cadr x))
          ⇒ t

The Lisp printer can produce this syntax to record circular and shared structure in a Lisp object, if you bind the variable print-circle to a non-nil value. See Output Variables.

2.6 Type Predicates

The Emacs Lisp interpreter itself does not perform type checking on the actual arguments passed to functions when they are called. It could not do so, since function arguments in Lisp do not have declared data types, as they do in other programming languages. It is therefore up to the individual function to test whether each actual argument belongs to a type that the function can use.

All built-in functions do check the types of their actual arguments when appropriate, and signal a wrong-type-argument error if an argument is of the wrong type. For example, here is what happens if you pass an argument to + that it cannot handle:

     (+ 2 'a)
          error--> Wrong type argument: number-or-marker-p, a

If you want your program to handle different types differently, you must do explicit type checking. The most common way to check the type of an object is to call a type predicate function. Emacs has a type predicate for each type, as well as some predicates for combinations of types.

A type predicate function takes one argument; it returns t if the argument belongs to the appropriate type, and nil otherwise. Following a general Lisp convention for predicate functions, most type predicates' names end with ‘p’.

Here is an example which uses the predicates listp to check for a list and symbolp to check for a symbol.

     (defun add-on (x)
       (cond ((symbolp x)
              ;; If X is a symbol, put it on LIST.
              (setq list (cons x list)))
             ((listp x)
              ;; If X is a list, add its elements to LIST.
              (setq list (append x list)))
             (t
              ;; We handle only symbols and lists.
              (error "Invalid argument %s in add-on" x))))

Here is a table of predefined type predicates, in alphabetical order, with references to further information.

atom

See atom.

arrayp

See arrayp.

bool-vector-p

See bool-vector-p.

bufferp

See bufferp.

byte-code-function-p

See byte-code-function-p.

case-table-p

See case-table-p.

char-or-string-p

See char-or-string-p.

char-table-p

See char-table-p.

commandp

See commandp.

consp

See consp.

display-table-p

See display-table-p.

floatp

See floatp.

fontp

See Low-Level Font.

frame-configuration-p

See frame-configuration-p.

frame-live-p

See frame-live-p.

framep

See framep.

functionp

See functionp.

hash-table-p

See hash-table-p.

integer-or-marker-p

See integer-or-marker-p.

integerp

See integerp.

keymapp

See keymapp.

keywordp

See Constant Variables.

listp

See listp.

markerp

See markerp.

wholenump

See wholenump.

nlistp

See nlistp.

numberp

See numberp.

number-or-marker-p

See number-or-marker-p.

overlayp

See overlayp.

processp

See processp.

sequencep

See sequencep.

stringp

See stringp.

subrp

See subrp.

symbolp

See symbolp.

syntax-table-p

See syntax-table-p.

user-variable-p

See user-variable-p.

vectorp

See vectorp.

window-configuration-p

See window-configuration-p.

window-live-p

See window-live-p.

windowp

See windowp.

booleanp

See booleanp.

string-or-null-p

See string-or-null-p.

The most general way to check the type of an object is to call the function type-of. Recall that each object belongs to one and only one primitive type; type-of tells you which one (see Lisp Data Types). But type-of knows nothing about non-primitive types. In most cases, it is more convenient to use type predicates than type-of.

2.7 Equality Predicates

Here we describe functions that test for equality between any two objects. Other functions test equality of contents between objects of specific types, e.g., strings. For these predicates, see the appropriate chapter describing the data type.

The test for equality is implemented recursively; for example, given two cons cells x and y, (equal x y) returns t if and only if both the expressions below return t:

     (equal (car x) (car y))
     (equal (cdr x) (cdr y))

Because of this recursive method, circular lists may therefore cause infinite recursion (leading to an error).

3 Numbers

GNU Emacs supports two numeric data types: integers and floating point numbers. Integers are whole numbers such as −3, 0, 7, 13, and 511. Their values are exact. Floating point numbers are numbers with fractional parts, such as −4.5, 0.0, or 2.71828. They can also be expressed in exponential notation: 1.5e2 equals 150; in this example, ‘e2’ stands for ten to the second power, and that is multiplied by 1.5. Floating point values are not exact; they have a fixed, limited amount of precision.

• Integer Basics: Representation and range of integers.
• Float Basics: Representation and range of floating point.
• Predicates on Numbers: Testing for numbers.
• Comparison of Numbers: Equality and inequality predicates.
• Numeric Conversions: Converting float to integer and vice versa.
• Arithmetic Operations: How to add, subtract, multiply and divide.
• Rounding Operations: Explicitly rounding floating point numbers.
• Bitwise Operations: Logical and, or, not, shifting.
• Math Functions: Trig, exponential and logarithmic functions.
• Random Numbers: Obtaining random integers, predictable or not.

3.1 Integer Basics

The range of values for an integer depends on the machine. The minimum range is −536870912 to 536870911 (30 bits; i.e., -2**29 to 2**29 - 1), but some machines may provide a wider range. Many examples in this chapter assume an integer has 30 bits. The Lisp reader reads an integer as a sequence of digits with optional initial sign and optional final period.

      1               ; The integer 1.
      1.              ; The integer 1.
     +1               ; Also the integer 1.
     -1               ; The integer −1.
      1073741825      ; Also the integer 1, due to overflow.
      0               ; The integer 0.
     -0               ; The integer 0.

The syntax for integers in bases other than 10 uses ‘#’ followed by a letter that specifies the radix: ‘b’ for binary, ‘o’ for octal, ‘x’ for hex, or ‘radixr’ to specify radix radix. Case is not significant for the letter that specifies the radix. Thus, ‘#binteger’ reads integer in binary, and ‘#radixrinteger’ reads integer in radix radix. Allowed values of radix run from 2 to 36. For example:

     #b101100 ⇒ 44
     #o54 ⇒ 44
     #x2c ⇒ 44
     #24r1k ⇒ 44

To understand how various functions work on integers, especially the bitwise operators (see Bitwise Operations), it is often helpful to view the numbers in their binary form.

In 30-bit binary, the decimal integer 5 looks like this:

     00 0000  0000 0000  0000 0000  0000 0101

(We have inserted spaces between groups of 4 bits, and two spaces between groups of 8 bits, to make the binary integer easier to read.)

The integer −1 looks like this:

     11 1111  1111 1111  1111 1111  1111 1111

−1 is represented as 30 ones. (This is called two's complement notation.)

The negative integer, −5, is creating by subtracting 4 from −1. In binary, the decimal integer 4 is 100. Consequently, −5 looks like this:

     11 1111  1111 1111  1111 1111  1111 1011

In this implementation, the largest 30-bit binary integer value is 536,870,911 in decimal. In binary, it looks like this:

     01 1111  1111 1111  1111 1111  1111 1111

Since the arithmetic functions do not check whether integers go outside their range, when you add 1 to 536,870,911, the value is the negative integer −536,870,912:

     (+ 1 536870911)
          ⇒ -536870912
          ⇒ 10 0000  0000 0000  0000 0000  0000 0000

Many of the functions described in this chapter accept markers for arguments in place of numbers. (See Markers.) Since the actual arguments to such functions may be either numbers or markers, we often give these arguments the name number-or-marker. When the argument value is a marker, its position value is used and its buffer is ignored.

See max-char, for the maximum value of a valid character codepoint.

3.2 Floating Point Basics

Floating point numbers are useful for representing numbers that are not integral. The precise range of floating point numbers is machine-specific; it is the same as the range of the C data type double on the machine you are using.

The read-syntax for floating point numbers requires either a decimal point (with at least one digit following), an exponent, or both. For example, ‘1500.0’, ‘15e2’, ‘15.0e2’, ‘1.5e3’, and ‘.15e4’ are five ways of writing a floating point number whose value is 1500. They are all equivalent. You can also use a minus sign to write negative floating point numbers, as in ‘-1.0’.

Most modern computers support the IEEE floating point standard, which provides for positive infinity and negative infinity as floating point values. It also provides for a class of values called NaN or “not-a-number”; numerical functions return such values in cases where there is no correct answer. For example, (/ 0.0 0.0) returns a NaN. For practical purposes, there's no significant difference between different NaN values in Emacs Lisp, and there's no rule for precisely which NaN value should be used in a particular case, so Emacs Lisp doesn't try to distinguish them (but it does report the sign, if you print it). Here are the read syntaxes for these special floating point values:

positive infinity

‘1.0e+INF’

negative infinity

‘-1.0e+INF’

Not-a-number

‘0.0e+NaN’ or ‘-0.0e+NaN’.

To test whether a floating point value is a NaN, compare it with itself using =. That returns nil for a NaN, and t for any other floating point value.

The value -0.0 is distinguishable from ordinary zero in IEEE floating point, but Emacs Lisp equal and = consider them equal values.

You can use logb to extract the binary exponent of a floating point number (or estimate the logarithm of an integer):

3.3 Type Predicates for Numbers

The functions in this section test for numbers, or for a specific type of number. The functions integerp and floatp can take any type of Lisp object as argument (they would not be of much use otherwise), but the zerop predicate requires a number as its argument. See also integer-or-marker-p and number-or-marker-p, in Predicates on Markers.

3.4 Comparison of Numbers

To test numbers for numerical equality, you should normally use =, not eq. There can be many distinct floating point number objects with the same numeric value. If you use eq to compare them, then you test whether two values are the same object. By contrast, = compares only the numeric values of the objects.

At present, each integer value has a unique Lisp object in Emacs Lisp. Therefore, eq is equivalent to = where integers are concerned. It is sometimes convenient to use eq for comparing an unknown value with an integer, because eq does not report an error if the unknown value is not a number—it accepts arguments of any type. By contrast, = signals an error if the arguments are not numbers or markers. However, it is a good idea to use = if you can, even for comparing integers, just in case we change the representation of integers in a future Emacs version.

Sometimes it is useful to compare numbers with equal; it treats two numbers as equal if they have the same data type (both integers, or both floating point) and the same value. By contrast, = can treat an integer and a floating point number as equal. See Equality Predicates.

There is another wrinkle: because floating point arithmetic is not exact, it is often a bad idea to check for equality of two floating point values. Usually it is better to test for approximate equality. Here's a function to do this:

     (defvar fuzz-factor 1.0e-6)
     (defun approx-equal (x y)
       (or (and (= x 0) (= y 0))
           (< (/ (abs (- x y))
                 (max (abs x) (abs y)))
              fuzz-factor)))

Common Lisp note: Comparing numbers in Common Lisp always requires = because Common Lisp implements multi-word integers, and two distinct integer objects can have the same numeric value. Emacs Lisp can have just one integer object for any given value because it has a limited range of integer values.

3.5 Numeric Conversions

To convert an integer to floating point, use the function float.

There are four functions to convert floating point numbers to integers; they differ in how they round. All accept an argument number and an optional argument divisor. Both arguments may be integers or floating point numbers. divisor may also be nil. If divisor is nil or omitted, these functions convert number to an integer, or return it unchanged if it already is an integer. If divisor is non-nil, they divide number by divisor and convert the result to an integer. An arith-error results if divisor is 0.

3.6 Arithmetic Operations

Emacs Lisp provides the traditional four arithmetic operations: addition, subtraction, multiplication, and division. Remainder and modulus functions supplement the division functions. The functions to add or subtract 1 are provided because they are traditional in Lisp and commonly used.

All of these functions except % return a floating point value if any argument is floating.

It is important to note that in Emacs Lisp, arithmetic functions do not check for overflow. Thus (1+ 268435455) may evaluate to −268435456, depending on your hardware.

3.7 Rounding Operations

The functions ffloor, fceiling, fround, and ftruncate take a floating point argument and return a floating point result whose value is a nearby integer. ffloor returns the nearest integer below; fceiling, the nearest integer above; ftruncate, the nearest integer in the direction towards zero; fround, the nearest integer.

3.8 Bitwise Operations on Integers

In a computer, an integer is represented as a binary number, a sequence of bits (digits which are either zero or one). A bitwise operation acts on the individual bits of such a sequence. For example, shifting moves the whole sequence left or right one or more places, reproducing the same pattern “moved over.”

The bitwise operations in Emacs Lisp apply only to integers.

3.9 Standard Mathematical Functions

These mathematical functions allow integers as well as floating point numbers as arguments.

3.10 Random Numbers

A deterministic computer program cannot generate true random numbers. For most purposes, pseudo-random numbers suffice. A series of pseudo-random numbers is generated in a deterministic fashion. The numbers are not truly random, but they have certain properties that mimic a random series. For example, all possible values occur equally often in a pseudo-random series.

In Emacs, pseudo-random numbers are generated from a “seed” number. Starting from any given seed, the random function always generates the same sequence of numbers. Emacs always starts with the same seed value, so the sequence of values of random is actually the same in each Emacs run! For example, in one operating system, the first call to (random) after you start Emacs always returns −1457731, and the second one always returns −7692030. This repeatability is helpful for debugging.

If you want random numbers that don't always come out the same, execute (random t). This chooses a new seed based on the current time of day and on Emacs's process ID number.

4 Strings and Characters

A string in Emacs Lisp is an array that contains an ordered sequence of characters. Strings are used as names of symbols, buffers, and files; to send messages to users; to hold text being copied between buffers; and for many other purposes. Because strings are so important, Emacs Lisp has many functions expressly for manipulating them. Emacs Lisp programs use strings more often than individual characters.

See Strings of Events, for special considerations for strings of keyboard character events.

• Basics: Basic properties of strings and characters.
• Predicates for Strings: Testing whether an object is a string or char.
• Creating Strings: Functions to allocate new strings.
• Modifying Strings: Altering the contents of an existing string.
• Text Comparison: Comparing characters or strings.
• String Conversion: Converting to and from characters and strings.
• Formatting Strings: format: Emacs's analogue of printf.
• Case Conversion: Case conversion functions.
• Case Tables: Customizing case conversion.

4.1 String and Character Basics

Characters are represented in Emacs Lisp as integers; whether an integer is a character or not is determined only by how it is used. Thus, strings really contain integers. See Character Codes, for details about character representation in Emacs.

The length of a string (like any array) is fixed, and cannot be altered once the string exists. Strings in Lisp are not terminated by a distinguished character code. (By contrast, strings in C are terminated by a character with ASCII code 0.)

Since strings are arrays, and therefore sequences as well, you can operate on them with the general array and sequence functions. (See Sequences Arrays Vectors.) For example, you can access or change individual characters in a string using the functions aref and aset (see Array Functions).

There are two text representations for non-ASCII characters in Emacs strings (and in buffers): unibyte and multibyte (see Text Representations). For most Lisp programming, you don't need to be concerned with these two representations.

Sometimes key sequences are represented as unibyte strings. When a unibyte string is a key sequence, string elements in the range 128 to 255 represent meta characters (which are large integers) rather than character codes in the range 128 to 255. Strings cannot hold characters that have the hyper, super or alt modifiers; they can hold ASCII control characters, but no other control characters. They do not distinguish case in ASCII control characters. If you want to store such characters in a sequence, such as a key sequence, you must use a vector instead of a string. See Character Type, for more information about keyboard input characters.

Strings are useful for holding regular expressions. You can also match regular expressions against strings with string-match (see Regexp Search). The functions match-string (see Simple Match Data) and replace-match (see Replacing Match) are useful for decomposing and modifying strings after matching regular expressions against them.

Like a buffer, a string can contain text properties for the characters in it, as well as the characters themselves. See Text Properties. All the Lisp primitives that copy text from strings to buffers or other strings also copy the properties of the characters being copied.

See Text, for information about functions that display strings or copy them into buffers. See Character Type, and String Type, for information about the syntax of characters and strings. See Non-ASCII Characters, for functions to convert between text representations and to encode and decode character codes.

4.2 The Predicates for Strings

For more information about general sequence and array predicates, see Sequences Arrays Vectors, and Arrays.

4.3 Creating Strings

The following functions create strings, either from scratch, or by putting strings together, or by taking them apart.

4.4 Modifying Strings

The most basic way to alter the contents of an existing string is with aset (see Array Functions). (aset string idx char) stores char into string at index idx. Each character occupies one or more bytes, and if char needs a different number of bytes from the character already present at that index, aset signals an error.

A more powerful function is store-substring:

To clear out a string that contained a password, use clear-string:

4.5 Comparison of Characters and Strings

See also the function compare-buffer-substrings in Comparing Text, for a way to compare text in buffers. The function string-match, which matches a regular expression against a string, can be used for a kind of string comparison; see Regexp Search.

4.6 Conversion of Characters and Strings

This section describes functions for converting between characters, strings and integers. format (see Formatting Strings) and prin1-to-string (see Output Functions) can also convert Lisp objects into strings. read-from-string (see Input Functions) can “convert” a string representation of a Lisp object into an object. The functions string-make-multibyte and string-make-unibyte convert the text representation of a string (see Converting Representations).

See Documentation, for functions that produce textual descriptions of text characters and general input events (single-key-description and text-char-description). These are used primarily for making help messages.

Here are some other functions that can convert to or from a string:

concat

This function converts a vector or a list into a string. See Creating Strings.

vconcat

This function converts a string into a vector. See Vector Functions.

append

This function converts a string into a list. See Building Lists.

byte-to-string

This function converts a byte of character data into a unibyte string. See Converting Representations.

4.7 Formatting Strings

Formatting means constructing a string by substituting computed values at various places in a constant string. This constant string controls how the other values are printed, as well as where they appear; it is called a format string.

Formatting is often useful for computing messages to be displayed. In fact, the functions message and error provide the same formatting feature described here; they differ from format only in how they use the result of formatting.

A format specification is a sequence of characters beginning with a ‘%’. Thus, if there is a ‘%d’ in string, the format function replaces it with the printed representation of one of the values to be formatted (one of the arguments objects). For example:

     (format "The value of fill-column is %d." fill-column)
          ⇒ "The value of fill-column is 72."

Since format interprets ‘%’ characters as format specifications, you should never pass an arbitrary string as the first argument. This is particularly true when the string is generated by some Lisp code. Unless the string is known to never include any ‘%’ characters, pass "%s", described below, as the first argument, and the string as the second, like this:

       (format "%s" arbitrary-string)

If string contains more than one format specification, the format specifications correspond to successive values from objects. Thus, the first format specification in string uses the first such value, the second format specification uses the second such value, and so on. Any extra format specifications (those for which there are no corresponding values) cause an error. Any extra values to be formatted are ignored.

Certain format specifications require values of particular types. If you supply a value that doesn't fit the requirements, an error is signaled.

Here is a table of valid format specifications:

‘%s’

Replace the specification with the printed representation of the object, made without quoting (that is, using princ, not prin1—see Output Functions). Thus, strings are represented by their contents alone, with no ‘"’ characters, and symbols appear without ‘\’ characters.

If the object is a string, its text properties are copied into the output. The text properties of the ‘%s’ itself are also copied, but those of the object take priority.

‘%S’

Replace the specification with the printed representation of the object, made with quoting (that is, using prin1—see Output Functions). Thus, strings are enclosed in ‘"’ characters, and ‘\’ characters appear where necessary before special characters.

‘%o’

Replace the specification with the base-eight representation of an integer.

‘%d’

Replace the specification with the base-ten representation of an integer.

‘%x’

‘%X’

Replace the specification with the base-sixteen representation of an integer. ‘%x’ uses lower case and ‘%X’ uses upper case.

‘%c’

Replace the specification with the character which is the value given.

‘%e’

Replace the specification with the exponential notation for a floating point number.

‘%f’

Replace the specification with the decimal-point notation for a floating point number.

‘%g’

Replace the specification with notation for a floating point number, using either exponential notation or decimal-point notation, whichever is shorter.

‘%%’

Replace the specification with a single ‘%’. This format specification is unusual in that it does not use a value. For example, (format "%% %d" 30) returns "% 30".

Any other format character results in an ‘Invalid format operation’ error.

Here are several examples:

     (format "The name of this buffer is %s." (buffer-name))
          ⇒ "The name of this buffer is strings.texi."
     
     (format "The buffer object prints as %s." (current-buffer))
          ⇒ "The buffer object prints as strings.texi."
     
     (format "The octal value of %d is %o,
              and the hex value is %x." 18 18 18)
          ⇒ "The octal value of 18 is 22,
              and the hex value is 12."

A specification can have a width, which is a decimal number between the ‘%’ and the specification character. If the printed representation of the object contains fewer characters than this width, format extends it with padding. The width specifier is ignored for the ‘%%’ specification. Any padding introduced by the width specifier normally consists of spaces inserted on the left:

     (format "%5d is padded on the left with spaces" 123)
          ⇒ "  123 is padded on the left with spaces"

If the width is too small, format does not truncate the object's printed representation. Thus, you can use a width to specify a minimum spacing between columns with no risk of losing information. In the following three examples, ‘%7s’ specifies a minimum width of 7. In the first case, the string inserted in place of ‘%7s’ has only 3 letters, and needs 4 blank spaces as padding. In the second case, the string "specification" is 13 letters wide but is not truncated.

     (format "The word `%7s' actually has %d letters in it."
             "foo" (length "foo"))
          ⇒ "The word `    foo' actually has 3 letters in it."
     (format "The word `%7s' actually has %d letters in it."
             "specification" (length "specification"))
          ⇒ "The word `specification' actually has 13 letters in it."

Immediately after the ‘%’ and before the optional width specifier, you can also put certain flag characters.

The flag ‘+’ inserts a plus sign before a positive number, so that it always has a sign. A space character as flag inserts a space before a positive number. (Otherwise, positive numbers start with the first digit.) These flags are useful for ensuring that positive numbers and negative numbers use the same number of columns. They are ignored except for ‘%d’, ‘%e’, ‘%f’, ‘%g’, and if both flags are used, ‘+’ takes precedence.

The flag ‘#’ specifies an “alternate form” which depends on the format in use. For ‘%o’, it ensures that the result begins with a ‘0’. For ‘%x’ and ‘%X’, it prefixes the result with ‘0x’ or ‘0X’. For ‘%e’, ‘%f’, and ‘%g’, the ‘#’ flag means include a decimal point even if the precision is zero.

The flag ‘-’ causes the padding inserted by the width specifier, if any, to be inserted on the right rather than the left. The flag ‘0’ ensures that the padding consists of ‘0’ characters instead of spaces, inserted on the left. These flags are ignored for specification characters for which they do not make sense: ‘%s’, ‘%S’ and ‘%c’ accept the ‘0’ flag, but still pad with spaces on the left. If both ‘-’ and ‘0’ are present and valid, ‘-’ takes precedence.

     (format "%06d is padded on the left with zeros" 123)
          ⇒ "000123 is padded on the left with zeros"
     
     (format "%-6d is padded on the right" 123)
          ⇒ "123    is padded on the right"
     
     (format "The word `%-7s' actually has %d letters in it."
             "foo" (length "foo"))
          ⇒ "The word `foo    ' actually has 3 letters in it."

All the specification characters allow an optional precision before the character (after the width, if present). The precision is a decimal-point ‘.’ followed by a digit-string. For the floating-point specifications (‘%e’, ‘%f’, ‘%g’), the precision specifies how many decimal places to show; if zero, the decimal-point itself is also omitted. For ‘%s’ and ‘%S’, the precision truncates the string to the given width, so ‘%.3s’ shows only the first three characters of the representation for object. Precision has no effect for other specification characters.

4.8 Case Conversion in Lisp

The character case functions change the case of single characters or of the contents of strings. The functions normally convert only alphabetic characters (the letters ‘A’ through ‘Z’ and ‘a’ through ‘z’, as well as non-ASCII letters); other characters are not altered. You can specify a different case conversion mapping by specifying a case table (see Case Tables).

These functions do not modify the strings that are passed to them as arguments.

The examples below use the characters ‘X’ and ‘x’ which have ASCII codes 88 and 120 respectively.

See Text Comparison, for functions that compare strings; some of them ignore case differences, or can optionally ignore case differences.

4.9 The Case Table

You can customize case conversion by installing a special case table. A case table specifies the mapping between upper case and lower case letters. It affects both the case conversion functions for Lisp objects (see the previous section) and those that apply to text in the buffer (see Case Changes). Each buffer has a case table; there is also a standard case table which is used to initialize the case table of new buffers.

A case table is a char-table (see Char-Tables) whose subtype is case-table. This char-table maps each character into the corresponding lower case character. It has three extra slots, which hold related tables:

upcase

The upcase table maps each character into the corresponding upper case character.

canonicalize

The canonicalize table maps all of a set of case-related characters into a particular member of that set.

equivalences

The equivalences table maps each one of a set of case-related characters into the next character in that set.

In simple cases, all you need to specify is the mapping to lower-case; the three related tables will be calculated automatically from that one.

For some languages, upper and lower case letters are not in one-to-one correspondence. There may be two different lower case letters with the same upper case equivalent. In these cases, you need to specify the maps for both lower case and upper case.

The extra table canonicalize maps each character to a canonical equivalent; any two characters that are related by case-conversion have the same canonical equivalent character. For example, since ‘a’ and ‘A’ are related by case-conversion, they should have the same canonical equivalent character (which should be either ‘a’ for both of them, or ‘A’ for both of them).

The extra table equivalences is a map that cyclically permutes each equivalence class (of characters with the same canonical equivalent). (For ordinary ASCII, this would map ‘a’ into ‘A’ and ‘A’ into ‘a’, and likewise for each set of equivalent characters.)

When constructing a case table, you can provide nil for canonicalize; then Emacs fills in this slot from the lower case and upper case mappings. You can also provide nil for equivalences; then Emacs fills in this slot from canonicalize. In a case table that is actually in use, those components are non-nil. Do not try to specify equivalences without also specifying canonicalize.

Here are the functions for working with case tables:

Some language environments modify the case conversions of ASCII characters; for example, in the Turkish language environment, the ASCII character ‘I’ is downcased into a Turkish “dotless i”. This can interfere with code that requires ordinary ASCII case conversion, such as implementations of ASCII-based network protocols. In that case, use the with-case-table macro with the variable ascii-case-table, which stores the unmodified case table for the ASCII character set.

The following three functions are convenient subroutines for packages that define non-ASCII character sets. They modify the specified case table case-table; they also modify the standard syntax table. See Syntax Tables. Normally you would use these functions to change the standard case table.

5 Lists

A list represents a sequence of zero or more elements (which may be any Lisp objects). The important difference between lists and vectors is that two or more lists can share part of their structure; in addition, you can insert or delete elements in a list without copying the whole list.

• Cons Cells: How lists are made out of cons cells.
• List-related Predicates: Is this object a list? Comparing two lists.
• List Elements: Extracting the pieces of a list.
• Building Lists: Creating list structure.
• List Variables: Modifying lists stored in variables.
• Modifying Lists: Storing new pieces into an existing list.
• Sets And Lists: A list can represent a finite mathematical set.
• Association Lists: A list can represent a finite relation or mapping.
• Rings: Managing a fixed-size ring of objects.

5.1 Lists and Cons Cells

Lists in Lisp are not a primitive data type; they are built up from cons cells. A cons cell is a data object that represents an ordered pair. That is, it has two slots, and each slot holds, or refers to, some Lisp object. One slot is known as the car, and the other is known as the cdr. (These names are traditional; see Cons Cell Type.) cdr is pronounced “could-er.”

We say that “the car of this cons cell is” whatever object its car slot currently holds, and likewise for the cdr.

A list is a series of cons cells “chained together,” so that each cell refers to the next one. There is one cons cell for each element of the list. By convention, the cars of the cons cells hold the elements of the list, and the cdrs are used to chain the list: the cdr slot of each cons cell refers to the following cons cell. The cdr of the last cons cell is nil. This asymmetry between the car and the cdr is entirely a matter of convention; at the level of cons cells, the car and cdr slots have the same characteristics.

Since nil is the conventional value to put in the cdr of the last cons cell in the list, we call that case a true list.

In Lisp, we consider the symbol nil a list as well as a symbol; it is the list with no elements. For convenience, the symbol nil is considered to have nil as its cdr (and also as its car). Therefore, the cdr of a true list is always a true list.

If the cdr of a list's last cons cell is some other value, neither nil nor another cons cell, we call the structure a dotted list, since its printed representation would use ‘.’. There is one other possibility: some cons cell's cdr could point to one of the previous cons cells in the list. We call that structure a circular list.

For some purposes, it does not matter whether a list is true, circular or dotted. If the program doesn't look far enough down the list to see the cdr of the final cons cell, it won't care. However, some functions that operate on lists demand true lists and signal errors if given a dotted list. Most functions that try to find the end of a list enter infinite loops if given a circular list.

Because most cons cells are used as part of lists, the phrase list structure has come to mean any structure made out of cons cells.

The cdr of any nonempty true list l is a list containing all the elements of l except the first.

See Cons Cell Type, for the read and print syntax of cons cells and lists, and for “box and arrow” illustrations of lists.

5.2 Predicates on Lists

The following predicates test whether a Lisp object is an atom, whether it is a cons cell or is a list, or whether it is the distinguished object nil. (Many of these predicates can be defined in terms of the others, but they are used so often that it is worth having all of them.)

5.3 Accessing Elements of Lists

The most common way to compute the length of a list, when you are not worried that it may be circular, is with length. See Sequence Functions.

5.4 Building Cons Cells and Lists

Many functions build lists, as lists reside at the very heart of Lisp. cons is the fundamental list-building function; however, it is interesting to note that list is used more times in the source code for Emacs than cons.

Here is an example of using append:

     (setq trees '(pine oak))
          ⇒ (pine oak)
     (setq more-trees (append '(maple birch) trees))
          ⇒ (maple birch pine oak)
     
     trees
          ⇒ (pine oak)
     more-trees
          ⇒ (maple birch pine oak)
     (eq trees (cdr (cdr more-trees)))
          ⇒ t

You can see how append works by looking at a box diagram. The variable trees is set to the list (pine oak) and then the variable more-trees is set to the list (maple birch pine oak). However, the variable trees continues to refer to the original list:

     more-trees                trees
     |                           |
     |     --- ---      --- ---   -> --- ---      --- ---
      --> |   |   |--> |   |   |--> |   |   |--> |   |   |--> nil
           --- ---      --- ---      --- ---      --- ---
            |            |            |            |
            |            |            |            |
             --> maple    -->birch     --> pine     --> oak

An empty sequence contributes nothing to the value returned by append. As a consequence of this, a final nil argument forces a copy of the previous argument:

     trees
          ⇒ (pine oak)
     (setq wood (append trees nil))
          ⇒ (pine oak)
     wood
          ⇒ (pine oak)
     (eq wood trees)
          ⇒ nil

This once was the usual way to copy a list, before the function copy-sequence was invented. See Sequences Arrays Vectors.

Here we show the use of vectors and strings as arguments to append:

     (append [a b] "cd" nil)
          ⇒ (a b 99 100)

With the help of apply (see Calling Functions), we can append all the lists in a list of lists:

     (apply 'append '((a b c) nil (x y z) nil))
          ⇒ (a b c x y z)

If no sequences are given, nil is returned:

     (append)
          ⇒ nil

Here are some examples where the final argument is not a list:

     (append '(x y) 'z)
          ⇒ (x y . z)
     (append '(x y) [z])
          ⇒ (x y . [z])

The second example shows that when the final argument is a sequence but not a list, the sequence's elements do not become elements of the resulting list. Instead, the sequence becomes the final cdr, like any other non-list final argument.

5.5 Modifying List Variables

These functions, and one macro, provide convenient ways to modify a list which is stored in a variable.

Two functions modify lists that are the values of variables.

Here's a scenario showing how to use add-to-list:

     (setq foo '(a b))
          ⇒ (a b)
     
     (add-to-list 'foo 'c)     ;; Add c.
          ⇒ (c a b)
     
     (add-to-list 'foo 'b)     ;; No effect.
          ⇒ (c a b)
     
     foo                       ;; foo was changed.
          ⇒ (c a b)

An equivalent expression for (add-to-list 'var value) is this:

     (or (member value var)
         (setq var (cons value var)))

Here's a scenario showing how to use add-to-ordered-list:

     (setq foo '())
          ⇒ nil
     
     (add-to-ordered-list 'foo 'a 1)     ;; Add a.
          ⇒ (a)
     
     (add-to-ordered-list 'foo 'c 3)     ;; Add c.
          ⇒ (a c)
     
     (add-to-ordered-list 'foo 'b 2)     ;; Add b.
          ⇒ (a b c)
     
     (add-to-ordered-list 'foo 'b 4)     ;; Move b.
          ⇒ (a c b)
     
     (add-to-ordered-list 'foo 'd)       ;; Append d.
          ⇒ (a c b d)
     
     (add-to-ordered-list 'foo 'e)       ;; Add e.
          ⇒ (a c b e d)
     
     foo                       ;; foo was changed.
          ⇒ (a c b e d)

5.6 Modifying Existing List Structure

You can modify the car and cdr contents of a cons cell with the primitives setcar and setcdr. We call these “destructive” operations because they change existing list structure.

Common Lisp note: Common Lisp uses functions rplaca and rplacd to alter list structure; they change structure the same way as setcar and setcdr, but the Common Lisp functions return the cons cell while setcar and setcdr return the new car or cdr.

• Setcar: Replacing an element in a list.
• Setcdr: Replacing part of the list backbone. This can be used to remove or add elements.
• Rearrangement: Reordering the elements in a list; combining lists.

5.6.1 Altering List Elements with setcar

Changing the car of a cons cell is done with setcar. When used on a list, setcar replaces one element of a list with a different element.

When a cons cell is part of the shared structure of several lists, storing a new car into the cons changes one element of each of these lists. Here is an example:

     ;; Create two lists that are partly shared.
     (setq x1 '(a b c))
          ⇒ (a b c)
     (setq x2 (cons 'z (cdr x1)))
          ⇒ (z b c)
     
     ;; Replace the car of a shared link.
     (setcar (cdr x1) 'foo)
          ⇒ foo
     x1                           ; Both lists are changed.
          ⇒ (a foo c)
     x2
          ⇒ (z foo c)
     
     ;; Replace the car of a link that is not shared.
     (setcar x1 'baz)
          ⇒ baz
     x1                           ; Only one list is changed.
          ⇒ (baz foo c)
     x2
          ⇒ (z foo c)

Here is a graphical depiction of the shared structure of the two lists in the variables x1 and x2, showing why replacing b changes them both:

             --- ---        --- ---      --- ---
     x1---> |   |   |----> |   |   |--> |   |   |--> nil
             --- ---        --- ---      --- ---
              |        -->   |            |
              |       |      |            |
               --> a  |       --> b        --> c
                      |
            --- ---   |
     x2--> |   |   |--
            --- ---
             |
             |
              --> z

Here is an alternative form of box diagram, showing the same relationship:

     x1:
      --------------       --------------       --------------
     | car   | cdr  |     | car   | cdr  |     | car   | cdr  |
     |   a   |   o------->|   b   |   o------->|   c   |  nil |
     |       |      |  -->|       |      |     |       |      |
      --------------  |    --------------       --------------
                      |
     x2:              |
      --------------  |
     | car   | cdr  | |
     |   z   |   o----
     |       |      |
      --------------

5.6.2 Altering the CDR of a List

The lowest-level primitive for modifying a cdr is setcdr:

Here is an example of replacing the cdr of a list with a different list. All but the first element of the list are removed in favor of a different sequence of elements. The first element is unchanged, because it resides in the car of the list, and is not reached via the cdr.

     (setq x '(1 2 3))
          ⇒ (1 2 3)
     (setcdr x '(4))
          ⇒ (4)
     x
          ⇒ (1 4)

You can delete elements from the middle of a list by altering the cdrs of the cons cells in the list. For example, here we delete the second element, b, from the list (a b c), by changing the cdr of the first cons cell:

     (setq x1 '(a b c))
          ⇒ (a b c)
     (setcdr x1 (cdr (cdr x1)))
          ⇒ (c)
     x1
          ⇒ (a c)

Here is the result in box notation:

                        --------------------
                       |                    |
      --------------   |   --------------   |    --------------
     | car   | cdr  |  |  | car   | cdr  |   -->| car   | cdr  |
     |   a   |   o-----   |   b   |   o-------->|   c   |  nil |
     |       |      |     |       |      |      |       |      |
      --------------       --------------        --------------

The second cons cell, which previously held the element b, still exists and its car is still b, but it no longer forms part of this list.

It is equally easy to insert a new element by changing cdrs:

     (setq x1 '(a b c))
          ⇒ (a b c)
     (setcdr x1 (cons 'd (cdr x1)))
          ⇒ (d b c)
     x1
          ⇒ (a d b c)

Here is this result in box notation:

      --------------        -------------       -------------
     | car  | cdr   |      | car  | cdr  |     | car  | cdr  |
     |   a  |   o   |   -->|   b  |   o------->|   c  |  nil |
     |      |   |   |  |   |      |      |     |      |      |
      --------- | --   |    -------------       -------------
                |      |
          -----         --------
         |                      |
         |    ---------------   |
         |   | car   | cdr   |  |
          -->|   d   |   o------
             |       |       |
              ---------------

5.6.3 Functions that Rearrange Lists

Here are some functions that rearrange lists “destructively” by modifying the cdrs of their component cons cells. We call these functions “destructive” because they chew up the original lists passed to them as arguments, relinking their cons cells to form a new list that is the returned value.

See delq, in Sets And Lists, for another function that modifies cons cells.

5.7 Using Lists as Sets

A list can represent an unordered mathematical set—simply consider a value an element of a set if it appears in the list, and ignore the order of the list. To form the union of two sets, use append (as long as you don't mind having duplicate elements). You can remove equal duplicates using delete-dups. Other useful functions for sets include memq and delq, and their equal versions, member and delete.

Common Lisp note: Common Lisp has functions union (which avoids duplicate elements) and intersection for set operations, but GNU Emacs Lisp does not have them. You can write them in Lisp if you wish.

When delq deletes elements from the front of the list, it does so simply by advancing down the list and returning a sublist that starts after those elements:

     (delq 'a '(a b c)) == (cdr '(a b c))

When an element to be deleted appears in the middle of the list, removing it involves changing the cdrs (see Setcdr).

     (setq sample-list '(a b c (4)))
          ⇒ (a b c (4))
     (delq 'a sample-list)
          ⇒ (b c (4))
     sample-list
          ⇒ (a b c (4))
     (delq 'c sample-list)
          ⇒ (a b (4))
     sample-list
          ⇒ (a b (4))

Note that (delq 'c sample-list) modifies sample-list to splice out the third element, but (delq 'a sample-list) does not splice anything—it just returns a shorter list. Don't assume that a variable which formerly held the argument list now has fewer elements, or that it still holds the original list! Instead, save the result of delq and use that. Most often we store the result back into the variable that held the original list:

     (setq flowers (delq 'rose flowers))

In the following example, the (4) that delq attempts to match and the (4) in the sample-list are not eq:

     (delq '(4) sample-list)
          ⇒ (a c (4))
     
     If you want to delete elements that are equal to a given value,
     use delete (see below).

The following three functions are like memq, delq and remq, but use equal rather than eq to compare elements. See Equality Predicates.

Common Lisp note: The functions member, delete and remove in GNU Emacs Lisp are derived from Maclisp, not Common Lisp. The Common Lisp versions do not use equal to compare elements.

See also the function add-to-list, in List Variables, for a way to add an element to a list stored in a variable and used as a set.

5.8 Association Lists

An association list, or alist for short, records a mapping from keys to values. It is a list of cons cells called associations: the car of each cons cell is the key, and the cdr is the associated value.²

Here is an example of an alist. The key pine is associated with the value cones; the key oak is associated with acorns; and the key maple is associated with seeds.

     ((pine . cones)
      (oak . acorns)
      (maple . seeds))

Both the values and the keys in an alist may be any Lisp objects. For example, in the following alist, the symbol a is associated with the number 1, and the string "b" is associated with the list (2 3), which is the cdr of the alist element:

     ((a . 1) ("b" 2 3))

Sometimes it is better to design an alist to store the associated value in the car of the cdr of the element. Here is an example of such an alist:

     ((rose red) (lily white) (buttercup yellow))

Here we regard red as the value associated with rose. One advantage of this kind of alist is that you can store other related information—even a list of other items—in the cdr of the cdr. One disadvantage is that you cannot use rassq (see below) to find the element containing a given value. When neither of these considerations is important, the choice is a matter of taste, as long as you are consistent about it for any given alist.

The same alist shown above could be regarded as having the associated value in the cdr of the element; the value associated with rose would be the list (red).

Association lists are often used to record information that you might otherwise keep on a stack, since new associations may be added easily to the front of the list. When searching an association list for an association with a given key, the first one found is returned, if there is more than one.

In Emacs Lisp, it is not an error if an element of an association list is not a cons cell. The alist search functions simply ignore such elements. Many other versions of Lisp signal errors in such cases.

Note that property lists are similar to association lists in several respects. A property list behaves like an association list in which each key can occur only once. See Property Lists, for a comparison of property lists and association lists.

The function assoc-string is much like assoc except that it ignores certain differences between strings. See Text Comparison.

5.9 Managing a Fixed-Size Ring of Objects

This section describes functions for operating on rings. A ring is a fixed-size data structure that supports insertion, deletion, rotation, and modulo-indexed reference and traversal.

The newest element in the ring always has index 0. Higher indices correspond to older elements. Indices are computed modulo the ring length. Index −1 corresponds to the oldest element, −2 to the next-oldest, and so forth.

If you are careful not to exceed the ring size, you can use the ring as a first-in-first-out queue. For example:

     (let ((fifo (make-ring 5)))
       (mapc (lambda (obj) (ring-insert fifo obj))
             '(0 one "two"))
       (list (ring-remove fifo) t
             (ring-remove fifo) t
             (ring-remove fifo)))
          ⇒ (0 t one t "two")

6 Sequences, Arrays, and Vectors

Recall that the sequence type is the union of two other Lisp types: lists and arrays. In other words, any list is a sequence, and any array is a sequence. The common property that all sequences have is that each is an ordered collection of elements.

An array is a fixed-length object with a slot for each of its elements. All the elements are accessible in constant time. The four types of arrays are strings, vectors, char-tables and bool-vectors.

A list is a sequence of elements, but it is not a single primitive object; it is made of cons cells, one cell per element. Finding the nth element requires looking through n cons cells, so elements farther from the beginning of the list take longer to access. But it is possible to add elements to the list, or remove elements.

The following diagram shows the relationship between these types:

               _____________________________________________
              |                                             |
              |          Sequence                           |
              |  ______   ________________________________  |
              | |      | |                                | |
              | | List | |             Array              | |
              | |      | |    ________       ________     | |
              | |______| |   |        |     |        |    | |
              |          |   | Vector |     | String |    | |
              |          |   |________|     |________|    | |
              |          |  ____________   _____________  | |
              |          | |            | |             | | |
              |          | | Char-table | | Bool-vector | | |
              |          | |____________| |_____________| | |
              |          |________________________________| |
              |_____________________________________________|

• Sequence Functions: Functions that accept any kind of sequence.
• Arrays: Characteristics of arrays in Emacs Lisp.
• Array Functions: Functions specifically for arrays.
• Vectors: Special characteristics of Emacs Lisp vectors.
• Vector Functions: Functions specifically for vectors.
• Char-Tables: How to work with char-tables.
• Bool-Vectors: How to work with bool-vectors.

6.1 Sequences

In Emacs Lisp, a sequence is either a list or an array. The common property of all sequences is that they are ordered collections of elements. This section describes functions that accept any kind of sequence.

See also string-bytes, in Text Representations.

6.2 Arrays

An array object has slots that hold a number of other Lisp objects, called the elements of the array. Any element of an array may be accessed in constant time. In contrast, the time to access an element of a list is proportional to the position of that element in the list.

Emacs defines four types of array, all one-dimensional: strings (see String Type), vectors (see Vector Type), bool-vectors (see Bool-Vector Type), and char-tables (see Char-Table Type). Vectors and char-tables can hold elements of any type, but strings can only hold characters, and bool-vectors can only hold t and nil.

All four kinds of array share these characteristics:

• The first element of an array has index zero, the second element has index 1, and so on. This is called zero-origin indexing. For example, an array of four elements has indices 0, 1, 2, and 3.
• The length of the array is fixed once you create it; you cannot change the length of an existing array.
• For purposes of evaluation, the array is a constant—in other words, it evaluates to itself.
• The elements of an array may be referenced or changed with the functions aref and aset, respectively (see Array Functions).

When you create an array, other than a char-table, you must specify its length. You cannot specify the length of a char-table, because that is determined by the range of character codes.

In principle, if you want an array of text characters, you could use either a string or a vector. In practice, we always choose strings for such applications, for four reasons:

• They occupy one-fourth the space of a vector of the same elements.
• Strings are printed in a way that shows the contents more clearly as text.
• Strings can hold text properties. See Text Properties.
• Many of the specialized editing and I/O facilities of Emacs accept only strings. For example, you cannot insert a vector of characters into a buffer the way you can insert a string. See Strings and Characters.

By contrast, for an array of keyboard input characters (such as a key sequence), a vector may be necessary, because many keyboard input characters are outside the range that will fit in a string. See Key Sequence Input.

6.3 Functions that Operate on Arrays

In this section, we describe the functions that accept all types of arrays.

The general sequence functions copy-sequence and length are often useful for objects known to be arrays. See Sequence Functions.

6.4 Vectors

A vector is a general-purpose array whose elements can be any Lisp objects. (By contrast, the elements of a string can only be characters. See Strings and Characters.) Vectors are used in Emacs for many purposes: as key sequences (see Key Sequences), as symbol-lookup tables (see Creating Symbols), as part of the representation of a byte-compiled function (see Byte Compilation), and more.

In Emacs Lisp, the indices of the elements of a vector start from zero and count up from there.

Vectors are printed with square brackets surrounding the elements. Thus, a vector whose elements are the symbols a, b and a is printed as [a b a]. You can write vectors in the same way in Lisp input.

A vector, like a string or a number, is considered a constant for evaluation: the result of evaluating it is the same vector. This does not evaluate or even examine the elements of the vector. See Self-Evaluating Forms.

Here are examples illustrating these principles:

     (setq avector [1 two '(three) "four" [five]])
          ⇒ [1 two (quote (three)) "four" [five]]
     (eval avector)
          ⇒ [1 two (quote (three)) "four" [five]]
     (eq avector (eval avector))
          ⇒ t

6.5 Functions for Vectors

Here are some functions that relate to vectors:

The append function also provides a way to convert a vector into a list with the same elements:

     (setq avector [1 two (quote (three)) "four" [five]])
          ⇒ [1 two (quote (three)) "four" [five]]
     (append avector nil)
          ⇒ (1 two (quote (three)) "four" [five])

6.6 Char-Tables

A char-table is much like a vector, except that it is indexed by character codes. Any valid character code, without modifiers, can be used as an index in a char-table. You can access a char-table's elements with aref and aset, as with any array. In addition, a char-table can have extra slots to hold additional data not associated with particular character codes. Like vectors, char-tables are constants when evaluated, and can hold elements of any type.

Each char-table has a subtype, a symbol, which serves two purposes:

• The subtype provides an easy way to tell what the char-table is for. For instance, display tables are char-tables with display-table as the subtype, and syntax tables are char-tables with syntax-table as the subtype. The subtype can be queried using the function char-table-subtype, described below.
• The subtype controls the number of extra slots in the char-table. This number is specified by the subtype's char-table-extra-slots symbol property, which should be an integer between 0 and 10. If the subtype has no such symbol property, the char-table has no extra slots. See Property Lists, for information about symbol properties.

A char-table can have a parent, which is another char-table. If it does, then whenever the char-table specifies nil for a particular character c, it inherits the value specified in the parent. In other words, (aref char-table c) returns the value from the parent of char-table if char-table itself specifies nil.

A char-table can also have a default value. If so, then (aref char-table c) returns the default value whenever the char-table does not specify any other non-nil value.

There is no special function to access default values in a char-table. To do that, use char-table-range (see below).

A char-table can specify an element value for a single character code; it can also specify a value for an entire character set.

6.7 Bool-vectors

A bool-vector is much like a vector, except that it stores only the values t and nil. If you try to store any non-nil value into an element of the bool-vector, the effect is to store t there. As with all arrays, bool-vector indices start from 0, and the length cannot be changed once the bool-vector is created. Bool-vectors are constants when evaluated.

There are two special functions for working with bool-vectors; aside from that, you manipulate them with same functions used for other kinds of arrays.

Here is an example of creating, examining, and updating a bool-vector. Note that the printed form represents up to 8 boolean values as a single character.

     (setq bv (make-bool-vector 5 t))
          ⇒ #&5"^_"
     (aref bv 1)
          ⇒ t
     (aset bv 3 nil)
          ⇒ nil
     bv
          ⇒ #&5"^W"

These results make sense because the binary codes for control-_ and control-W are 11111 and 10111, respectively.

7 Hash Tables

A hash table is a very fast kind of lookup table, somewhat like an alist (see Association Lists) in that it maps keys to corresponding values. It differs from an alist in these ways:

• Lookup in a hash table is extremely fast for large tables—in fact, the time required is essentially independent of how many elements are stored in the table. For smaller tables (a few tens of elements) alists may still be faster because hash tables have a more-or-less constant overhead.
• The correspondences in a hash table are in no particular order.
• There is no way to share structure between two hash tables, the way two alists can share a common tail.

Emacs Lisp provides a general-purpose hash table data type, along with a series of functions for operating on them. Hash tables have a special printed representation, which consists of ‘#s’ followed by a list specifying the hash table properties and contents. See Creating Hash. (Note that the term “hash notation”, which refers to the initial ‘#’ character used in the printed representations of objects with no read representation, has nothing to do with the term “hash table”. See Printed Representation.)

Obarrays are also a kind of hash table, but they are a different type of object and are used only for recording interned symbols (see Creating Symbols).

• Creating Hash: Functions to create hash tables.
• Hash Access: Reading and writing the hash table contents.
• Defining Hash: Defining new comparison methods.
• Other Hash: Miscellaneous.

7.1 Creating Hash Tables

The principal function for creating a hash table is make-hash-table.

You can also create a new hash table using the printed representation for hash tables. The Lisp reader can read this printed representation, provided each element in the specified hash table has a valid read syntax (see Printed Representation). For instance, the following specifies a new hash table containing the keys key1 and key2 (both symbols) associated with val1 (a symbol) and 300 (a number) respectively.

     #s(hash-table size 30 data (key1 val1 key2 300))

The printed representation for a hash table consists of ‘#s’ followed by a list beginning with ‘hash-table’. The rest of the list should consist of zero or more property-value pairs specifying the hash table's properties and initial contents. The properties and values are read literally. Valid property names are size, test, weakness, rehash-size, rehash-threshold, and data. The data property should be a list of key-value pairs for the initial contents; the other properties have the same meanings as the matching make-hash-table keywords (:size, :test, etc.), described above.

Note that you cannot specify a hash table whose initial contents include objects that have no read syntax, such as buffers and frames. Such objects may be added to the hash table after it is created.

7.2 Hash Table Access

This section describes the functions for accessing and storing associations in a hash table. In general, any Lisp object can be used as a hash key, unless the comparison method imposes limits. Any Lisp object can also be used as the value.

7.3 Defining Hash Comparisons

You can define new methods of key lookup by means of define-hash-table-test. In order to use this feature, you need to understand how hash tables work, and what a hash code means.

You can think of a hash table conceptually as a large array of many slots, each capable of holding one association. To look up a key, gethash first computes an integer, the hash code, from the key. It reduces this integer modulo the length of the array, to produce an index in the array. Then it looks in that slot, and if necessary in other nearby slots, to see if it has found the key being sought.

Thus, to define a new method of key lookup, you need to specify both a function to compute the hash code from a key, and a function to compare two keys directly.

This example creates a hash table whose keys are strings that are compared case-insensitively.

     (defun case-fold-string= (a b)
       (compare-strings a nil nil b nil nil t))
     (defun case-fold-string-hash (a)
       (sxhash (upcase a)))
     
     (define-hash-table-test 'case-fold
       'case-fold-string= 'case-fold-string-hash)
     
     (make-hash-table :test 'case-fold)

Here is how you could define a hash table test equivalent to the predefined test value equal. The keys can be any Lisp object, and equal-looking objects are considered the same key.

     (define-hash-table-test 'contents-hash 'equal 'sxhash)
     
     (make-hash-table :test 'contents-hash)

7.4 Other Hash Table Functions

Here are some other functions for working with hash tables.

8 Symbols

A symbol is an object with a unique name. This chapter describes symbols, their components, their property lists, and how they are created and interned. Separate chapters describe the use of symbols as variables and as function names; see Variables, and Functions. For the precise read syntax for symbols, see Symbol Type.

You can test whether an arbitrary Lisp object is a symbol with symbolp:

• Symbol Components: Symbols have names, values, function definitions and property lists.
• Definitions: A definition says how a symbol will be used.
• Creating Symbols: How symbols are kept unique.
• Property Lists: Each symbol has a property list for recording miscellaneous information.

8.1 Symbol Components

Each symbol has four components (or “cells”), each of which references another object:

Print name

The print name cell holds a string that names the symbol for reading and printing. See symbol-name in Creating Symbols.

Value

The value cell holds the current value of the symbol as a variable. When a symbol is used as a form, the value of the form is the contents of the symbol's value cell. See symbol-value in Accessing Variables.

Function

The function cell holds the function definition of the symbol. When a symbol is used as a function, its function definition is used in its place. This cell is also used to make a symbol stand for a keymap or a keyboard macro, for editor command execution. Because each symbol has separate value and function cells, variables names and function names do not conflict. See symbol-function in Function Cells.

Property list

The property list cell holds the property list of the symbol. See symbol-plist in Property Lists.

The print name cell always holds a string, and cannot be changed. The other three cells can be set individually to any specified Lisp object.

The print name cell holds the string that is the name of the symbol. Since symbols are represented textually by their names, it is important not to have two symbols with the same name. The Lisp reader ensures this: every time it reads a symbol, it looks for an existing symbol with the specified name before it creates a new one. (In GNU Emacs Lisp, this lookup uses a hashing algorithm and an obarray; see Creating Symbols.)

The value cell holds the symbol's value as a variable (see Variables). That is what you get if you evaluate the symbol as a Lisp expression (see Evaluation). Any Lisp object is a legitimate value. Certain symbols have values that cannot be changed; these include nil and t, and any symbol whose name starts with ‘:’ (those are called keywords). See Constant Variables.

We often refer to “the function foo” when we really mean the function stored in the function cell of the symbol foo. We make the distinction explicit only when necessary. In normal usage, the function cell usually contains a function (see Functions) or a macro (see Macros), as that is what the Lisp interpreter expects to see there (see Evaluation). Keyboard macros (see Keyboard Macros), keymaps (see Keymaps) and autoload objects (see Autoloading) are also sometimes stored in the function cells of symbols.

The property list cell normally should hold a correctly formatted property list (see Property Lists), as a number of functions expect to see a property list there.

The function cell or the value cell may be void, which means that the cell does not reference any object. (This is not the same thing as holding the symbol void, nor the same as holding the symbol nil.) Examining a function or value cell that is void results in an error, such as ‘Symbol's value as variable is void’.

The four functions symbol-name, symbol-value, symbol-plist, and symbol-function return the contents of the four cells of a symbol. Here as an example we show the contents of the four cells of the symbol buffer-file-name:

     (symbol-name 'buffer-file-name)
          ⇒ "buffer-file-name"
     (symbol-value 'buffer-file-name)
          ⇒ "/gnu/elisp/symbols.texi"
     (symbol-function 'buffer-file-name)
          ⇒ #<subr buffer-file-name>
     (symbol-plist 'buffer-file-name)
          ⇒ (variable-documentation 29529)

Because this symbol is the variable which holds the name of the file being visited in the current buffer, the value cell contents we see are the name of the source file of this chapter of the Emacs Lisp Manual. The property list cell contains the list (variable-documentation 29529) which tells the documentation functions where to find the documentation string for the variable buffer-file-name in the DOC-version file. (29529 is the offset from the beginning of the DOC-version file to where that documentation string begins—see Documentation Basics.) The function cell contains the function for returning the name of the file. buffer-file-name names a primitive function, which has no read syntax and prints in hash notation (see Primitive Function Type). A symbol naming a function written in Lisp would have a lambda expression (or a byte-code object) in this cell.

8.2 Defining Symbols

A definition in Lisp is a special form that announces your intention to use a certain symbol in a particular way. In Emacs Lisp, you can define a symbol as a variable, or define it as a function (or macro), or both independently.

A definition construct typically specifies a value or meaning for the symbol for one kind of use, plus documentation for its meaning when used in this way. Thus, when you define a symbol as a variable, you can supply an initial value for the variable, plus documentation for the variable.

defvar and defconst are special forms that define a symbol as a global variable. They are documented in detail in Defining Variables. For defining user option variables that can be customized, use defcustom (see Customization).

defun defines a symbol as a function, creating a lambda expression and storing it in the function cell of the symbol. This lambda expression thus becomes the function definition of the symbol. (The term “function definition,” meaning the contents of the function cell, is derived from the idea that defun gives the symbol its definition as a function.) defsubst and defalias are two other ways of defining a function. See Functions.

defmacro defines a symbol as a macro. It creates a macro object and stores it in the function cell of the symbol. Note that a given symbol can be a macro or a function, but not both at once, because both macro and function definitions are kept in the function cell, and that cell can hold only one Lisp object at any given time. See Macros.

In Emacs Lisp, a definition is not required in order to use a symbol as a variable or function. Thus, you can make a symbol a global variable with setq, whether you define it first or not. The real purpose of definitions is to guide programmers and programming tools. They inform programmers who read the code that certain symbols are intended to be used as variables, or as functions. In addition, utilities such as etags and make-docfile recognize definitions, and add appropriate information to tag tables and the DOC-version file. See Accessing Documentation.

8.3 Creating and Interning Symbols

To understand how symbols are created in GNU Emacs Lisp, you must know how Lisp reads them. Lisp must ensure that it finds the same symbol every time it reads the same set of characters. Failure to do so would cause complete confusion.

When the Lisp reader encounters a symbol, it reads all the characters of the name. Then it “hashes” those characters to find an index in a table called an obarray. Hashing is an efficient method of looking something up. For example, instead of searching a telephone book cover to cover when looking up Jan Jones, you start with the J's and go from there. That is a simple version of hashing. Each element of the obarray is a bucket which holds all the symbols with a given hash code; to look for a given name, it is sufficient to look through all the symbols in the bucket for that name's hash code. (The same idea is used for general Emacs hash tables, but they are a different data type; see Hash Tables.)

If a symbol with the desired name is found, the reader uses that symbol. If the obarray does not contain a symbol with that name, the reader makes a new symbol and adds it to the obarray. Finding or adding a symbol with a certain name is called interning it, and the symbol is then called an interned symbol.

Interning ensures that each obarray has just one symbol with any particular name. Other like-named symbols may exist, but not in the same obarray. Thus, the reader gets the same symbols for the same names, as long as you keep reading with the same obarray.

Interning usually happens automatically in the reader, but sometimes other programs need to do it. For example, after the M-x command obtains the command name as a string using the minibuffer, it then interns the string, to get the interned symbol with that name.

No obarray contains all symbols; in fact, some symbols are not in any obarray. They are called uninterned symbols. An uninterned symbol has the same four cells as other symbols; however, the only way to gain access to it is by finding it in some other object or as the value of a variable.

Creating an uninterned symbol is useful in generating Lisp code, because an uninterned symbol used as a variable in the code you generate cannot clash with any variables used in other Lisp programs.

In Emacs Lisp, an obarray is actually a vector. Each element of the vector is a bucket; its value is either an interned symbol whose name hashes to that bucket, or 0 if the bucket is empty. Each interned symbol has an internal link (invisible to the user) to the next symbol in the bucket. Because these links are invisible, there is no way to find all the symbols in an obarray except using mapatoms (below). The order of symbols in a bucket is not significant.

In an empty obarray, every element is 0, so you can create an obarray with (make-vector length 0). This is the only valid way to create an obarray. Prime numbers as lengths tend to result in good hashing; lengths one less than a power of two are also good.

Do not try to put symbols in an obarray yourself. This does not work—only intern can enter a symbol in an obarray properly.

Common Lisp note: In Common Lisp, a single symbol may be interned in several obarrays.

Most of the functions below take a name and sometimes an obarray as arguments. A wrong-type-argument error is signaled if the name is not a string, or if the obarray is not a vector.

Common Lisp note: In Common Lisp, you can intern an existing symbol in an obarray. In Emacs Lisp, you cannot do this, because the argument to intern must be a string, not a symbol.

8.4 Property Lists

A property list (plist for short) is a list of paired elements. Each of the pairs associates a property name (usually a symbol) with a property or value.

Every symbol has a cell that stores a property list (see Symbol Components). This property list is used to record information about the symbol, such as its variable documentation and the name of the file where it was defined.

Property lists can also be used in other contexts. For instance, you can assign property lists to character positions in a string or buffer. See Text Properties.

The property names and values in a property list can be any Lisp objects, but the names are usually symbols. Property list functions compare the property names using eq. Here is an example of a property list, found on the symbol progn when the compiler is loaded:

     (lisp-indent-function 0 byte-compile byte-compile-progn)

Here lisp-indent-function and byte-compile are property names, and the other two elements are the corresponding values.

• Plists and Alists: Comparison of the advantages of property lists and association lists.
• Symbol Plists: Functions to access symbols' property lists.
• Other Plists: Accessing property lists stored elsewhere.

8.4.1 Property Lists and Association Lists

Association lists (see Association Lists) are very similar to property lists. In contrast to association lists, the order of the pairs in the property list is not significant since the property names must be distinct.

Property lists are better than association lists for attaching information to various Lisp function names or variables. If your program keeps all of its associations in one association list, it will typically need to search that entire list each time it checks for an association. This could be slow. By contrast, if you keep the same information in the property lists of the function names or variables themselves, each search will scan only the length of one property list, which is usually short. This is why the documentation for a variable is recorded in a property named variable-documentation. The byte compiler likewise uses properties to record those functions needing special treatment.

However, association lists have their own advantages. Depending on your application, it may be faster to add an association to the front of an association list than to update a property. All properties for a symbol are stored in the same property list, so there is a possibility of a conflict between different uses of a property name. (For this reason, it is a good idea to choose property names that are probably unique, such as by beginning the property name with the program's usual name-prefix for variables and functions.) An association list may be used like a stack where associations are pushed on the front of the list and later discarded; this is not possible with a property list.

8.4.2 Property List Functions for Symbols

8.4.3 Property Lists Outside Symbols

These functions are useful for manipulating property lists that are stored in places other than symbols:

You could define put in terms of plist-put as follows:

     (defun put (symbol prop value)
       (setplist symbol
                 (plist-put (symbol-plist symbol) prop value)))

9 Evaluation

The evaluation of expressions in Emacs Lisp is performed by the Lisp interpreter—a program that receives a Lisp object as input and computes its value as an expression. How it does this depends on the data type of the object, according to rules described in this chapter. The interpreter runs automatically to evaluate portions of your program, but can also be called explicitly via the Lisp primitive function eval.

• Intro Eval: Evaluation in the scheme of things.
• Forms: How various sorts of objects are evaluated.
• Quoting: Avoiding evaluation (to put constants in the program).
• Eval: How to invoke the Lisp interpreter explicitly.

9.1 Introduction to Evaluation

The Lisp interpreter, or evaluator, is the part of Emacs that computes the value of an expression that is given to it. When a function written in Lisp is called, the evaluator computes the value of the function by evaluating the expressions in the function body. Thus, running any Lisp program really means running the Lisp interpreter.

A Lisp object that is intended for evaluation is called an expression or a form. The fact that forms are data objects and not merely text is one of the fundamental differences between Lisp-like languages and typical programming languages. Any object can be evaluated, but in practice only numbers, symbols, lists and strings are evaluated very often.

In subsequent sections, we will describe the details of what evaluation means for each kind of form.

It is very common to read a Lisp form and then evaluate the form, but reading and evaluation are separate activities, and either can be performed alone. Reading per se does not evaluate anything; it converts the printed representation of a Lisp object to the object itself. It is up to the caller of read to specify whether this object is a form to be evaluated, or serves some entirely different purpose. See Input Functions.

Evaluation is a recursive process, and evaluating a form often involves evaluating parts within that form. For instance, when you evaluate a function call form such as (car x), Emacs first evaluates the argument (the subform x). After evaluating the argument, Emacs executes the function (car), and if the function is written in Lisp, execution works by evaluating the body of the function. (In this example, however, car is not a Lisp function; it is a primitive function implemented in C.) See Functions, for more information about functions and function calls.

Evaluation takes place in a context called the environment, which consists of the current values and bindings of all Lisp variables (see Variables).³ Whenever a form refers to a variable without creating a new binding for it, the variable evaluates to the value given by the current environment. Evaluating a form may create a new environment for recursive evaluation, by binding variables (see Local Variables). Such environments are temporary, and vanish when the evaluation of the form is complete.

Evaluating a form may also make changes that persist; these changes are called side effects. An example of a form that produces a side effect is (setq foo 1).

Do not confuse evaluation with command key interpretation. The editor command loop translates keyboard input into a command (an interactively callable function) using the active keymaps, and then uses call-interactively to execute that command. Executing the command usually involves evaluation, if the command is written in Lisp; however, this step is not considered a part of command key interpretation. See Command Loop.

9.2 Kinds of Forms

A Lisp object that is intended to be evaluated is called a form. How Emacs evaluates a form depends on its data type. Emacs has three different kinds of form that are evaluated differently: symbols, lists, and “all other types.” This section describes all three kinds, one by one, starting with the “all other types” which are self-evaluating forms.

• Self-Evaluating Forms: Forms that evaluate to themselves.
• Symbol Forms: Symbols evaluate as variables.
• Classifying Lists: How to distinguish various sorts of list forms.
• Function Indirection: When a symbol appears as the car of a list, we find the real function via the symbol.
• Function Forms: Forms that call functions.
• Macro Forms: Forms that call macros.
• Special Forms: "Special forms" are idiosyncratic primitives, most of them extremely important.
• Autoloading: Functions set up to load files containing their real definitions.

9.2.1 Self-Evaluating Forms

A self-evaluating form is any form that is not a list or symbol. Self-evaluating forms evaluate to themselves: the result of evaluation is the same object that was evaluated. Thus, the number 25 evaluates to 25, and the string "foo" evaluates to the string "foo". Likewise, evaluating a vector does not cause evaluation of the elements of the vector—it returns the same vector with its contents unchanged.

     '123               ; A number, shown without evaluation.
          ⇒ 123
     123                ; Evaluated as usual---result is the same.
          ⇒ 123
     (eval '123)        ; Evaluated ``by hand''---result is the same.
          ⇒ 123
     (eval (eval '123)) ; Evaluating twice changes nothing.
          ⇒ 123

It is common to write numbers, characters, strings, and even vectors in Lisp code, taking advantage of the fact that they self-evaluate. However, it is quite unusual to do this for types that lack a read syntax, because there's no way to write them textually. It is possible to construct Lisp expressions containing these types by means of a Lisp program. Here is an example:

     ;; Build an expression containing a buffer object.
     (setq print-exp (list 'print (current-buffer)))
          ⇒ (print #<buffer eval.texi>)
     ;; Evaluate it.
     (eval print-exp)
          -| #<buffer eval.texi>
          ⇒ #<buffer eval.texi>

9.2.2 Symbol Forms

When a symbol is evaluated, it is treated as a variable. The result is the variable's value, if it has one. If it has none (if its value cell is void), an error is signaled. For more information on the use of variables, see Variables.

In the following example, we set the value of a symbol with setq. Then we evaluate the symbol, and get back the value that setq stored.

     (setq a 123)
          ⇒ 123
     (eval 'a)
          ⇒ 123
     a
          ⇒ 123

The symbols nil and t are treated specially, so that the value of nil is always nil, and the value of t is always t; you cannot set or bind them to any other values. Thus, these two symbols act like self-evaluating forms, even though eval treats them like any other symbol. A symbol whose name starts with ‘:’ also self-evaluates in the same way; likewise, its value ordinarily cannot be changed. See Constant Variables.

9.2.3 Classification of List Forms

A form that is a nonempty list is either a function call, a macro call, or a special form, according to its first element. These three kinds of forms are evaluated in different ways, described below. The remaining list elements constitute the arguments for the function, macro, or special form.

The first step in evaluating a nonempty list is to examine its first element. This element alone determines what kind of form the list is and how the rest of the list is to be processed. The first element is not evaluated, as it would be in some Lisp dialects such as Scheme.

9.2.4 Symbol Function Indirection

If the first element of the list is a symbol then evaluation examines the symbol's function cell, and uses its contents instead of the original symbol. If the contents are another symbol, this process, called symbol function indirection, is repeated until it obtains a non-symbol. See Function Names, for more information about symbol function indirection.

One possible consequence of this process is an infinite loop, in the event that a symbol's function cell refers to the same symbol. Or a symbol may have a void function cell, in which case the subroutine symbol-function signals a void-function error. But if neither of these things happens, we eventually obtain a non-symbol, which ought to be a function or other suitable object.

More precisely, we should now have a Lisp function (a lambda expression), a byte-code function, a primitive function, a Lisp macro, a special form, or an autoload object. Each of these types is a case described in one of the following sections. If the object is not one of these types, Emacs signals an invalid-function error.

The following example illustrates the symbol indirection process. We use fset to set the function cell of a symbol and symbol-function to get the function cell contents (see Function Cells). Specifically, we store the symbol car into the function cell of first, and the symbol first into the function cell of erste.

     ;; Build this function cell linkage:
     ;;   -------------       -----        -------        -------
     ;;  | #<subr car> | <-- | car |  <-- | first |  <-- | erste |
     ;;   -------------       -----        -------        -------

     (symbol-function 'car)
          ⇒ #<subr car>
     (fset 'first 'car)
          ⇒ car
     (fset 'erste 'first)
          ⇒ first
     (erste '(1 2 3))   ; Call the function referenced by erste.
          ⇒ 1

By contrast, the following example calls a function without any symbol function indirection, because the first element is an anonymous Lisp function, not a symbol.

     ((lambda (arg) (erste arg))
      '(1 2 3))
          ⇒ 1

Executing the function itself evaluates its body; this does involve symbol function indirection when calling erste.

The built-in function indirect-function provides an easy way to perform symbol function indirection explicitly.

9.2.5 Evaluation of Function Forms

If the first element of a list being evaluated is a Lisp function object, byte-code object or primitive function object, then that list is a function call. For example, here is a call to the function +:

     (+ 1 x)

The first step in evaluating a function call is to evaluate the remaining elements of the list from left to right. The results are the actual argument values, one value for each list element. The next step is to call the function with this list of arguments, effectively using the function apply (see Calling Functions). If the function is written in Lisp, the arguments are used to bind the argument variables of the function (see Lambda Expressions); then the forms in the function body are evaluated in order, and the value of the last body form becomes the value of the function call.

9.2.6 Lisp Macro Evaluation

If the first element of a list being evaluated is a macro object, then the list is a macro call. When a macro call is evaluated, the elements of the rest of the list are not initially evaluated. Instead, these elements themselves are used as the arguments of the macro. The macro definition computes a replacement form, called the expansion of the macro, to be evaluated in place of the original form. The expansion may be any sort of form: a self-evaluating constant, a symbol, or a list. If the expansion is itself a macro call, this process of expansion repeats until some other sort of form results.

Ordinary evaluation of a macro call finishes by evaluating the expansion. However, the macro expansion is not necessarily evaluated right away, or at all, because other programs also expand macro calls, and they may or may not evaluate the expansions.

Normally, the argument expressions are not evaluated as part of computing the macro expansion, but instead appear as part of the expansion, so they are computed when the expansion is evaluated.

For example, given a macro defined as follows:

     (defmacro cadr (x)
       (list 'car (list 'cdr x)))

an expression such as (cadr (assq 'handler list)) is a macro call, and its expansion is:

     (car (cdr (assq 'handler list)))

Note that the argument (assq 'handler list) appears in the expansion.

See Macros, for a complete description of Emacs Lisp macros.

9.2.7 Special Forms

A special form is a primitive function specially marked so that its arguments are not all evaluated. Most special forms define control structures or perform variable bindings—things which functions cannot do.

Each special form has its own rules for which arguments are evaluated and which are used without evaluation. Whether a particular argument is evaluated may depend on the results of evaluating other arguments.

Here is a list, in alphabetical order, of all of the special forms in Emacs Lisp with a reference to where each is described.

and

see Combining Conditions

catch

see Catch and Throw

cond

see Conditionals

condition-case

see Handling Errors

defconst

see Defining Variables

defmacro

see Defining Macros

defun

see Defining Functions

defvar

see Defining Variables

function

see Anonymous Functions

if

see Conditionals

interactive

see Interactive Call

let

let*

see Local Variables

or

see Combining Conditions

prog1

prog2

progn

see Sequencing

quote

see Quoting

save-current-buffer

see Current Buffer

save-excursion

see Excursions

save-restriction

see Narrowing

save-window-excursion

see Window Configurations

setq

see Setting Variables

setq-default

see Creating Buffer-Local

track-mouse

see Mouse Tracking

unwind-protect

see Nonlocal Exits

while

see Iteration

with-output-to-temp-buffer

see Temporary Displays

Common Lisp note: Here are some comparisons of special forms in GNU Emacs Lisp and Common Lisp. setq, if, and catch are special forms in both Emacs Lisp and Common Lisp. defun is a special form in Emacs Lisp, but a macro in Common Lisp. save-excursion is a special form in Emacs Lisp, but doesn't exist in Common Lisp. throw is a special form in Common Lisp (because it must be able to throw multiple values), but it is a function in Emacs Lisp (which doesn't have multiple values).

9.2.8 Autoloading

The autoload feature allows you to call a function or macro whose function definition has not yet been loaded into Emacs. It specifies which file contains the definition. When an autoload object appears as a symbol's function definition, calling that symbol as a function automatically loads the specified file; then it calls the real definition loaded from that file. See Autoload.

9.3 Quoting

The special form quote returns its single argument, as written, without evaluating it. This provides a way to include constant symbols and lists, which are not self-evaluating objects, in a program. (It is not necessary to quote self-evaluating objects such as numbers, strings, and vectors.)

Because quote is used so often in programs, Lisp provides a convenient read syntax for it. An apostrophe character (‘'’) followed by a Lisp object (in read syntax) expands to a list whose first element is quote, and whose second element is the object. Thus, the read syntax 'x is an abbreviation for (quote x).

Here are some examples of expressions that use quote:

     (quote (+ 1 2))
          ⇒ (+ 1 2)
     (quote foo)
          ⇒ foo
     'foo
          ⇒ foo
     ''foo
          ⇒ (quote foo)
     '(quote foo)
          ⇒ (quote foo)
     ['foo]
          ⇒ [(quote foo)]

Other quoting constructs include function (see Anonymous Functions), which causes an anonymous lambda expression written in Lisp to be compiled, and ‘`’ (see Backquote), which is used to quote only part of a list, while computing and substituting other parts.

9.4 Eval

Most often, forms are evaluated automatically, by virtue of their occurrence in a program being run. On rare occasions, you may need to write code that evaluates a form that is computed at run time, such as after reading a form from text being edited or getting one from a property list. On these occasions, use the eval function.

The functions and variables described in this section evaluate forms, specify limits to the evaluation process, or record recently returned values. Loading a file also does evaluation (see Loading).

It is generally cleaner and more flexible to store a function in a data structure, and call it with funcall or apply, than to store an expression in the data structure and evaluate it. Using functions provides the ability to pass information to them as arguments.

10 Control Structures

A Lisp program consists of expressions or forms (see Forms). We control the order of execution of these forms by enclosing them in control structures. Control structures are special forms which control when, whether, or how many times to execute the forms they contain.

The simplest order of execution is sequential execution: first form a, then form b, and so on. This is what happens when you write several forms in succession in the body of a function, or at top level in a file of Lisp code—the forms are executed in the order written. We call this textual order. For example, if a function body consists of two forms a and b, evaluation of the function evaluates first a and then b. The result of evaluating b becomes the value of the function.

Explicit control structures make possible an order of execution other than sequential.

Emacs Lisp provides several kinds of control structure, including other varieties of sequencing, conditionals, iteration, and (controlled) jumps—all discussed below. The built-in control structures are special forms since their subforms are not necessarily evaluated or not evaluated sequentially. You can use macros to define your own control structure constructs (see Macros).

• Sequencing: Evaluation in textual order.
• Conditionals: if, cond, when, unless.
• Combining Conditions: and, or, not.
• Iteration: while loops.
• Nonlocal Exits: Jumping out of a sequence.

10.1 Sequencing

Evaluating forms in the order they appear is the most common way control passes from one form to another. In some contexts, such as in a function body, this happens automatically. Elsewhere you must use a control structure construct to do this: progn, the simplest control construct of Lisp.

A progn special form looks like this:

     (progn a b c ...)

and it says to execute the forms a, b, c, and so on, in that order. These forms are called the body of the progn form. The value of the last form in the body becomes the value of the entire progn. (progn) returns nil.

In the early days of Lisp, progn was the only way to execute two or more forms in succession and use the value of the last of them. But programmers found they often needed to use a progn in the body of a function, where (at that time) only one form was allowed. So the body of a function was made into an “implicit progn”: several forms are allowed just as in the body of an actual progn. Many other control structures likewise contain an implicit progn. As a result, progn is not used as much as it was many years ago. It is needed now most often inside an unwind-protect, and, or, or in the then-part of an if.

Two other control constructs likewise evaluate a series of forms but return a different value:

10.2 Conditionals

Conditional control structures choose among alternatives. Emacs Lisp has four conditional forms: if, which is much the same as in other languages; when and unless, which are variants of if; and cond, which is a generalized case statement.

Any conditional construct can be expressed with cond or with if. Therefore, the choice between them is a matter of style. For example:

     (if a b c)
     ==
     (cond (a b) (t c))

10.3 Constructs for Combining Conditions

This section describes three constructs that are often used together with if and cond to express complicated conditions. The constructs and and or can also be used individually as kinds of multiple conditional constructs.

10.4 Iteration

Iteration means executing part of a program repetitively. For example, you might want to repeat some computation once for each element of a list, or once for each integer from 0 to n. You can do this in Emacs Lisp with the special form while:

The dolist and dotimes macros provide convenient ways to write two common kinds of loops.

10.5 Nonlocal Exits

A nonlocal exit is a transfer of control from one point in a program to another remote point. Nonlocal exits can occur in Emacs Lisp as a result of errors; you can also use them under explicit control. Nonlocal exits unbind all variable bindings made by the constructs being exited.

• Catch and Throw: Nonlocal exits for the program's own purposes.
• Examples of Catch: Showing how such nonlocal exits can be written.
• Errors: How errors are signaled and handled.
• Cleanups: Arranging to run a cleanup form if an error happens.

10.5.1 Explicit Nonlocal Exits: catch and throw

Most control constructs affect only the flow of control within the construct itself. The function throw is the exception to this rule of normal program execution: it performs a nonlocal exit on request. (There are other exceptions, but they are for error handling only.) throw is used inside a catch, and jumps back to that catch. For example:

     (defun foo-outer ()
       (catch 'foo
         (foo-inner)))
     
     (defun foo-inner ()
       ...
       (if x
           (throw 'foo t))
       ...)

The throw form, if executed, transfers control straight back to the corresponding catch, which returns immediately. The code following the throw is not executed. The second argument of throw is used as the return value of the catch.

The function throw finds the matching catch based on the first argument: it searches for a catch whose first argument is eq to the one specified in the throw. If there is more than one applicable catch, the innermost one takes precedence. Thus, in the above example, the throw specifies foo, and the catch in foo-outer specifies the same symbol, so that catch is the applicable one (assuming there is no other matching catch in between).

Executing throw exits all Lisp constructs up to the matching catch, including function calls. When binding constructs such as let or function calls are exited in this way, the bindings are unbound, just as they are when these constructs exit normally (see Local Variables). Likewise, throw restores the buffer and position saved by save-excursion (see Excursions), and the narrowing status saved by save-restriction and the window selection saved by save-window-excursion (see Window Configurations). It also runs any cleanups established with the unwind-protect special form when it exits that form (see Cleanups).

The throw need not appear lexically within the catch that it jumps to. It can equally well be called from another function called within the catch. As long as the throw takes place chronologically after entry to the catch, and chronologically before exit from it, it has access to that catch. This is why throw can be used in commands such as exit-recursive-edit that throw back to the editor command loop (see Recursive Editing).

Common Lisp note: Most other versions of Lisp, including Common Lisp, have several ways of transferring control nonsequentially: return, return-from, and go, for example. Emacs Lisp has only throw.

10.5.2 Examples of catch and throw

One way to use catch and throw is to exit from a doubly nested loop. (In most languages, this would be done with a “go to.”) Here we compute (foo i j) for i and j varying from 0 to 9:

     (defun search-foo ()
       (catch 'loop
         (let ((i 0))
           (while (< i 10)
             (let ((j 0))
               (while (< j 10)
                 (if (foo i j)
                     (throw 'loop (list i j)))
                 (setq j (1+ j))))
             (setq i (1+ i))))))

If foo ever returns non-nil, we stop immediately and return a list of i and j. If foo always returns nil, the catch returns normally, and the value is nil, since that is the result of the while.

Here are two tricky examples, slightly different, showing two return points at once. First, two return points with the same tag, hack:

     (defun catch2 (tag)
       (catch tag
         (throw 'hack 'yes)))
     ⇒ catch2
     
     (catch 'hack
       (print (catch2 'hack))
       'no)
     -| yes
     ⇒ no

Since both return points have tags that match the throw, it goes to the inner one, the one established in catch2. Therefore, catch2 returns normally with value yes, and this value is printed. Finally the second body form in the outer catch, which is 'no, is evaluated and returned from the outer catch.

Now let's change the argument given to catch2:

     (catch 'hack
       (print (catch2 'quux))
       'no)
     ⇒ yes

We still have two return points, but this time only the outer one has the tag hack; the inner one has the tag quux instead. Therefore, throw makes the outer catch return the value yes. The function print is never called, and the body-form 'no is never evaluated.

10.5.3 Errors

When Emacs Lisp attempts to evaluate a form that, for some reason, cannot be evaluated, it signals an error.

When an error is signaled, Emacs's default reaction is to print an error message and terminate execution of the current command. This is the right thing to do in most cases, such as if you type C-f at the end of the buffer.

In complicated programs, simple termination may not be what you want. For example, the program may have made temporary changes in data structures, or created temporary buffers that should be deleted before the program is finished. In such cases, you would use unwind-protect to establish cleanup expressions to be evaluated in case of error. (See Cleanups.) Occasionally, you may wish the program to continue execution despite an error in a subroutine. In these cases, you would use condition-case to establish error handlers to recover control in case of error.

Resist the temptation to use error handling to transfer control from one part of the program to another; use catch and throw instead. See Catch and Throw.

• Signaling Errors: How to report an error.
• Processing of Errors: What Emacs does when you report an error.
• Handling Errors: How you can trap errors and continue execution.
• Error Symbols: How errors are classified for trapping them.

10.5.3.1 How to Signal an Error

Signaling an error means beginning error processing. Error processing normally aborts all or part of the running program and returns to a point that is set up to handle the error (see Processing of Errors). Here we describe how to signal an error.

Most errors are signaled “automatically” within Lisp primitives which you call for other purposes, such as if you try to take the car of an integer or move forward a character at the end of the buffer. You can also signal errors explicitly with the functions error and signal.

Quitting, which happens when the user types C-g, is not considered an error, but it is handled almost like an error. See Quitting.

Every error specifies an error message, one way or another. The message should state what is wrong (“File does not exist”), not how things ought to be (“File must exist”). The convention in Emacs Lisp is that error messages should start with a capital letter, but should not end with any sort of punctuation.

Common Lisp note: Emacs Lisp has nothing like the Common Lisp concept of continuable errors.

10.5.3.2 How Emacs Processes Errors

When an error is signaled, signal searches for an active handler for the error. A handler is a sequence of Lisp expressions designated to be executed if an error happens in part of the Lisp program. If the error has an applicable handler, the handler is executed, and control resumes following the handler. The handler executes in the environment of the condition-case that established it; all functions called within that condition-case have already been exited, and the handler cannot return to them.

If there is no applicable handler for the error, it terminates the current command and returns control to the editor command loop. (The command loop has an implicit handler for all kinds of errors.) The command loop's handler uses the error symbol and associated data to print an error message. You can use the variable command-error-function to control how this is done:

An error that has no explicit handler may call the Lisp debugger. The debugger is enabled if the variable debug-on-error (see Error Debugging) is non-nil. Unlike error handlers, the debugger runs in the environment of the error, so that you can examine values of variables precisely as they were at the time of the error.

10.5.3.3 Writing Code to Handle Errors

The usual effect of signaling an error is to terminate the command that is running and return immediately to the Emacs editor command loop. You can arrange to trap errors occurring in a part of your program by establishing an error handler, with the special form condition-case. A simple example looks like this:

     (condition-case nil
         (delete-file filename)
       (error nil))

This deletes the file named filename, catching any error and returning nil if an error occurs⁴.

The condition-case construct is often used to trap errors that are predictable, such as failure to open a file in a call to insert-file-contents. It is also used to trap errors that are totally unpredictable, such as when the program evaluates an expression read from the user.

The second argument of condition-case is called the protected form. (In the example above, the protected form is a call to delete-file.) The error handlers go into effect when this form begins execution and are deactivated when this form returns. They remain in effect for all the intervening time. In particular, they are in effect during the execution of functions called by this form, in their subroutines, and so on. This is a good thing, since, strictly speaking, errors can be signaled only by Lisp primitives (including signal and error) called by the protected form, not by the protected form itself.

The arguments after the protected form are handlers. Each handler lists one or more condition names (which are symbols) to specify which errors it will handle. The error symbol specified when an error is signaled also defines a list of condition names. A handler applies to an error if they have any condition names in common. In the example above, there is one handler, and it specifies one condition name, error, which covers all errors.

The search for an applicable handler checks all the established handlers starting with the most recently established one. Thus, if two nested condition-case forms offer to handle the same error, the inner of the two gets to handle it.

If an error is handled by some condition-case form, this ordinarily prevents the debugger from being run, even if debug-on-error says this error should invoke the debugger.

If you want to be able to debug errors that are caught by a condition-case, set the variable debug-on-signal to a non-nil value. You can also specify that a particular handler should let the debugger run first, by writing debug among the conditions, like this:

     (condition-case nil
         (delete-file filename)
       ((debug error) nil))

The effect of debug here is only to prevent condition-case from suppressing the call to the debugger. Any given error will invoke the debugger only if debug-on-error and the other usual filtering mechanisms say it should. See Error Debugging.

Once Emacs decides that a certain handler handles the error, it returns control to that handler. To do so, Emacs unbinds all variable bindings made by binding constructs that are being exited, and executes the cleanups of all unwind-protect forms that are being exited. Once control arrives at the handler, the body of the handler executes normally.

After execution of the handler body, execution returns from the condition-case form. Because the protected form is exited completely before execution of the handler, the handler cannot resume execution at the point of the error, nor can it examine variable bindings that were made within the protected form. All it can do is clean up and proceed.

Error signaling and handling have some resemblance to throw and catch (see Catch and Throw), but they are entirely separate facilities. An error cannot be caught by a catch, and a throw cannot be handled by an error handler (though using throw when there is no suitable catch signals an error that can be handled).

Here is an example of using condition-case to handle the error that results from dividing by zero. The handler displays the error message (but without a beep), then returns a very large number.

     (defun safe-divide (dividend divisor)
       (condition-case err
           ;; Protected form.
           (/ dividend divisor)
         ;; The handler.
         (arith-error                        ; Condition.
          ;; Display the usual message for this error.
          (message "%s" (error-message-string err))
          1000000)))
     ⇒ safe-divide
     
     (safe-divide 5 0)
          -| Arithmetic error: (arith-error)
     ⇒ 1000000

The handler specifies condition name arith-error so that it will handle only division-by-zero errors. Other kinds of errors will not be handled, at least not by this condition-case. Thus,

     (safe-divide nil 3)
          error--> Wrong type argument: number-or-marker-p, nil

Here is a condition-case that catches all kinds of errors, including those signaled with error:

     (setq baz 34)
          ⇒ 34
     
     (condition-case err
         (if (eq baz 35)
             t
           ;; This is a call to the function error.
           (error "Rats!  The variable %s was %s, not 35" 'baz baz))
       ;; This is the handler; it is not a form.
       (error (princ (format "The error was: %s" err))
              2))
     -| The error was: (error "Rats!  The variable baz was 34, not 35")
     ⇒ 2

10.5.3.4 Error Symbols and Condition Names

When you signal an error, you specify an error symbol to specify the kind of error you have in mind. Each error has one and only one error symbol to categorize it. This is the finest classification of errors defined by the Emacs Lisp language.

These narrow classifications are grouped into a hierarchy of wider classes called error conditions, identified by condition names. The narrowest such classes belong to the error symbols themselves: each error symbol is also a condition name. There are also condition names for more extensive classes, up to the condition name error which takes in all kinds of errors (but not quit). Thus, each error has one or more condition names: error, the error symbol if that is distinct from error, and perhaps some intermediate classifications.

In order for a symbol to be an error symbol, it must have an error-conditions property which gives a list of condition names. This list defines the conditions that this kind of error belongs to. (The error symbol itself, and the symbol error, should always be members of this list.) Thus, the hierarchy of condition names is defined by the error-conditions properties of the error symbols. Because quitting is not considered an error, the value of the error-conditions property of quit is just (quit).

In addition to the error-conditions list, the error symbol should have an error-message property whose value is a string to be printed when that error is signaled but not handled. If the error symbol has no error-message property or if the error-message property exists, but is not a string, the error message ‘peculiar error’ is used. See Definition of signal.

Here is how we define a new error symbol, new-error:

     (put 'new-error
          'error-conditions
          '(error my-own-errors new-error))
     ⇒ (error my-own-errors new-error)
     (put 'new-error 'error-message "A new error")
     ⇒ "A new error"

This error has three condition names: new-error, the narrowest classification; my-own-errors, which we imagine is a wider classification; and error, which is the widest of all.

The error string should start with a capital letter but it should not end with a period. This is for consistency with the rest of Emacs.

Naturally, Emacs will never signal new-error on its own; only an explicit call to signal (see Definition of signal) in your code can do this:

     (signal 'new-error '(x y))
          error--> A new error: x, y

This error can be handled through any of the three condition names. This example handles new-error and any other errors in the class my-own-errors:

     (condition-case foo
         (bar nil t)
       (my-own-errors nil))

The significant way that errors are classified is by their condition names—the names used to match errors with handlers. An error symbol serves only as a convenient way to specify the intended error message and list of condition names. It would be cumbersome to give signal a list of condition names rather than one error symbol.

By contrast, using only error symbols without condition names would seriously decrease the power of condition-case. Condition names make it possible to categorize errors at various levels of generality when you write an error handler. Using error symbols alone would eliminate all but the narrowest level of classification.

See Standard Errors, for a list of all the standard error symbols and their conditions.

10.5.4 Cleaning Up from Nonlocal Exits

The unwind-protect construct is essential whenever you temporarily put a data structure in an inconsistent state; it permits you to make the data consistent again in the event of an error or throw. (Another more specific cleanup construct that is used only for changes in buffer contents is the atomic change group; Atomic Changes.)

For example, here we make an invisible buffer for temporary use, and make sure to kill it before finishing:

     (let ((buffer (get-buffer-create " *temp*")))
       (with-current-buffer buffer
         (unwind-protect
             body-form
           (kill-buffer buffer))))

You might think that we could just as well write (kill-buffer (current-buffer)) and dispense with the variable buffer. However, the way shown above is safer, if body-form happens to get an error after switching to a different buffer! (Alternatively, you could write a save-current-buffer around body-form, to ensure that the temporary buffer becomes current again in time to kill it.)

Emacs includes a standard macro called with-temp-buffer which expands into more or less the code shown above (see Current Buffer). Several of the macros defined in this manual use unwind-protect in this way.

Here is an actual example derived from an FTP package. It creates a process (see Processes) to try to establish a connection to a remote machine. As the function ftp-login is highly susceptible to numerous problems that the writer of the function cannot anticipate, it is protected with a form that guarantees deletion of the process in the event of failure. Otherwise, Emacs might fill up with useless subprocesses.

     (let ((win nil))
       (unwind-protect
           (progn
             (setq process (ftp-setup-buffer host file))
             (if (setq win (ftp-login process host user password))
                 (message "Logged in")
               (error "Ftp login failed")))
         (or win (and process (delete-process process)))))

This example has a small bug: if the user types C-g to quit, and the quit happens immediately after the function ftp-setup-buffer returns but before the variable process is set, the process will not be killed. There is no easy way to fix this bug, but at least it is very unlikely.

11 Variables

A variable is a name used in a program to stand for a value. Nearly all programming languages have variables of some sort. In the text of a Lisp program, variables are written using the syntax for symbols.

In Lisp, unlike most programming languages, programs are represented primarily as Lisp objects and only secondarily as text. The Lisp objects used for variables are symbols: the symbol name is the variable name, and the variable's value is stored in the value cell of the symbol. The use of a symbol as a variable is independent of its use as a function name. See Symbol Components.

The textual form of a Lisp program is given by the read syntax of the Lisp objects that constitute the program. Hence, a variable in a textual Lisp program is written using the read syntax for the symbol representing the variable.

• Global Variables: Variable values that exist permanently, everywhere.
• Constant Variables: Certain "variables" have values that never change.
• Local Variables: Variable values that exist only temporarily.
• Void Variables: Symbols that lack values.
• Defining Variables: A definition says a symbol is used as a variable.
• Tips for Defining: Things you should think about when you define a variable.
• Accessing Variables: Examining values of variables whose names are known only at run time.
• Setting Variables: Storing new values in variables.
• Variable Scoping: How Lisp chooses among local and global values.
• Buffer-Local Variables: Variable values in effect only in one buffer.
• File Local Variables: Handling local variable lists in files.
• Directory Local Variables: Local variables common to all files in a directory.
• Frame-Local Variables: Frame-local bindings for variables.
• Variable Aliases: Variables that are aliases for other variables.
• Variables with Restricted Values: Non-constant variables whose value can not be an arbitrary Lisp object.

11.1 Global Variables

The simplest way to use a variable is globally. This means that the variable has just one value at a time, and this value is in effect (at least for the moment) throughout the Lisp system. The value remains in effect until you specify a new one. When a new value replaces the old one, no trace of the old value remains in the variable.

You specify a value for a symbol with setq. For example,

     (setq x '(a b))

gives the variable x the value (a b). Note that setq is a special form (see Special Forms); it does not evaluate its first argument, the name of the variable, but it does evaluate the second argument, the new value.

Once the variable has a value, you can refer to it by using the symbol itself as an expression. Thus,

     x ⇒ (a b)

assuming the setq form shown above has already been executed.

If you do set the same variable again, the new value replaces the old one:

     x
          ⇒ (a b)
     (setq x 4)
          ⇒ 4
     x
          ⇒ 4

11.2 Variables that Never Change

In Emacs Lisp, certain symbols normally evaluate to themselves. These include nil and t, as well as any symbol whose name starts with ‘:’ (these are called keywords). These symbols cannot be rebound, nor can their values be changed. Any attempt to set or bind nil or t signals a setting-constant error. The same is true for a keyword (a symbol whose name starts with ‘:’), if it is interned in the standard obarray, except that setting such a symbol to itself is not an error.

     nil == 'nil
          ⇒ nil
     (setq nil 500)
     error--> Attempt to set constant symbol: nil

These constants are fundamentally different from the “constants” defined using the defconst special form (see Defining Variables). A defconst form serves to inform human readers that you do not intend to change the value of a variable, but Emacs does not raise an error if you actually change it.

11.3 Local Variables

Global variables have values that last until explicitly superseded with new values. Sometimes it is useful to create variable values that exist temporarily—only until a certain part of the program finishes. These values are called local, and the variables so used are called local variables.

For example, when a function is called, its argument variables receive new local values that last until the function exits. The let special form explicitly establishes new local values for specified variables; these last until exit from the let form.

Establishing a local value saves away the variable's previous value (or lack of one). We say that the previous value is shadowed and not visible. Both global and local values may be shadowed (see Scope). After the life span of the local value is over, the previous value (or lack of one) is restored.

If you set a variable (such as with setq) while it is local, this replaces the local value; it does not alter the global value, or previous local values, that are shadowed. To model this behavior, we speak of a local binding of the variable as well as a local value.

The local binding is a conceptual place that holds a local value. Entering a function, or a special form such as let, creates the local binding; exiting the function or the let removes the local binding. While the local binding lasts, the variable's value is stored within it. Using setq or set while there is a local binding stores a different value into the local binding; it does not create a new binding.

We also speak of the global binding, which is where (conceptually) the global value is kept.

A variable can have more than one local binding at a time (for example, if there are nested let forms that bind it). In such a case, the most recently created local binding that still exists is the current binding of the variable. (This rule is called dynamic scoping; see Variable Scoping.) If there are no local bindings, the variable's global binding is its current binding. We sometimes call the current binding the most-local existing binding, for emphasis. Ordinary evaluation of a symbol always returns the value of its current binding.

The special forms let and let* exist to create local bindings.

Here is a complete list of the other facilities that create local bindings:

• Function calls (see Functions).
• Macro calls (see Macros).
• condition-case (see Errors).

Variables can also have buffer-local bindings (see Buffer-Local Variables); a few variables have terminal-local bindings (see Multiple Terminals). These kinds of bindings work somewhat like ordinary local bindings, but they are localized depending on “where” you are in Emacs, rather than localized in time.

11.4 When a Variable is “Void”

If you have never given a symbol any value as a global variable, we say that that symbol's global value is void. In other words, the symbol's value cell does not have any Lisp object in it. If you try to evaluate the symbol, you get a void-variable error rather than a value.

Note that a value of nil is not the same as void. The symbol nil is a Lisp object and can be the value of a variable just as any other object can be; but it is a value. A void variable does not have any value.

After you have given a variable a value, you can make it void once more using makunbound.

A variable that has been made void with makunbound is indistinguishable from one that has never received a value and has always been void.

You can use the function boundp to test whether a variable is currently void.

11.5 Defining Global Variables

You may announce your intention to use a symbol as a global variable with a variable definition: a special form, either defconst or defvar.

In Emacs Lisp, definitions serve three purposes. First, they inform people who read the code that certain symbols are intended to be used a certain way (as variables). Second, they inform the Lisp system of these things, supplying a value and documentation. Third, they provide information to utilities such as etags and make-docfile, which create data bases of the functions and variables in a program.

The difference between defconst and defvar is primarily a matter of intent, serving to inform human readers of whether the value should ever change. Emacs Lisp does not restrict the ways in which a variable can be used based on defconst or defvar declarations. However, it does make a difference for initialization: defconst unconditionally initializes the variable, while defvar initializes it only if it is void.

If a user option variable has a variable-interactive property, the set-variable command uses that value to control reading the new value for the variable. The property's value is used as if it were specified in interactive (see Using Interactive). However, this feature is largely obsoleted by defcustom (see Customization).

Warning: If the defconst and defvar special forms are used while the variable has a local binding (made with let, or a function argument), they set the local-binding's value; the top-level binding is not changed. This is not what you usually want. To prevent it, use these special forms at top level in a file, where normally no local binding is in effect, and make sure to load the file before making a local binding for the variable.

11.6 Tips for Defining Variables Robustly

When you define a variable whose value is a function, or a list of functions, use a name that ends in ‘-function’ or ‘-functions’, respectively.

There are several other variable name conventions; here is a complete list:

‘...-hook’

The variable is a normal hook (see Hooks).

‘...-function’

The value is a function.

‘...-functions’

The value is a list of functions.

‘...-form’

The value is a form (an expression).

‘...-forms’

The value is a list of forms (expressions).

‘...-predicate’

The value is a predicate—a function of one argument that returns non-nil for “good” arguments and nil for “bad” arguments.

‘...-flag’

The value is significant only as to whether it is nil or not. Since such variables often end up acquiring more values over time, this convention is not strongly recommended.

‘...-program’

The value is a program name.

‘...-command’

The value is a whole shell command.

‘...-switches’

The value specifies options for a command.

When you define a variable, always consider whether you should mark it as “safe” or “risky”; see File Local Variables.

When defining and initializing a variable that holds a complicated value (such as a keymap with bindings in it), it's best to put the entire computation of the value into the defvar, like this:

     (defvar my-mode-map
       (let ((map (make-sparse-keymap)))
         (define-key map "\C-c\C-a" 'my-command)
         ...
         map)
       docstring)

This method has several benefits. First, if the user quits while loading the file, the variable is either still uninitialized or initialized properly, never in-between. If it is still uninitialized, reloading the file will initialize it properly. Second, reloading the file once the variable is initialized will not alter it; that is important if the user has run hooks to alter part of the contents (such as, to rebind keys). Third, evaluating the defvar form with C-M-x will reinitialize the map completely.

Putting so much code in the defvar form has one disadvantage: it puts the documentation string far away from the line which names the variable. Here's a safe way to avoid that:

     (defvar my-mode-map nil
       docstring)
     (unless my-mode-map
       (let ((map (make-sparse-keymap)))
         (define-key map "\C-c\C-a" 'my-command)
         ...
         (setq my-mode-map map)))

This has all the same advantages as putting the initialization inside the defvar, except that you must type C-M-x twice, once on each form, if you do want to reinitialize the variable.

But be careful not to write the code like this:

     (defvar my-mode-map nil
       docstring)
     (unless my-mode-map
       (setq my-mode-map (make-sparse-keymap))
       (define-key my-mode-map "\C-c\C-a" 'my-command)
       ...)

This code sets the variable, then alters it, but it does so in more than one step. If the user quits just after the setq, that leaves the variable neither correctly initialized nor void nor nil. Once that happens, reloading the file will not initialize the variable; it will remain incomplete.

11.7 Accessing Variable Values

The usual way to reference a variable is to write the symbol which names it (see Symbol Forms). This requires you to specify the variable name when you write the program. Usually that is exactly what you want to do. Occasionally you need to choose at run time which variable to reference; then you can use symbol-value.

11.8 How to Alter a Variable Value

The usual way to change the value of a variable is with the special form setq. When you need to compute the choice of variable at run time, use the function set.

11.9 Scoping Rules for Variable Bindings

A given symbol foo can have several local variable bindings, established at different places in the Lisp program, as well as a global binding. The most recently established binding takes precedence over the others.

Local bindings in Emacs Lisp have indefinite scope and dynamic extent. Scope refers to where textually in the source code the binding can be accessed. “Indefinite scope” means that any part of the program can potentially access the variable binding. Extent refers to when, as the program is executing, the binding exists. “Dynamic extent” means that the binding lasts as long as the activation of the construct that established it.

The combination of dynamic extent and indefinite scope is called dynamic scoping. By contrast, most programming languages use lexical scoping, in which references to a local variable must be located textually within the function or block that binds the variable.

Common Lisp note: Variables declared “special” in Common Lisp are dynamically scoped, like all variables in Emacs Lisp.

• Scope: Scope means where in the program a value is visible. Comparison with other languages.
• Extent: Extent means how long in time a value exists.
• Impl of Scope: Two ways to implement dynamic scoping.
• Using Scoping: How to use dynamic scoping carefully and avoid problems.

11.9.1 Scope

Emacs Lisp uses indefinite scope for local variable bindings. This means that any function anywhere in the program text might access a given binding of a variable. Consider the following function definitions:

     (defun binder (x)   ; x is bound in binder.
        (foo 5))         ; foo is some other function.
     
     (defun user ()      ; x is used ``free'' in user.
       (list x))

In a lexically scoped language, the binding of x in binder would never be accessible in user, because user is not textually contained within the function binder. However, in dynamically-scoped Emacs Lisp, user may or may not refer to the binding of x established in binder, depending on the circumstances:

• If we call user directly without calling binder at all, then whatever binding of x is found, it cannot come from binder.
• If we define foo as follows and then call binder, then the binding made in binder will be seen in user:

            (defun foo (lose)
              (user))
  

• However, if we define foo as follows and then call binder, then the binding made in binder will not be seen in user:

            (defun foo (x)
              (user))
  

  Here, when foo is called by binder, it binds x. (The binding in foo is said to shadow the one made in binder.) Therefore, user will access the x bound by foo instead of the one bound by binder.

Emacs Lisp uses dynamic scoping because simple implementations of lexical scoping are slow. In addition, every Lisp system needs to offer dynamic scoping at least as an option; if lexical scoping is the norm, there must be a way to specify dynamic scoping instead for a particular variable. It might not be a bad thing for Emacs to offer both, but implementing it with dynamic scoping only was much easier.

11.9.2 Extent

Extent refers to the time during program execution that a variable name is valid. In Emacs Lisp, a variable is valid only while the form that bound it is executing. This is called dynamic extent. “Local” or “automatic” variables in most languages, including C and Pascal, have dynamic extent.

One alternative to dynamic extent is indefinite extent. This means that a variable binding can live on past the exit from the form that made the binding. Common Lisp and Scheme, for example, support this, but Emacs Lisp does not.

To illustrate this, the function below, make-add, returns a function that purports to add n to its own argument m. This would work in Common Lisp, but it does not do the job in Emacs Lisp, because after the call to make-add exits, the variable n is no longer bound to the actual argument 2.

     (defun make-add (n)
         (function (lambda (m) (+ n m))))  ; Return a function.
          ⇒ make-add
     (fset 'add2 (make-add 2))  ; Define function add2
                                ;   with (make-add 2).
          ⇒ (lambda (m) (+ n m))
     (add2 4)                   ; Try to add 2 to 4.
     error--> Symbol's value as variable is void: n

Some Lisp dialects have “closures,” objects that are like functions but record additional variable bindings. Emacs Lisp does not have closures.

11.9.3 Implementation of Dynamic Scoping

A simple sample implementation (which is not how Emacs Lisp actually works) may help you understand dynamic binding. This technique is called deep binding and was used in early Lisp systems.

Suppose there is a stack of bindings, which are variable-value pairs. At entry to a function or to a let form, we can push bindings onto the stack for the arguments or local variables created there. We can pop those bindings from the stack at exit from the binding construct.

We can find the value of a variable by searching the stack from top to bottom for a binding for that variable; the value from that binding is the value of the variable. To set the variable, we search for the current binding, then store the new value into that binding.

As you can see, a function's bindings remain in effect as long as it continues execution, even during its calls to other functions. That is why we say the extent of the binding is dynamic. And any other function can refer to the bindings, if it uses the same variables while the bindings are in effect. That is why we say the scope is indefinite.

The actual implementation of variable scoping in GNU Emacs Lisp uses a technique called shallow binding. Each variable has a standard place in which its current value is always found—the value cell of the symbol.

In shallow binding, setting the variable works by storing a value in the value cell. Creating a new binding works by pushing the old value (belonging to a previous binding) onto a stack, and storing the new local value in the value cell. Eliminating a binding works by popping the old value off the stack, into the value cell.

We use shallow binding because it has the same results as deep binding, but runs faster, since there is never a need to search for a binding.

11.9.4 Proper Use of Dynamic Scoping

Binding a variable in one function and using it in another is a powerful technique, but if used without restraint, it can make programs hard to understand. There are two clean ways to use this technique:

• Use or bind the variable only in a few related functions, written close together in one file. Such a variable is used for communication within one program.

  You should write comments to inform other programmers that they can see all uses of the variable before them, and to advise them not to add uses elsewhere.

• Give the variable a well-defined, documented meaning, and make all appropriate functions refer to it (but not bind it or set it) wherever that meaning is relevant. For example, the variable case-fold-search is defined as “non-nil means ignore case when searching”; various search and replace functions refer to it directly or through their subroutines, but do not bind or set it.

  Then you can bind the variable in other programs, knowing reliably what the effect will be.

In either case, you should define the variable with defvar. This helps other people understand your program by telling them to look for inter-function usage. It also avoids a warning from the byte compiler. Choose the variable's name to avoid name conflicts—don't use short names like x.

11.10 Buffer-Local Variables

Global and local variable bindings are found in most programming languages in one form or another. Emacs, however, also supports additional, unusual kinds of variable binding, such as buffer-local bindings, which apply only in one buffer. Having different values for a variable in different buffers is an important customization method. (Variables can also have bindings that are local to each terminal, or to each frame. See Multiple Terminals, and See Frame-Local Variables.)

• Intro to Buffer-Local: Introduction and concepts.
• Creating Buffer-Local: Creating and destroying buffer-local bindings.
• Default Value: The default value is seen in buffers that don't have their own buffer-local values.

11.10.1 Introduction to Buffer-Local Variables

A buffer-local variable has a buffer-local binding associated with a particular buffer. The binding is in effect when that buffer is current; otherwise, it is not in effect. If you set the variable while a buffer-local binding is in effect, the new value goes in that binding, so its other bindings are unchanged. This means that the change is visible only in the buffer where you made it.

The variable's ordinary binding, which is not associated with any specific buffer, is called the default binding. In most cases, this is the global binding.

A variable can have buffer-local bindings in some buffers but not in other buffers. The default binding is shared by all the buffers that don't have their own bindings for the variable. (This includes all newly-created buffers.) If you set the variable in a buffer that does not have a buffer-local binding for it, this sets the default binding, so the new value is visible in all the buffers that see the default binding.

The most common use of buffer-local bindings is for major modes to change variables that control the behavior of commands. For example, C mode and Lisp mode both set the variable paragraph-start to specify that only blank lines separate paragraphs. They do this by making the variable buffer-local in the buffer that is being put into C mode or Lisp mode, and then setting it to the new value for that mode. See Major Modes.

The usual way to make a buffer-local binding is with make-local-variable, which is what major mode commands typically use. This affects just the current buffer; all other buffers (including those yet to be created) will continue to share the default value unless they are explicitly given their own buffer-local bindings.

A more powerful operation is to mark the variable as automatically buffer-local by calling make-variable-buffer-local. You can think of this as making the variable local in all buffers, even those yet to be created. More precisely, the effect is that setting the variable automatically makes the variable local to the current buffer if it is not already so. All buffers start out by sharing the default value of the variable as usual, but setting the variable creates a buffer-local binding for the current buffer. The new value is stored in the buffer-local binding, leaving the default binding untouched. This means that the default value cannot be changed with setq in any buffer; the only way to change it is with setq-default.

Warning: When a variable has buffer-local bindings in one or more buffers, let rebinds the binding that's currently in effect. For instance, if the current buffer has a buffer-local value, let temporarily rebinds that. If no buffer-local bindings are in effect, let rebinds the default value. If inside the let you then change to a different current buffer in which a different binding is in effect, you won't see the let binding any more. And if you exit the let while still in the other buffer, you won't see the unbinding occur (though it will occur properly). Here is an example to illustrate:

     (setq foo 'g)
     (set-buffer "a")
     (make-local-variable 'foo)
     (setq foo 'a)
     (let ((foo 'temp))
       ;; foo ⇒ 'temp  ; let binding in buffer ‘a’
       (set-buffer "b")
       ;; foo ⇒ 'g     ; the global value since foo is not local in ‘b’
       body...)
     foo ⇒ 'g        ; exiting restored the local value in buffer ‘a’,
                      ; but we don't see that in buffer ‘b’
     (set-buffer "a") ; verify the local value was restored
     foo ⇒ 'a

Note that references to foo in body access the buffer-local binding of buffer ‘b’.

When a file specifies local variable values, these become buffer-local values when you visit the file. See File Variables.

A buffer-local variable cannot be made frame-local (see Frame-Local Variables) or terminal-local (see Multiple Terminals).

11.10.2 Creating and Deleting Buffer-Local Bindings

A buffer-local variable is permanent if the variable name (a symbol) has a permanent-local property that is non-nil. Such variables are unaffected by kill-all-local-variables, and their local bindings are therefore not cleared by changing major modes. Permanent locals are appropriate for data pertaining to where the file came from or how to save it, rather than with how to edit the contents.

11.10.3 The Default Value of a Buffer-Local Variable

The global value of a variable with buffer-local bindings is also called the default value, because it is the value that is in effect whenever neither the current buffer nor the selected frame has its own binding for the variable.

The functions default-value and setq-default access and change a variable's default value regardless of whether the current buffer has a buffer-local binding. For example, you could use setq-default to change the default setting of paragraph-start for most buffers; and this would work even when you are in a C or Lisp mode buffer that has a buffer-local value for this variable.

The special forms defvar and defconst also set the default value (if they set the variable at all), rather than any buffer-local value.

11.11 File Local Variables

A file can specify local variable values; Emacs uses these to create buffer-local bindings for those variables in the buffer visiting that file. See Local Variables in Files, for basic information about file-local variables. This section describes the functions and variables that affect how file-local variables are processed.

If a file-local variable could specify an arbitrary function or Lisp expression that would be called later, visiting a file could take over your Emacs. Emacs protects against this by automatically setting only those file-local variables whose specified values are known to be safe. Other file-local variables are set only if the user agrees.

For additional safety, read-circle is temporarily bound to nil when Emacs reads file-local variables (see Input Functions). This prevents the Lisp reader from recognizing circular and shared Lisp structures (see Circular Objects).

You can specify safe values for a variable with a safe-local-variable property. The property has to be a function of one argument; any value is safe if the function returns non-nil given that value. Many commonly-encountered file variables have safe-local-variable properties; these include fill-column, fill-prefix, and indent-tabs-mode. For boolean-valued variables that are safe, use booleanp as the property value. Lambda expressions should be quoted so that describe-variable can display the predicate.

Some variables are considered risky. A variable whose name ends in any of ‘-command’, ‘-frame-alist’, ‘-function’, ‘-functions’, ‘-hook’, ‘-hooks’, ‘-form’, ‘-forms’, ‘-map’, ‘-map-alist’, ‘-mode-alist’, ‘-program’, or ‘-predicate’ is considered risky. The variables ‘font-lock-keywords’, ‘font-lock-keywords’ followed by a digit, and ‘font-lock-syntactic-keywords’ are also considered risky. Finally, any variable whose name has a non-nil risky-local-variable property is considered risky.

If a variable is risky, it will not be entered automatically into safe-local-variable-values as described above. Therefore, Emacs will always query before setting a risky variable, unless the user explicitly allows the setting by customizing safe-local-variable-values directly.

The ‘Eval:’ “variable” is also a potential loophole, so Emacs normally asks for confirmation before handling it.

If the expression is a function call and the function has a safe-local-eval-function property, the property value determines whether the expression is safe to evaluate. The property value can be a predicate to call to test the expression, a list of such predicates (it's safe if any predicate succeeds), or t (always safe provided the arguments are constant).

Text properties are also potential loopholes, since their values could include functions to call. So Emacs discards all text properties from string values specified for file-local variables.

11.12 Directory Local Variables

A directory can specify local variable values common to all files in that directory; Emacs uses these to create buffer-local bindings for those variables in buffers visiting any file in that directory. This is useful when the files in the directory belong to some project and therefore share the same local variables.

There are two different methods for specifying directory local variables: by putting them in a special file, or by defining a project class for that directory.

11.13 Frame-Local Values for Variables

In addition to buffer-local variable bindings (see Buffer-Local Variables), Emacs supports frame-local bindings. A frame-local binding for a variable is in effect in a frame for which it was defined.

In practice, frame-local variables have not proven very useful. Ordinary frame parameters are generally used instead (see Frame Parameters). The function make-variable-frame-local, which was used to define frame-local variables, has been deprecated since Emacs 22.2. However, you can still define a frame-specific binding for a variable var in frame frame, by setting the var frame parameter for that frame:

       (modify-frame-parameters frame '((var . value)))

This causes the variable var to be bound to the specified value in the named frame. To check the frame-specific values of such variables, use frame-parameter. See Parameter Access.

Note that you cannot have a frame-local binding for a variable that has a buffer-local binding.

11.14 Variable Aliases

It is sometimes useful to make two variables synonyms, so that both variables always have the same value, and changing either one also changes the other. Whenever you change the name of a variable—either because you realize its old name was not well chosen, or because its meaning has partly changed—it can be useful to keep the old name as an alias of the new one for compatibility. You can do this with defvaralias.

Variable aliases are convenient for replacing an old name for a variable with a new name. make-obsolete-variable declares that the old name is obsolete and therefore that it may be removed at some stage in the future.

You can make two variables synonyms and declare one obsolete at the same time using the macro define-obsolete-variable-alias.

     (defvaralias 'foo 'bar)
     (indirect-variable 'foo)
          ⇒ bar
     (indirect-variable 'bar)
          ⇒ bar
     (setq bar 2)
     bar
          ⇒ 2
     foo
          ⇒ 2
     (setq foo 0)
     bar
          ⇒ 0
     foo
          ⇒ 0

11.15 Variables with Restricted Values

Ordinary Lisp variables can be assigned any value that is a valid Lisp object. However, certain Lisp variables are not defined in Lisp, but in C. Most of these variables are defined in the C code using DEFVAR_LISP. Like variables defined in Lisp, these can take on any value. However, some variables are defined using DEFVAR_INT or DEFVAR_BOOL. See Writing Emacs Primitives, in particular the description of functions of the type syms_of_filename, for a brief discussion of the C implementation.

Variables of type DEFVAR_BOOL can only take on the values nil or t. Attempting to assign them any other value will set them to t:

     (let ((display-hourglass 5))
       display-hourglass)
          ⇒ t

Variables of type DEFVAR_INT can only take on integer values. Attempting to assign them any other value will result in an error:

     (setq window-min-height 5.0)
     error--> Wrong type argument: integerp, 5.0

12 Functions

A Lisp program is composed mainly of Lisp functions. This chapter explains what functions are, how they accept arguments, and how to define them.

• What Is a Function: Lisp functions vs. primitives; terminology.
• Lambda Expressions: How functions are expressed as Lisp objects.
• Function Names: A symbol can serve as the name of a function.
• Defining Functions: Lisp expressions for defining functions.
• Calling Functions: How to use an existing function.
• Mapping Functions: Applying a function to each element of a list, etc.
• Anonymous Functions: Lambda expressions are functions with no names.
• Function Cells: Accessing or setting the function definition of a symbol.
• Obsolete Functions: Declaring functions obsolete.
• Inline Functions: Defining functions that the compiler will open code.
• Declaring Functions: Telling the compiler that a function is defined.
• Function Safety: Determining whether a function is safe to call.
• Related Topics: Cross-references to specific Lisp primitives that have a special bearing on how functions work.

12.1 What Is a Function?

In a general sense, a function is a rule for carrying on a computation given several values called arguments. The result of the computation is called the value of the function. The computation can also have side effects: lasting changes in the values of variables or the contents of data structures.

Here are important terms for functions in Emacs Lisp and for other function-like objects.

function

In Emacs Lisp, a function is anything that can be applied to arguments in a Lisp program. In some cases, we use it more specifically to mean a function written in Lisp. Special forms and macros are not functions.

primitive

A primitive is a function callable from Lisp that is written in C, such as car or append. These functions are also called built-in functions, or subrs. (Special forms are also considered primitives.)

Usually the reason we implement a function as a primitive is either because it is fundamental, because it provides a low-level interface to operating system services, or because it needs to run fast. Primitives can be modified or added only by changing the C sources and recompiling the editor. See Writing Emacs Primitives.

lambda expression

A lambda expression is a function written in Lisp. These are described in the following section. See Lambda Expressions.

special form

A special form is a primitive that is like a function but does not evaluate all of its arguments in the usual way. It may evaluate only some of the arguments, or may evaluate them in an unusual order, or several times. Many special forms are described in Control Structures.

macro

A macro is a construct defined in Lisp by the programmer. It differs from a function in that it translates a Lisp expression that you write into an equivalent expression to be evaluated instead of the original expression. Macros enable Lisp programmers to do the sorts of things that special forms can do. See Macros, for how to define and use macros.

command

A command is an object that command-execute can invoke; it is a possible definition for a key sequence. Some functions are commands; a function written in Lisp is a command if it contains an interactive declaration (see Defining Commands). Such a function can be called from Lisp expressions like other functions; in this case, the fact that the function is a command makes no difference.

Keyboard macros (strings and vectors) are commands also, even though they are not functions. A symbol is a command if its function definition is a command; such symbols can be invoked with M-x. The symbol is a function as well if the definition is a function. See Interactive Call.

keystroke command

A keystroke command is a command that is bound to a key sequence (typically one to three keystrokes). The distinction is made here merely to avoid confusion with the meaning of “command” in non-Emacs editors; for Lisp programs, the distinction is normally unimportant.

byte-code function

A byte-code function is a function that has been compiled by the byte compiler. See Byte-Code Type.

Unlike functionp, the next three functions do not treat a symbol as its function definition.

12.2 Lambda Expressions

A function written in Lisp is a list that looks like this:

     (lambda (arg-variables...)
       [documentation-string]
       [interactive-declaration]
       body-forms...)

Such a list is called a lambda expression. In Emacs Lisp, it actually is valid as an expression—it evaluates to itself. In some other Lisp dialects, a lambda expression is not a valid expression at all. In either case, its main use is not to be evaluated as an expression, but to be called as a function.

• Lambda Components: The parts of a lambda expression.
• Simple Lambda: A simple example.
• Argument List: Details and special features of argument lists.
• Function Documentation: How to put documentation in a function.

12.2.1 Components of a Lambda Expression

A function written in Lisp (a “lambda expression”) is a list that looks like this:

     (lambda (arg-variables...)
       [documentation-string]
       [interactive-declaration]
       body-forms...)

The first element of a lambda expression is always the symbol lambda. This indicates that the list represents a function. The reason functions are defined to start with lambda is so that other lists, intended for other uses, will not accidentally be valid as functions.

The second element is a list of symbols—the argument variable names. This is called the lambda list. When a Lisp function is called, the argument values are matched up against the variables in the lambda list, which are given local bindings with the values provided. See Local Variables.

The documentation string is a Lisp string object placed within the function definition to describe the function for the Emacs help facilities. See Function Documentation.

The interactive declaration is a list of the form (interactive code-string). This declares how to provide arguments if the function is used interactively. Functions with this declaration are called commands; they can be called using M-x or bound to a key. Functions not intended to be called in this way should not have interactive declarations. See Defining Commands, for how to write an interactive declaration.

The rest of the elements are the body of the function: the Lisp code to do the work of the function (or, as a Lisp programmer would say, “a list of Lisp forms to evaluate”). The value returned by the function is the value returned by the last element of the body.

12.2.2 A Simple Lambda-Expression Example

Consider for example the following function:

     (lambda (a b c) (+ a b c))

We can call this function by writing it as the car of an expression, like this:

     ((lambda (a b c) (+ a b c))
      1 2 3)

This call evaluates the body of the lambda expression with the variable a bound to 1, b bound to 2, and c bound to 3. Evaluation of the body adds these three numbers, producing the result 6; therefore, this call to the function returns the value 6.

Note that the arguments can be the results of other function calls, as in this example:

     ((lambda (a b c) (+ a b c))
      1 (* 2 3) (- 5 4))

This evaluates the arguments 1, (* 2 3), and (- 5 4) from left to right. Then it applies the lambda expression to the argument values 1, 6 and 1 to produce the value 8.

It is not often useful to write a lambda expression as the car of a form in this way. You can get the same result, of making local variables and giving them values, using the special form let (see Local Variables). And let is clearer and easier to use. In practice, lambda expressions are either stored as the function definitions of symbols, to produce named functions, or passed as arguments to other functions (see Anonymous Functions).

However, calls to explicit lambda expressions were very useful in the old days of Lisp, before the special form let was invented. At that time, they were the only way to bind and initialize local variables.

12.2.3 Other Features of Argument Lists

Our simple sample function, (lambda (a b c) (+ a b c)), specifies three argument variables, so it must be called with three arguments: if you try to call it with only two arguments or four arguments, you get a wrong-number-of-arguments error.

It is often convenient to write a function that allows certain arguments to be omitted. For example, the function substring accepts three arguments—a string, the start index and the end index—but the third argument defaults to the length of the string if you omit it. It is also convenient for certain functions to accept an indefinite number of arguments, as the functions list and + do.

To specify optional arguments that may be omitted when a function is called, simply include the keyword &optional before the optional arguments. To specify a list of zero or more extra arguments, include the keyword &rest before one final argument.

Thus, the complete syntax for an argument list is as follows:

     (required-vars...
      [&optional optional-vars...]
      [&rest rest-var])

The square brackets indicate that the &optional and &rest clauses, and the variables that follow them, are optional.

A call to the function requires one actual argument for each of the required-vars. There may be actual arguments for zero or more of the optional-vars, and there cannot be any actual arguments beyond that unless the lambda list uses &rest. In that case, there may be any number of extra actual arguments.

If actual arguments for the optional and rest variables are omitted, then they always default to nil. There is no way for the function to distinguish between an explicit argument of nil and an omitted argument. However, the body of the function is free to consider nil an abbreviation for some other meaningful value. This is what substring does; nil as the third argument to substring means to use the length of the string supplied.

Common Lisp note: Common Lisp allows the function to specify what default value to use when an optional argument is omitted; Emacs Lisp always uses nil. Emacs Lisp does not support “supplied-p” variables that tell you whether an argument was explicitly passed.

For example, an argument list that looks like this:

     (a b &optional c d &rest e)

binds a and b to the first two actual arguments, which are required. If one or two more arguments are provided, c and d are bound to them respectively; any arguments after the first four are collected into a list and e is bound to that list. If there are only two arguments, c is nil; if two or three arguments, d is nil; if four arguments or fewer, e is nil.

There is no way to have required arguments following optional ones—it would not make sense. To see why this must be so, suppose that c in the example were optional and d were required. Suppose three actual arguments are given; which variable would the third argument be for? Would it be used for the c, or for d? One can argue for both possibilities. Similarly, it makes no sense to have any more arguments (either required or optional) after a &rest argument.

Here are some examples of argument lists and proper calls:

     ((lambda (n) (1+ n))                ; One required:
      1)                                 ; requires exactly one argument.
          ⇒ 2
     ((lambda (n &optional n1)           ; One required and one optional:
              (if n1 (+ n n1) (1+ n)))   ; 1 or 2 arguments.
      1 2)
          ⇒ 3
     ((lambda (n &rest ns)               ; One required and one rest:
              (+ n (apply '+ ns)))       ; 1 or more arguments.
      1 2 3 4 5)
          ⇒ 15

12.2.4 Documentation Strings of Functions

A lambda expression may optionally have a documentation string just after the lambda list. This string does not affect execution of the function; it is a kind of comment, but a systematized comment which actually appears inside the Lisp world and can be used by the Emacs help facilities. See Documentation, for how the documentation-string is accessed.

It is a good idea to provide documentation strings for all the functions in your program, even those that are called only from within your program. Documentation strings are like comments, except that they are easier to access.

The first line of the documentation string should stand on its own, because apropos displays just this first line. It should consist of one or two complete sentences that summarize the function's purpose.

The start of the documentation string is usually indented in the source file, but since these spaces come before the starting double-quote, they are not part of the string. Some people make a practice of indenting any additional lines of the string so that the text lines up in the program source. That is a mistake. The indentation of the following lines is inside the string; what looks nice in the source code will look ugly when displayed by the help commands.

You may wonder how the documentation string could be optional, since there are required components of the function that follow it (the body). Since evaluation of a string returns that string, without any side effects, it has no effect if it is not the last form in the body. Thus, in practice, there is no confusion between the first form of the body and the documentation string; if the only body form is a string then it serves both as the return value and as the documentation.

The last line of the documentation string can specify calling conventions different from the actual function arguments. Write text like this:

     \(fn arglist)

following a blank line, at the beginning of the line, with no newline following it inside the documentation string. (The ‘\’ is used to avoid confusing the Emacs motion commands.) The calling convention specified in this way appears in help messages in place of the one derived from the actual arguments of the function.

This feature is particularly useful for macro definitions, since the arguments written in a macro definition often do not correspond to the way users think of the parts of the macro call.

12.3 Naming a Function

In most computer languages, every function has a name; the idea of a function without a name is nonsensical. In Lisp, a function in the strictest sense has no name. It is simply a list whose first element is lambda, a byte-code function object, or a primitive subr-object.

However, a symbol can serve as the name of a function. This happens when you put the function in the symbol's function cell (see Symbol Components). Then the symbol itself becomes a valid, callable function, equivalent to the list or subr-object that its function cell refers to. The contents of the function cell are also called the symbol's function definition. The procedure of using a symbol's function definition in place of the symbol is called symbol function indirection; see Function Indirection.

In practice, nearly all functions are given names in this way and referred to through their names. For example, the symbol car works as a function and does what it does because the primitive subr-object #<subr car> is stored in its function cell.

We give functions names because it is convenient to refer to them by their names in Lisp expressions. For primitive subr-objects such as #<subr car>, names are the only way you can refer to them: there is no read syntax for such objects. For functions written in Lisp, the name is more convenient to use in a call than an explicit lambda expression. Also, a function with a name can refer to itself—it can be recursive. Writing the function's name in its own definition is much more convenient than making the function definition point to itself (something that is not impossible but that has various disadvantages in practice).

We often identify functions with the symbols used to name them. For example, we often speak of “the function car,” not distinguishing between the symbol car and the primitive subr-object that is its function definition. For most purposes, the distinction is not important.

Even so, keep in mind that a function need not have a unique name. While a given function object usually appears in the function cell of only one symbol, this is just a matter of convenience. It is easy to store it in several symbols using fset; then each of the symbols is equally well a name for the same function.

A symbol used as a function name may also be used as a variable; these two uses of a symbol are independent and do not conflict. (Some Lisp dialects, such as Scheme, do not distinguish between a symbol's value and its function definition; a symbol's value as a variable is also its function definition.) If you have not given a symbol a function definition, you cannot use it as a function; whether the symbol has a value as a variable makes no difference to this.

12.4 Defining Functions

We usually give a name to a function when it is first created. This is called defining a function, and it is done with the defun special form.

You cannot create a new primitive function with defun or defalias, but you can use them to change the function definition of any symbol, even one such as car or x-popup-menu whose normal definition is a primitive. However, this is risky: for instance, it is next to impossible to redefine car without breaking Lisp completely. Redefining an obscure function such as x-popup-menu is less dangerous, but it still may not work as you expect. If there are calls to the primitive from C code, they call the primitive's C definition directly, so changing the symbol's definition will have no effect on them.

See also defsubst, which defines a function like defun and tells the Lisp compiler to open-code it. See Inline Functions.

12.5 Calling Functions

Defining functions is only half the battle. Functions don't do anything until you call them, i.e., tell them to run. Calling a function is also known as invocation.

The most common way of invoking a function is by evaluating a list. For example, evaluating the list (concat "a" "b") calls the function concat with arguments "a" and "b". See Evaluation, for a description of evaluation.

When you write a list as an expression in your program, you specify which function to call, and how many arguments to give it, in the text of the program. Usually that's just what you want. Occasionally you need to compute at run time which function to call. To do that, use the function funcall. When you also need to determine at run time how many arguments to pass, use apply.

Sometimes it is useful to fix some of the function's arguments at certain values, and leave the rest of arguments for when the function is actually called. The act of fixing some of the function's arguments is called partial application of the function⁷. The result is a new function that accepts the rest of arguments and calls the original function with all the arguments combined.

Here's how to do partial application in Emacs Lisp:

It is common for Lisp functions to accept functions as arguments or find them in data structures (especially in hook variables and property lists) and call them using funcall or apply. Functions that accept function arguments are often called functionals.

Sometimes, when you call a functional, it is useful to supply a no-op function as the argument. Here are two different kinds of no-op function:

12.6 Mapping Functions

A mapping function applies a given function (not a special form or macro) to each element of a list or other collection. Emacs Lisp has several such functions; mapcar and mapconcat, which scan a list, are described here. See Definition of mapatoms, for the function mapatoms which maps over the symbols in an obarray. See Definition of maphash, for the function maphash which maps over key/value associations in a hash table.

These mapping functions do not allow char-tables because a char-table is a sparse array whose nominal range of indices is very large. To map over a char-table in a way that deals properly with its sparse nature, use the function map-char-table (see Char-Tables).

12.7 Anonymous Functions

In Lisp, a function is a list that starts with lambda, a byte-code function compiled from such a list, or alternatively a primitive subr-object; names are “extra.” Although functions are usually defined with defun and given names at the same time, it is occasionally more concise to use an explicit lambda expression—an anonymous function. Such a list is valid wherever a function name is.

Any method of creating such a list makes a valid function. Even this:

     (setq silly (append '(lambda (x)) (list (list '+ (* 3 4) 'x))))
     ⇒ (lambda (x) (+ 12 x))

This computes a list that looks like (lambda (x) (+ 12 x)) and makes it the value (not the function definition!) of silly.

Here is how we might call this function:

     (funcall silly 1)
     ⇒ 13

It does not work to write (silly 1), because this function is not the function definition of silly. We have not given silly any function definition, just a value as a variable.

Most of the time, anonymous functions are constants that appear in your program. For instance, you might want to pass one as an argument to the function mapcar, which applies any given function to each element of a list (see Mapping Functions). See describe-symbols example, for a realistic example of this.

In the following example, we define a change-property function that takes a function as its third argument, followed by a double-property function that makes use of change-property by passing it an anonymous function:

     (defun change-property (symbol prop function)
       (let ((value (get symbol prop)))
         (put symbol prop (funcall function value))))
     
     (defun double-property (symbol prop)
       (change-property symbol prop (lambda (x) (* 2 x))))

In the double-property function, we did not quote the lambda form. This is permissible, because a lambda form is self-quoting: evaluating the form yields the form itself.

Whether or not you quote a lambda form makes a difference if you compile the code (see Byte Compilation). If the lambda form is unquoted, as in the above example, the anonymous function is also compiled. Suppose, however, that we quoted the lambda form:

     (defun double-property (symbol prop)
       (change-property symbol prop '(lambda (x) (* 2 x))))

If you compile this, the argument passed to change-property is the precise list shown:

     (lambda (x) (* x 2))

The Lisp compiler cannot assume this list is a function, even though it looks like one, since it does not know what change-property will do with the list. Perhaps it will check whether the car of the third element is the symbol *!

The function special form explicitly tells the byte-compiler that its argument is a function:

The read syntax #' is a short-hand for using function. Generally, it is not necessary to use either #' or function; just use an unquoted lambda form instead. (Actually, lambda is a macro defined using function.) The following forms are all equivalent:

     #'(lambda (x) (* x x))
     (function (lambda (x) (* x x)))
     (lambda (x) (* x x))

We sometimes write function instead of quote when quoting the name of a function, but this usage is just a sort of comment:

     (function symbol) == (quote symbol) == 'symbol

12.8 Accessing Function Cell Contents

The function definition of a symbol is the object stored in the function cell of the symbol. The functions described here access, test, and set the function cell of symbols.

See also the function indirect-function. See Definition of indirect-function.

If you have never given a symbol any function definition, we say that that symbol's function cell is void. In other words, the function cell does not have any Lisp object in it. If you try to call such a symbol as a function, it signals a void-function error.

Note that void is not the same as nil or the symbol void. The symbols nil and void are Lisp objects, and can be stored into a function cell just as any other object can be (and they can be valid functions if you define them in turn with defun). A void function cell contains no object whatsoever.

You can test the voidness of a symbol's function definition with fboundp. After you have given a symbol a function definition, you can make it void once more using fmakunbound.

fset is sometimes used to save the old definition of a function before redefining it. That permits the new definition to invoke the old definition. But it is unmodular and unclean for a Lisp file to redefine a function defined elsewhere. If you want to modify a function defined by another package, it is cleaner to use defadvice (see Advising Functions).

12.9 Declaring Functions Obsolete

You can use make-obsolete to declare a function obsolete. This indicates that the function may be removed at some stage in the future.

You can define a function as an alias and declare it obsolete at the same time using the macro define-obsolete-function-alias:

In addition, you can mark a certain a particular calling convention for a function as obsolete:

12.10 Inline Functions

You can define an inline function by using defsubst instead of defun. An inline function works just like an ordinary function except for one thing: when you compile a call to the function, the function's definition is open-coded into the caller.

Making a function inline makes explicit calls run faster. But it also has disadvantages. For one thing, it reduces flexibility; if you change the definition of the function, calls already inlined still use the old definition until you recompile them.

Another disadvantage is that making a large function inline can increase the size of compiled code both in files and in memory. Since the speed advantage of inline functions is greatest for small functions, you generally should not make large functions inline.

Also, inline functions do not behave well with respect to debugging, tracing, and advising (see Advising Functions). Since ease of debugging and the flexibility of redefining functions are important features of Emacs, you should not make a function inline, even if it's small, unless its speed is really crucial, and you've timed the code to verify that using defun actually has performance problems.

It's possible to define a macro to expand into the same code that an inline function would execute. (See Macros.) But the macro would be limited to direct use in expressions—a macro cannot be called with apply, mapcar and so on. Also, it takes some work to convert an ordinary function into a macro. To convert it into an inline function is very easy; simply replace defun with defsubst. Since each argument of an inline function is evaluated exactly once, you needn't worry about how many times the body uses the arguments, as you do for macros. (See Argument Evaluation.)

Inline functions can be used and open-coded later on in the same file, following the definition, just like macros.

12.11 Telling the Compiler that a Function is Defined

Byte-compiling a file often produces warnings about functions that the compiler doesn't know about (see Compiler Errors). Sometimes this indicates a real problem, but usually the functions in question are defined in other files which would be loaded if that code is run. For example, byte-compiling fortran.el used to warn:

     In end of data:
     fortran.el:2152:1:Warning: the function `gud-find-c-expr' is not known
         to be defined.

In fact, gud-find-c-expr is only used in the function that Fortran mode uses for the local value of gud-find-expr-function, which is a callback from GUD; if it is called, the GUD functions will be loaded. When you know that such a warning does not indicate a real problem, it is good to suppress the warning. That makes new warnings which might mean real problems more visible. You do that with declare-function.

All you need to do is add a declare-function statement before the first use of the function in question:

     (declare-function gud-find-c-expr "gud.el" nil)

This says that gud-find-c-expr is defined in gud.el (the ‘.el’ can be omitted). The compiler takes for granted that that file really defines the function, and does not check.

The optional third argument specifies the argument list of gud-find-c-expr. In this case, it takes no arguments (nil is different from not specifying a value). In other cases, this might be something like (file &optional overwrite). You don't have to specify the argument list, but if you do the byte compiler can check that the calls match the declaration.

To verify that these functions really are declared where declare-function says they are, use check-declare-file to check all declare-function calls in one source file, or use check-declare-directory check all the files in and under a certain directory.

These commands find the file that ought to contain a function's definition using locate-library; if that finds no file, they expand the definition file name relative to the directory of the file that contains the declare-function call.

You can also say that a function is defined by C code by specifying a file name ending in ‘.c’ or ‘.m’. check-declare-file looks for these files in the C source code directory. This is useful only when you call a function that is defined only on certain systems. Most of the primitive functions of Emacs are always defined so they will never give you a warning.

Sometimes a file will optionally use functions from an external package. If you prefix the filename in the declare-function statement with ‘ext:’, then it will be checked if it is found, otherwise skipped without error.

There are some function definitions that ‘check-declare’ does not understand (e.g. defstruct and some other macros). In such cases, you can pass a non-nil fileonly argument to declare-function, meaning to only check that the file exists, not that it actually defines the function. Note that to do this without having to specify an argument list, you should set the arglist argument to t (because nil means an empty argument list, as opposed to an unspecified one).

12.12 Determining whether a Function is Safe to Call

Some major modes such as SES call functions that are stored in user files. (see Top, for more information on SES.) User files sometimes have poor pedigrees—you can get a spreadsheet from someone you've just met, or you can get one through email from someone you've never met. So it is risky to call a function whose source code is stored in a user file until you have determined that it is safe.

Being quick and simple, unsafep does a very light analysis and rejects many Lisp expressions that are actually safe. There are no known cases where unsafep returns nil for an unsafe expression. However, a “safe” Lisp expression can return a string with a display property, containing an associated Lisp expression to be executed after the string is inserted into a buffer. This associated expression can be a virus. In order to be safe, you must delete properties from all strings calculated by user code before inserting them into buffers.

12.13 Other Topics Related to Functions

Here is a table of several functions that do things related to function calling and function definitions. They are documented elsewhere, but we provide cross references here.

apply

See Calling Functions.

autoload

See Autoload.

call-interactively

See Interactive Call.

called-interactively-p

See Distinguish Interactive.

commandp

See Interactive Call.

documentation

See Accessing Documentation.

eval

See Eval.

funcall

See Calling Functions.

function

See Anonymous Functions.

ignore

See Calling Functions.

indirect-function

See Function Indirection.

interactive

See Using Interactive.

interactive-p

See Distinguish Interactive.

mapatoms

See Creating Symbols.

mapcar

See Mapping Functions.

map-char-table

See Char-Tables.

mapconcat

See Mapping Functions.

undefined

See Functions for Key Lookup.

13 Macros

Macros enable you to define new control constructs and other language features. A macro is defined much like a function, but instead of telling how to compute a value, it tells how to compute another Lisp expression which will in turn compute the value. We call this expression the expansion of the macro.

Macros can do this because they operate on the unevaluated expressions for the arguments, not on the argument values as functions do. They can therefore construct an expansion containing these argument expressions or parts of them.

If you are using a macro to do something an ordinary function could do, just for the sake of speed, consider using an inline function instead. See Inline Functions.

• Simple Macro: A basic example.
• Expansion: How, when and why macros are expanded.
• Compiling Macros: How macros are expanded by the compiler.
• Defining Macros: How to write a macro definition.
• Backquote: Easier construction of list structure.
• Problems with Macros: Don't evaluate the macro arguments too many times. Don't hide the user's variables.
• Indenting Macros: Specifying how to indent macro calls.

13.1 A Simple Example of a Macro

Suppose we would like to define a Lisp construct to increment a variable value, much like the ++ operator in C. We would like to write (inc x) and have the effect of (setq x (1+ x)). Here's a macro definition that does the job:

     (defmacro inc (var)
        (list 'setq var (list '1+ var)))

When this is called with (inc x), the argument var is the symbol x—not the value of x, as it would be in a function. The body of the macro uses this to construct the expansion, which is (setq x (1+ x)). Once the macro definition returns this expansion, Lisp proceeds to evaluate it, thus incrementing x.

13.2 Expansion of a Macro Call

A macro call looks just like a function call in that it is a list which starts with the name of the macro. The rest of the elements of the list are the arguments of the macro.

Evaluation of the macro call begins like evaluation of a function call except for one crucial difference: the macro arguments are the actual expressions appearing in the macro call. They are not evaluated before they are given to the macro definition. By contrast, the arguments of a function are results of evaluating the elements of the function call list.

Having obtained the arguments, Lisp invokes the macro definition just as a function is invoked. The argument variables of the macro are bound to the argument values from the macro call, or to a list of them in the case of a &rest argument. And the macro body executes and returns its value just as a function body does.

The second crucial difference between macros and functions is that the value returned by the macro body is not the value of the macro call. Instead, it is an alternate expression for computing that value, also known as the expansion of the macro. The Lisp interpreter proceeds to evaluate the expansion as soon as it comes back from the macro.

Since the expansion is evaluated in the normal manner, it may contain calls to other macros. It may even be a call to the same macro, though this is unusual.

You can see the expansion of a given macro call by calling macroexpand.

13.3 Macros and Byte Compilation

You might ask why we take the trouble to compute an expansion for a macro and then evaluate the expansion. Why not have the macro body produce the desired results directly? The reason has to do with compilation.

When a macro call appears in a Lisp program being compiled, the Lisp compiler calls the macro definition just as the interpreter would, and receives an expansion. But instead of evaluating this expansion, it compiles the expansion as if it had appeared directly in the program. As a result, the compiled code produces the value and side effects intended for the macro, but executes at full compiled speed. This would not work if the macro body computed the value and side effects itself—they would be computed at compile time, which is not useful.

In order for compilation of macro calls to work, the macros must already be defined in Lisp when the calls to them are compiled. The compiler has a special feature to help you do this: if a file being compiled contains a defmacro form, the macro is defined temporarily for the rest of the compilation of that file.

Byte-compiling a file also executes any require calls at top-level in the file, so you can ensure that necessary macro definitions are available during compilation by requiring the files that define them (see Named Features). To avoid loading the macro definition files when someone runs the compiled program, write eval-when-compile around the require calls (see Eval During Compile).

13.4 Defining Macros

A Lisp macro is a list whose car is macro. Its cdr should be a function; expansion of the macro works by applying the function (with apply) to the list of unevaluated argument-expressions from the macro call.

It is possible to use an anonymous Lisp macro just like an anonymous function, but this is never done, because it does not make sense to pass an anonymous macro to functionals such as mapcar. In practice, all Lisp macros have names, and they are usually defined with the special form defmacro.

The body of the macro definition can include a declare form, which can specify how <TAB> should indent macro calls, and how to step through them for Edebug.

No macro absolutely needs a declare form, because that form has no effect on how the macro expands, on what the macro means in the program. It only affects secondary features: indentation and Edebug.

13.5 Backquote

Macros often need to construct large list structures from a mixture of constants and nonconstant parts. To make this easier, use the ‘`’ syntax (usually called backquote).

Backquote allows you to quote a list, but selectively evaluate elements of that list. In the simplest case, it is identical to the special form quote (see Quoting). For example, these two forms yield identical results:

     `(a list of (+ 2 3) elements)
          ⇒ (a list of (+ 2 3) elements)
     '(a list of (+ 2 3) elements)
          ⇒ (a list of (+ 2 3) elements)

The special marker ‘,’ inside of the argument to backquote indicates a value that isn't constant. Backquote evaluates the argument of ‘,’ and puts the value in the list structure:

     (list 'a 'list 'of (+ 2 3) 'elements)
          ⇒ (a list of 5 elements)
     `(a list of ,(+ 2 3) elements)
          ⇒ (a list of 5 elements)

Substitution with ‘,’ is allowed at deeper levels of the list structure also. For example:

     (defmacro t-becomes-nil (variable)
       `(if (eq ,variable t)
            (setq ,variable nil)))
     
     (t-becomes-nil foo)
          == (if (eq foo t) (setq foo nil))

You can also splice an evaluated value into the resulting list, using the special marker ‘,@’. The elements of the spliced list become elements at the same level as the other elements of the resulting list. The equivalent code without using ‘`’ is often unreadable. Here are some examples:

     (setq some-list '(2 3))
          ⇒ (2 3)
     (cons 1 (append some-list '(4) some-list))
          ⇒ (1 2 3 4 2 3)
     `(1 ,@some-list 4 ,@some-list)
          ⇒ (1 2 3 4 2 3)
     
     (setq list '(hack foo bar))
          ⇒ (hack foo bar)
     (cons 'use
       (cons 'the
         (cons 'words (append (cdr list) '(as elements)))))
          ⇒ (use the words foo bar as elements)
     `(use the words ,@(cdr list) as elements)
          ⇒ (use the words foo bar as elements)

13.6 Common Problems Using Macros

The basic facts of macro expansion have counterintuitive consequences. This section describes some important consequences that can lead to trouble, and rules to follow to avoid trouble.

• Wrong Time: Do the work in the expansion, not in the macro.
• Argument Evaluation: The expansion should evaluate each macro arg once.
• Surprising Local Vars: Local variable bindings in the expansion require special care.
• Eval During Expansion: Don't evaluate them; put them in the expansion.
• Repeated Expansion: Avoid depending on how many times expansion is done.

13.6.1 Wrong Time

The most common problem in writing macros is doing some of the real work prematurely—while expanding the macro, rather than in the expansion itself. For instance, one real package had this macro definition:

     (defmacro my-set-buffer-multibyte (arg)
       (if (fboundp 'set-buffer-multibyte)
           (set-buffer-multibyte arg)))

With this erroneous macro definition, the program worked fine when interpreted but failed when compiled. This macro definition called set-buffer-multibyte during compilation, which was wrong, and then did nothing when the compiled package was run. The definition that the programmer really wanted was this:

     (defmacro my-set-buffer-multibyte (arg)
       (if (fboundp 'set-buffer-multibyte)
           `(set-buffer-multibyte ,arg)))

This macro expands, if appropriate, into a call to set-buffer-multibyte that will be executed when the compiled program is actually run.

13.6.2 Evaluating Macro Arguments Repeatedly

When defining a macro you must pay attention to the number of times the arguments will be evaluated when the expansion is executed. The following macro (used to facilitate iteration) illustrates the problem. This macro allows us to write a simple “for” loop such as one might find in Pascal.

     (defmacro for (var from init to final do &rest body)
       "Execute a simple \"for\" loop.
     For example, (for i from 1 to 10 do (print i))."
       (list 'let (list (list var init))
             (cons 'while (cons (list '<= var final)
                                (append body (list (list 'inc var)))))))
     ⇒ for
     
     (for i from 1 to 3 do
        (setq square (* i i))
        (princ (format "\n%d %d" i square)))
     ==>
     (let ((i 1))
       (while (<= i 3)
         (setq square (* i i))
         (princ (format "\n%d %d" i square))
         (inc i)))
     
          -|1       1
          -|2       4
          -|3       9
     ⇒ nil

The arguments from, to, and do in this macro are “syntactic sugar”; they are entirely ignored. The idea is that you will write noise words (such as from, to, and do) in those positions in the macro call.

Here's an equivalent definition simplified through use of backquote:

     (defmacro for (var from init to final do &rest body)
       "Execute a simple \"for\" loop.
     For example, (for i from 1 to 10 do (print i))."
       `(let ((,var ,init))
          (while (<= ,var ,final)
            ,@body
            (inc ,var))))

Both forms of this definition (with backquote and without) suffer from the defect that final is evaluated on every iteration. If final is a constant, this is not a problem. If it is a more complex form, say (long-complex-calculation x), this can slow down the execution significantly. If final has side effects, executing it more than once is probably incorrect.

A well-designed macro definition takes steps to avoid this problem by producing an expansion that evaluates the argument expressions exactly once unless repeated evaluation is part of the intended purpose of the macro. Here is a correct expansion for the for macro:

     (let ((i 1)
           (max 3))
       (while (<= i max)
         (setq square (* i i))
         (princ (format "%d      %d" i square))
         (inc i)))

Here is a macro definition that creates this expansion:

     (defmacro for (var from init to final do &rest body)
       "Execute a simple for loop: (for i from 1 to 10 do (print i))."
       `(let ((,var ,init)
              (max ,final))
          (while (<= ,var max)
            ,@body
            (inc ,var))))

Unfortunately, this fix introduces another problem, described in the following section.

13.6.3 Local Variables in Macro Expansions

In the previous section, the definition of for was fixed as follows to make the expansion evaluate the macro arguments the proper number of times:

     (defmacro for (var from init to final do &rest body)
       "Execute a simple for loop: (for i from 1 to 10 do (print i))."
       `(let ((,var ,init)
              (max ,final))
          (while (<= ,var max)
            ,@body
            (inc ,var))))

The new definition of for has a new problem: it introduces a local variable named max which the user does not expect. This causes trouble in examples such as the following:

     (let ((max 0))
       (for x from 0 to 10 do
         (let ((this (frob x)))
           (if (< max this)
               (setq max this)))))

The references to max inside the body of the for, which are supposed to refer to the user's binding of max, really access the binding made by for.

The way to correct this is to use an uninterned symbol instead of max (see Creating Symbols). The uninterned symbol can be bound and referred to just like any other symbol, but since it is created by for, we know that it cannot already appear in the user's program. Since it is not interned, there is no way the user can put it into the program later. It will never appear anywhere except where put by for. Here is a definition of for that works this way:

     (defmacro for (var from init to final do &rest body)
       "Execute a simple for loop: (for i from 1 to 10 do (print i))."
       (let ((tempvar (make-symbol "max")))
         `(let ((,var ,init)
                (,tempvar ,final))
            (while (<= ,var ,tempvar)
              ,@body
              (inc ,var)))))

This creates an uninterned symbol named max and puts it in the expansion instead of the usual interned symbol max that appears in expressions ordinarily.

13.6.4 Evaluating Macro Arguments in Expansion

Another problem can happen if the macro definition itself evaluates any of the macro argument expressions, such as by calling eval (see Eval). If the argument is supposed to refer to the user's variables, you may have trouble if the user happens to use a variable with the same name as one of the macro arguments. Inside the macro body, the macro argument binding is the most local binding of this variable, so any references inside the form being evaluated do refer to it. Here is an example:

     (defmacro foo (a)
       (list 'setq (eval a) t))
          ⇒ foo
     (setq x 'b)
     (foo x) ==> (setq b t)
          ⇒ t                  ; and b has been set.
     ;; but
     (setq a 'c)
     (foo a) ==> (setq a t)
          ⇒ t                  ; but this set a, not c.

It makes a difference whether the user's variable is named a or x, because a conflicts with the macro argument variable a.

Another problem with calling eval in a macro definition is that it probably won't do what you intend in a compiled program. The byte compiler runs macro definitions while compiling the program, when the program's own computations (which you might have wished to access with eval) don't occur and its local variable bindings don't exist.

To avoid these problems, don't evaluate an argument expression while computing the macro expansion. Instead, substitute the expression into the macro expansion, so that its value will be computed as part of executing the expansion. This is how the other examples in this chapter work.

13.6.5 How Many Times is the Macro Expanded?

Occasionally problems result from the fact that a macro call is expanded each time it is evaluated in an interpreted function, but is expanded only once (during compilation) for a compiled function. If the macro definition has side effects, they will work differently depending on how many times the macro is expanded.

Therefore, you should avoid side effects in computation of the macro expansion, unless you really know what you are doing.

One special kind of side effect can't be avoided: constructing Lisp objects. Almost all macro expansions include constructed lists; that is the whole point of most macros. This is usually safe; there is just one case where you must be careful: when the object you construct is part of a quoted constant in the macro expansion.

If the macro is expanded just once, in compilation, then the object is constructed just once, during compilation. But in interpreted execution, the macro is expanded each time the macro call runs, and this means a new object is constructed each time.

In most clean Lisp code, this difference won't matter. It can matter only if you perform side-effects on the objects constructed by the macro definition. Thus, to avoid trouble, avoid side effects on objects constructed by macro definitions. Here is an example of how such side effects can get you into trouble:

     (defmacro empty-object ()
       (list 'quote (cons nil nil)))
     
     (defun initialize (condition)
       (let ((object (empty-object)))
         (if condition
             (setcar object condition))
         object))

If initialize is interpreted, a new list (nil) is constructed each time initialize is called. Thus, no side effect survives between calls. If initialize is compiled, then the macro empty-object is expanded during compilation, producing a single “constant” (nil) that is reused and altered each time initialize is called.

One way to avoid pathological cases like this is to think of empty-object as a funny kind of constant, not as a memory allocation construct. You wouldn't use setcar on a constant such as '(nil), so naturally you won't use it on (empty-object) either.

13.7 Indenting Macros

You can use the declare form in the macro definition to specify how to <TAB> should indent calls to the macro. You write it like this:

     (declare (indent indent-spec))

Here are the possibilities for indent-spec:

nil

This is the same as no property—use the standard indentation pattern.

defun

Handle this function like a ‘def’ construct: treat the second line as the start of a body.

an integer, number

The first number arguments of the function are distinguished arguments; the rest are considered the body of the expression. A line in the expression is indented according to whether the first argument on it is distinguished or not. If the argument is part of the body, the line is indented lisp-body-indent more columns than the open-parenthesis starting the containing expression. If the argument is distinguished and is either the first or second argument, it is indented twice that many extra columns. If the argument is distinguished and not the first or second argument, the line uses the standard pattern.

a symbol, symbol

symbol should be a function name; that function is called to calculate the indentation of a line within this expression. The function receives two arguments:

state

The value returned by parse-partial-sexp (a Lisp primitive for indentation and nesting computation) when it parses up to the beginning of this line.

pos

The position at which the line being indented begins.

It should return either a number, which is the number of columns of indentation for that line, or a list whose car is such a number. The difference between returning a number and returning a list is that a number says that all following lines at the same nesting level should be indented just like this one; a list says that following lines might call for different indentations. This makes a difference when the indentation is being computed by C-M-q; if the value is a number, C-M-q need not recalculate indentation for the following lines until the end of the list.

14 Writing Customization Definitions

This chapter describes how to declare user options for customization, and also customization groups for classifying them. We use the term customization item to include both kinds of customization definitions—as well as face definitions (see Defining Faces).

• Common Keywords: Common keyword arguments for all kinds of customization declarations.
• Group Definitions: Writing customization group definitions.
• Variable Definitions: Declaring user options.
• Customization Types: Specifying the type of a user option.

14.1 Common Item Keywords

All kinds of customization declarations (for variables and groups, and for faces) accept keyword arguments for specifying various information. This section describes some keywords that apply to all kinds.

All of these keywords, except :tag, can be used more than once in a given item. Each use of the keyword has an independent effect. The keyword :tag is an exception because any given item can only display one name.

:tag label

Use label, a string, instead of the item's name, to label the item in customization menus and buffers. Don't use a tag which is substantially different from the item's real name; that would cause confusion.

:group group

Put this customization item in group group. When you use :group in a defgroup, it makes the new group a subgroup of group.

If you use this keyword more than once, you can put a single item into more than one group. Displaying any of those groups will show this item. Please don't overdo this, since the result would be annoying.

:link link-data

Include an external link after the documentation string for this item. This is a sentence containing an active field which references some other documentation.

There are several alternatives you can use for link-data:

(custom-manual info-node)

Link to an Info node; info-node is a string which specifies the node name, as in "(emacs)Top". The link appears as ‘[Manual]’ in the customization buffer and enters the built-in Info reader on info-node.

(info-link info-node)

Like custom-manual except that the link appears in the customization buffer with the Info node name.

(url-link url)

Link to a web page; url is a string which specifies the URL. The link appears in the customization buffer as url and invokes the WWW browser specified by browse-url-browser-function.

(emacs-commentary-link library)

Link to the commentary section of a library; library is a string which specifies the library name.

(emacs-library-link library)

Link to an Emacs Lisp library file; library is a string which specifies the library name.

(file-link file)

Link to a file; file is a string which specifies the name of the file to visit with find-file when the user invokes this link.

(function-link function)

Link to the documentation of a function; function is a string which specifies the name of the function to describe with describe-function when the user invokes this link.

(variable-link variable)

Link to the documentation of a variable; variable is a string which specifies the name of the variable to describe with describe-variable when the user invokes this link.

(custom-group-link group)

Link to another customization group. Invoking it creates a new customization buffer for group.

You can specify the text to use in the customization buffer by adding :tag name after the first element of the link-data; for example, (info-link :tag "foo" "(emacs)Top") makes a link to the Emacs manual which appears in the buffer as ‘foo’.

An item can have more than one external link; however, most items have none at all.

:load file

Load file file (a string) before displaying this customization item (see Loading). Loading is done with load, and only if the file is not already loaded.

:require feature

Execute (require 'feature) when your saved customizations set the value of this item. feature should be a symbol.

The most common reason to use :require is when a variable enables a feature such as a minor mode, and just setting the variable won't have any effect unless the code which implements the mode is loaded.

:version version

This keyword specifies that the item was first introduced in Emacs version version, or that its default value was changed in that version. The value version must be a string.

:package-version '(package . version)

This keyword specifies that the item was first introduced in package version version, or that its meaning or default value was changed in that version. The value of package is a symbol and version is a string.

This keyword takes priority over :version.

package should be the official name of the package, such as MH-E or Gnus. If the package package is released as part of Emacs, package and version should appear in the value of customize-package-emacs-version-alist.

Packages distributed as part of Emacs that use the :package-version keyword must also update the customize-package-emacs-version-alist variable.

14.2 Defining Customization Groups

Each Emacs Lisp package should have one main customization group which contains all the options, faces and other groups in the package. If the package has a small number of options and faces, use just one group and put everything in it. When there are more than twelve or so options and faces, then you should structure them into subgroups, and put the subgroups under the package's main customization group. It is OK to put some of the options and faces in the package's main group alongside the subgroups.

The package's main or only group should be a member of one or more of the standard customization groups. (To display the full list of them, use M-x customize.) Choose one or more of them (but not too many), and add your group to each of them using the :group keyword.

The way to declare new customization groups is with defgroup.

The prefix-discarding feature is currently turned off, which means that :prefix currently has no effect. We did this because we found that discarding the specified prefixes often led to confusing names for options. This happened because the people who wrote the defgroup definitions for various groups added :prefix keywords whenever they make logical sense—that is, whenever the variables in the library have a common prefix.

In order to obtain good results with :prefix, it would be necessary to check the specific effects of discarding a particular prefix, given the specific items in a group and their names and documentation. If the resulting text is not clear, then :prefix should not be used in that case.

It should be possible to recheck all the customization groups, delete the :prefix specifications which give unclear results, and then turn this feature back on, if someone would like to do the work.

14.3 Defining Customization Variables

Use defcustom to declare user-customizable variables.

defcustom accepts the following additional keywords:

:type type

Use type as the data type for this option. It specifies which values are legitimate, and how to display the value. See Customization Types, for more information.

:options value-list

Specify the list of reasonable values for use in this option. The user is not restricted to using only these values, but they are offered as convenient alternatives.

This is meaningful only for certain types, currently including hook, plist and alist. See the definition of the individual types for a description of how to use :options.

:set setfunction

Specify setfunction as the way to change the value of this option. The function setfunction should take two arguments, a symbol (the option name) and the new value, and should do whatever is necessary to update the value properly for this option (which may not mean simply setting the option as a Lisp variable). The default for setfunction is set-default.

:get getfunction

Specify getfunction as the way to extract the value of this option. The function getfunction should take one argument, a symbol, and should return whatever customize should use as the “current value” for that symbol (which need not be the symbol's Lisp value). The default is default-value.

You have to really understand the workings of Custom to use :get correctly. It is meant for values that are treated in Custom as variables but are not actually stored in Lisp variables. It is almost surely a mistake to specify getfunction for a value that really is stored in a Lisp variable.

:initialize function

function should be a function used to initialize the variable when the defcustom is evaluated. It should take two arguments, the option name (a symbol) and the value. Here are some predefined functions meant for use in this way:

custom-initialize-set

Use the variable's :set function to initialize the variable, but do not reinitialize it if it is already non-void.

custom-initialize-default

Like custom-initialize-set, but use the function set-default to set the variable, instead of the variable's :set function. This is the usual choice for a variable whose :set function enables or disables a minor mode; with this choice, defining the variable will not call the minor mode function, but customizing the variable will do so.

custom-initialize-reset

Always use the :set function to initialize the variable. If the variable is already non-void, reset it by calling the :set function using the current value (returned by the :get method). This is the default :initialize function.

custom-initialize-changed

Use the :set function to initialize the variable, if it is already set or has been customized; otherwise, just use set-default.

custom-initialize-safe-set

custom-initialize-safe-default

These functions behave like custom-initialize-set (custom-initialize-default, respectively), but catch errors. If an error occurs during initialization, they set the variable to nil using set-default, and throw no error.

These two functions are only meant for options defined in pre-loaded files, where some variables or functions used to compute the option's value may not yet be defined. The option normally gets updated in startup.el, ignoring the previously computed value. Because of this typical usage, the value which these two functions compute normally only matters when, after startup, one unsets the option's value and then reevaluates the defcustom. By that time, the necessary variables and functions will be defined, so there will not be an error.

:risky value

Set the variable's risky-local-variable property to value (see File Local Variables).

:safe function

Set the variable's safe-local-variable property to function (see File Local Variables).

:set-after variables

When setting variables according to saved customizations, make sure to set the variables variables before this one; in other words, delay setting this variable until after those others have been handled. Use :set-after if setting this variable won't work properly unless those other variables already have their intended values.

It is useful to specify the :require keyword for an option that “turns on” a certain feature. This causes Emacs to load the feature, if it is not already loaded, whenever the option is set. See Common Keywords. Here is an example, from the library saveplace.el:

     (defcustom save-place nil
       "Non-nil means automatically save place in each file..."
       :type 'boolean
       :require 'saveplace
       :group 'save-place)

If a customization item has a type such as hook or alist, which supports :options, you can add additional values to the list from outside the defcustom declaration by calling custom-add-frequent-value. For example, if you define a function my-lisp-mode-initialization intended to be called from emacs-lisp-mode-hook, you might want to add that to the list of reasonable values for emacs-lisp-mode-hook, but not by editing its definition. You can do it thus:

     (custom-add-frequent-value 'emacs-lisp-mode-hook
        'my-lisp-mode-initialization)

Internally, defcustom uses the symbol property standard-value to record the expression for the standard value, saved-value to record the value saved by the user with the customization buffer, and customized-value to record the value set by the user with the customization buffer, but not saved. See Property Lists. These properties are lists, the car of which is an expression that evaluates to the value.

14.4 Customization Types

When you define a user option with defcustom, you must specify its customization type. That is a Lisp object which describes (1) which values are legitimate and (2) how to display the value in the customization buffer for editing.

You specify the customization type in defcustom with the :type keyword. The argument of :type is evaluated, but only once when the defcustom is executed, so it isn't useful for the value to vary. Normally we use a quoted constant. For example:

     (defcustom diff-command "diff"
       "The command to use to run diff."
       :type '(string)
       :group 'diff)

In general, a customization type is a list whose first element is a symbol, one of the customization type names defined in the following sections. After this symbol come a number of arguments, depending on the symbol. Between the type symbol and its arguments, you can optionally write keyword-value pairs (see Type Keywords).

Some type symbols do not use any arguments; those are called simple types. For a simple type, if you do not use any keyword-value pairs, you can omit the parentheses around the type symbol. For example just string as a customization type is equivalent to (string).

All customization types are implemented as widgets; see Introduction, for details.

• Simple Types: Simple customization types: sexp, integer, number, string, file, directory, alist.
• Composite Types: Build new types from other types or data.
• Splicing into Lists: Splice elements into list with :inline.
• Type Keywords: Keyword-argument pairs in a customization type.
• Defining New Types: Give your type a name.

14.4.1 Simple Types

This section describes all the simple customization types.

sexp

The value may be any Lisp object that can be printed and read back. You can use sexp as a fall-back for any option, if you don't want to take the time to work out a more specific type to use.

integer

The value must be an integer, and is represented textually in the customization buffer.

number

The value must be a number (floating point or integer), and is represented textually in the customization buffer.

float

The value must be a floating point number, and is represented textually in the customization buffer.

string

The value must be a string, and the customization buffer shows just the contents, with no delimiting ‘"’ characters and no quoting with ‘\’.

regexp

Like string except that the string must be a valid regular expression.

character

The value must be a character code. A character code is actually an integer, but this type shows the value by inserting the character in the buffer, rather than by showing the number.

file

The value must be a file name, and you can do completion with M-<TAB>.

(file :must-match t)

The value must be a file name for an existing file, and you can do completion with M-<TAB>.

directory

The value must be a directory name, and you can do completion with M-<TAB>.

hook

The value must be a list of functions (or a single function, but that is obsolete usage). This customization type is used for hook variables. You can use the :options keyword in a hook variable's defcustom to specify a list of functions recommended for use in the hook; see Variable Definitions.

alist

The value must be a list of cons-cells, the car of each cell representing a key, and the cdr of the same cell representing an associated value. The user can add and delete key/value pairs, and edit both the key and the value of each pair.

You can specify the key and value types like this:

          (alist :key-type key-type :value-type value-type)

where key-type and value-type are customization type specifications. The default key type is sexp, and the default value type is sexp.

The user can add any key matching the specified key type, but you can give some keys a preferential treatment by specifying them with the :options (see Variable Definitions). The specified keys will always be shown in the customize buffer (together with a suitable value), with a checkbox to include or exclude or disable the key/value pair from the alist. The user will not be able to edit the keys specified by the :options keyword argument.

The argument to the :options keywords should be a list of specifications for reasonable keys in the alist. Ordinarily, they are simply atoms, which stand for themselves as. For example:

          :options '("foo" "bar" "baz")

specifies that there are three “known” keys, namely "foo", "bar" and "baz", which will always be shown first.

You may want to restrict the value type for specific keys, for example, the value associated with the "bar" key can only be an integer. You can specify this by using a list instead of an atom in the list. The first element will specify the key, like before, while the second element will specify the value type. For example:

          :options '("foo" ("bar" integer) "baz")

Finally, you may want to change how the key is presented. By default, the key is simply shown as a const, since the user cannot change the special keys specified with the :options keyword. However, you may want to use a more specialized type for presenting the key, like function-item if you know it is a symbol with a function binding. This is done by using a customization type specification instead of a symbol for the key.

          :options '("foo" ((function-item some-function) integer)
                     "baz")

Many alists use lists with two elements, instead of cons cells. For example,

          (defcustom list-alist '(("foo" 1) ("bar" 2) ("baz" 3))
            "Each element is a list of the form (KEY VALUE).")

instead of

          (defcustom cons-alist '(("foo" . 1) ("bar" . 2) ("baz" . 3))
            "Each element is a cons-cell (KEY . VALUE).")

Because of the way lists are implemented on top of cons cells, you can treat list-alist in the example above as a cons cell alist, where the value type is a list with a single element containing the real value.

          (defcustom list-alist '(("foo" 1) ("bar" 2) ("baz" 3))
            "Each element is a list of the form (KEY VALUE)."
            :type '(alist :value-type (group integer)))

The group widget is used here instead of list only because the formatting is better suited for the purpose.

Similarly, you can have alists with more values associated with each key, using variations of this trick:

          (defcustom person-data '(("brian"  50 t)
                                   ("dorith" 55 nil)
                                   ("ken"    52 t))
            "Alist of basic info about people.
          Each element has the form (NAME AGE MALE-FLAG)."
            :type '(alist :value-type (group integer boolean)))
          
          (defcustom pets '(("brian")
                            ("dorith" "dog" "guppy")
                            ("ken" "cat"))
            "Alist of people's pets.
          In an element (KEY . VALUE), KEY is the person's name,
          and the VALUE is a list of that person's pets."
            :type '(alist :value-type (repeat string)))

plist

The plist custom type is similar to the alist (see above), except that the information is stored as a property list, i.e. a list of this form:

          (key value key value key value ...)

The default :key-type for plist is symbol, rather than sexp.

symbol

The value must be a symbol. It appears in the customization buffer as the name of the symbol.

function

The value must be either a lambda expression or a function name. When it is a function name, you can do completion with M-<TAB>.

variable

The value must be a variable name, and you can do completion with M-<TAB>.

face

The value must be a symbol which is a face name, and you can do completion with M-<TAB>.

boolean

The value is boolean—either nil or t. Note that by using choice and const together (see the next section), you can specify that the value must be nil or t, but also specify the text to describe each value in a way that fits the specific meaning of the alternative.

coding-system

The value must be a coding-system name, and you can do completion with M-<TAB>.

color

The value must be a valid color name, and you can do completion with M-<TAB>. A sample is provided.

14.4.2 Composite Types

When none of the simple types is appropriate, you can use composite types, which build new types from other types or from specified data. The specified types or data are called the arguments of the composite type. The composite type normally looks like this:

     (constructor arguments...)

but you can also add keyword-value pairs before the arguments, like this:

     (constructor {keyword value}... arguments...)

Here is a table of constructors and how to use them to write composite types:

(cons car-type cdr-type)

The value must be a cons cell, its car must fit car-type, and its cdr must fit cdr-type. For example, (cons string symbol) is a customization type which matches values such as ("foo" . foo).

In the customization buffer, the car and the cdr are displayed and edited separately, each according to the type that you specify for it.

(list element-types...)

The value must be a list with exactly as many elements as the element-types given; and each element must fit the corresponding element-type.

For example, (list integer string function) describes a list of three elements; the first element must be an integer, the second a string, and the third a function.

In the customization buffer, each element is displayed and edited separately, according to the type specified for it.

(group element-types...)

This works like list except for the formatting of text in the Custom buffer. list labels each element value with its tag; group does not.

(vector element-types...)

Like list except that the value must be a vector instead of a list. The elements work the same as in list.

(choice alternative-types...)

The value must fit at least one of alternative-types. For example, (choice integer string) allows either an integer or a string.

In the customization buffer, the user selects an alternative using a menu, and can then edit the value in the usual way for that alternative.

Normally the strings in this menu are determined automatically from the choices; however, you can specify different strings for the menu by including the :tag keyword in the alternatives. For example, if an integer stands for a number of spaces, while a string is text to use verbatim, you might write the customization type this way,

          (choice (integer :tag "Number of spaces")
                  (string :tag "Literal text"))

so that the menu offers ‘Number of spaces’ and ‘Literal text’.

In any alternative for which nil is not a valid value, other than a const, you should specify a valid default for that alternative using the :value keyword. See Type Keywords.

If some values are covered by more than one of the alternatives, customize will choose the first alternative that the value fits. This means you should always list the most specific types first, and the most general last. Here's an example of proper usage:

          (choice (const :tag "Off" nil)
                  symbol (sexp :tag "Other"))

This way, the special value nil is not treated like other symbols, and symbols are not treated like other Lisp expressions.

(radio element-types...)

This is similar to choice, except that the choices are displayed using `radio buttons' rather than a menu. This has the advantage of displaying documentation for the choices when applicable and so is often a good choice for a choice between constant functions (function-item customization types).

(const value)

The value must be value—nothing else is allowed.

The main use of const is inside of choice. For example, (choice integer (const nil)) allows either an integer or nil.

:tag is often used with const, inside of choice. For example,

          (choice (const :tag "Yes" t)
                  (const :tag "No" nil)
                  (const :tag "Ask" foo))

describes a variable for which t means yes, nil means no, and foo means “ask.”

(other value)

This alternative can match any Lisp value, but if the user chooses this alternative, that selects the value value.

The main use of other is as the last element of choice. For example,

          (choice (const :tag "Yes" t)
                  (const :tag "No" nil)
                  (other :tag "Ask" foo))

describes a variable for which t means yes, nil means no, and anything else means “ask.” If the user chooses ‘Ask’ from the menu of alternatives, that specifies the value foo; but any other value (not t, nil or foo) displays as ‘Ask’, just like foo.

(function-item function)

Like const, but used for values which are functions. This displays the documentation string as well as the function name. The documentation string is either the one you specify with :doc, or function's own documentation string.

(variable-item variable)

Like const, but used for values which are variable names. This displays the documentation string as well as the variable name. The documentation string is either the one you specify with :doc, or variable's own documentation string.

(set types...)

The value must be a list, and each element of the list must match one of the types specified.

This appears in the customization buffer as a checklist, so that each of types may have either one corresponding element or none. It is not possible to specify two different elements that match the same one of types. For example, (set integer symbol) allows one integer and/or one symbol in the list; it does not allow multiple integers or multiple symbols. As a result, it is rare to use nonspecific types such as integer in a set.

Most often, the types in a set are const types, as shown here:

          (set (const :bold) (const :italic))

Sometimes they describe possible elements in an alist:

          (set (cons :tag "Height" (const height) integer)
               (cons :tag "Width" (const width) integer))

That lets the user specify a height value optionally and a width value optionally.

(repeat element-type)

The value must be a list and each element of the list must fit the type element-type. This appears in the customization buffer as a list of elements, with ‘[INS]’ and ‘[DEL]’ buttons for adding more elements or removing elements.

(restricted-sexp :match-alternatives criteria)

This is the most general composite type construct. The value may be any Lisp object that satisfies one of criteria. criteria should be a list, and each element should be one of these possibilities:

• A predicate—that is, a function of one argument that has no side effects, and returns either nil or non-nil according to the argument. Using a predicate in the list says that objects for which the predicate returns non-nil are acceptable.
• A quoted constant—that is, 'object. This sort of element in the list says that object itself is an acceptable value.

For example,

          (restricted-sexp :match-alternatives
                           (integerp 't 'nil))

allows integers, t and nil as legitimate values.

The customization buffer shows all legitimate values using their read syntax, and the user edits them textually.

Here is a table of the keywords you can use in keyword-value pairs in a composite type:

:tag tag

Use tag as the name of this alternative, for user communication purposes. This is useful for a type that appears inside of a choice.

:match-alternatives criteria

Use criteria to match possible values. This is used only in restricted-sexp.

:args argument-list

Use the elements of argument-list as the arguments of the type construct. For instance, (const :args (foo)) is equivalent to (const foo). You rarely need to write :args explicitly, because normally the arguments are recognized automatically as whatever follows the last keyword-value pair.

14.4.3 Splicing into Lists

The :inline feature lets you splice a variable number of elements into the middle of a list or vector. You use it in a set, choice or repeat type which appears among the element-types of a list or vector.

Normally, each of the element-types in a list or vector describes one and only one element of the list or vector. Thus, if an element-type is a repeat, that specifies a list of unspecified length which appears as one element.

But when the element-type uses :inline, the value it matches is merged directly into the containing sequence. For example, if it matches a list with three elements, those become three elements of the overall sequence. This is analogous to using ‘,@’ in the backquote construct.

For example, to specify a list whose first element must be baz and whose remaining arguments should be zero or more of foo and bar, use this customization type:

     (list (const baz) (set :inline t (const foo) (const bar)))

This matches values such as (baz), (baz foo), (baz bar) and (baz foo bar).

When the element-type is a choice, you use :inline not in the choice itself, but in (some of) the alternatives of the choice. For example, to match a list which must start with a file name, followed either by the symbol t or two strings, use this customization type:

     (list file
           (choice (const t)
                   (list :inline t string string)))

If the user chooses the first alternative in the choice, then the overall list has two elements and the second element is t. If the user chooses the second alternative, then the overall list has three elements and the second and third must be strings.

14.4.4 Type Keywords

You can specify keyword-argument pairs in a customization type after the type name symbol. Here are the keywords you can use, and their meanings:

:value default

This is used for a type that appears as an alternative inside of choice; it specifies the default value to use, at first, if and when the user selects this alternative with the menu in the customization buffer.

Of course, if the actual value of the option fits this alternative, it will appear showing the actual value, not default.

If nil is not a valid value for the alternative, then it is essential to specify a valid default with :value.

:format format-string

This string will be inserted in the buffer to represent the value corresponding to the type. The following ‘%’ escapes are available for use in format-string:

‘%[button%]’

Display the text button marked as a button. The :action attribute specifies what the button will do if the user invokes it; its value is a function which takes two arguments—the widget which the button appears in, and the event.

There is no way to specify two different buttons with different actions.

‘%{sample%}’

Show sample in a special face specified by :sample-face.

‘%v’

Substitute the item's value. How the value is represented depends on the kind of item, and (for variables) on the customization type.

‘%d’

Substitute the item's documentation string.

‘%h’

Like ‘%d’, but if the documentation string is more than one line, add an active field to control whether to show all of it or just the first line.

‘%t’

Substitute the tag here. You specify the tag with the :tag keyword.

‘%%’

Display a literal ‘%’.

:action action

Perform action if the user clicks on a button.

:button-face face

Use the face face (a face name or a list of face names) for button text displayed with ‘%[...%]’.

:button-prefix prefix

:button-suffix suffix

These specify the text to display before and after a button. Each can be:

nil

No text is inserted.

a string

The string is inserted literally.

a symbol

The symbol's value is used.

:tag tag

Use tag (a string) as the tag for the value (or part of the value) that corresponds to this type.

:doc doc

Use doc as the documentation string for this value (or part of the value) that corresponds to this type. In order for this to work, you must specify a value for :format, and use ‘%d’ or ‘%h’ in that value.

The usual reason to specify a documentation string for a type is to provide more information about the meanings of alternatives inside a :choice type or the parts of some other composite type.

:help-echo motion-doc

When you move to this item with widget-forward or widget-backward, it will display the string motion-doc in the echo area. In addition, motion-doc is used as the mouse help-echo string and may actually be a function or form evaluated to yield a help string. If it is a function, it is called with one argument, the widget.

:match function

Specify how to decide whether a value matches the type. The corresponding value, function, should be a function that accepts two arguments, a widget and a value; it should return non-nil if the value is acceptable.

:validate function

Specify a validation function for input. function takes a widget as an argument, and should return nil if the widget's current value is valid for the widget. Otherwise, it should return the widget containing the invalid data, and set that widget's :error property to a string explaining the error.

14.4.5 Defining New Types

In the previous sections we have described how to construct elaborate type specifications for defcustom. In some cases you may want to give such a type specification a name. The obvious case is when you are using the same type for many user options: rather than repeat the specification for each option, you can give the type specification a name, and use that name each defcustom. The other case is when a user option's value is a recursive data structure. To make it possible for a datatype to refer to itself, it needs to have a name.

Since custom types are implemented as widgets, the way to define a new customize type is to define a new widget. We are not going to describe the widget interface here in details, see Introduction, for that. Instead we are going to demonstrate the minimal functionality needed for defining new customize types by a simple example.

     (define-widget 'binary-tree-of-string 'lazy
       "A binary tree made of cons-cells and strings."
       :offset 4
       :tag "Node"
       :type '(choice (string :tag "Leaf" :value "")
                      (cons :tag "Interior"
                            :value ("" . "")
                            binary-tree-of-string
                            binary-tree-of-string)))
     
     (defcustom foo-bar ""
       "Sample variable holding a binary tree of strings."
       :type 'binary-tree-of-string)

The function to define a new widget is called define-widget. The first argument is the symbol we want to make a new widget type. The second argument is a symbol representing an existing widget, the new widget is going to be defined in terms of difference from the existing widget. For the purpose of defining new customization types, the lazy widget is perfect, because it accepts a :type keyword argument with the same syntax as the keyword argument to defcustom with the same name. The third argument is a documentation string for the new widget. You will be able to see that string with the M-x widget-browse <RET> binary-tree-of-string <RET> command.

After these mandatory arguments follow the keyword arguments. The most important is :type, which describes the data type we want to match with this widget. Here a binary-tree-of-string is described as being either a string, or a cons-cell whose car and cdr are themselves both binary-tree-of-string. Note the reference to the widget type we are currently in the process of defining. The :tag attribute is a string to name the widget in the user interface, and the :offset argument is there to ensure that child nodes are indented four spaces relative to the parent node, making the tree structure apparent in the customization buffer.

The defcustom shows how the new widget can be used as an ordinary customization type.

The reason for the name lazy is that the other composite widgets convert their inferior widgets to internal form when the widget is instantiated in a buffer. This conversion is recursive, so the inferior widgets will convert their inferior widgets. If the data structure is itself recursive, this conversion is an infinite recursion. The lazy widget prevents the recursion: it convert its :type argument only when needed.

15 Loading

Loading a file of Lisp code means bringing its contents into the Lisp environment in the form of Lisp objects. Emacs finds and opens the file, reads the text, evaluates each form, and then closes the file.

The load functions evaluate all the expressions in a file just as the eval-buffer function evaluates all the expressions in a buffer. The difference is that the load functions read and evaluate the text in the file as found on disk, not the text in an Emacs buffer.

The loaded file must contain Lisp expressions, either as source code or as byte-compiled code. Each form in the file is called a top-level form. There is no special format for the forms in a loadable file; any form in a file may equally well be typed directly into a buffer and evaluated there. (Indeed, most code is tested this way.) Most often, the forms are function definitions and variable definitions.

A file containing Lisp code is often called a library. Thus, the “Rmail library” is a file containing code for Rmail mode. Similarly, a “Lisp library directory” is a directory of files containing Lisp code.

• How Programs Do Loading: The load function and others.
• Load Suffixes: Details about the suffixes that load tries.
• Library Search: Finding a library to load.
• Loading Non-ASCII: Non-ASCII characters in Emacs Lisp files.
• Autoload: Setting up a function to autoload.
• Repeated Loading: Precautions about loading a file twice.
• Named Features: Loading a library if it isn't already loaded.
• Where Defined: Finding which file defined a certain symbol.
• Unloading: How to "unload" a library that was loaded.
• Hooks for Loading: Providing code to be run when particular libraries are loaded.

15.1 How Programs Do Loading

Emacs Lisp has several interfaces for loading. For example, autoload creates a placeholder object for a function defined in a file; trying to call the autoloading function loads the file to get the function's real definition (see Autoload). require loads a file if it isn't already loaded (see Named Features). Ultimately, all these facilities call the load function to do the work.

For information about how load is used in building Emacs, see Building Emacs.

15.2 Load Suffixes

We now describe some technical details about the exact suffixes that load tries.

To summarize, load normally first tries the suffixes in the value of (get-load-suffixes) and then those in load-file-rep-suffixes. If nosuffix is non-nil, it skips the former group, and if must-suffix is non-nil, it skips the latter group.

15.3 Library Search

When Emacs loads a Lisp library, it searches for the library in a list of directories specified by the variable load-path.

The value of load-path is initialized from the environment variable EMACSLOADPATH, if that exists; otherwise its default value is specified in emacs/src/epaths.h when Emacs is built. Then the list is expanded by adding subdirectories of the directories in the list.

The syntax of EMACSLOADPATH is the same as used for PATH; ‘:’ (or ‘;’, according to the operating system) separates directory names, and ‘.’ is used for the current default directory. Here is an example of how to set your EMACSLOADPATH variable from a csh .login file:

     setenv EMACSLOADPATH .:/user/bil/emacs:/usr/local/share/emacs/20.3/lisp

Here is how to set it using sh:

     export EMACSLOADPATH
     EMACSLOADPATH=.:/user/bil/emacs:/usr/local/share/emacs/20.3/lisp

Here is an example of code you can place in your init file (see Init File) to add several directories to the front of your default load-path:

     (setq load-path
           (append (list nil "/user/bil/emacs"
                         "/usr/local/lisplib"
                         "~/emacs")
                   load-path))

In this example, the path searches the current working directory first, followed then by the /user/bil/emacs directory, the /usr/local/lisplib directory, and the ~/emacs directory, which are then followed by the standard directories for Lisp code.

Dumping Emacs uses a special value of load-path. If the value of load-path at the end of dumping is unchanged (that is, still the same special value), the dumped Emacs switches to the ordinary load-path value when it starts up, as described above. But if load-path has any other value at the end of dumping, that value is used for execution of the dumped Emacs also.

Therefore, if you want to change load-path temporarily for loading a few libraries in site-init.el or site-load.el, you should bind load-path locally with let around the calls to load.

The default value of load-path, when running an Emacs which has been installed on the system, includes two special directories (and their subdirectories as well):

     "/usr/local/share/emacs/version/site-lisp"

and

     "/usr/local/share/emacs/site-lisp"

The first one is for locally installed packages for a particular Emacs version; the second is for locally installed packages meant for use with all installed Emacs versions.

There are several reasons why a Lisp package that works well in one Emacs version can cause trouble in another. Sometimes packages need updating for incompatible changes in Emacs; sometimes they depend on undocumented internal Emacs data that can change without notice; sometimes a newer Emacs version incorporates a version of the package, and should be used only with that version.

Emacs finds these directories' subdirectories and adds them to load-path when it starts up. Both immediate subdirectories and subdirectories multiple levels down are added to load-path.

Not all subdirectories are included, though. Subdirectories whose names do not start with a letter or digit are excluded. Subdirectories named RCS or CVS are excluded. Also, a subdirectory which contains a file named .nosearch is excluded. You can use these methods to prevent certain subdirectories of the site-lisp directories from being searched.

If you run Emacs from the directory where it was built—that is, an executable that has not been formally installed—then load-path normally contains two additional directories. These are the lisp and site-lisp subdirectories of the main build directory. (Both are represented as absolute file names.)

15.4 Loading Non-ASCII Characters

When Emacs Lisp programs contain string constants with non-ASCII characters, these can be represented within Emacs either as unibyte strings or as multibyte strings (see Text Representations). Which representation is used depends on how the file is read into Emacs. If it is read with decoding into multibyte representation, the text of the Lisp program will be multibyte text, and its string constants will be multibyte strings. If a file containing Latin-1 characters (for example) is read without decoding, the text of the program will be unibyte text, and its string constants will be unibyte strings. See Coding Systems.

To make the results more predictable, Emacs always performs decoding into the multibyte representation when loading Lisp files, even if it was started with the ‘--unibyte’ option. This means that string constants with non-ASCII characters translate into multibyte strings. The only exception is when a particular file specifies no decoding.

The reason Emacs is designed this way is so that Lisp programs give predictable results, regardless of how Emacs was started. In addition, this enables programs that depend on using multibyte text to work even in a unibyte Emacs.

In most Emacs Lisp programs, the fact that non-ASCII strings are multibyte strings should not be noticeable, since inserting them in unibyte buffers converts them to unibyte automatically. However, if this does make a difference, you can force a particular Lisp file to be interpreted as unibyte by writing ‘-*-unibyte: t;-*-’ in a comment on the file's first line. With that designator, the file will unconditionally be interpreted as unibyte, even in an ordinary multibyte Emacs session. This can matter when making keybindings to non-ASCII characters written as ?vliteral.

15.5 Autoload

The autoload facility allows you to make a function or macro known in Lisp, but put off loading the file that defines it. The first call to the function automatically reads the proper file to install the real definition and other associated code, then runs the real definition as if it had been loaded all along.

There are two ways to set up an autoloaded function: by calling autoload, and by writing a special “magic” comment in the source before the real definition. autoload is the low-level primitive for autoloading; any Lisp program can call autoload at any time. Magic comments are the most convenient way to make a function autoload, for packages installed along with Emacs. These comments do nothing on their own, but they serve as a guide for the command update-file-autoloads, which constructs calls to autoload and arranges to execute them when Emacs is built.

The autoloaded file usually contains other definitions and may require or provide one or more features. If the file is not completely loaded (due to an error in the evaluation of its contents), any function definitions or provide calls that occurred during the load are undone. This is to ensure that the next attempt to call any function autoloading from this file will try again to load the file. If not for this, then some of the functions in the file might be defined by the aborted load, but fail to work properly for the lack of certain subroutines not loaded successfully because they come later in the file.

If the autoloaded file fails to define the desired Lisp function or macro, then an error is signaled with data "Autoloading failed to define function function-name".

A magic autoload comment (often called an autoload cookie) consists of ‘;;;###autoload’, on a line by itself, just before the real definition of the function in its autoloadable source file. The command M-x update-file-autoloads writes a corresponding autoload call into loaddefs.el. (The string that serves as the autoload cookie and the name of the file generated by update-file-autoloads can be changed from the above defaults, see below.) Building Emacs loads loaddefs.el and thus calls autoload. M-x update-directory-autoloads is even more powerful; it updates autoloads for all files in the current directory.

The same magic comment can copy any kind of form into loaddefs.el. If the form following the magic comment is not a function-defining form or a defcustom form, it is copied verbatim. “Function-defining forms” include define-skeleton, define-derived-mode, define-generic-mode and define-minor-mode as well as defun and defmacro. To save space, a defcustom form is converted to a defvar in loaddefs.el, with some additional information if it uses :require.

You can also use a magic comment to execute a form at build time without executing it when the file itself is loaded. To do this, write the form on the same line as the magic comment. Since it is in a comment, it does nothing when you load the source file; but M-x update-file-autoloads copies it to loaddefs.el, where it is executed while building Emacs.

The following example shows how doctor is prepared for autoloading with a magic comment:

     ;;;###autoload
     (defun doctor ()
       "Switch to *doctor* buffer and start giving psychotherapy."
       (interactive)
       (switch-to-buffer "*doctor*")
       (doctor-mode))

Here's what that produces in loaddefs.el:

     (autoload (quote doctor) "doctor" "\
     Switch to *doctor* buffer and start giving psychotherapy.
     
     \(fn)" t nil)

The backslash and newline immediately following the double-quote are a convention used only in the preloaded uncompiled Lisp files such as loaddefs.el; they tell make-docfile to put the documentation string in the etc/DOC file. See Building Emacs. See also the commentary in lib-src/make-docfile.c. ‘(fn)’ in the usage part of the documentation string is replaced with the function's name when the various help functions (see Help Functions) display it.

If you write a function definition with an unusual macro that is not one of the known and recognized function definition methods, use of an ordinary magic autoload comment would copy the whole definition into loaddefs.el. That is not desirable. You can put the desired autoload call into loaddefs.el instead by writing this:

     ;;;###autoload (autoload 'foo "myfile")
     (mydefunmacro foo
       ...)

You can use a non-default string as the autoload cookie and have the corresponding autoload calls written into a file whose name is different from the default loaddefs.el. Emacs provides two variables to control this:

15.6 Repeated Loading

You can load a given file more than once in an Emacs session. For example, after you have rewritten and reinstalled a function definition by editing it in a buffer, you may wish to return to the original version; you can do this by reloading the file it came from.

When you load or reload files, bear in mind that the load and load-library functions automatically load a byte-compiled file rather than a non-compiled file of similar name. If you rewrite a file that you intend to save and reinstall, you need to byte-compile the new version; otherwise Emacs will load the older, byte-compiled file instead of your newer, non-compiled file! If that happens, the message displayed when loading the file includes, ‘(compiled; note, source is newer)’, to remind you to recompile it.

When writing the forms in a Lisp library file, keep in mind that the file might be loaded more than once. For example, think about whether each variable should be reinitialized when you reload the library; defvar does not change the value if the variable is already initialized. (See Defining Variables.)

The simplest way to add an element to an alist is like this:

     (push '(leif-mode " Leif") minor-mode-alist)

But this would add multiple elements if the library is reloaded. To avoid the problem, use add-to-list (see List Variables):

     (add-to-list 'minor-mode-alist '(leif-mode " Leif"))

Occasionally you will want to test explicitly whether a library has already been loaded. If the library uses provide to provide a named feature, you can use featurep earlier in the file to test whether the provide call has been executed before (see Named Features). Alternatively, you could use something like this:

     (defvar foo-was-loaded nil)
     
     (unless foo-was-loaded
       execute-first-time-only
       (setq foo-was-loaded t))

15.7 Features

provide and require are an alternative to autoload for loading files automatically. They work in terms of named features. Autoloading is triggered by calling a specific function, but a feature is loaded the first time another program asks for it by name.

A feature name is a symbol that stands for a collection of functions, variables, etc. The file that defines them should provide the feature. Another program that uses them may ensure they are defined by requiring the feature. This loads the file of definitions if it hasn't been loaded already.

To require the presence of a feature, call require with the feature name as argument. require looks in the global variable features to see whether the desired feature has been provided already. If not, it loads the feature from the appropriate file. This file should call provide at the top level to add the feature to features; if it fails to do so, require signals an error.

For example, in emacs/lisp/prolog.el, the definition for run-prolog includes the following code:

     (defun run-prolog ()
       "Run an inferior Prolog process, with I/O via buffer *prolog*."
       (interactive)
       (require 'comint)
       (switch-to-buffer (make-comint "prolog" prolog-program-name))
       (inferior-prolog-mode))

The expression (require 'comint) loads the file comint.el if it has not yet been loaded. This ensures that make-comint is defined. Features are normally named after the files that provide them, so that require need not be given the file name.

The comint.el file contains the following top-level expression:

     (provide 'comint)

This adds comint to the global features list, so that (require 'comint) will henceforth know that nothing needs to be done.

When require is used at top level in a file, it takes effect when you byte-compile that file (see Byte Compilation) as well as when you load it. This is in case the required package contains macros that the byte compiler must know about. It also avoids byte compiler warnings for functions and variables defined in the file loaded with require.

Although top-level calls to require are evaluated during byte compilation, provide calls are not. Therefore, you can ensure that a file of definitions is loaded before it is byte-compiled by including a provide followed by a require for the same feature, as in the following example.

     (provide 'my-feature)  ; Ignored by byte compiler,
                            ;   evaluated by load.
     (require 'my-feature)  ; Evaluated by byte compiler.

The compiler ignores the provide, then processes the require by loading the file in question. Loading the file does execute the provide call, so the subsequent require call does nothing when the file is loaded.

15.8 Which File Defined a Certain Symbol

The basis for symbol-file is the data in the variable load-history.

The command eval-region updates load-history, but does so by adding the symbols defined to the element for the file being visited, rather than replacing that element. See Eval.

15.9 Unloading

You can discard the functions and variables loaded by a library to reclaim memory for other Lisp objects. To do this, use the function unload-feature:

The unload-feature function is written in Lisp; its actions are based on the variable load-history.

15.10 Hooks for Loading

You can ask for code to be executed each time Emacs loads a library, by using the variable after-load-functions:

If you want code to be executed when a particular library is loaded, use the function eval-after-load:

Normally, well-designed Lisp programs should not use eval-after-load. If you need to examine and set the variables defined in another library (those meant for outside use), you can do it immediately—there is no need to wait until the library is loaded. If you need to call functions defined by that library, you should load the library, preferably with require (see Named Features).

But it is OK to use eval-after-load in your personal customizations if you don't feel that they must meet the design standards for programs meant for wider use.

16 Byte Compilation

Emacs Lisp has a compiler that translates functions written in Lisp into a special representation called byte-code that can be executed more efficiently. The compiler replaces Lisp function definitions with byte-code. When a byte-code function is called, its definition is evaluated by the byte-code interpreter.

Because the byte-compiled code is evaluated by the byte-code interpreter, instead of being executed directly by the machine's hardware (as true compiled code is), byte-code is completely transportable from machine to machine without recompilation. It is not, however, as fast as true compiled code.

Compiling a Lisp file with the Emacs byte compiler always reads the file as multibyte text, even if Emacs was started with ‘--unibyte’, unless the file specifies otherwise. This is so that compilation gives results compatible with running the same file without compilation. See Loading Non-ASCII.

In general, any version of Emacs can run byte-compiled code produced by recent earlier versions of Emacs, but the reverse is not true.

If you do not want a Lisp file to be compiled, ever, put a file-local variable binding for no-byte-compile into it, like this:

     ;; -*-no-byte-compile: t; -*-

See Compilation Errors, for how to investigate errors occurring in byte compilation.

• Speed of Byte-Code: An example of speedup from byte compilation.
• Compilation Functions: Byte compilation functions.
• Docs and Compilation: Dynamic loading of documentation strings.
• Dynamic Loading: Dynamic loading of individual functions.
• Eval During Compile: Code to be evaluated when you compile.
• Compiler Errors: Handling compiler error messages.
• Byte-Code Objects: The data type used for byte-compiled functions.
• Disassembly: Disassembling byte-code; how to read byte-code.

16.1 Performance of Byte-Compiled Code

A byte-compiled function is not as efficient as a primitive function written in C, but runs much faster than the version written in Lisp. Here is an example:

     (defun silly-loop (n)
       "Return time before and after N iterations of a loop."
       (let ((t1 (current-time-string)))
         (while (> (setq n (1- n))
                   0))
         (list t1 (current-time-string))))
     ⇒ silly-loop
     
     (silly-loop 50000000)
     ⇒ ("Wed Mar 11 21:10:19 2009"
         "Wed Mar 11 21:10:41 2009")  ; 22 seconds
     
     (byte-compile 'silly-loop)
     ⇒ [Compiled code not shown]
     
     (silly-loop 50000000)
     ⇒ ("Wed Mar 11 21:12:26 2009"
         "Wed Mar 11 21:12:32 2009")  ; 6 seconds

In this example, the interpreted code required 22 seconds to run, whereas the byte-compiled code required 6 seconds. These results are representative, but actual results will vary greatly.

16.2 The Compilation Functions

You can byte-compile an individual function or macro definition with the byte-compile function. You can compile a whole file with byte-compile-file, or several files with byte-recompile-directory or batch-byte-compile.

The byte compiler produces error messages and warnings about each file in a buffer called ‘*Compile-Log*’. These report things in your program that suggest a problem but are not necessarily erroneous.

Be careful when writing macro calls in files that you may someday byte-compile. Macro calls are expanded when they are compiled, so the macros must already be defined for proper compilation. For more details, see Compiling Macros. If a program does not work the same way when compiled as it does when interpreted, erroneous macro definitions are one likely cause (see Problems with Macros). Inline (defsubst) functions are less troublesome; if you compile a call to such a function before its definition is known, the call will still work right, it will just run slower.

Normally, compiling a file does not evaluate the file's contents or load the file. But it does execute any require calls at top level in the file. One way to ensure that necessary macro definitions are available during compilation is to require the file that defines them (see Named Features). To avoid loading the macro definition files when someone runs the compiled program, write eval-when-compile around the require calls (see Eval During Compile).

16.3 Documentation Strings and Compilation

Functions and variables loaded from a byte-compiled file access their documentation strings dynamically from the file whenever needed. This saves space within Emacs, and makes loading faster because the documentation strings themselves need not be processed while loading the file. Actual access to the documentation strings becomes slower as a result, but this normally is not enough to bother users.

Dynamic access to documentation strings does have drawbacks:

• If you delete or move the compiled file after loading it, Emacs can no longer access the documentation strings for the functions and variables in the file.
• If you alter the compiled file (such as by compiling a new version), then further access to documentation strings in this file will probably give nonsense results.

If your site installs Emacs following the usual procedures, these problems will never normally occur. Installing a new version uses a new directory with a different name; as long as the old version remains installed, its files will remain unmodified in the places where they are expected to be.

However, if you have built Emacs yourself and use it from the directory where you built it, you will experience this problem occasionally if you edit and recompile Lisp files. When it happens, you can cure the problem by reloading the file after recompiling it.

You can turn off this feature at compile time by setting byte-compile-dynamic-docstrings to nil; this is useful mainly if you expect to change the file, and you want Emacs processes that have already loaded it to keep working when the file changes. You can do this globally, or for one source file by specifying a file-local binding for the variable. One way to do that is by adding this string to the file's first line:

     -*-byte-compile-dynamic-docstrings: nil;-*-

The dynamic documentation string feature writes compiled files that use a special Lisp reader construct, ‘#@count’. This construct skips the next count characters. It also uses the ‘#$’ construct, which stands for “the name of this file, as a string.” It is usually best not to use these constructs in Lisp source files, since they are not designed to be clear to humans reading the file.

16.4 Dynamic Loading of Individual Functions

When you compile a file, you can optionally enable the dynamic function loading feature (also known as lazy loading). With dynamic function loading, loading the file doesn't fully read the function definitions in the file. Instead, each function definition contains a place-holder which refers to the file. The first time each function is called, it reads the full definition from the file, to replace the place-holder.

The advantage of dynamic function loading is that loading the file becomes much faster. This is a good thing for a file which contains many separate user-callable functions, if using one of them does not imply you will probably also use the rest. A specialized mode which provides many keyboard commands often has that usage pattern: a user may invoke the mode, but use only a few of the commands it provides.

The dynamic loading feature has certain disadvantages:

• If you delete or move the compiled file after loading it, Emacs can no longer load the remaining function definitions not already loaded.
• If you alter the compiled file (such as by compiling a new version), then trying to load any function not already loaded will usually yield nonsense results.

These problems will never happen in normal circumstances with installed Emacs files. But they are quite likely to happen with Lisp files that you are changing. The easiest way to prevent these problems is to reload the new compiled file immediately after each recompilation.

The byte compiler uses the dynamic function loading feature if the variable byte-compile-dynamic is non-nil at compilation time. Do not set this variable globally, since dynamic loading is desirable only for certain files. Instead, enable the feature for specific source files with file-local variable bindings. For example, you could do it by writing this text in the source file's first line:

     -*-byte-compile-dynamic: t;-*-

16.5 Evaluation During Compilation

These features permit you to write code to be evaluated during compilation of a program.

16.6 Compiler Errors

Byte compilation outputs all errors and warnings into the buffer ‘*Compile-Log*’. The messages include file names and line numbers that identify the location of the problem. The usual Emacs commands for operating on compiler diagnostics work properly on these messages.

However, the warnings about functions that were used but not defined are always “located” at the end of the file, so these commands won't find the places they are really used. To do that, you must search for the function names.

You can suppress the compiler warning for calling an undefined function func by conditionalizing the function call on an fboundp test, like this:

     (if (fboundp 'func) ...(func ...)...)

The call to func must be in the then-form of the if, and func must appear quoted in the call to fboundp. (This feature operates for cond as well.)

You can tell the compiler that a function is defined using declare-function (see Declaring Functions). Likewise, you can tell the compiler that a variable is defined using defvar with no initial value.

You can suppress the compiler warning for a specific use of an undefined variable variable by conditionalizing its use on a boundp test, like this:

     (if (boundp 'variable) ...variable...)

The reference to variable must be in the then-form of the if, and variable must appear quoted in the call to boundp.

You can suppress any and all compiler warnings within a certain expression using the construct with-no-warnings:

More precise control of warnings is possible by setting the variable byte-compile-warnings.

16.7 Byte-Code Function Objects

Byte-compiled functions have a special data type: they are byte-code function objects.

Internally, a byte-code function object is much like a vector; however, the evaluator handles this data type specially when it appears as a function to be called. The printed representation for a byte-code function object is like that for a vector, with an additional ‘#’ before the opening ‘[’.

A byte-code function object must have at least four elements; there is no maximum number, but only the first six elements have any normal use. They are:

arglist

The list of argument symbols.

byte-code

The string containing the byte-code instructions.

constants

The vector of Lisp objects referenced by the byte code. These include symbols used as function names and variable names.

stacksize

The maximum stack size this function needs.

docstring

The documentation string (if any); otherwise, nil. The value may be a number or a list, in case the documentation string is stored in a file. Use the function documentation to get the real documentation string (see Accessing Documentation).

interactive

The interactive spec (if any). This can be a string or a Lisp expression. It is nil for a function that isn't interactive.

Here's an example of a byte-code function object, in printed representation. It is the definition of the command backward-sexp.

     #[(&optional arg)
       "^H\204^F^@\301^P\302^H[!\207"
       [arg 1 forward-sexp]
       2
       254435
       "p"]

The primitive way to create a byte-code object is with make-byte-code:

You should not try to come up with the elements for a byte-code function yourself, because if they are inconsistent, Emacs may crash when you call the function. Always leave it to the byte compiler to create these objects; it makes the elements consistent (we hope).

You can access the elements of a byte-code object using aref; you can also use vconcat to create a vector with the same elements.

16.8 Disassembled Byte-Code

People do not write byte-code; that job is left to the byte compiler. But we provide a disassembler to satisfy a cat-like curiosity. The disassembler converts the byte-compiled code into human-readable form.

The byte-code interpreter is implemented as a simple stack machine. It pushes values onto a stack of its own, then pops them off to use them in calculations whose results are themselves pushed back on the stack. When a byte-code function returns, it pops a value off the stack and returns it as the value of the function.

In addition to the stack, byte-code functions can use, bind, and set ordinary Lisp variables, by transferring values between variables and the stack.

Here are two examples of using the disassemble function. We have added explanatory comments to help you relate the byte-code to the Lisp source; these do not appear in the output of disassemble.

     (defun factorial (integer)
       "Compute factorial of an integer."
       (if (= 1 integer) 1
         (* integer (factorial (1- integer)))))
          ⇒ factorial
     
     (factorial 4)
          ⇒ 24
     
     (disassemble 'factorial)
          -| byte-code for factorial:
      doc: Compute factorial of an integer.
      args: (integer)
     
     0   varref   integer        ; Get the value of integer
                                 ;   and push it onto the stack.
     1   constant 1              ; Push 1 onto stack.
     
     2   eqlsign                 ; Pop top two values off stack, compare
                                 ;   them, and push result onto stack.
     
     3   goto-if-nil 1           ; Pop and test top of stack;
                                 ;   if nil, go to 1,
                                 ;   else continue.
     6   constant 1              ; Push 1 onto top of stack.
     7   return                  ; Return the top element
                                 ;   of the stack.
     
     8:1 varref   integer        ; Push value of integer onto stack.
     9   constant factorial      ; Push factorial onto stack.
     10  varref   integer        ; Push value of integer onto stack.
     11  sub1                    ; Pop integer, decrement value,
                                 ;   push new value onto stack.
     12  call     1              ; Call function factorial using
                                 ;   the first (i.e., the top) element
                                 ;   of the stack as the argument;
                                 ;   push returned value onto stack.
     
     13 mult                     ; Pop top two values off stack, multiply
                                 ;   them, and push result onto stack.
     14 return                   ; Return the top element of stack.

The silly-loop function is somewhat more complex:

     (defun silly-loop (n)
       "Return time before and after N iterations of a loop."
       (let ((t1 (current-time-string)))
         (while (> (setq n (1- n))
                   0))
         (list t1 (current-time-string))))
          ⇒ silly-loop
     
     (disassemble 'silly-loop)
          -| byte-code for silly-loop:
      doc: Return time before and after N iterations of a loop.
      args: (n)
     
     0   constant current-time-string  ; Push
                                       ;   current-time-string
                                       ;   onto top of stack.
     
     1   call     0              ; Call current-time-string
                                 ;   with no argument,
                                 ;   pushing result onto stack.
     
     2   varbind  t1             ; Pop stack and bind t1
                                 ;   to popped value.
     
     3:1 varref   n              ; Get value of n from
                                 ;   the environment and push
                                 ;   the value onto the stack.
     4   sub1                    ; Subtract 1 from top of stack.
     
     5   dup                     ; Duplicate the top of the stack;
                                 ;   i.e., copy the top of
                                 ;   the stack and push the
                                 ;   copy onto the stack.
     6   varset   n              ; Pop the top of the stack,
                                 ;   and bind n to the value.
     
                                 ; In effect, the sequence dup varset
                                 ;   copies the top of the stack
                                 ;   into the value of n
                                 ;   without popping it.
     
     7   constant 0              ; Push 0 onto stack.
     8   gtr                     ; Pop top two values off stack,
                                 ;   test if n is greater than 0
                                 ;   and push result onto stack.
     
     9   goto-if-not-nil 1       ; Goto 1 if n > 0
                                 ;   (this continues the while loop)
                                 ;   else continue.
     
     12  varref   t1             ; Push value of t1 onto stack.
     13  constant current-time-string  ; Push current-time-string
                                       ;   onto top of stack.
     14  call     0              ; Call current-time-string again.
     
     15  unbind   1              ; Unbind t1 in local environment.
     16  list2                   ; Pop top two elements off stack,
                                 ;   create a list of them,
                                 ;   and push list onto stack.
     17  return                  ; Return value of the top of stack.

17 Advising Emacs Lisp Functions

The advice feature lets you add to the existing definition of a function, by advising the function. This is a cleaner method for a library to customize functions defined within Emacs—cleaner than redefining the whole function.

Each function can have multiple pieces of advice, separately defined. Each defined piece of advice can be enabled or disabled explicitly. All the enabled pieces of advice for any given function actually take effect when you activate advice for that function, or when you define or redefine the function. Note that enabling a piece of advice and activating advice for a function are not the same thing.

Usage Note: Advice is useful for altering the behavior of existing calls to an existing function. If you want the new behavior for new calls, or for key bindings, you should define a new function (or a new command) which uses the existing function.

Usage note: Advising a function can cause confusion in debugging, since people who debug calls to the original function may not notice that it has been modified with advice. Therefore, if you have the possibility to change the code of that function (or ask someone to do so) to run a hook, please solve the problem that way. Advice should be reserved for the cases where you cannot get the function changed.

In particular, this means that a file in Emacs should not put advice on a function in Emacs. There are currently a few exceptions to this convention, but we aim to correct them.

• Simple Advice: A simple example to explain the basics of advice.
• Defining Advice: Detailed description of defadvice.
• Around-Advice: Wrapping advice around a function's definition.
• Computed Advice: ...is to defadvice as fset is to defun.
• Activation of Advice: Advice doesn't do anything until you activate it.
• Enabling Advice: You can enable or disable each piece of advice.
• Preactivation: Preactivation is a way of speeding up the loading of compiled advice.
• Argument Access in Advice: How advice can access the function's arguments.
• Advising Primitives: Accessing arguments when advising a primitive.
• Combined Definition: How advice is implemented.

17.1 A Simple Advice Example

The command next-line moves point down vertically one or more lines; it is the standard binding of C-n. When used on the last line of the buffer, this command inserts a newline to create a line to move to if next-line-add-newlines is non-nil (its default is nil.)

Suppose you wanted to add a similar feature to previous-line, which would insert a new line at the beginning of the buffer for the command to move to (when next-line-add-newlines is non-nil). How could you do this?

You could do it by redefining the whole function, but that is not modular. The advice feature provides a cleaner alternative: you can effectively add your code to the existing function definition, without actually changing or even seeing that definition. Here is how to do this:

     (defadvice previous-line (before next-line-at-end
                                      (&optional arg try-vscroll))
       "Insert an empty line when moving up from the top line."
       (if (and next-line-add-newlines (= arg 1)
                (save-excursion (beginning-of-line) (bobp)))
           (progn
             (beginning-of-line)
             (newline))))

This expression defines a piece of advice for the function previous-line. This piece of advice is named next-line-at-end, and the symbol before says that it is before-advice which should run before the regular definition of previous-line. (&optional arg try-vscroll) specifies how the advice code can refer to the function's arguments.

When this piece of advice runs, it creates an additional line, in the situation where that is appropriate, but does not move point to that line. This is the correct way to write the advice, because the normal definition will run afterward and will move back to the newly inserted line.

Defining the advice doesn't immediately change the function previous-line. That happens when you activate the advice, like this:

     (ad-activate 'previous-line)

This is what actually begins to use the advice that has been defined so far for the function previous-line. Henceforth, whenever that function is run, whether invoked by the user with C-p or M-x, or called from Lisp, it runs the advice first, and its regular definition second.

This example illustrates before-advice, which is one class of advice: it runs before the function's base definition. There are two other advice classes: after-advice, which runs after the base definition, and around-advice, which lets you specify an expression to wrap around the invocation of the base definition.

17.2 Defining Advice

To define a piece of advice, use the macro defadvice. A call to defadvice has the following syntax, which is based on the syntax of defun and defmacro, but adds more:

     (defadvice function (class name
                              [position] [arglist]
                              flags...)
       [documentation-string]
       [interactive-form]
       body-forms...)

Here, function is the name of the function (or macro or special form) to be advised. From now on, we will write just “function” when describing the entity being advised, but this always includes macros and special forms.

In place of the argument list in an ordinary definition, an advice definition calls for several different pieces of information.

class specifies the class of the advice—one of before, after, or around. Before-advice runs before the function itself; after-advice runs after the function itself; around-advice is wrapped around the execution of the function itself. After-advice and around-advice can override the return value by setting ad-return-value.

The argument name is the name of the advice, a non-nil symbol. The advice name uniquely identifies one piece of advice, within all the pieces of advice in a particular class for a particular function. The name allows you to refer to the piece of advice—to redefine it, or to enable or disable it.

The optional position specifies where, in the current list of advice of the specified class, this new advice should be placed. It should be either first, last or a number that specifies a zero-based position (first is equivalent to 0). If no position is specified, the default is first. Position values outside the range of existing positions in this class are mapped to the beginning or the end of the range, whichever is closer. The position value is ignored when redefining an existing piece of advice.

The optional arglist can be used to define the argument list for the sake of advice. This becomes the argument list of the combined definition that is generated in order to run the advice (see Combined Definition). Therefore, the advice expressions can use the argument variables in this list to access argument values.

The argument list used in advice need not be the same as the argument list used in the original function, but must be compatible with it, so that it can handle the ways the function is actually called. If two pieces of advice for a function both specify an argument list, they must specify the same argument list.

See Argument Access in Advice, for more information about argument lists and advice, and a more flexible way for advice to access the arguments.

The remaining elements, flags, are symbols that specify further information about how to use this piece of advice. Here are the valid symbols and their meanings:

activate

Activate the advice for function now. Changes in a function's advice always take effect the next time you activate advice for the function; this flag says to do so, for function, immediately after defining this piece of advice.

This flag has no immediate effect if function itself is not defined yet (a situation known as forward advice), because it is impossible to activate an undefined function's advice. However, defining function will automatically activate its advice.

protect

Protect this piece of advice against non-local exits and errors in preceding code and advice. Protecting advice places it as a cleanup in an unwind-protect form, so that it will execute even if the previous code gets an error or uses throw. See Cleanups.

compile

Compile the combined definition that is used to run the advice. This flag is ignored unless activate is also specified. See Combined Definition.

disable

Initially disable this piece of advice, so that it will not be used unless subsequently explicitly enabled. See Enabling Advice.

preactivate

Activate advice for function when this defadvice is compiled or macroexpanded. This generates a compiled advised definition according to the current advice state, which will be used during activation if appropriate. See Preactivation.

This is useful only if this defadvice is byte-compiled.

The optional documentation-string serves to document this piece of advice. When advice is active for function, the documentation for function (as returned by documentation) combines the documentation strings of all the advice for function with the documentation string of its original function definition.

The optional interactive-form form can be supplied to change the interactive behavior of the original function. If more than one piece of advice has an interactive-form, then the first one (the one with the smallest position) found among all the advice takes precedence.

The possibly empty list of body-forms specifies the body of the advice. The body of an advice can access or change the arguments, the return value, the binding environment, and perform any other kind of side effect.

Warning: When you advise a macro, keep in mind that macros are expanded when a program is compiled, not when a compiled program is run. All subroutines used by the advice need to be available when the byte compiler expands the macro.

17.3 Around-Advice

Around-advice lets you “wrap” a Lisp expression “around” the original function definition. You specify where the original function definition should go by means of the special symbol ad-do-it. Where this symbol occurs inside the around-advice body, it is replaced with a progn containing the forms of the surrounded code. Here is an example:

     (defadvice foo (around foo-around)
       "Ignore case in `foo'."
       (let ((case-fold-search t))
         ad-do-it))

Its effect is to make sure that case is ignored in searches when the original definition of foo is run.

If the around-advice does not use ad-do-it, then it does not run the original function definition. This provides a way to override the original definition completely. (It also overrides lower-positioned pieces of around-advice).

If the around-advice uses ad-do-it more than once, the original definition is run at each place. In this way, around-advice can execute the original definition (and lower-positioned pieces of around-advice) several times. Another way to do that is by using ad-do-it inside of a loop.

17.4 Computed Advice

The macro defadvice resembles defun in that the code for the advice, and all other information about it, are explicitly stated in the source code. You can also create advice whose details are computed, using the function ad-add-advice.

17.5 Activation of Advice

By default, advice does not take effect when you define it—only when you activate advice for the function that was advised. However, the advice will be activated automatically if you define or redefine the function later. You can request the activation of advice for a function when you define the advice, by specifying the activate flag in the defadvice. But normally you activate the advice for a function by calling the function ad-activate or one of the other activation commands listed below.

Separating the activation of advice from the act of defining it permits you to add several pieces of advice to one function efficiently, without redefining the function over and over as each advice is added. More importantly, it permits defining advice for a function before that function is actually defined.

When a function's advice is first activated, the function's original definition is saved, and all enabled pieces of advice for that function are combined with the original definition to make a new definition. (Pieces of advice that are currently disabled are not used; see Enabling Advice.) This definition is installed, and optionally byte-compiled as well, depending on conditions described below.

In all of the commands to activate advice, if compile is t (or anything but nil or a negative number), the command also compiles the combined definition which implements the advice. If it is nil or a negative number, what happens depends on ad-default-compilation-action as described below.

Activating advice does nothing if function's advice is already active. But if there is new advice, added since the previous time you activated advice for function, it activates the new advice.

If the advised definition was constructed during “preactivation” (see Preactivation), then that definition must already be compiled, because it was constructed during byte-compilation of the file that contained the defadvice with the preactivate flag.

17.6 Enabling and Disabling Advice

Each piece of advice has a flag that says whether it is enabled or not. By enabling or disabling a piece of advice, you can turn it on and off without having to undefine and redefine it. For example, here is how to disable a particular piece of advice named my-advice for the function foo:

     (ad-disable-advice 'foo 'before 'my-advice)

This function by itself only changes the enable flag for a piece of advice. To make the change take effect in the advised definition, you must activate the advice for foo again:

     (ad-activate 'foo)

You can also disable many pieces of advice at once, for various functions, using a regular expression. As always, the changes take real effect only when you next reactivate advice for the functions in question.

17.7 Preactivation

Constructing a combined definition to execute advice is moderately expensive. When a library advises many functions, this can make loading the library slow. In that case, you can use preactivation to construct suitable combined definitions in advance.

To use preactivation, specify the preactivate flag when you define the advice with defadvice. This defadvice call creates a combined definition which embodies this piece of advice (whether enabled or not) plus any other currently enabled advice for the same function, and the function's own definition. If the defadvice is compiled, that compiles the combined definition also.

When the function's advice is subsequently activated, if the enabled advice for the function matches what was used to make this combined definition, then the existing combined definition is used, thus avoiding the need to construct one. Thus, preactivation never causes wrong results—but it may fail to do any good, if the enabled advice at the time of activation doesn't match what was used for preactivation.

Here are some symptoms that can indicate that a preactivation did not work properly, because of a mismatch.

• Activation of the advised function takes longer than usual.
• The byte compiler gets loaded while an advised function gets activated.
• byte-compile is included in the value of features even though you did not ever explicitly use the byte compiler.

Compiled preactivated advice works properly even if the function itself is not defined until later; however, the function needs to be defined when you compile the preactivated advice.

There is no elegant way to find out why preactivated advice is not being used. What you can do is to trace the function ad-cache-id-verification-code (with the function trace-function-background) before the advised function's advice is activated. After activation, check the value returned by ad-cache-id-verification-code for that function: verified means that the preactivated advice was used, while other values give some information about why they were considered inappropriate.

Warning: There is one known case that can make preactivation fail, in that a preconstructed combined definition is used even though it fails to match the current state of advice. This can happen when two packages define different pieces of advice with the same name, in the same class, for the same function. But you should avoid that anyway.

17.8 Argument Access in Advice

The simplest way to access the arguments of an advised function in the body of a piece of advice is to use the same names that the function definition uses. To do this, you need to know the names of the argument variables of the original function.

While this simple method is sufficient in many cases, it has a disadvantage: it is not robust, because it hard-codes the argument names into the advice. If the definition of the original function changes, the advice might break.

Another method is to specify an argument list in the advice itself. This avoids the need to know the original function definition's argument names, but it has a limitation: all the advice on any particular function must use the same argument list, because the argument list actually used for all the advice comes from the first piece of advice for that function.

A more robust method is to use macros that are translated into the proper access forms at activation time, i.e., when constructing the advised definition. Access macros access actual arguments by their (zero-based) position, regardless of how these actual arguments get distributed onto the argument variables of a function. This is robust because in Emacs Lisp the meaning of an argument is strictly determined by its position in the argument list.

Now an example. Suppose the function foo is defined as

     (defun foo (x y &optional z &rest r) ...)

and is then called with

     (foo 0 1 2 3 4 5 6)

which means that x is 0, y is 1, z is 2 and r is (3 4 5 6) within the body of foo. Here is what ad-get-arg and ad-get-args return in this case:

     (ad-get-arg 0) ⇒ 0
     (ad-get-arg 1) ⇒ 1
     (ad-get-arg 2) ⇒ 2
     (ad-get-arg 3) ⇒ 3
     (ad-get-args 2) ⇒ (2 3 4 5 6)
     (ad-get-args 4) ⇒ (4 5 6)

Setting arguments also makes sense in this example:

     (ad-set-arg 5 "five")

has the effect of changing the sixth argument to "five". If this happens in advice executed before the body of foo is run, then r will be (3 4 "five" 6) within that body.

Here is an example of setting a tail of the argument list:

     (ad-set-args 0 '(5 4 3 2 1 0))

If this happens in advice executed before the body of foo is run, then within that body, x will be 5, y will be 4, z will be 3, and r will be (2 1 0) inside the body of foo.

These argument constructs are not really implemented as Lisp macros. Instead they are implemented specially by the advice mechanism.

17.9 Advising Primitives

Advising a primitive function (see What Is a Function) is risky. Some primitive functions are used by the advice mechanism; advising them could cause an infinite recursion. Also, many primitive functions are called directly from C code. Calls to the primitive from Lisp code will take note of the advice, but calls from C code will ignore the advice.

When the advice facility constructs the combined definition, it needs to know the argument list of the original function. This is not always possible for primitive functions. When advice cannot determine the argument list, it uses (&rest ad-subr-args), which always works but is inefficient because it constructs a list of the argument values. You can use ad-define-subr-args to declare the proper argument names for a primitive function:

For example,

     (ad-define-subr-args 'fset '(sym newdef))

specifies the argument list for the function fset.

17.10 The Combined Definition

Suppose that a function has n pieces of before-advice (numbered from 0 through n−1), m pieces of around-advice and k pieces of after-advice. Assuming no piece of advice is protected, the combined definition produced to implement the advice for a function looks like this:

     (lambda arglist
       [ [advised-docstring] [(interactive ...)] ]
       (let (ad-return-value)
         before-0-body-form...
              ....
         before-n−1-body-form...
         around-0-body-form...
            around-1-body-form...
                  ....
               around-m−1-body-form...
                  (setq ad-return-value
                        apply original definition to arglist)
               end-of-around-m−1-body-form...
                  ....
            end-of-around-1-body-form...
         end-of-around-0-body-form...
         after-0-body-form...
               ....
         after-k−1-body-form...
         ad-return-value))

Macros are redefined as macros, which means adding macro to the beginning of the combined definition.

The interactive form is present if the original function or some piece of advice specifies one. When an interactive primitive function is advised, advice uses a special method: it calls the primitive with call-interactively so that it will read its own arguments. In this case, the advice cannot access the arguments.

The body forms of the various advice in each class are assembled according to their specified order. The forms of around-advice l are included in one of the forms of around-advice l − 1.

The innermost part of the around advice onion is

     apply original definition to arglist

whose form depends on the type of the original function. The variable ad-return-value is set to whatever this returns. The variable is visible to all pieces of advice, which can access and modify it before it is actually returned from the advised function.

The semantic structure of advised functions that contain protected pieces of advice is the same. The only difference is that unwind-protect forms ensure that the protected advice gets executed even if some previous piece of advice had an error or a non-local exit. If any around-advice is protected, then the whole around-advice onion is protected as a result.

18 Debugging Lisp Programs

There are three ways to investigate a problem in an Emacs Lisp program, depending on what you are doing with the program when the problem appears.

• If the problem occurs when you run the program, you can use a Lisp debugger to investigate what is happening during execution. In addition to the ordinary debugger, Emacs comes with a source-level debugger, Edebug. This chapter describes both of them.
• If the problem is syntactic, so that Lisp cannot even read the program, you can use the Emacs facilities for editing Lisp to localize it.
• If the problem occurs when trying to compile the program with the byte compiler, you need to know how to examine the compiler's input buffer.

• Debugger: How the Emacs Lisp debugger is implemented.
• Edebug: A source-level Emacs Lisp debugger.
• Syntax Errors: How to find syntax errors.
• Test Coverage: Ensuring you have tested all branches in your code.
• Compilation Errors: How to find errors that show up in byte compilation.

Another useful debugging tool is the dribble file. When a dribble file is open, Emacs copies all keyboard input characters to that file. Afterward, you can examine the file to find out what input was used. See Terminal Input.

For debugging problems in terminal descriptions, the open-termscript function can be useful. See Terminal Output.

18.1 The Lisp Debugger

The ordinary Lisp debugger provides the ability to suspend evaluation of a form. While evaluation is suspended (a state that is commonly known as a break), you may examine the run time stack, examine the values of local or global variables, or change those values. Since a break is a recursive edit, all the usual editing facilities of Emacs are available; you can even run programs that will enter the debugger recursively. See Recursive Editing.

• Error Debugging: Entering the debugger when an error happens.
• Infinite Loops: Stopping and debugging a program that doesn't exit.
• Function Debugging: Entering it when a certain function is called.
• Explicit Debug: Entering it at a certain point in the program.
• Using Debugger: What the debugger does; what you see while in it.
• Debugger Commands: Commands used while in the debugger.
• Invoking the Debugger: How to call the function debug.
• Internals of Debugger: Subroutines of the debugger, and global variables.

18.1.1 Entering the Debugger on an Error

The most important time to enter the debugger is when a Lisp error happens. This allows you to investigate the immediate causes of the error.

However, entry to the debugger is not a normal consequence of an error. Many commands frequently cause Lisp errors when invoked inappropriately, and during ordinary editing it would be very inconvenient to enter the debugger each time this happens. So if you want errors to enter the debugger, set the variable debug-on-error to non-nil. (The command toggle-debug-on-error provides an easy way to do this.)

To debug an error that happens during loading of the init file, use the option ‘--debug-init’. This binds debug-on-error to t while loading the init file, and bypasses the condition-case which normally catches errors in the init file.

18.1.2 Debugging Infinite Loops

When a program loops infinitely and fails to return, your first problem is to stop the loop. On most operating systems, you can do this with C-g, which causes a quit.

Ordinary quitting gives no information about why the program was looping. To get more information, you can set the variable debug-on-quit to non-nil. Quitting with C-g is not considered an error, and debug-on-error has no effect on the handling of C-g. Likewise, debug-on-quit has no effect on errors.

Once you have the debugger running in the middle of the infinite loop, you can proceed from the debugger using the stepping commands. If you step through the entire loop, you will probably get enough information to solve the problem.

18.1.3 Entering the Debugger on a Function Call

To investigate a problem that happens in the middle of a program, one useful technique is to enter the debugger whenever a certain function is called. You can do this to the function in which the problem occurs, and then step through the function, or you can do this to a function called shortly before the problem, step quickly over the call to that function, and then step through its caller.

18.1.4 Explicit Entry to the Debugger

You can cause the debugger to be called at a certain point in your program by writing the expression (debug) at that point. To do this, visit the source file, insert the text ‘(debug)’ at the proper place, and type C-M-x (eval-defun, a Lisp mode key binding). Warning: if you do this for temporary debugging purposes, be sure to undo this insertion before you save the file!

The place where you insert ‘(debug)’ must be a place where an additional form can be evaluated and its value ignored. (If the value of (debug) isn't ignored, it will alter the execution of the program!) The most common suitable places are inside a progn or an implicit progn (see Sequencing).

18.1.5 Using the Debugger

When the debugger is entered, it displays the previously selected buffer in one window and a buffer named ‘*Backtrace*’ in another window. The backtrace buffer contains one line for each level of Lisp function execution currently going on. At the beginning of this buffer is a message describing the reason that the debugger was invoked (such as the error message and associated data, if it was invoked due to an error).

The backtrace buffer is read-only and uses a special major mode, Debugger mode, in which letters are defined as debugger commands. The usual Emacs editing commands are available; thus, you can switch windows to examine the buffer that was being edited at the time of the error, switch buffers, visit files, or do any other sort of editing. However, the debugger is a recursive editing level (see Recursive Editing) and it is wise to go back to the backtrace buffer and exit the debugger (with the q command) when you are finished with it. Exiting the debugger gets out of the recursive edit and kills the backtrace buffer.

The backtrace buffer shows you the functions that are executing and their argument values. It also allows you to specify a stack frame by moving point to the line describing that frame. (A stack frame is the place where the Lisp interpreter records information about a particular invocation of a function.) The frame whose line point is on is considered the current frame. Some of the debugger commands operate on the current frame. If a line starts with a star, that means that exiting that frame will call the debugger again. This is useful for examining the return value of a function.

If a function name is underlined, that means the debugger knows where its source code is located. You can click Mouse-2 on that name, or move to it and type <RET>, to visit the source code.

The debugger itself must be run byte-compiled, since it makes assumptions about how many stack frames are used for the debugger itself. These assumptions are false if the debugger is running interpreted.

18.1.6 Debugger Commands

The debugger buffer (in Debugger mode) provides special commands in addition to the usual Emacs commands. The most important use of debugger commands is for stepping through code, so that you can see how control flows. The debugger can step through the control structures of an interpreted function, but cannot do so in a byte-compiled function. If you would like to step through a byte-compiled function, replace it with an interpreted definition of the same function. (To do this, visit the source for the function and type C-M-x on its definition.) You cannot use the Lisp debugger to step through a primitive function.

Here is a list of Debugger mode commands:

c

Exit the debugger and continue execution. When continuing is possible, it resumes execution of the program as if the debugger had never been entered (aside from any side-effects that you caused by changing variable values or data structures while inside the debugger).

Continuing is possible after entry to the debugger due to function entry or exit, explicit invocation, or quitting. You cannot continue if the debugger was entered because of an error.

d

Continue execution, but enter the debugger the next time any Lisp function is called. This allows you to step through the subexpressions of an expression, seeing what values the subexpressions compute, and what else they do.

The stack frame made for the function call which enters the debugger in this way will be flagged automatically so that the debugger will be called again when the frame is exited. You can use the u command to cancel this flag.

b

Flag the current frame so that the debugger will be entered when the frame is exited. Frames flagged in this way are marked with stars in the backtrace buffer.

u

Don't enter the debugger when the current frame is exited. This cancels a b command on that frame. The visible effect is to remove the star from the line in the backtrace buffer.

j

Flag the current frame like b. Then continue execution like c, but temporarily disable break-on-entry for all functions that are set up to do so by debug-on-entry.

e

Read a Lisp expression in the minibuffer, evaluate it, and print the value in the echo area. The debugger alters certain important variables, and the current buffer, as part of its operation; e temporarily restores their values from outside the debugger, so you can examine and change them. This makes the debugger more transparent. By contrast, M-: does nothing special in the debugger; it shows you the variable values within the debugger.

R

Like e, but also save the result of evaluation in the buffer ‘*Debugger-record*’.

q

Terminate the program being debugged; return to top-level Emacs command execution.

If the debugger was entered due to a C-g but you really want to quit, and not debug, use the q command.

r

Return a value from the debugger. The value is computed by reading an expression with the minibuffer and evaluating it.

The r command is useful when the debugger was invoked due to exit from a Lisp call frame (as requested with b or by entering the frame with d); then the value specified in the r command is used as the value of that frame. It is also useful if you call debug and use its return value. Otherwise, r has the same effect as c, and the specified return value does not matter.

You can't use r when the debugger was entered due to an error.

l

Display a list of functions that will invoke the debugger when called. This is a list of functions that are set to break on entry by means of debug-on-entry. Warning: if you redefine such a function and thus cancel the effect of debug-on-entry, it may erroneously show up in this list.

18.1.7 Invoking the Debugger

Here we describe in full detail the function debug that is used to invoke the debugger.

18.1.8 Internals of the Debugger

This section describes functions and variables used internally by the debugger.

18.2 Edebug

Edebug is a source-level debugger for Emacs Lisp programs, with which you can:

• Step through evaluation, stopping before and after each expression.
• Set conditional or unconditional breakpoints.
• Stop when a specified condition is true (the global break event).
• Trace slow or fast, stopping briefly at each stop point, or at each breakpoint.
• Display expression results and evaluate expressions as if outside of Edebug.
• Automatically re-evaluate a list of expressions and display their results each time Edebug updates the display.
• Output trace information on function calls and returns.
• Stop when an error occurs.
• Display a backtrace, omitting Edebug's own frames.
• Specify argument evaluation for macros and defining forms.
• Obtain rudimentary coverage testing and frequency counts.

The first three sections below should tell you enough about Edebug to start using it.

• Using Edebug: Introduction to use of Edebug.
• Instrumenting: You must instrument your code in order to debug it with Edebug.
• Modes: Execution modes, stopping more or less often.
• Jumping: Commands to jump to a specified place.
• Misc: Miscellaneous commands.
• Breaks: Setting breakpoints to make the program stop.
• Trapping Errors: Trapping errors with Edebug.
• Views: Views inside and outside of Edebug.
• Eval: Evaluating expressions within Edebug.
• Eval List: Expressions whose values are displayed each time you enter Edebug.
• Printing in Edebug: Customization of printing.
• Trace Buffer: How to produce trace output in a buffer.
• Coverage Testing: How to test evaluation coverage.
• The Outside Context: Data that Edebug saves and restores.
• Edebug and Macros: Specifying how to handle macro calls.
• Options: Option variables for customizing Edebug.

18.2.1 Using Edebug

To debug a Lisp program with Edebug, you must first instrument the Lisp code that you want to debug. A simple way to do this is to first move point into the definition of a function or macro and then do C-u C-M-x (eval-defun with a prefix argument). See Instrumenting, for alternative ways to instrument code.

Once a function is instrumented, any call to the function activates Edebug. Depending on which Edebug execution mode you have selected, activating Edebug may stop execution and let you step through the function, or it may update the display and continue execution while checking for debugging commands. The default execution mode is step, which stops execution. See Edebug Execution Modes.

Within Edebug, you normally view an Emacs buffer showing the source of the Lisp code you are debugging. This is referred to as the source code buffer, and it is temporarily read-only.

An arrow in the left fringe indicates the line where the function is executing. Point initially shows where within the line the function is executing, but this ceases to be true if you move point yourself.

If you instrument the definition of fac (shown below) and then execute (fac 3), here is what you would normally see. Point is at the open-parenthesis before if.

     (defun fac (n)
     =>-!-(if (< 0 n)
           (* n (fac (1- n)))
         1))

The places within a function where Edebug can stop execution are called stop points. These occur both before and after each subexpression that is a list, and also after each variable reference. Here we use periods to show the stop points in the function fac:

     (defun fac (n)
       .(if .(< 0 n.).
           .(* n. .(fac .(1- n.).).).
         1).)

The special commands of Edebug are available in the source code buffer in addition to the commands of Emacs Lisp mode. For example, you can type the Edebug command <SPC> to execute until the next stop point. If you type <SPC> once after entry to fac, here is the display you will see:

     (defun fac (n)
     =>(if -!-(< 0 n)
           (* n (fac (1- n)))
         1))

When Edebug stops execution after an expression, it displays the expression's value in the echo area.

Other frequently used commands are b to set a breakpoint at a stop point, g to execute until a breakpoint is reached, and q to exit Edebug and return to the top-level command loop. Type ? to display a list of all Edebug commands.

18.2.2 Instrumenting for Edebug

In order to use Edebug to debug Lisp code, you must first instrument the code. Instrumenting code inserts additional code into it, to invoke Edebug at the proper places.

When you invoke command C-M-x (eval-defun) with a prefix argument on a function definition, it instruments the definition before evaluating it. (This does not modify the source code itself.) If the variable edebug-all-defs is non-nil, that inverts the meaning of the prefix argument: in this case, C-M-x instruments the definition unless it has a prefix argument. The default value of edebug-all-defs is nil. The command M-x edebug-all-defs toggles the value of the variable edebug-all-defs.

If edebug-all-defs is non-nil, then the commands eval-region, eval-current-buffer, and eval-buffer also instrument any definitions they evaluate. Similarly, edebug-all-forms controls whether eval-region should instrument any form, even non-defining forms. This doesn't apply to loading or evaluations in the minibuffer. The command M-x edebug-all-forms toggles this option.

Another command, M-x edebug-eval-top-level-form, is available to instrument any top-level form regardless of the values of edebug-all-defs and edebug-all-forms.

While Edebug is active, the command I (edebug-instrument-callee) instruments the definition of the function or macro called by the list form after point, if it is not already instrumented. This is possible only if Edebug knows where to find the source for that function; for this reason, after loading Edebug, eval-region records the position of every definition it evaluates, even if not instrumenting it. See also the i command (see Jumping), which steps into the call after instrumenting the function.

Edebug knows how to instrument all the standard special forms, interactive forms with an expression argument, anonymous lambda expressions, and other defining forms. However, Edebug cannot determine on its own what a user-defined macro will do with the arguments of a macro call, so you must provide that information using Edebug specifications; for details, see Edebug and Macros.

When Edebug is about to instrument code for the first time in a session, it runs the hook edebug-setup-hook, then sets it to nil. You can use this to load Edebug specifications associated with a package you are using, but only when you use Edebug.

To remove instrumentation from a definition, simply re-evaluate its definition in a way that does not instrument. There are two ways of evaluating forms that never instrument them: from a file with load, and from the minibuffer with eval-expression (M-:).

If Edebug detects a syntax error while instrumenting, it leaves point at the erroneous code and signals an invalid-read-syntax error.

See Edebug Eval, for other evaluation functions available inside of Edebug.

18.2.3 Edebug Execution Modes

Edebug supports several execution modes for running the program you are debugging. We call these alternatives Edebug execution modes; do not confuse them with major or minor modes. The current Edebug execution mode determines how far Edebug continues execution before stopping—whether it stops at each stop point, or continues to the next breakpoint, for example—and how much Edebug displays the progress of the evaluation before it stops.

Normally, you specify the Edebug execution mode by typing a command to continue the program in a certain mode. Here is a table of these commands; all except for S resume execution of the program, at least for a certain distance.

S

Stop: don't execute any more of the program, but wait for more Edebug commands (edebug-stop).

<SPC>

Step: stop at the next stop point encountered (edebug-step-mode).

n

Next: stop at the next stop point encountered after an expression (edebug-next-mode). Also see edebug-forward-sexp in Jumping.

t

Trace: pause (normally one second) at each Edebug stop point (edebug-trace-mode).

T

Rapid trace: update the display at each stop point, but don't actually pause (edebug-Trace-fast-mode).

g

Go: run until the next breakpoint (edebug-go-mode). See Breakpoints.

c

Continue: pause one second at each breakpoint, and then continue (edebug-continue-mode).

C

Rapid continue: move point to each breakpoint, but don't pause (edebug-Continue-fast-mode).

G

Go non-stop: ignore breakpoints (edebug-Go-nonstop-mode). You can still stop the program by typing S, or any editing command.

In general, the execution modes earlier in the above list run the program more slowly or stop sooner than the modes later in the list.

While executing or tracing, you can interrupt the execution by typing any Edebug command. Edebug stops the program at the next stop point and then executes the command you typed. For example, typing t during execution switches to trace mode at the next stop point. You can use S to stop execution without doing anything else.

If your function happens to read input, a character you type intending to interrupt execution may be read by the function instead. You can avoid such unintended results by paying attention to when your program wants input.

Keyboard macros containing the commands in this section do not completely work: exiting from Edebug, to resume the program, loses track of the keyboard macro. This is not easy to fix. Also, defining or executing a keyboard macro outside of Edebug does not affect commands inside Edebug. This is usually an advantage. See also the edebug-continue-kbd-macro option in Edebug Options.

When you enter a new Edebug level, the initial execution mode comes from the value of the variable edebug-initial-mode (see Edebug Options). By default, this specifies step mode. Note that you may reenter the same Edebug level several times if, for example, an instrumented function is called several times from one command.

18.2.4 Jumping

The commands described in this section execute until they reach a specified location. All except i make a temporary breakpoint to establish the place to stop, then switch to go mode. Any other breakpoint reached before the intended stop point will also stop execution. See Breakpoints, for the details on breakpoints.

These commands may fail to work as expected in case of nonlocal exit, as that can bypass the temporary breakpoint where you expected the program to stop.

h

Proceed to the stop point near where point is (edebug-goto-here).

f

Run the program for one expression (edebug-forward-sexp).

o

Run the program until the end of the containing sexp (edebug-step-out).

i

Step into the function or macro called by the form after point.

The h command proceeds to the stop point at or after the current location of point, using a temporary breakpoint.

The f command runs the program forward over one expression. More precisely, it sets a temporary breakpoint at the position that forward-sexp would reach, then executes in go mode so that the program will stop at breakpoints.

With a prefix argument n, the temporary breakpoint is placed n sexps beyond point. If the containing list ends before n more elements, then the place to stop is after the containing expression.

You must check that the position forward-sexp finds is a place that the program will really get to. In cond, for example, this may not be true.

For flexibility, the f command does forward-sexp starting at point, rather than at the stop point. If you want to execute one expression from the current stop point, first type w (edebug-where) to move point there, and then type f.

The o command continues “out of” an expression. It places a temporary breakpoint at the end of the sexp containing point. If the containing sexp is a function definition itself, o continues until just before the last sexp in the definition. If that is where you are now, it returns from the function and then stops. In other words, this command does not exit the currently executing function unless you are positioned after the last sexp.

The i command steps into the function or macro called by the list form after point, and stops at its first stop point. Note that the form need not be the one about to be evaluated. But if the form is a function call about to be evaluated, remember to use this command before any of the arguments are evaluated, since otherwise it will be too late.

The i command instruments the function or macro it's supposed to step into, if it isn't instrumented already. This is convenient, but keep in mind that the function or macro remains instrumented unless you explicitly arrange to deinstrument it.

18.2.5 Miscellaneous Edebug Commands

Some miscellaneous Edebug commands are described here.

?

Display the help message for Edebug (edebug-help).

C-]

Abort one level back to the previous command level (abort-recursive-edit).

q

Return to the top level editor command loop (top-level). This exits all recursive editing levels, including all levels of Edebug activity. However, instrumented code protected with unwind-protect or condition-case forms may resume debugging.

Q

Like q, but don't stop even for protected code (edebug-top-level-nonstop).

r

Redisplay the most recently known expression result in the echo area (edebug-previous-result).

d

Display a backtrace, excluding Edebug's own functions for clarity (edebug-backtrace).

You cannot use debugger commands in the backtrace buffer in Edebug as you would in the standard debugger.

The backtrace buffer is killed automatically when you continue execution.

You can invoke commands from Edebug that activate Edebug again recursively. Whenever Edebug is active, you can quit to the top level with q or abort one recursive edit level with C-]. You can display a backtrace of all the pending evaluations with d.

18.2.6 Breaks

Edebug's step mode stops execution when the next stop point is reached. There are three other ways to stop Edebug execution once it has started: breakpoints, the global break condition, and source breakpoints.

• Breakpoints: Breakpoints at stop points.
• Global Break Condition: Breaking on an event.
• Source Breakpoints: Embedding breakpoints in source code.

18.2.6.1 Edebug Breakpoints

While using Edebug, you can specify breakpoints in the program you are testing: these are places where execution should stop. You can set a breakpoint at any stop point, as defined in Using Edebug. For setting and unsetting breakpoints, the stop point that is affected is the first one at or after point in the source code buffer. Here are the Edebug commands for breakpoints:

b

Set a breakpoint at the stop point at or after point (edebug-set-breakpoint). If you use a prefix argument, the breakpoint is temporary—it turns off the first time it stops the program.

u

Unset the breakpoint (if any) at the stop point at or after point (edebug-unset-breakpoint).

x condition <RET>

Set a conditional breakpoint which stops the program only if evaluating condition produces a non-nil value (edebug-set-conditional-breakpoint). With a prefix argument, the breakpoint is temporary.

B

Move point to the next breakpoint in the current definition (edebug-next-breakpoint).

While in Edebug, you can set a breakpoint with b and unset one with u. First move point to the Edebug stop point of your choice, then type b or u to set or unset a breakpoint there. Unsetting a breakpoint where none has been set has no effect.

Re-evaluating or reinstrumenting a definition removes all of its previous breakpoints.

A conditional breakpoint tests a condition each time the program gets there. Any errors that occur as a result of evaluating the condition are ignored, as if the result were nil. To set a conditional breakpoint, use x, and specify the condition expression in the minibuffer. Setting a conditional breakpoint at a stop point that has a previously established conditional breakpoint puts the previous condition expression in the minibuffer so you can edit it.

You can make a conditional or unconditional breakpoint temporary by using a prefix argument with the command to set the breakpoint. When a temporary breakpoint stops the program, it is automatically unset.

Edebug always stops or pauses at a breakpoint, except when the Edebug mode is Go-nonstop. In that mode, it ignores breakpoints entirely.

To find out where your breakpoints are, use the B command, which moves point to the next breakpoint following point, within the same function, or to the first breakpoint if there are no following breakpoints. This command does not continue execution—it just moves point in the buffer.

18.2.6.2 Global Break Condition

A global break condition stops execution when a specified condition is satisfied, no matter where that may occur. Edebug evaluates the global break condition at every stop point; if it evaluates to a non-nil value, then execution stops or pauses depending on the execution mode, as if a breakpoint had been hit. If evaluating the condition gets an error, execution does not stop.

The condition expression is stored in edebug-global-break-condition. You can specify a new expression using the X command from the source code buffer while Edebug is active, or using C-x X X from any buffer at any time, as long as Edebug is loaded (edebug-set-global-break-condition).

The global break condition is the simplest way to find where in your code some event occurs, but it makes code run much more slowly. So you should reset the condition to nil when not using it.

18.2.6.3 Source Breakpoints

All breakpoints in a definition are forgotten each time you reinstrument it. If you wish to make a breakpoint that won't be forgotten, you can write a source breakpoint, which is simply a call to the function edebug in your source code. You can, of course, make such a call conditional. For example, in the fac function, you can insert the first line as shown below, to stop when the argument reaches zero:

     (defun fac (n)
       (if (= n 0) (edebug))
       (if (< 0 n)
           (* n (fac (1- n)))
         1))

When the fac definition is instrumented and the function is called, the call to edebug acts as a breakpoint. Depending on the execution mode, Edebug stops or pauses there.

If no instrumented code is being executed when edebug is called, that function calls debug.

18.2.7 Trapping Errors

Emacs normally displays an error message when an error is signaled and not handled with condition-case. While Edebug is active and executing instrumented code, it normally responds to all unhandled errors. You can customize this with the options edebug-on-error and edebug-on-quit; see Edebug Options.

When Edebug responds to an error, it shows the last stop point encountered before the error. This may be the location of a call to a function which was not instrumented, and within which the error actually occurred. For an unbound variable error, the last known stop point might be quite distant from the offending variable reference. In that case, you might want to display a full backtrace (see Edebug Misc).

If you change debug-on-error or debug-on-quit while Edebug is active, these changes will be forgotten when Edebug becomes inactive. Furthermore, during Edebug's recursive edit, these variables are bound to the values they had outside of Edebug.

18.2.8 Edebug Views

These Edebug commands let you view aspects of the buffer and window status as they were before entry to Edebug. The outside window configuration is the collection of windows and contents that were in effect outside of Edebug.

v

Switch to viewing the outside window configuration (edebug-view-outside). Type C-x X w to return to Edebug.

p

Temporarily display the outside current buffer with point at its outside position (edebug-bounce-point), pausing for one second before returning to Edebug. With a prefix argument n, pause for n seconds instead.

w

Move point back to the current stop point in the source code buffer (edebug-where).

If you use this command in a different window displaying the same buffer, that window will be used instead to display the current definition in the future.

W

Toggle whether Edebug saves and restores the outside window configuration (edebug-toggle-save-windows).

With a prefix argument, W only toggles saving and restoring of the selected window. To specify a window that is not displaying the source code buffer, you must use C-x X W from the global keymap.

You can view the outside window configuration with v or just bounce to the point in the current buffer with p, even if it is not normally displayed.

After moving point, you may wish to jump back to the stop point. You can do that with w from a source code buffer. You can jump back to the stop point in the source code buffer from any buffer using C-x X w.

Each time you use W to turn saving off, Edebug forgets the saved outside window configuration—so that even if you turn saving back on, the current window configuration remains unchanged when you next exit Edebug (by continuing the program). However, the automatic redisplay of ‘*edebug*’ and ‘*edebug-trace*’ may conflict with the buffers you wish to see unless you have enough windows open.

18.2.9 Evaluation

While within Edebug, you can evaluate expressions “as if” Edebug were not running. Edebug tries to be invisible to the expression's evaluation and printing. Evaluation of expressions that cause side effects will work as expected, except for changes to data that Edebug explicitly saves and restores. See The Outside Context, for details on this process.

e exp <RET>

Evaluate expression exp in the context outside of Edebug (edebug-eval-expression). That is, Edebug tries to minimize its interference with the evaluation.

M-: exp <RET>

Evaluate expression exp in the context of Edebug itself (eval-expression).

C-x C-e

Evaluate the expression before point, in the context outside of Edebug (edebug-eval-last-sexp).

Edebug supports evaluation of expressions containing references to lexically bound symbols created by the following constructs in cl.el: lexical-let, macrolet, and symbol-macrolet.

18.2.10 Evaluation List Buffer

You can use the evaluation list buffer, called ‘*edebug*’, to evaluate expressions interactively. You can also set up the evaluation list of expressions to be evaluated automatically each time Edebug updates the display.

E

Switch to the evaluation list buffer ‘*edebug*’ (edebug-visit-eval-list).

In the ‘*edebug*’ buffer you can use the commands of Lisp Interaction mode (see Lisp Interaction) as well as these special commands:

C-j

Evaluate the expression before point, in the outside context, and insert the value in the buffer (edebug-eval-print-last-sexp).

C-x C-e

Evaluate the expression before point, in the context outside of Edebug (edebug-eval-last-sexp).

C-c C-u

Build a new evaluation list from the contents of the buffer (edebug-update-eval-list).

C-c C-d

Delete the evaluation list group that point is in (edebug-delete-eval-item).

C-c C-w

Switch back to the source code buffer at the current stop point (edebug-where).

You can evaluate expressions in the evaluation list window with C-j or C-x C-e, just as you would in ‘*scratch*’; but they are evaluated in the context outside of Edebug.

The expressions you enter interactively (and their results) are lost when you continue execution; but you can set up an evaluation list consisting of expressions to be evaluated each time execution stops.

To do this, write one or more evaluation list groups in the evaluation list buffer. An evaluation list group consists of one or more Lisp expressions. Groups are separated by comment lines.

The command C-c C-u (edebug-update-eval-list) rebuilds the evaluation list, scanning the buffer and using the first expression of each group. (The idea is that the second expression of the group is the value previously computed and displayed.)

Each entry to Edebug redisplays the evaluation list by inserting each expression in the buffer, followed by its current value. It also inserts comment lines so that each expression becomes its own group. Thus, if you type C-c C-u again without changing the buffer text, the evaluation list is effectively unchanged.

If an error occurs during an evaluation from the evaluation list, the error message is displayed in a string as if it were the result. Therefore, expressions using variables that are not currently valid do not interrupt your debugging.

Here is an example of what the evaluation list window looks like after several expressions have been added to it:

     (current-buffer)
     #<buffer *scratch*>
     ;---------------------------------------------------------------
     (selected-window)
     #<window 16 on *scratch*>
     ;---------------------------------------------------------------
     (point)
     196
     ;---------------------------------------------------------------
     bad-var
     "Symbol's value as variable is void: bad-var"
     ;---------------------------------------------------------------
     (recursion-depth)
     0
     ;---------------------------------------------------------------
     this-command
     eval-last-sexp
     ;---------------------------------------------------------------

To delete a group, move point into it and type C-c C-d, or simply delete the text for the group and update the evaluation list with C-c C-u. To add a new expression to the evaluation list, insert the expression at a suitable place, insert a new comment line, then type C-c C-u. You need not insert dashes in the comment line—its contents don't matter.

After selecting ‘*edebug*’, you can return to the source code buffer with C-c C-w. The ‘*edebug*’ buffer is killed when you continue execution, and recreated next time it is needed.

18.2.11 Printing in Edebug

If an expression in your program produces a value containing circular list structure, you may get an error when Edebug attempts to print it.

One way to cope with circular structure is to set print-length or print-level to truncate the printing. Edebug does this for you; it binds print-length and print-level to the values of the variables edebug-print-length and edebug-print-level (so long as they have non-nil values). See Output Variables.

You can also print circular structures and structures that share elements more informatively by binding print-circle to a non-nil value.

Here is an example of code that creates a circular structure:

     (setq a '(x y))
     (setcar a a)

Custom printing prints this as ‘Result: #1=(#1# y)’. The ‘#1=’ notation labels the structure that follows it with the label ‘1’, and the ‘#1#’ notation references the previously labeled structure. This notation is used for any shared elements of lists or vectors.

Other programs can also use custom printing; see cust-print.el for details.

18.2.12 Trace Buffer

Edebug can record an execution trace, storing it in a buffer named ‘*edebug-trace*’. This is a log of function calls and returns, showing the function names and their arguments and values. To enable trace recording, set edebug-trace to a non-nil value.

Making a trace buffer is not the same thing as using trace execution mode (see Edebug Execution Modes).

When trace recording is enabled, each function entry and exit adds lines to the trace buffer. A function entry record consists of ‘::::{’, followed by the function name and argument values. A function exit record consists of ‘::::}’, followed by the function name and result of the function.

The number of ‘:’s in an entry shows its recursion depth. You can use the braces in the trace buffer to find the matching beginning or end of function calls.

You can customize trace recording for function entry and exit by redefining the functions edebug-print-trace-before and edebug-print-trace-after.

edebug-tracing and edebug-trace insert lines in the trace buffer whenever they are called, even if Edebug is not active. Adding text to the trace buffer also scrolls its window to show the last lines inserted.

18.2.13 Coverage Testing

Edebug provides rudimentary coverage testing and display of execution frequency.

Coverage testing works by comparing the result of each expression with the previous result; each form in the program is considered “covered” if it has returned two different values since you began testing coverage in the current Emacs session. Thus, to do coverage testing on your program, execute it under various conditions and note whether it behaves correctly; Edebug will tell you when you have tried enough different conditions that each form has returned two different values.

Coverage testing makes execution slower, so it is only done if edebug-test-coverage is non-nil. Frequency counting is performed for all executions of an instrumented function, even if the execution mode is Go-nonstop, and regardless of whether coverage testing is enabled.

Use C-x X = (edebug-display-freq-count) to display both the coverage information and the frequency counts for a definition. Just = (edebug-temp-display-freq-count) displays the same information temporarily, only until you type another key.

For example, after evaluating (fac 5) with a source breakpoint, and setting edebug-test-coverage to t, when the breakpoint is reached, the frequency data looks like this:

     (defun fac (n)
       (if (= n 0) (edebug))
     ;#6           1      = =5
       (if (< 0 n)
     ;#5         =
           (* n (fac (1- n)))
     ;#    5               0
         1))
     ;#   0

The comment lines show that fac was called 6 times. The first if statement returned 5 times with the same result each time; the same is true of the condition on the second if. The recursive call of fac did not return at all.

18.2.14 The Outside Context

Edebug tries to be transparent to the program you are debugging, but it does not succeed completely. Edebug also tries to be transparent when you evaluate expressions with e or with the evaluation list buffer, by temporarily restoring the outside context. This section explains precisely what context Edebug restores, and how Edebug fails to be completely transparent.

• Checking Whether to Stop: When Edebug decides what to do.
• Edebug Display Update: When Edebug updates the display.
• Edebug Recursive Edit: When Edebug stops execution.

18.2.14.1 Checking Whether to Stop

Whenever Edebug is entered, it needs to save and restore certain data before even deciding whether to make trace information or stop the program.

• max-lisp-eval-depth and max-specpdl-size are both increased to reduce Edebug's impact on the stack. You could, however, still run out of stack space when using Edebug.
• The state of keyboard macro execution is saved and restored. While Edebug is active, executing-kbd-macro is bound to nil unless edebug-continue-kbd-macro is non-nil.

18.2.14.2 Edebug Display Update

When Edebug needs to display something (e.g., in trace mode), it saves the current window configuration from “outside” Edebug (see Window Configurations). When you exit Edebug (by continuing the program), it restores the previous window configuration.

Emacs redisplays only when it pauses. Usually, when you continue execution, the program re-enters Edebug at a breakpoint or after stepping, without pausing or reading input in between. In such cases, Emacs never gets a chance to redisplay the “outside” configuration. Consequently, what you see is the same window configuration as the last time Edebug was active, with no interruption.

Entry to Edebug for displaying something also saves and restores the following data (though some of them are deliberately not restored if an error or quit signal occurs).

• Which buffer is current, and the positions of point and the mark in the current buffer, are saved and restored.
• The outside window configuration is saved and restored if edebug-save-windows is non-nil (see Edebug Options).

  The window configuration is not restored on error or quit, but the outside selected window is reselected even on error or quit in case a save-excursion is active. If the value of edebug-save-windows is a list, only the listed windows are saved and restored.

  The window start and horizontal scrolling of the source code buffer are not restored, however, so that the display remains coherent within Edebug.

• The value of point in each displayed buffer is saved and restored if edebug-save-displayed-buffer-points is non-nil.
• The variables overlay-arrow-position and overlay-arrow-string are saved and restored, so you can safely invoke Edebug from the recursive edit elsewhere in the same buffer.
• cursor-in-echo-area is locally bound to nil so that the cursor shows up in the window.

18.2.14.3 Edebug Recursive Edit

When Edebug is entered and actually reads commands from the user, it saves (and later restores) these additional data:

• The current match data. See Match Data.
• The variables last-command, this-command, last-input-event, last-command-event, last-event-frame, last-nonmenu-event, and track-mouse. Commands used within Edebug do not affect these variables outside of Edebug.

  Executing commands within Edebug can change the key sequence that would be returned by this-command-keys, and there is no way to reset the key sequence from Lisp.

  Edebug cannot save and restore the value of unread-command-events. Entering Edebug while this variable has a nontrivial value can interfere with execution of the program you are debugging.

• Complex commands executed while in Edebug are added to the variable command-history. In rare cases this can alter execution.
• Within Edebug, the recursion depth appears one deeper than the recursion depth outside Edebug. This is not true of the automatically updated evaluation list window.
• standard-output and standard-input are bound to nil by the recursive-edit, but Edebug temporarily restores them during evaluations.
• The state of keyboard macro definition is saved and restored. While Edebug is active, defining-kbd-macro is bound to edebug-continue-kbd-macro.

18.2.15 Edebug and Macros

To make Edebug properly instrument expressions that call macros, some extra care is needed. This subsection explains the details.

• Instrumenting Macro Calls: The basic problem.
• Specification List: How to specify complex patterns of evaluation.
• Backtracking: What Edebug does when matching fails.
• Specification Examples: To help understand specifications.

18.2.15.1 Instrumenting Macro Calls

When Edebug instruments an expression that calls a Lisp macro, it needs additional information about the macro to do the job properly. This is because there is no a-priori way to tell which subexpressions of the macro call are forms to be evaluated. (Evaluation may occur explicitly in the macro body, or when the resulting expansion is evaluated, or any time later.)

Therefore, you must define an Edebug specification for each macro that Edebug will encounter, to explain the format of calls to that macro. To do this, add a debug declaration to the macro definition. Here is a simple example that shows the specification for the for example macro (see Argument Evaluation).

     (defmacro for (var from init to final do &rest body)
       "Execute a simple \"for\" loop.
     For example, (for i from 1 to 10 do (print i))."
       (declare (debug (symbolp "from" form "to" form "do" &rest form)))
       ...)

The Edebug specification says which parts of a call to the macro are forms to be evaluated. For simple macros, the specification often looks very similar to the formal argument list of the macro definition, but specifications are much more general than macro arguments. See Defining Macros, for more explanation of the declare form.

You can also define an edebug specification for a macro separately from the macro definition with def-edebug-spec. Adding debug declarations is preferred, and more convenient, for macro definitions in Lisp, but def-edebug-spec makes it possible to define Edebug specifications for special forms implemented in C.

Here is a table of the possibilities for specification and how each directs processing of arguments.

t

All arguments are instrumented for evaluation.

0

None of the arguments is instrumented.

a symbol

The symbol must have an Edebug specification, which is used instead. This indirection is repeated until another kind of specification is found. This allows you to inherit the specification from another macro.

a list

The elements of the list describe the types of the arguments of a calling form. The possible elements of a specification list are described in the following sections.

If a macro has no Edebug specification, neither through a debug declaration nor through a def-edebug-spec call, the variable edebug-eval-macro-args comes into play.

18.2.15.2 Specification List

A specification list is required for an Edebug specification if some arguments of a macro call are evaluated while others are not. Some elements in a specification list match one or more arguments, but others modify the processing of all following elements. The latter, called specification keywords, are symbols beginning with ‘&’ (such as &optional).

A specification list may contain sublists which match arguments that are themselves lists, or it may contain vectors used for grouping. Sublists and groups thus subdivide the specification list into a hierarchy of levels. Specification keywords apply only to the remainder of the sublist or group they are contained in.

When a specification list involves alternatives or repetition, matching it against an actual macro call may require backtracking. For more details, see Backtracking.

Edebug specifications provide the power of regular expression matching, plus some context-free grammar constructs: the matching of sublists with balanced parentheses, recursive processing of forms, and recursion via indirect specifications.

Here's a table of the possible elements of a specification list, with their meanings (see Specification Examples, for the referenced examples):

sexp

A single unevaluated Lisp object, which is not instrumented.

form

A single evaluated expression, which is instrumented.

place

A place to store a value, as in the Common Lisp setf construct.

body

Short for &rest form. See &rest below.

function-form

A function form: either a quoted function symbol, a quoted lambda expression, or a form (that should evaluate to a function symbol or lambda expression). This is useful when an argument that's a lambda expression might be quoted with quote rather than function, since it instruments the body of the lambda expression either way.

lambda-expr

A lambda expression with no quoting.

&optional

All following elements in the specification list are optional; as soon as one does not match, Edebug stops matching at this level.

To make just a few elements optional followed by non-optional elements, use [&optional specs...]. To specify that several elements must all match or none, use &optional [specs...]. See the defun example.

&rest

All following elements in the specification list are repeated zero or more times. In the last repetition, however, it is not a problem if the expression runs out before matching all of the elements of the specification list.

To repeat only a few elements, use [&rest specs...]. To specify several elements that must all match on every repetition, use &rest [specs...].

&or

Each of the following elements in the specification list is an alternative. One of the alternatives must match, or the &or specification fails.

Each list element following &or is a single alternative. To group two or more list elements as a single alternative, enclose them in [...].

&not

Each of the following elements is matched as alternatives as if by using &or, but if any of them match, the specification fails. If none of them match, nothing is matched, but the &not specification succeeds.

&define

Indicates that the specification is for a defining form. The defining form itself is not instrumented (that is, Edebug does not stop before and after the defining form), but forms inside it typically will be instrumented. The &define keyword should be the first element in a list specification.

nil

This is successful when there are no more arguments to match at the current argument list level; otherwise it fails. See sublist specifications and the backquote example.

gate

No argument is matched but backtracking through the gate is disabled while matching the remainder of the specifications at this level. This is primarily used to generate more specific syntax error messages. See Backtracking, for more details. Also see the let example.

other-symbol

Any other symbol in a specification list may be a predicate or an indirect specification.

If the symbol has an Edebug specification, this indirect specification should be either a list specification that is used in place of the symbol, or a function that is called to process the arguments. The specification may be defined with def-edebug-spec just as for macros. See the defun example.

Otherwise, the symbol should be a predicate. The predicate is called with the argument and the specification fails if the predicate returns nil, and the argument is not instrumented.

Some suitable predicates include symbolp, integerp, stringp, vectorp, and atom.

[elements...]

A vector of elements groups the elements into a single group specification. Its meaning has nothing to do with vectors.

"string"

The argument should be a symbol named string. This specification is equivalent to the quoted symbol, 'symbol, where the name of symbol is the string, but the string form is preferred.

(vector elements...)

The argument should be a vector whose elements must match the elements in the specification. See the backquote example.

(elements...)

Any other list is a sublist specification and the argument must be a list whose elements match the specification elements.

A sublist specification may be a dotted list and the corresponding list argument may then be a dotted list. Alternatively, the last cdr of a dotted list specification may be another sublist specification (via a grouping or an indirect specification, e.g., (spec . [(more specs...)])) whose elements match the non-dotted list arguments. This is useful in recursive specifications such as in the backquote example. Also see the description of a nil specification above for terminating such recursion.

Note that a sublist specification written as (specs . nil) is equivalent to (specs), and (specs . (sublist-elements...)) is equivalent to (specs sublist-elements...).

Here is a list of additional specifications that may appear only after &define. See the defun example.

name

The argument, a symbol, is the name of the defining form.

A defining form is not required to have a name field; and it may have multiple name fields.

:name

This construct does not actually match an argument. The element following :name should be a symbol; it is used as an additional name component for the definition. You can use this to add a unique, static component to the name of the definition. It may be used more than once.

arg

The argument, a symbol, is the name of an argument of the defining form. However, lambda-list keywords (symbols starting with ‘&’) are not allowed.

lambda-list

This matches a lambda list—the argument list of a lambda expression.

def-body

The argument is the body of code in a definition. This is like body, described above, but a definition body must be instrumented with a different Edebug call that looks up information associated with the definition. Use def-body for the highest level list of forms within the definition.

def-form

The argument is a single, highest-level form in a definition. This is like def-body, except it is used to match a single form rather than a list of forms. As a special case, def-form also means that tracing information is not output when the form is executed. See the interactive example.

18.2.15.3 Backtracking in Specifications

If a specification fails to match at some point, this does not necessarily mean a syntax error will be signaled; instead, backtracking will take place until all alternatives have been exhausted. Eventually every element of the argument list must be matched by some element in the specification, and every required element in the specification must match some argument.

When a syntax error is detected, it might not be reported until much later, after higher-level alternatives have been exhausted, and with the point positioned further from the real error. But if backtracking is disabled when an error occurs, it can be reported immediately. Note that backtracking is also reenabled automatically in several situations; when a new alternative is established by &optional, &rest, or &or, or at the start of processing a sublist, group, or indirect specification. The effect of enabling or disabling backtracking is limited to the remainder of the level currently being processed and lower levels.

Backtracking is disabled while matching any of the form specifications (that is, form, body, def-form, and def-body). These specifications will match any form so any error must be in the form itself rather than at a higher level.

Backtracking is also disabled after successfully matching a quoted symbol or string specification, since this usually indicates a recognized construct. But if you have a set of alternative constructs that all begin with the same symbol, you can usually work around this constraint by factoring the symbol out of the alternatives, e.g., ["foo" &or [first case] [second case] ...].

Most needs are satisfied by these two ways that backtracking is automatically disabled, but occasionally it is useful to explicitly disable backtracking by using the gate specification. This is useful when you know that no higher alternatives could apply. See the example of the let specification.

18.2.15.4 Specification Examples

It may be easier to understand Edebug specifications by studying the examples provided here.

A let special form has a sequence of bindings and a body. Each of the bindings is either a symbol or a sublist with a symbol and optional expression. In the specification below, notice the gate inside of the sublist to prevent backtracking once a sublist is found.

     (def-edebug-spec let
       ((&rest
         &or symbolp (gate symbolp &optional form))
        body))

Edebug uses the following specifications for defun and the associated argument list and interactive specifications. It is necessary to handle interactive forms specially since an expression argument is actually evaluated outside of the function body. (The specification for defmacro is very similar to that for defun, but allows for the declare statement.)

     (def-edebug-spec defun
       (&define name lambda-list
                [&optional stringp]   ; Match the doc string, if present.
                [&optional ("interactive" interactive)]
                def-body))
     
     (def-edebug-spec lambda-list
       (([&rest arg]
         [&optional ["&optional" arg &rest arg]]
         &optional ["&rest" arg]
         )))
     
     (def-edebug-spec interactive
       (&optional &or stringp def-form))    ; Notice: def-form

The specification for backquote below illustrates how to match dotted lists and use nil to terminate recursion. It also illustrates how components of a vector may be matched. (The actual specification defined by Edebug is a little different, and does not support dotted lists because doing so causes very deep recursion that could fail.)

     (def-edebug-spec \` (backquote-form))   ; Alias just for clarity.
     
     (def-edebug-spec backquote-form
       (&or ([&or "," ",@"] &or ("quote" backquote-form) form)
            (backquote-form . [&or nil backquote-form])
            (vector &rest backquote-form)
            sexp))

18.2.16 Edebug Options

These options affect the behavior of Edebug:

If you change the values of edebug-on-error or edebug-on-quit while Edebug is active, their values won't be used until the next time Edebug is invoked via a new command.

18.3 Debugging Invalid Lisp Syntax

The Lisp reader reports invalid syntax, but cannot say where the real problem is. For example, the error “End of file during parsing” in evaluating an expression indicates an excess of open parentheses (or square brackets). The reader detects this imbalance at the end of the file, but it cannot figure out where the close parenthesis should have been. Likewise, “Invalid read syntax: ")"” indicates an excess close parenthesis or missing open parenthesis, but does not say where the missing parenthesis belongs. How, then, to find what to change?

If the problem is not simply an imbalance of parentheses, a useful technique is to try C-M-e at the beginning of each defun, and see if it goes to the place where that defun appears to end. If it does not, there is a problem in that defun.

However, unmatched parentheses are the most common syntax errors in Lisp, and we can give further advice for those cases. (In addition, just moving point through the code with Show Paren mode enabled might find the mismatch.)

• Excess Open: How to find a spurious open paren or missing close.
• Excess Close: How to find a spurious close paren or missing open.

18.3.1 Excess Open Parentheses

The first step is to find the defun that is unbalanced. If there is an excess open parenthesis, the way to do this is to go to the end of the file and type C-u C-M-u. This will move you to the beginning of the first defun that is unbalanced.

The next step is to determine precisely what is wrong. There is no way to be sure of this except by studying the program, but often the existing indentation is a clue to where the parentheses should have been. The easiest way to use this clue is to reindent with C-M-q and see what moves. But don't do this yet! Keep reading, first.

Before you do this, make sure the defun has enough close parentheses. Otherwise, C-M-q will get an error, or will reindent all the rest of the file until the end. So move to the end of the defun and insert a close parenthesis there. Don't use C-M-e to move there, since that too will fail to work until the defun is balanced.

Now you can go to the beginning of the defun and type C-M-q. Usually all the lines from a certain point to the end of the function will shift to the right. There is probably a missing close parenthesis, or a superfluous open parenthesis, near that point. (However, don't assume this is true; study the code to make sure.) Once you have found the discrepancy, undo the C-M-q with C-_, since the old indentation is probably appropriate to the intended parentheses.

After you think you have fixed the problem, use C-M-q again. If the old indentation actually fit the intended nesting of parentheses, and you have put back those parentheses, C-M-q should not change anything.

18.3.2 Excess Close Parentheses

To deal with an excess close parenthesis, first go to the beginning of the file, then type C-u -1 C-M-u to find the end of the first unbalanced defun.

Then find the actual matching close parenthesis by typing C-M-f at the beginning of that defun. This will leave you somewhere short of the place where the defun ought to end. It is possible that you will find a spurious close parenthesis in that vicinity.

If you don't see a problem at that point, the next thing to do is to type C-M-q at the beginning of the defun. A range of lines will probably shift left; if so, the missing open parenthesis or spurious close parenthesis is probably near the first of those lines. (However, don't assume this is true; study the code to make sure.) Once you have found the discrepancy, undo the C-M-q with C-_, since the old indentation is probably appropriate to the intended parentheses.

After you think you have fixed the problem, use C-M-q again. If the old indentation actually fits the intended nesting of parentheses, and you have put back those parentheses, C-M-q should not change anything.

18.4 Test Coverage

You can do coverage testing for a file of Lisp code by loading the testcover library and using the command M-x testcover-start <RET> file <RET> to instrument the code. Then test your code by calling it one or more times. Then use the command M-x testcover-mark-all to display colored highlights on the code to show where coverage is insufficient. The command M-x testcover-next-mark will move point forward to the next highlighted spot.

Normally, a red highlight indicates the form was never completely evaluated; a brown highlight means it always evaluated to the same value (meaning there has been little testing of what is done with the result). However, the red highlight is skipped for forms that can't possibly complete their evaluation, such as error. The brown highlight is skipped for forms that are expected to always evaluate to the same value, such as (setq x 14).

For difficult cases, you can add do-nothing macros to your code to give advice to the test coverage tool.

Edebug also has a coverage testing feature (see Coverage Testing). These features partly duplicate each other, and it would be cleaner to combine them.

18.5 Debugging Problems in Compilation

When an error happens during byte compilation, it is normally due to invalid syntax in the program you are compiling. The compiler prints a suitable error message in the ‘*Compile-Log*’ buffer, and then stops. The message may state a function name in which the error was found, or it may not. Either way, here is how to find out where in the file the error occurred.

What you should do is switch to the buffer ‘ *Compiler Input*’. (Note that the buffer name starts with a space, so it does not show up in M-x list-buffers.) This buffer contains the program being compiled, and point shows how far the byte compiler was able to read.

If the error was due to invalid Lisp syntax, point shows exactly where the invalid syntax was detected. The cause of the error is not necessarily near by! Use the techniques in the previous section to find the error.

If the error was detected while compiling a form that had been read successfully, then point is located at the end of the form. In this case, this technique can't localize the error precisely, but can still show you which function to check.

19 Reading and Printing Lisp Objects

Printing and reading are the operations of converting Lisp objects to textual form and vice versa. They use the printed representations and read syntax described in Lisp Data Types.

This chapter describes the Lisp functions for reading and printing. It also describes streams, which specify where to get the text (if reading) or where to put it (if printing).

• Streams Intro: Overview of streams, reading and printing.
• Input Streams: Various data types that can be used as input streams.
• Input Functions: Functions to read Lisp objects from text.
• Output Streams: Various data types that can be used as output streams.
• Output Functions: Functions to print Lisp objects as text.
• Output Variables: Variables that control what the printing functions do.

19.1 Introduction to Reading and Printing

Reading a Lisp object means parsing a Lisp expression in textual form and producing a corresponding Lisp object. This is how Lisp programs get into Lisp from files of Lisp code. We call the text the read syntax of the object. For example, the text ‘(a . 5)’ is the read syntax for a cons cell whose car is a and whose cdr is the number 5.

Printing a Lisp object means producing text that represents that object—converting the object to its printed representation (see Printed Representation). Printing the cons cell described above produces the text ‘(a . 5)’.

Reading and printing are more or less inverse operations: printing the object that results from reading a given piece of text often produces the same text, and reading the text that results from printing an object usually produces a similar-looking object. For example, printing the symbol foo produces the text ‘foo’, and reading that text returns the symbol foo. Printing a list whose elements are a and b produces the text ‘(a b)’, and reading that text produces a list (but not the same list) with elements a and b.

However, these two operations are not precisely inverse to each other. There are three kinds of exceptions:

• Printing can produce text that cannot be read. For example, buffers, windows, frames, subprocesses and markers print as text that starts with ‘#’; if you try to read this text, you get an error. There is no way to read those data types.
• One object can have multiple textual representations. For example, ‘1’ and ‘01’ represent the same integer, and ‘(a b)’ and ‘(a . (b))’ represent the same list. Reading will accept any of the alternatives, but printing must choose one of them.
• Comments can appear at certain points in the middle of an object's read sequence without affecting the result of reading it.

19.2 Input Streams

Most of the Lisp functions for reading text take an input stream as an argument. The input stream specifies where or how to get the characters of the text to be read. Here are the possible types of input stream:

buffer

The input characters are read from buffer, starting with the character directly after point. Point advances as characters are read.

marker

The input characters are read from the buffer that marker is in, starting with the character directly after the marker. The marker position advances as characters are read. The value of point in the buffer has no effect when the stream is a marker.

string

The input characters are taken from string, starting at the first character in the string and using as many characters as required.

function

The input characters are generated by function, which must support two kinds of calls:

• When it is called with no arguments, it should return the next character.
• When it is called with one argument (always a character), function should save the argument and arrange to return it on the next call. This is called unreading the character; it happens when the Lisp reader reads one character too many and wants to “put it back where it came from.” In this case, it makes no difference what value function returns.

t

t used as a stream means that the input is read from the minibuffer. In fact, the minibuffer is invoked once and the text given by the user is made into a string that is then used as the input stream. If Emacs is running in batch mode, standard input is used instead of the minibuffer. For example,

          (message "%s" (read t))

will read a Lisp expression from standard input and print the result to standard output.

nil

nil supplied as an input stream means to use the value of standard-input instead; that value is the default input stream, and must be a non-nil input stream.

symbol

A symbol as input stream is equivalent to the symbol's function definition (if any).

Here is an example of reading from a stream that is a buffer, showing where point is located before and after:

     ---------- Buffer: foo ----------
     This-!- is the contents of foo.
     ---------- Buffer: foo ----------
     
     (read (get-buffer "foo"))
          ⇒ is
     (read (get-buffer "foo"))
          ⇒ the
     
     ---------- Buffer: foo ----------
     This is the-!- contents of foo.
     ---------- Buffer: foo ----------

Note that the first read skips a space. Reading skips any amount of whitespace preceding the significant text.

Here is an example of reading from a stream that is a marker, initially positioned at the beginning of the buffer shown. The value read is the symbol This.

     
     ---------- Buffer: foo ----------
     This is the contents of foo.
     ---------- Buffer: foo ----------
     
     (setq m (set-marker (make-marker) 1 (get-buffer "foo")))
          ⇒ #<marker at 1 in foo>
     (read m)
          ⇒ This
     m
          ⇒ #<marker at 5 in foo>   ;; Before the first space.

Here we read from the contents of a string:

     (read "(When in) the course")
          ⇒ (When in)

The following example reads from the minibuffer. The prompt is: ‘Lisp expression: ’. (That is always the prompt used when you read from the stream t.) The user's input is shown following the prompt.

     (read t)
          ⇒ 23
     ---------- Buffer: Minibuffer ----------
     Lisp expression: 23 <RET>
     ---------- Buffer: Minibuffer ----------

Finally, here is an example of a stream that is a function, named useless-stream. Before we use the stream, we initialize the variable useless-list to a list of characters. Then each call to the function useless-stream obtains the next character in the list or unreads a character by adding it to the front of the list.

     (setq useless-list (append "XY()" nil))
          ⇒ (88 89 40 41)
     
     (defun useless-stream (&optional unread)
       (if unread
           (setq useless-list (cons unread useless-list))
         (prog1 (car useless-list)
                (setq useless-list (cdr useless-list)))))
          ⇒ useless-stream

Now we read using the stream thus constructed:

     (read 'useless-stream)
          ⇒ XY
     
     useless-list
          ⇒ (40 41)

Note that the open and close parentheses remain in the list. The Lisp reader encountered the open parenthesis, decided that it ended the input, and unread it. Another attempt to read from the stream at this point would read ‘()’ and return nil.

19.3 Input Functions

This section describes the Lisp functions and variables that pertain to reading.

In the functions below, stream stands for an input stream (see the previous section). If stream is nil or omitted, it defaults to the value of standard-input.

An end-of-file error is signaled if reading encounters an unterminated list, vector, or string.

19.4 Output Streams

An output stream specifies what to do with the characters produced by printing. Most print functions accept an output stream as an optional argument. Here are the possible types of output stream:

buffer

The output characters are inserted into buffer at point. Point advances as characters are inserted.

marker

The output characters are inserted into the buffer that marker points into, at the marker position. The marker position advances as characters are inserted. The value of point in the buffer has no effect on printing when the stream is a marker, and this kind of printing does not move point (except that if the marker points at or before the position of point, point advances with the surrounding text, as usual).

function

The output characters are passed to function, which is responsible for storing them away. It is called with a single character as argument, as many times as there are characters to be output, and is responsible for storing the characters wherever you want to put them.

t

The output characters are displayed in the echo area.

nil

nil specified as an output stream means to use the value of standard-output instead; that value is the default output stream, and must not be nil.

symbol

A symbol as output stream is equivalent to the symbol's function definition (if any).

Many of the valid output streams are also valid as input streams. The difference between input and output streams is therefore more a matter of how you use a Lisp object, than of different types of object.

Here is an example of a buffer used as an output stream. Point is initially located as shown immediately before the ‘h’ in ‘the’. At the end, point is located directly before that same ‘h’.

     ---------- Buffer: foo ----------
     This is t-!-he contents of foo.
     ---------- Buffer: foo ----------
     
     (print "This is the output" (get-buffer "foo"))
          ⇒ "This is the output"
     
     ---------- Buffer: foo ----------
     This is t
     "This is the output"
     -!-he contents of foo.
     ---------- Buffer: foo ----------

Now we show a use of a marker as an output stream. Initially, the marker is in buffer foo, between the ‘t’ and the ‘h’ in the word ‘the’. At the end, the marker has advanced over the inserted text so that it remains positioned before the same ‘h’. Note that the location of point, shown in the usual fashion, has no effect.

     ---------- Buffer: foo ----------
     This is the -!-output
     ---------- Buffer: foo ----------
     
     (setq m (copy-marker 10))
          ⇒ #<marker at 10 in foo>
     
     (print "More output for foo." m)
          ⇒ "More output for foo."
     
     ---------- Buffer: foo ----------
     This is t
     "More output for foo."
     he -!-output
     ---------- Buffer: foo ----------
     
     m
          ⇒ #<marker at 34 in foo>

The following example shows output to the echo area:

     (print "Echo Area output" t)
          ⇒ "Echo Area output"
     ---------- Echo Area ----------
     "Echo Area output"
     ---------- Echo Area ----------

Finally, we show the use of a function as an output stream. The function eat-output takes each character that it is given and conses it onto the front of the list last-output (see Building Lists). At the end, the list contains all the characters output, but in reverse order.

     (setq last-output nil)
          ⇒ nil
     
     (defun eat-output (c)
       (setq last-output (cons c last-output)))
          ⇒ eat-output
     
     (print "This is the output" 'eat-output)
          ⇒ "This is the output"
     
     last-output
          ⇒ (10 34 116 117 112 116 117 111 32 101 104
         116 32 115 105 32 115 105 104 84 34 10)

Now we can put the output in the proper order by reversing the list:

     (concat (nreverse last-output))
          ⇒ "
     \"This is the output\"
     "

Calling concat converts the list to a string so you can see its contents more clearly.

19.5 Output Functions

This section describes the Lisp functions for printing Lisp objects—converting objects into their printed representation.

Some of the Emacs printing functions add quoting characters to the output when necessary so that it can be read properly. The quoting characters used are ‘"’ and ‘\’; they distinguish strings from symbols, and prevent punctuation characters in strings and symbols from being taken as delimiters when reading. See Printed Representation, for full details. You specify quoting or no quoting by the choice of printing function.

If the text is to be read back into Lisp, then you should print with quoting characters to avoid ambiguity. Likewise, if the purpose is to describe a Lisp object clearly for a Lisp programmer. However, if the purpose of the output is to look nice for humans, then it is usually better to print without quoting.

Lisp objects can refer to themselves. Printing a self-referential object in the normal way would require an infinite amount of text, and the attempt could cause infinite recursion. Emacs detects such recursion and prints ‘#level’ instead of recursively printing an object already being printed. For example, here ‘#0’ indicates a recursive reference to the object at level 0 of the current print operation:

     (setq foo (list nil))
          ⇒ (nil)
     (setcar foo foo)
          ⇒ (#0)

In the functions below, stream stands for an output stream. (See the previous section for a description of output streams.) If stream is nil or omitted, it defaults to the value of standard-output.

19.6 Variables Affecting Output

These variables are used for detecting and reporting circular and shared structure:

20 Minibuffers

A minibuffer is a special buffer that Emacs commands use to read arguments more complicated than the single numeric prefix argument. These arguments include file names, buffer names, and command names (as in M-x). The minibuffer is displayed on the bottom line of the frame, in the same place as the echo area (see The Echo Area), but only while it is in use for reading an argument.

• Intro to Minibuffers: Basic information about minibuffers.
• Text from Minibuffer: How to read a straight text string.
• Object from Minibuffer: How to read a Lisp object or expression.
• Minibuffer History: Recording previous minibuffer inputs so the user can reuse them.
• Initial Input: Specifying initial contents for the minibuffer.
• Completion: How to invoke and customize completion.
• Yes-or-No Queries: Asking a question with a simple answer.
• Multiple Queries: Asking a series of similar questions.
• Reading a Password: Reading a password from the terminal.
• Minibuffer Commands: Commands used as key bindings in minibuffers.
• Minibuffer Contents: How such commands access the minibuffer text.
• Minibuffer Windows: Operating on the special minibuffer windows.
• Recursive Mini: Whether recursive entry to minibuffer is allowed.
• Minibuffer Misc: Various customization hooks and variables.

20.1 Introduction to Minibuffers

In most ways, a minibuffer is a normal Emacs buffer. Most operations within a buffer, such as editing commands, work normally in a minibuffer. However, many operations for managing buffers do not apply to minibuffers. The name of a minibuffer always has the form ‘ *Minibuf-number*’, and it cannot be changed. Minibuffers are displayed only in special windows used only for minibuffers; these windows always appear at the bottom of a frame. (Sometimes frames have no minibuffer window, and sometimes a special kind of frame contains nothing but a minibuffer window; see Minibuffers and Frames.)

The text in the minibuffer always starts with the prompt string, the text that was specified by the program that is using the minibuffer to tell the user what sort of input to type. This text is marked read-only so you won't accidentally delete or change it. It is also marked as a field (see Fields), so that certain motion functions, including beginning-of-line, forward-word, forward-sentence, and forward-paragraph, stop at the boundary between the prompt and the actual text.

The minibuffer's window is normally a single line; it grows automatically if the contents require more space. You can explicitly resize it temporarily with the window sizing commands; it reverts to its normal size when the minibuffer is exited. You can resize it permanently by using the window sizing commands in the frame's other window, when the minibuffer is not active. If the frame contains just a minibuffer, you can change the minibuffer's size by changing the frame's size.

Use of the minibuffer reads input events, and that alters the values of variables such as this-command and last-command (see Command Loop Info). Your program should bind them around the code that uses the minibuffer, if you do not want that to change them.

Under some circumstances, a command can use a minibuffer even if there is an active minibuffer; such minibuffers are called a recursive minibuffer. The first minibuffer is named ‘ *Minibuf-0*’. Recursive minibuffers are named by incrementing the number at the end of the name. (The names begin with a space so that they won't show up in normal buffer lists.) Of several recursive minibuffers, the innermost (or most recently entered) is the active minibuffer. We usually call this “the” minibuffer. You can permit or forbid recursive minibuffers by setting the variable enable-recursive-minibuffers, or by putting properties of that name on command symbols (See Recursive Mini.)

Like other buffers, a minibuffer uses a local keymap (see Keymaps) to specify special key bindings. The function that invokes the minibuffer also sets up its local map according to the job to be done. See Text from Minibuffer, for the non-completion minibuffer local maps. See Completion Commands, for the minibuffer local maps for completion.

When Emacs is running in batch mode, any request to read from the minibuffer actually reads a line from the standard input descriptor that was supplied when Emacs was started.

20.2 Reading Text Strings with the Minibuffer

The most basic primitive for minibuffer input is read-from-minibuffer, which can be used to read either a string or a Lisp object in textual form. The function read-regexp is used for reading regular expressions (see Regular Expressions), which are a special kind of string. There are also specialized functions for reading commands, variables, file names, etc. (see Completion).

In most cases, you should not call minibuffer input functions in the middle of a Lisp function. Instead, do all minibuffer input as part of reading the arguments for a command, in the interactive specification. See Defining Commands.

20.3 Reading Lisp Objects with the Minibuffer

This section describes functions for reading Lisp objects with the minibuffer.

20.4 Minibuffer History

A minibuffer history list records previous minibuffer inputs so the user can reuse them conveniently. A history list is actually a symbol, not a list; it is a variable whose value is a list of strings (previous inputs), most recent first.

There are many separate history lists, used for different kinds of inputs. It's the Lisp programmer's job to specify the right history list for each use of the minibuffer.

You specify the history list with the optional hist argument to either read-from-minibuffer or completing-read. Here are the possible values for it:

variable

Use variable (a symbol) as the history list.

(variable . startpos)

Use variable (a symbol) as the history list, and assume that the initial history position is startpos (a nonnegative integer).

Specifying 0 for startpos is equivalent to just specifying the symbol variable. previous-history-element will display the most recent element of the history list in the minibuffer. If you specify a positive startpos, the minibuffer history functions behave as if (elt variable (1- STARTPOS)) were the history element currently shown in the minibuffer.

For consistency, you should also specify that element of the history as the initial minibuffer contents, using the initial argument to the minibuffer input function (see Initial Input).

If you don't specify hist, then the default history list minibuffer-history is used. For other standard history lists, see below. You can also create your own history list variable; just initialize it to nil before the first use.

Both read-from-minibuffer and completing-read add new elements to the history list automatically, and provide commands to allow the user to reuse items on the list. The only thing your program needs to do to use a history list is to initialize it and to pass its name to the input functions when you wish. But it is safe to modify the list by hand when the minibuffer input functions are not using it.

Emacs functions that add a new element to a history list can also delete old elements if the list gets too long. The variable history-length specifies the maximum length for most history lists. To specify a different maximum length for a particular history list, put the length in the history-length property of the history list symbol. The variable history-delete-duplicates specifies whether to delete duplicates in history.

Here are some of the standard minibuffer history list variables:

20.5 Initial Input

Several of the functions for minibuffer input have an argument called initial or initial-contents. This is a mostly-deprecated feature for specifying that the minibuffer should start out with certain text, instead of empty as usual.

If initial is a string, the minibuffer starts out containing the text of the string, with point at the end, when the user starts to edit the text. If the user simply types <RET> to exit the minibuffer, it will use the initial input string to determine the value to return.

We discourage use of a non-nil value for initial, because initial input is an intrusive interface. History lists and default values provide a much more convenient method to offer useful default inputs to the user.

There is just one situation where you should specify a string for an initial argument. This is when you specify a cons cell for the hist or history argument. See Minibuffer History.

initial can also be a cons cell of the form (string . position). This means to insert string in the minibuffer but put point at position within the string's text.

As a historical accident, position was implemented inconsistently in different functions. In completing-read, position's value is interpreted as origin-zero; that is, a value of 0 means the beginning of the string, 1 means after the first character, etc. In read-minibuffer, and the other non-completion minibuffer input functions that support this argument, 1 means the beginning of the string 2 means after the first character, etc.

Use of a cons cell as the value for initial arguments is deprecated in user code.

20.6 Completion

Completion is a feature that fills in the rest of a name starting from an abbreviation for it. Completion works by comparing the user's input against a list of valid names and determining how much of the name is determined uniquely by what the user has typed. For example, when you type C-x b (switch-to-buffer) and then type the first few letters of the name of the buffer to which you wish to switch, and then type <TAB> (minibuffer-complete), Emacs extends the name as far as it can.

Standard Emacs commands offer completion for names of symbols, files, buffers, and processes; with the functions in this section, you can implement completion for other kinds of names.

The try-completion function is the basic primitive for completion: it returns the longest determined completion of a given initial string, with a given set of strings to match against.

The function completing-read provides a higher-level interface for completion. A call to completing-read specifies how to determine the list of valid names. The function then activates the minibuffer with a local keymap that binds a few keys to commands useful for completion. Other functions provide convenient simple interfaces for reading certain kinds of names with completion.

• Basic Completion: Low-level functions for completing strings.
• Minibuffer Completion: Invoking the minibuffer with completion.
• Completion Commands: Minibuffer commands that do completion.
• High-Level Completion: Convenient special cases of completion (reading buffer name, file name, etc.).
• Reading File Names: Using completion to read file names and shell commands.
• Completion Styles: Specifying rules for performing completion.
• Programmed Completion: Writing your own completion-function.

20.6.1 Basic Completion Functions

The following completion functions have nothing in themselves to do with minibuffers. We describe them here to keep them near the higher-level completion features that do use the minibuffer.

If you store a completion alist in a variable, you should mark the variable as “risky” with a non-nil risky-local-variable property. See File Local Variables.

The function completion-in-region provides a convenient way to perform completion on an arbitrary stretch of text in an Emacs buffer:

20.6.2 Completion and the Minibuffer

This section describes the basic interface for reading from the minibuffer with completion.

20.6.3 Minibuffer Commands that Do Completion

This section describes the keymaps, commands and user options used in the minibuffer to do completion.

20.6.4 High-Level Completion Functions

This section describes the higher-level convenient functions for reading certain sorts of names with completion.

In most cases, you should not call these functions in the middle of a Lisp function. When possible, do all minibuffer input as part of reading the arguments for a command, in the interactive specification. See Defining Commands.

See also the functions read-coding-system and read-non-nil-coding-system, in User-Chosen Coding Systems, and read-input-method-name, in Input Methods.

20.6.5 Reading File Names

The high-level completion functions read-file-name, read-directory-name, and read-shell-command are designed to read file names, directory names, and shell commands respectively. They provide special features, including automatic insertion of the default directory.

20.6.6 Completion Styles

A completion style is a set of rules for generating completions. The user option completion-styles stores a list of completion styles, which are represented by symbols.

By default, completion-styles-alist contains five pre-defined completion styles: basic, a basic completion style; partial-completion, which does partial completion (completing each word in the input separately); emacs22, which performs completion according to the rules used in Emacs 22; emacs21, which performs completion according to the rules used in Emacs 21; and initials, which completes acronyms and initialisms.

20.6.7 Programmed Completion

Sometimes it is not possible to create an alist or an obarray containing all the intended possible completions. In such a case, you can supply your own function to compute the completion of a given string. This is called programmed completion. Emacs uses programmed completion when completing file names (see File Name Completion), among many other cases.

To use this feature, pass a function as the collection argument to completing-read. The function completing-read arranges to pass your completion function along to try-completion, all-completions, and other basic completion functions, which will then let your function do all the work.

The completion function should accept three arguments:

• The string to be completed.
• The predicate function to filter possible matches, or nil if none. Your function should call the predicate for each possible match, and ignore the possible match if the predicate returns nil.
• A flag specifying the type of operation. The best way to think about it is that the function stands for an object (in the “object-oriented” sense of the word), and this third argument specifies which method to run.

There are currently four methods, i.e. four flag values, one for each of the four different basic operations:

• nil specifies try-completion. The completion function should return the completion of the specified string, or t if the string is a unique and exact match already, or nil if the string matches no possibility.

  If the string is an exact match for one possibility, but also matches other longer possibilities, the function should return the string, not t.

• t specifies all-completions. The completion function should return a list of all possible completions of the specified string.
• lambda specifies test-completion. The completion function should return t if the specified string is an exact match for some possibility; nil otherwise.
• (boundaries . SUFFIX) specifies completion-boundaries. The function should return a value of the form (boundaries START . END) where START is the position of the beginning boundary in in the string to complete, and END is the position of the end boundary in SUFFIX.

It would be consistent and clean for completion functions to allow lambda expressions (lists that are functions) as well as function symbols as collection, but this is impossible. Lists as completion tables already have other meanings, and it would be unreliable to treat one differently just because it is also a possible function. So you must arrange for any function you wish to use for completion to be encapsulated in a symbol.

20.7 Yes-or-No Queries

This section describes functions used to ask the user a yes-or-no question. The function y-or-n-p can be answered with a single character; it is useful for questions where an inadvertent wrong answer will not have serious consequences. yes-or-no-p is suitable for more momentous questions, since it requires three or four characters to answer.

If either of these functions is called in a command that was invoked using the mouse—more precisely, if last-nonmenu-event (see Command Loop Info) is either nil or a list—then it uses a dialog box or pop-up menu to ask the question. Otherwise, it uses keyboard input. You can force use of the mouse or use of keyboard input by binding last-nonmenu-event to a suitable value around the call.

Strictly speaking, yes-or-no-p uses the minibuffer and y-or-n-p does not; but it seems best to describe them together.

20.8 Asking Multiple Y-or-N Questions

When you have a series of similar questions to ask, such as “Do you want to save this buffer” for each buffer in turn, you should use map-y-or-n-p to ask the collection of questions, rather than asking each question individually. This gives the user certain convenient facilities such as the ability to answer the whole series at once.

20.9 Reading a Password

To read a password to pass to another program, you can use the function read-passwd.

20.10 Minibuffer Commands

This section describes some commands meant for use in the minibuffer.

20.11 Minibuffer Windows

These functions access and select minibuffer windows and test whether they are active.

It is not correct to determine whether a given window is a minibuffer by comparing it with the result of (minibuffer-window), because there can be more than one minibuffer window if there is more than one frame.

20.12 Minibuffer Contents

These functions access the minibuffer prompt and contents.

20.13 Recursive Minibuffers

These functions and variables deal with recursive minibuffers (see Recursive Editing):

If a command name has a property enable-recursive-minibuffers that is non-nil, then the command can use the minibuffer to read arguments even if it is invoked from the minibuffer. A command can also achieve this by binding enable-recursive-minibuffers to t in the interactive declaration (see Using Interactive). The minibuffer command next-matching-history-element (normally M-s in the minibuffer) does the latter.

20.14 Minibuffer Miscellany

21 Command Loop

When you run Emacs, it enters the editor command loop almost immediately. This loop reads key sequences, executes their definitions, and displays the results. In this chapter, we describe how these things are done, and the subroutines that allow Lisp programs to do them.

• Command Overview: How the command loop reads commands.
• Defining Commands: Specifying how a function should read arguments.
• Interactive Call: Calling a command, so that it will read arguments.
• Distinguish Interactive: Making a command distinguish interactive calls.
• Command Loop Info: Variables set by the command loop for you to examine.
• Adjusting Point: Adjustment of point after a command.
• Input Events: What input looks like when you read it.
• Reading Input: How to read input events from the keyboard or mouse.
• Special Events: Events processed immediately and individually.
• Waiting: Waiting for user input or elapsed time.
• Quitting: How C-g works. How to catch or defer quitting.
• Prefix Command Arguments: How the commands to set prefix args work.
• Recursive Editing: Entering a recursive edit, and why you usually shouldn't.
• Disabling Commands: How the command loop handles disabled commands.
• Command History: How the command history is set up, and how accessed.
• Keyboard Macros: How keyboard macros are implemented.

21.1 Command Loop Overview

The first thing the command loop must do is read a key sequence, which is a sequence of events that translates into a command. It does this by calling the function read-key-sequence. Your Lisp code can also call this function (see Key Sequence Input). Lisp programs can also do input at a lower level with read-event (see Reading One Event) or discard pending input with discard-input (see Event Input Misc).

The key sequence is translated into a command through the currently active keymaps. See Key Lookup, for information on how this is done. The result should be a keyboard macro or an interactively callable function. If the key is M-x, then it reads the name of another command, which it then calls. This is done by the command execute-extended-command (see Interactive Call).

Prior to executing the command, Emacs runs undo-boundary to create an undo boundary. See Maintaining Undo.

To execute a command, Emacs first reads its arguments by calling command-execute (see Interactive Call). For commands written in Lisp, the interactive specification says how to read the arguments. This may use the prefix argument (see Prefix Command Arguments) or may read with prompting in the minibuffer (see Minibuffers). For example, the command find-file has an interactive specification which says to read a file name using the minibuffer. The function body of find-file does not use the minibuffer, so if you call find-file as a function from Lisp code, you must supply the file name string as an ordinary Lisp function argument.

If the command is a string or vector (i.e., a keyboard macro) then execute-kbd-macro is used to execute it. You can call this function yourself (see Keyboard Macros).

To terminate the execution of a running command, type C-g. This character causes quitting (see Quitting).

Quitting is suppressed while running pre-command-hook and post-command-hook. If an error happens while executing one of these hooks, it terminates execution of the hook, and clears the hook variable to nil so as to prevent an infinite loop of errors.

A request coming into the Emacs server (see Emacs Server) runs these two hooks just as a keyboard command does.

21.2 Defining Commands

The special form interactive turns a Lisp function into a command. The interactive form must be located at top-level in the function body (usually as the first form in the body), or in the interactive-form property of the function symbol. When the interactive form is located in the function body, it does nothing when actually executed. Its presence serves as a flag, which tells the Emacs command loop that the function can be called interactively. The argument of the interactive form controls the reading of arguments for an interactive call.

• Using Interactive: General rules for interactive.
• Interactive Codes: The standard letter-codes for reading arguments in various ways.
• Interactive Examples: Examples of how to read interactive arguments.

21.2.1 Using interactive

This section describes how to write the interactive form that makes a Lisp function an interactively-callable command, and how to examine a command's interactive form.

There are three possibilities for the argument arg-descriptor:

• It may be omitted or nil; then the command is called with no arguments. This leads quickly to an error if the command requires one or more arguments.
• It may be a string; its contents are a sequence of elements separated by newlines, one for each parameter⁸. Each element consists of a code character (see Interactive Codes) optionally followed by a prompt (which some code characters use and some ignore). Here is an example:

            (interactive "P\nbFrobnicate buffer: ")
  

  The code letter ‘P’ sets the command's first argument to the raw command prefix (see Prefix Command Arguments). ‘bFrobnicate buffer: ’ prompts the user with ‘Frobnicate buffer: ’ to enter the name of an existing buffer, which becomes the second and final argument.

  The prompt string can use ‘%’ to include previous argument values (starting with the first argument) in the prompt. This is done using format (see Formatting Strings). For example, here is how you could read the name of an existing buffer followed by a new name to give to that buffer:

            (interactive "bBuffer to rename: \nsRename buffer %s to: ")
  

  If ‘*’ appears at the beginning of the string, then an error is signaled if the buffer is read-only.

  If ‘@’ appears at the beginning of the string, and if the key sequence used to invoke the command includes any mouse events, then the window associated with the first of those events is selected before the command is run.

  If ‘^’ appears at the beginning of the string, and if the command was invoked through shift-translation, set the mark and activate the region temporarily, or extend an already active region, before the command is run. If the command was invoked without shift-translation, and the region is temporarily active, deactivate the region before the command is run. Shift-translation is controlled on the user level by shift-select-mode; see Shift Selection.

  You can use ‘*’, ‘@’, and ^ together; the order does not matter. Actual reading of arguments is controlled by the rest of the prompt string (starting with the first character that is not ‘*’, ‘@’, or ‘^’).

• It may be a Lisp expression that is not a string; then it should be a form that is evaluated to get a list of arguments to pass to the command. Usually this form will call various functions to read input from the user, most often through the minibuffer (see Minibuffers) or directly from the keyboard (see Reading Input).

  Providing point or the mark as an argument value is also common, but if you do this and read input (whether using the minibuffer or not), be sure to get the integer values of point or the mark after reading. The current buffer may be receiving subprocess output; if subprocess output arrives while the command is waiting for input, it could relocate point and the mark.

  Here's an example of what not to do:

            (interactive
             (list (region-beginning) (region-end)
                   (read-string "Foo: " nil 'my-history)))
  

  Here's how to avoid the problem, by examining point and the mark after reading the keyboard input:

            (interactive
             (let ((string (read-string "Foo: " nil 'my-history)))
               (list (region-beginning) (region-end) string)))
  

  Warning: the argument values should not include any data types that can't be printed and then read. Some facilities save command-history in a file to be read in the subsequent sessions; if a command's arguments contain a data type that prints using ‘#<...>’ syntax, those facilities won't work.

  There are, however, a few exceptions: it is ok to use a limited set of expressions such as (point), (mark), (region-beginning), and (region-end), because Emacs recognizes them specially and puts the expression (rather than its value) into the command history. To see whether the expression you wrote is one of these exceptions, run the command, then examine (car command-history).

21.2.2 Code Characters for interactive

The code character descriptions below contain a number of key words, defined here as follows:

Completion

Provide completion. <TAB>, <SPC>, and <RET> perform name completion because the argument is read using completing-read (see Completion). ? displays a list of possible completions.

Existing

Require the name of an existing object. An invalid name is not accepted; the commands to exit the minibuffer do not exit if the current input is not valid.

Default

A default value of some sort is used if the user enters no text in the minibuffer. The default depends on the code character.

No I/O

This code letter computes an argument without reading any input. Therefore, it does not use a prompt string, and any prompt string you supply is ignored.

Even though the code letter doesn't use a prompt string, you must follow it with a newline if it is not the last code character in the string.

Prompt

A prompt immediately follows the code character. The prompt ends either with the end of the string or with a newline.

Special

This code character is meaningful only at the beginning of the interactive string, and it does not look for a prompt or a newline. It is a single, isolated character.

Here are the code character descriptions for use with interactive:

‘*’

Signal an error if the current buffer is read-only. Special.

‘@’

Select the window mentioned in the first mouse event in the key sequence that invoked this command. Special.

‘^’

If the command was invoked through shift-translation, set the mark and activate the region temporarily, or extend an already active region, before the command is run. If the command was invoked without shift-translation, and the region is temporarily active, deactivate the region before the command is run. Special.

‘a’

A function name (i.e., a symbol satisfying fboundp). Existing, Completion, Prompt.

‘b’

The name of an existing buffer. By default, uses the name of the current buffer (see Buffers). Existing, Completion, Default, Prompt.

‘B’

A buffer name. The buffer need not exist. By default, uses the name of a recently used buffer other than the current buffer. Completion, Default, Prompt.

‘c’

A character. The cursor does not move into the echo area. Prompt.

‘C’

A command name (i.e., a symbol satisfying commandp). Existing, Completion, Prompt.

‘d’

The position of point, as an integer (see Point). No I/O.

‘D’

A directory name. The default is the current default directory of the current buffer, default-directory (see File Name Expansion). Existing, Completion, Default, Prompt.

‘e’

The first or next mouse event in the key sequence that invoked the command. More precisely, ‘e’ gets events that are lists, so you can look at the data in the lists. See Input Events. No I/O.

You can use ‘e’ more than once in a single command's interactive specification. If the key sequence that invoked the command has n events that are lists, the nth ‘e’ provides the nth such event. Events that are not lists, such as function keys and ASCII characters, do not count where ‘e’ is concerned.

‘f’

A file name of an existing file (see File Names). The default directory is default-directory. Existing, Completion, Default, Prompt.

‘F’

A file name. The file need not exist. Completion, Default, Prompt.

‘G’

A file name. The file need not exist. If the user enters just a directory name, then the value is just that directory name, with no file name within the directory added. Completion, Default, Prompt.

‘i’

An irrelevant argument. This code always supplies nil as the argument's value. No I/O.

‘k’

A key sequence (see Key Sequences). This keeps reading events until a command (or undefined command) is found in the current key maps. The key sequence argument is represented as a string or vector. The cursor does not move into the echo area. Prompt.

If ‘k’ reads a key sequence that ends with a down-event, it also reads and discards the following up-event. You can get access to that up-event with the ‘U’ code character.

This kind of input is used by commands such as describe-key and global-set-key.

‘K’

A key sequence, whose definition you intend to change. This works like ‘k’, except that it suppresses, for the last input event in the key sequence, the conversions that are normally used (when necessary) to convert an undefined key into a defined one.

‘m’

The position of the mark, as an integer. No I/O.

‘M’

Arbitrary text, read in the minibuffer using the current buffer's input method, and returned as a string (see Input Methods). Prompt.

‘n’

A number, read with the minibuffer. If the input is not a number, the user has to try again. ‘n’ never uses the prefix argument. Prompt.

‘N’

The numeric prefix argument; but if there is no prefix argument, read a number as with n. The value is always a number. See Prefix Command Arguments. Prompt.

‘p’

The numeric prefix argument. (Note that this ‘p’ is lower case.) No I/O.

‘P’

The raw prefix argument. (Note that this ‘P’ is upper case.) No I/O.

‘r’

Point and the mark, as two numeric arguments, smallest first. This is the only code letter that specifies two successive arguments rather than one. No I/O.

‘s’

Arbitrary text, read in the minibuffer and returned as a string (see Text from Minibuffer). Terminate the input with either C-j or <RET>. (C-q may be used to include either of these characters in the input.) Prompt.

‘S’

An interned symbol whose name is read in the minibuffer. Any whitespace character terminates the input. (Use C-q to include whitespace in the string.) Other characters that normally terminate a symbol (e.g., parentheses and brackets) do not do so here. Prompt.

‘U’

A key sequence or nil. Can be used after a ‘k’ or ‘K’ argument to get the up-event that was discarded (if any) after ‘k’ or ‘K’ read a down-event. If no up-event has been discarded, ‘U’ provides nil as the argument. No I/O.

‘v’

A variable declared to be a user option (i.e., satisfying the predicate user-variable-p). This reads the variable using read-variable. See Definition of read-variable. Existing, Completion, Prompt.

‘x’

A Lisp object, specified with its read syntax, terminated with a C-j or <RET>. The object is not evaluated. See Object from Minibuffer. Prompt.

‘X’

A Lisp form's value. ‘X’ reads as ‘x’ does, then evaluates the form so that its value becomes the argument for the command. Prompt.

‘z’

A coding system name (a symbol). If the user enters null input, the argument value is nil. See Coding Systems. Completion, Existing, Prompt.

‘Z’

A coding system name (a symbol)—but only if this command has a prefix argument. With no prefix argument, ‘Z’ provides nil as the argument value. Completion, Existing, Prompt.

21.2.3 Examples of Using interactive

Here are some examples of interactive:

     (defun foo1 ()              ; foo1 takes no arguments,
         (interactive)           ;   just moves forward two words.
         (forward-word 2))
          ⇒ foo1
     
     (defun foo2 (n)             ; foo2 takes one argument,
         (interactive "^p")      ;   which is the numeric prefix.
                                 ; under shift-select-mode,
                                 ;   will activate or extend region.
         (forward-word (* 2 n)))
          ⇒ foo2
     
     (defun foo3 (n)             ; foo3 takes one argument,
         (interactive "nCount:") ;   which is read with the Minibuffer.
         (forward-word (* 2 n)))
          ⇒ foo3
     
     (defun three-b (b1 b2 b3)
       "Select three existing buffers.
     Put them into three windows, selecting the last one."
         (interactive "bBuffer1:\nbBuffer2:\nbBuffer3:")
         (delete-other-windows)
         (split-window (selected-window) 8)
         (switch-to-buffer b1)
         (other-window 1)
         (split-window (selected-window) 8)
         (switch-to-buffer b2)
         (other-window 1)
         (switch-to-buffer b3))
          ⇒ three-b
     (three-b "*scratch*" "declarations.texi" "*mail*")
          ⇒ nil

21.3 Interactive Call

After the command loop has translated a key sequence into a command, it invokes that command using the function command-execute. If the command is a function, command-execute calls call-interactively, which reads the arguments and calls the command. You can also call these functions yourself.

21.4 Distinguish Interactive Calls

Sometimes a command should display additional visual feedback (such as an informative message in the echo area) for interactive calls only. There are three ways to do this. The recommended way to test whether the function was called using call-interactively is to give it an optional argument print-message and use the interactive spec to make it non-nil in interactive calls. Here's an example:

     (defun foo (&optional print-message)
       (interactive "p")
       (when print-message
         (message "foo")))

We use "p" because the numeric prefix argument is never nil. Defined in this way, the function does display the message when called from a keyboard macro.

The above method with the additional argument is usually best, because it allows callers to say “treat this call as interactive.” But you can also do the job by testing called-interactively-p.

Here is an example of using called-interactively-p:

     (defun foo ()
       (interactive)
       (when (called-interactively-p 'any)
         (message "Interactive!")
         'foo-called-interactively))
     
     ;; Type M-x foo.
          -| Interactive!
     
     (foo)
          ⇒ nil

Here is another example that contrasts direct and indirect calls to called-interactively-p.

     (defun bar ()
       (interactive)
       (message "%s" (list (foo) (called-interactively-p 'any))))
     
     ;; Type M-x bar.
          -| (nil t)

21.5 Information from the Command Loop

The editor command loop sets several Lisp variables to keep status records for itself and for commands that are run. With the exception of this-command and last-command it's generally a bad idea to change any of these variables in a Lisp program.

If you do not want a particular command to be recognized as the previous command in the case where it got an error, you must code that command to prevent this. One way is to set this-command to t at the beginning of the command, and set this-command back to its proper value at the end, like this:

     (defun foo (args...)
       (interactive ...)
       (let ((old-this-command this-command))
         (setq this-command t)
         ...do the work...
         (setq this-command old-this-command)))

We do not bind this-command with let because that would restore the old value in case of error—a feature of let which in this case does precisely what we want to avoid.

21.6 Adjusting Point After Commands

It is not easy to display a value of point in the middle of a sequence of text that has the display, composition or intangible property, or is invisible. Therefore, after a command finishes and returns to the command loop, if point is within such a sequence, the command loop normally moves point to the edge of the sequence.

A command can inhibit this feature by setting the variable disable-point-adjustment:

21.7 Input Events

The Emacs command loop reads a sequence of input events that represent keyboard or mouse activity. The events for keyboard activity are characters or symbols; mouse events are always lists. This section describes the representation and meaning of input events in detail.

• Keyboard Events: Ordinary characters--keys with symbols on them.
• Function Keys: Function keys--keys with names, not symbols.
• Mouse Events: Overview of mouse events.
• Click Events: Pushing and releasing a mouse button.
• Drag Events: Moving the mouse before releasing the button.
• Button-Down Events: A button was pushed and not yet released.
• Repeat Events: Double and triple click (or drag, or down).
• Motion Events: Just moving the mouse, not pushing a button.
• Focus Events: Moving the mouse between frames.
• Misc Events: Other events the system can generate.
• Event Examples: Examples of the lists for mouse events.
• Classifying Events: Finding the modifier keys in an event symbol. Event types.
• Accessing Mouse: Functions to extract info from mouse events.
• Accessing Scroll: Functions to get info from scroll bar events.
• Strings of Events: Special considerations for putting keyboard character events in a string.

21.7.1 Keyboard Events

There are two kinds of input you can get from the keyboard: ordinary keys, and function keys. Ordinary keys correspond to characters; the events they generate are represented in Lisp as characters. The event type of a character event is the character itself (an integer); see Classifying Events.

An input character event consists of a basic code between 0 and 524287, plus any or all of these modifier bits:

meta

The 2**27 bit in the character code indicates a character typed with the meta key held down.

control

The 2**26 bit in the character code indicates a non-ASCII control character.

ascii control characters such as C-a have special basic codes of their own, so Emacs needs no special bit to indicate them. Thus, the code for C-a is just 1.

But if you type a control combination not in ASCII, such as % with the control key, the numeric value you get is the code for % plus 2**26 (assuming the terminal supports non-ASCII control characters).

shift

The 2**25 bit in the character code indicates an ASCII control character typed with the shift key held down.

For letters, the basic code itself indicates upper versus lower case; for digits and punctuation, the shift key selects an entirely different character with a different basic code. In order to keep within the ASCII character set whenever possible, Emacs avoids using the 2**25 bit for those characters.

However, ASCII provides no way to distinguish C-A from C-a, so Emacs uses the 2**25 bit in C-A and not in C-a.

hyper

The 2**24 bit in the character code indicates a character typed with the hyper key held down.

super

The 2**23 bit in the character code indicates a character typed with the super key held down.

alt

The 2**22 bit in the character code indicates a character typed with the alt key held down. (On some terminals, the key labeled <ALT> is actually the meta key.)

It is best to avoid mentioning specific bit numbers in your program. To test the modifier bits of a character, use the function event-modifiers (see Classifying Events). When making key bindings, you can use the read syntax for characters with modifier bits (‘\C-’, ‘\M-’, and so on). For making key bindings with define-key, you can use lists such as (control hyper ?x) to specify the characters (see Changing Key Bindings). The function event-convert-list converts such a list into an event type (see Classifying Events).

21.7.2 Function Keys

Most keyboards also have function keys—keys that have names or symbols that are not characters. Function keys are represented in Emacs Lisp as symbols; the symbol's name is the function key's label, in lower case. For example, pressing a key labeled <F1> places the symbol f1 in the input stream.

The event type of a function key event is the event symbol itself. See Classifying Events.

Here are a few special cases in the symbol-naming convention for function keys:

backspace, tab, newline, return, delete

These keys correspond to common ASCII control characters that have special keys on most keyboards.

In ASCII, C-i and <TAB> are the same character. If the terminal can distinguish between them, Emacs conveys the distinction to Lisp programs by representing the former as the integer 9, and the latter as the symbol tab.

Most of the time, it's not useful to distinguish the two. So normally local-function-key-map (see Translation Keymaps) is set up to map tab into 9. Thus, a key binding for character code 9 (the character C-i) also applies to tab. Likewise for the other symbols in this group. The function read-char likewise converts these events into characters.

In ASCII, <BS> is really C-h. But backspace converts into the character code 127 (<DEL>), not into code 8 (<BS>). This is what most users prefer.

left, up, right, down

Cursor arrow keys

kp-add, kp-decimal, kp-divide, ...

Keypad keys (to the right of the regular keyboard).

kp-0, kp-1, ...

Keypad keys with digits.

kp-f1, kp-f2, kp-f3, kp-f4

Keypad PF keys.

kp-home, kp-left, kp-up, kp-right, kp-down

Keypad arrow keys. Emacs normally translates these into the corresponding non-keypad keys home, left, ...

kp-prior, kp-next, kp-end, kp-begin, kp-insert, kp-delete

Additional keypad duplicates of keys ordinarily found elsewhere. Emacs normally translates these into the like-named non-keypad keys.

You can use the modifier keys <ALT>, <CTRL>, <HYPER>, <META>, <SHIFT>, and <SUPER> with function keys. The way to represent them is with prefixes in the symbol name:

‘A-’

The alt modifier.

‘C-’

The control modifier.

‘H-’

The hyper modifier.

‘M-’

The meta modifier.

‘S-’

The shift modifier.

‘s-’

The super modifier.

Thus, the symbol for the key <F3> with <META> held down is M-f3. When you use more than one prefix, we recommend you write them in alphabetical order; but the order does not matter in arguments to the key-binding lookup and modification functions.

21.7.3 Mouse Events

Emacs supports four kinds of mouse events: click events, drag events, button-down events, and motion events. All mouse events are represented as lists. The car of the list is the event type; this says which mouse button was involved, and which modifier keys were used with it. The event type can also distinguish double or triple button presses (see Repeat Events). The rest of the list elements give position and time information.

For key lookup, only the event type matters: two events of the same type necessarily run the same command. The command can access the full values of these events using the ‘e’ interactive code. See Interactive Codes.

A key sequence that starts with a mouse event is read using the keymaps of the buffer in the window that the mouse was in, not the current buffer. This does not imply that clicking in a window selects that window or its buffer—that is entirely under the control of the command binding of the key sequence.

21.7.4 Click Events

When the user presses a mouse button and releases it at the same location, that generates a click event. All mouse click event share the same format:

     (event-type position click-count)

event-type

This is a symbol that indicates which mouse button was used. It is one of the symbols mouse-1, mouse-2, ..., where the buttons are numbered left to right.

You can also use prefixes ‘A-’, ‘C-’, ‘H-’, ‘M-’, ‘S-’ and ‘s-’ for modifiers alt, control, hyper, meta, shift and super, just as you would with function keys.

This symbol also serves as the event type of the event. Key bindings describe events by their types; thus, if there is a key binding for mouse-1, that binding would apply to all events whose event-type is mouse-1.

position

This is the position where the mouse click occurred. The actual format of position depends on what part of a window was clicked on.

For mouse click events in the text area, mode line, header line, or in the marginal areas, position has this form:

          (window pos-or-area (x . y) timestamp
           object text-pos (col . row)
           image (dx . dy) (width . height))

window

This is the window in which the click occurred.

pos-or-area

This is the buffer position of the character clicked on in the text area, or if clicked outside the text area, it is the window area in which the click occurred. It is one of the symbols mode-line, header-line, vertical-line, left-margin, right-margin, left-fringe, or right-fringe.

In one special case, pos-or-area is a list containing a symbol (one of the symbols listed above) instead of just the symbol. This happens after the imaginary prefix keys for the event are inserted into the input stream. See Key Sequence Input.

x, y

These are the pixel coordinates of the click, relative to the top left corner of window, which is (0 . 0). For the mode or header line, y does not have meaningful data. For the vertical line, x does not have meaningful data.

timestamp

This is the time at which the event occurred, in milliseconds.

object

This is the object on which the click occurred. It is either nil if there is no string property, or it has the form (string . string-pos) when there is a string-type text property at the click position.

string

This is the string on which the click occurred, including any properties.

string-pos

This is the position in the string on which the click occurred, relevant if properties at the click need to be looked up.

text-pos

For clicks on a marginal area or on a fringe, this is the buffer position of the first visible character in the corresponding line in the window. For other events, it is the current buffer position in the window.

col, row

These are the actual coordinates of the glyph under the x, y position, possibly padded with default character width glyphs if x is beyond the last glyph on the line.

image

This is the image object on which the click occurred. It is either nil if there is no image at the position clicked on, or it is an image object as returned by find-image if click was in an image.

dx, dy

These are the pixel coordinates of the click, relative to the top left corner of object, which is (0 . 0). If object is nil, the coordinates are relative to the top left corner of the character glyph clicked on.

width, height

These are the pixel width and height of object or, if this is nil, those of the character glyph clicked on.

For mouse clicks on a scroll-bar, position has this form:

          (window area (portion . whole) timestamp part)

window

This is the window whose scroll-bar was clicked on.

area

This is the scroll bar where the click occurred. It is one of the symbols vertical-scroll-bar or horizontal-scroll-bar.

portion

This is the distance of the click from the top or left end of the scroll bar.

whole

This is the length of the entire scroll bar.

timestamp

This is the time at which the event occurred, in milliseconds.

part

This is the part of the scroll-bar which was clicked on. It is one of the symbols above-handle, handle, below-handle, up, down, top, bottom, and end-scroll.

click-count

This is the number of rapid repeated presses so far of the same mouse button. See Repeat Events.

21.7.5 Drag Events

With Emacs, you can have a drag event without even changing your clothes. A drag event happens every time the user presses a mouse button and then moves the mouse to a different character position before releasing the button. Like all mouse events, drag events are represented in Lisp as lists. The lists record both the starting mouse position and the final position, like this:

     (event-type
      (window1 START-POSITION)
      (window2 END-POSITION))

For a drag event, the name of the symbol event-type contains the prefix ‘drag-’. For example, dragging the mouse with button 2 held down generates a drag-mouse-2 event. The second and third elements of the event give the starting and ending position of the drag. They have the same form as position in a click event (see Click Events) that is not on the scroll bar part of the window. You can access the second element of any mouse event in the same way, with no need to distinguish drag events from others.

The ‘drag-’ prefix follows the modifier key prefixes such as ‘C-’ and ‘M-’.

If read-key-sequence receives a drag event that has no key binding, and the corresponding click event does have a binding, it changes the drag event into a click event at the drag's starting position. This means that you don't have to distinguish between click and drag events unless you want to.

21.7.6 Button-Down Events

Click and drag events happen when the user releases a mouse button. They cannot happen earlier, because there is no way to distinguish a click from a drag until the button is released.

If you want to take action as soon as a button is pressed, you need to handle button-down events.⁹ These occur as soon as a button is pressed. They are represented by lists that look exactly like click events (see Click Events), except that the event-type symbol name contains the prefix ‘down-’. The ‘down-’ prefix follows modifier key prefixes such as ‘C-’ and ‘M-’.

The function read-key-sequence ignores any button-down events that don't have command bindings; therefore, the Emacs command loop ignores them too. This means that you need not worry about defining button-down events unless you want them to do something. The usual reason to define a button-down event is so that you can track mouse motion (by reading motion events) until the button is released. See Motion Events.

21.7.7 Repeat Events

If you press the same mouse button more than once in quick succession without moving the mouse, Emacs generates special repeat mouse events for the second and subsequent presses.

The most common repeat events are double-click events. Emacs generates a double-click event when you click a button twice; the event happens when you release the button (as is normal for all click events).

The event type of a double-click event contains the prefix ‘double-’. Thus, a double click on the second mouse button with <meta> held down comes to the Lisp program as M-double-mouse-2. If a double-click event has no binding, the binding of the corresponding ordinary click event is used to execute it. Thus, you need not pay attention to the double click feature unless you really want to.

When the user performs a double click, Emacs generates first an ordinary click event, and then a double-click event. Therefore, you must design the command binding of the double click event to assume that the single-click command has already run. It must produce the desired results of a double click, starting from the results of a single click.

This is convenient, if the meaning of a double click somehow “builds on” the meaning of a single click—which is recommended user interface design practice for double clicks.

If you click a button, then press it down again and start moving the mouse with the button held down, then you get a double-drag event when you ultimately release the button. Its event type contains ‘double-drag’ instead of just ‘drag’. If a double-drag event has no binding, Emacs looks for an alternate binding as if the event were an ordinary drag.

Before the double-click or double-drag event, Emacs generates a double-down event when the user presses the button down for the second time. Its event type contains ‘double-down’ instead of just ‘down’. If a double-down event has no binding, Emacs looks for an alternate binding as if the event were an ordinary button-down event. If it finds no binding that way either, the double-down event is ignored.

To summarize, when you click a button and then press it again right away, Emacs generates a down event and a click event for the first click, a double-down event when you press the button again, and finally either a double-click or a double-drag event.

If you click a button twice and then press it again, all in quick succession, Emacs generates a triple-down event, followed by either a triple-click or a triple-drag. The event types of these events contain ‘triple’ instead of ‘double’. If any triple event has no binding, Emacs uses the binding that it would use for the corresponding double event.

If you click a button three or more times and then press it again, the events for the presses beyond the third are all triple events. Emacs does not have separate event types for quadruple, quintuple, etc. events. However, you can look at the event list to find out precisely how many times the button was pressed.

21.7.8 Motion Events

Emacs sometimes generates mouse motion events to describe motion of the mouse without any button activity. Mouse motion events are represented by lists that look like this:

     (mouse-movement POSITION)

The second element of the list describes the current position of the mouse, just as in a click event (see Click Events).

The special form track-mouse enables generation of motion events within its body. Outside of track-mouse forms, Emacs does not generate events for mere motion of the mouse, and these events do not appear. See Mouse Tracking.

21.7.9 Focus Events

Window systems provide general ways for the user to control which window gets keyboard input. This choice of window is called the focus. When the user does something to switch between Emacs frames, that generates a focus event. The normal definition of a focus event, in the global keymap, is to select a new frame within Emacs, as the user would expect. See Input Focus.

Focus events are represented in Lisp as lists that look like this:

     (switch-frame new-frame)

where new-frame is the frame switched to.

Some X window managers are set up so that just moving the mouse into a window is enough to set the focus there. Usually, there is no need for a Lisp program to know about the focus change until some other kind of input arrives. Emacs generates a focus event only when the user actually types a keyboard key or presses a mouse button in the new frame; just moving the mouse between frames does not generate a focus event.

A focus event in the middle of a key sequence would garble the sequence. So Emacs never generates a focus event in the middle of a key sequence. If the user changes focus in the middle of a key sequence—that is, after a prefix key—then Emacs reorders the events so that the focus event comes either before or after the multi-event key sequence, and not within it.

21.7.10 Miscellaneous System Events

A few other event types represent occurrences within the system.

(delete-frame (frame))

This kind of event indicates that the user gave the window manager a command to delete a particular window, which happens to be an Emacs frame.

The standard definition of the delete-frame event is to delete frame.

(iconify-frame (frame))

This kind of event indicates that the user iconified frame using the window manager. Its standard definition is ignore; since the frame has already been iconified, Emacs has no work to do. The purpose of this event type is so that you can keep track of such events if you want to.

(make-frame-visible (frame))

This kind of event indicates that the user deiconified frame using the window manager. Its standard definition is ignore; since the frame has already been made visible, Emacs has no work to do.

(wheel-up position)

(wheel-down position)

These kinds of event are generated by moving a mouse wheel. Their usual meaning is a kind of scroll or zoom.

The element position is a list describing the position of the event, in the same format as used in a mouse-click event.

This kind of event is generated only on some kinds of systems. On some systems, mouse-4 and mouse-5 are used instead. For portable code, use the variables mouse-wheel-up-event and mouse-wheel-down-event defined in mwheel.el to determine what event types to expect for the mouse wheel.

(drag-n-drop position files)

This kind of event is generated when a group of files is selected in an application outside of Emacs, and then dragged and dropped onto an Emacs frame.

The element position is a list describing the position of the event, in the same format as used in a mouse-click event, and files is the list of file names that were dragged and dropped. The usual way to handle this event is by visiting these files.

This kind of event is generated, at present, only on some kinds of systems.

help-echo

This kind of event is generated when a mouse pointer moves onto a portion of buffer text which has a help-echo text property. The generated event has this form:

          (help-echo frame help window object pos)

The precise meaning of the event parameters and the way these parameters are used to display the help-echo text are described in Text help-echo.

sigusr1

sigusr2

These events are generated when the Emacs process receives the signals SIGUSR1 and SIGUSR2. They contain no additional data because signals do not carry additional information.

To catch a user signal, bind the corresponding event to an interactive command in the special-event-map (see Active Keymaps). The command is called with no arguments, and the specific signal event is available in last-input-event. For example:

          (defun sigusr-handler ()
            (interactive)
            (message "Caught signal %S" last-input-event))
          
          (define-key special-event-map [sigusr1] 'sigusr-handler)

To test the signal handler, you can make Emacs send a signal to itself:

          (signal-process (emacs-pid) 'sigusr1)

If one of these events arrives in the middle of a key sequence—that is, after a prefix key—then Emacs reorders the events so that this event comes either before or after the multi-event key sequence, not within it.

21.7.11 Event Examples

If the user presses and releases the left mouse button over the same location, that generates a sequence of events like this:

     (down-mouse-1 (#<window 18 on NEWS> 2613 (0 . 38) -864320))
     (mouse-1      (#<window 18 on NEWS> 2613 (0 . 38) -864180))

While holding the control key down, the user might hold down the second mouse button, and drag the mouse from one line to the next. That produces two events, as shown here:

     (C-down-mouse-2 (#<window 18 on NEWS> 3440 (0 . 27) -731219))
     (C-drag-mouse-2 (#<window 18 on NEWS> 3440 (0 . 27) -731219)
                     (#<window 18 on NEWS> 3510 (0 . 28) -729648))

While holding down the meta and shift keys, the user might press the second mouse button on the window's mode line, and then drag the mouse into another window. That produces a pair of events like these:

     (M-S-down-mouse-2 (#<window 18 on NEWS> mode-line (33 . 31) -457844))
     (M-S-drag-mouse-2 (#<window 18 on NEWS> mode-line (33 . 31) -457844)
                       (#<window 20 on carlton-sanskrit.tex> 161 (33 . 3)
                        -453816))

To handle a SIGUSR1 signal, define an interactive function, and bind it to the signal usr1 event sequence:

     (defun usr1-handler ()
       (interactive)
       (message "Got USR1 signal"))
     (global-set-key [signal usr1] 'usr1-handler)

21.7.12 Classifying Events

Every event has an event type, which classifies the event for key binding purposes. For a keyboard event, the event type equals the event value; thus, the event type for a character is the character, and the event type for a function key symbol is the symbol itself. For events that are lists, the event type is the symbol in the car of the list. Thus, the event type is always a symbol or a character.

Two events of the same type are equivalent where key bindings are concerned; thus, they always run the same command. That does not necessarily mean they do the same things, however, as some commands look at the whole event to decide what to do. For example, some commands use the location of a mouse event to decide where in the buffer to act.

Sometimes broader classifications of events are useful. For example, you might want to ask whether an event involved the <META> key, regardless of which other key or mouse button was used.

The functions event-modifiers and event-basic-type are provided to get such information conveniently.

21.7.13 Accessing Mouse Events

This section describes convenient functions for accessing the data in a mouse button or motion event.

These two functions return the starting or ending position of a mouse-button event, as a list of this form:

     (window pos-or-area (x . y) timestamp
      object text-pos (col . row)
      image (dx . dy) (width . height))

These functions take a position list as described above, and return various parts of it.

These functions compute a position list given particular buffer position or screen position. You can access the data in this position list with the functions described above.

21.7.14 Accessing Scroll Bar Events

These functions are useful for decoding scroll bar events.

21.7.15 Putting Keyboard Events in Strings

In most of the places where strings are used, we conceptualize the string as containing text characters—the same kind of characters found in buffers or files. Occasionally Lisp programs use strings that conceptually contain keyboard characters; for example, they may be key sequences or keyboard macro definitions. However, storing keyboard characters in a string is a complex matter, for reasons of historical compatibility, and it is not always possible.

We recommend that new programs avoid dealing with these complexities by not storing keyboard events in strings. Here is how to do that:

• Use vectors instead of strings for key sequences, when you plan to use them for anything other than as arguments to lookup-key and define-key. For example, you can use read-key-sequence-vector instead of read-key-sequence, and this-command-keys-vector instead of this-command-keys.
• Use vectors to write key sequence constants containing meta characters, even when passing them directly to define-key.
• When you have to look at the contents of a key sequence that might be a string, use listify-key-sequence (see Event Input Misc) first, to convert it to a list.

The complexities stem from the modifier bits that keyboard input characters can include. Aside from the Meta modifier, none of these modifier bits can be included in a string, and the Meta modifier is allowed only in special cases.

The earliest GNU Emacs versions represented meta characters as codes in the range of 128 to 255. At that time, the basic character codes ranged from 0 to 127, so all keyboard character codes did fit in a string. Many Lisp programs used ‘\M-’ in string constants to stand for meta characters, especially in arguments to define-key and similar functions, and key sequences and sequences of events were always represented as strings.

When we added support for larger basic character codes beyond 127, and additional modifier bits, we had to change the representation of meta characters. Now the flag that represents the Meta modifier in a character is 2**27 and such numbers cannot be included in a string.

To support programs with ‘\M-’ in string constants, there are special rules for including certain meta characters in a string. Here are the rules for interpreting a string as a sequence of input characters:

• If the keyboard character value is in the range of 0 to 127, it can go in the string unchanged.
• The meta variants of those characters, with codes in the range of 2**27 to 2**27+127, can also go in the string, but you must change their numeric values. You must set the 2**7 bit instead of the 2**27 bit, resulting in a value between 128 and 255. Only a unibyte string can include these codes.
• Non-ASCII characters above 256 can be included in a multibyte string.
• Other keyboard character events cannot fit in a string. This includes keyboard events in the range of 128 to 255.

Functions such as read-key-sequence that construct strings of keyboard input characters follow these rules: they construct vectors instead of strings, when the events won't fit in a string.

When you use the read syntax ‘\M-’ in a string, it produces a code in the range of 128 to 255—the same code that you get if you modify the corresponding keyboard event to put it in the string. Thus, meta events in strings work consistently regardless of how they get into the strings.

However, most programs would do well to avoid these issues by following the recommendations at the beginning of this section.

21.8 Reading Input

The editor command loop reads key sequences using the function read-key-sequence, which uses read-event. These and other functions for event input are also available for use in Lisp programs. See also momentary-string-display in Temporary Displays, and sit-for in Waiting. See Terminal Input, for functions and variables for controlling terminal input modes and debugging terminal input.

For higher-level input facilities, see Minibuffers.

• Key Sequence Input: How to read one key sequence.
• Reading One Event: How to read just one event.
• Event Mod: How Emacs modifies events as they are read.
• Invoking the Input Method: How reading an event uses the input method.
• Quoted Character Input: Asking the user to specify a character.
• Event Input Misc: How to reread or throw away input events.

21.8.1 Key Sequence Input

The command loop reads input a key sequence at a time, by calling read-key-sequence. Lisp programs can also call this function; for example, describe-key uses it to read the key to describe.

If an input character is upper-case (or has the shift modifier) and has no key binding, but its lower-case equivalent has one, then read-key-sequence converts the character to lower case. Note that lookup-key does not perform case conversion in this way.

When reading input results in such a shift-translation, Emacs sets the variable this-command-keys-shift-translated to a non-nil value. Lisp programs can examine this variable if they need to modify their behavior when invoked by shift-translated keys. For example, the function handle-shift-selection examines the value of this variable to determine how to activate or deactivate the region (see handle-shift-selection).

The function read-key-sequence also transforms some mouse events. It converts unbound drag events into click events, and discards unbound button-down events entirely. It also reshuffles focus events and miscellaneous window events so that they never appear in a key sequence with any other events.

When mouse events occur in special parts of a window, such as a mode line or a scroll bar, the event type shows nothing special—it is the same symbol that would normally represent that combination of mouse button and modifier keys. The information about the window part is kept elsewhere in the event—in the coordinates. But read-key-sequence translates this information into imaginary “prefix keys,” all of which are symbols: header-line, horizontal-scroll-bar, menu-bar, mode-line, vertical-line, and vertical-scroll-bar. You can define meanings for mouse clicks in special window parts by defining key sequences using these imaginary prefix keys.

For example, if you call read-key-sequence and then click the mouse on the window's mode line, you get two events, like this:

     (read-key-sequence "Click on the mode line: ")
          ⇒ [mode-line
              (mouse-1
               (#<window 6 on NEWS> mode-line
                (40 . 63) 5959987))]

21.8.2 Reading One Event

The lowest level functions for command input are read-event, read-char, and read-char-exclusive.

None of the above functions suppress quitting.

We emphasize that, unlike read-key-sequence, the functions read-event, read-char, and read-char-exclusive do not perform the translations described in Translation Keymaps. If you wish to read a single key taking these translations into account, use the function read-key:

21.8.3 Modifying and Translating Input Events

Emacs modifies every event it reads according to extra-keyboard-modifiers, then translates it through keyboard-translate-table (if applicable), before returning it from read-event.

Here's an example of using the keyboard-translate-table to make C-x, C-c and C-v perform the cut, copy and paste operations:

     (keyboard-translate ?\C-x 'control-x)
     (keyboard-translate ?\C-c 'control-c)
     (keyboard-translate ?\C-v 'control-v)
     (global-set-key [control-x] 'kill-region)
     (global-set-key [control-c] 'kill-ring-save)
     (global-set-key [control-v] 'yank)

On a graphical terminal that supports extended ASCII input, you can still get the standard Emacs meanings of one of those characters by typing it with the shift key. That makes it a different character as far as keyboard translation is concerned, but it has the same usual meaning.

See Translation Keymaps, for mechanisms that translate event sequences at the level of read-key-sequence.

21.8.4 Invoking the Input Method

The event-reading functions invoke the current input method, if any (see Input Methods). If the value of input-method-function is non-nil, it should be a function; when read-event reads a printing character (including <SPC>) with no modifier bits, it calls that function, passing the character as an argument.

The input method function should return a list of events which should be used as input. (If the list is nil, that means there is no input, so read-event waits for another event.) These events are processed before the events in unread-command-events (see Event Input Misc). Events returned by the input method function are not passed to the input method function again, even if they are printing characters with no modifier bits.

If the input method function calls read-event or read-key-sequence, it should bind input-method-function to nil first, to prevent recursion.

The input method function is not called when reading the second and subsequent events of a key sequence. Thus, these characters are not subject to input method processing. The input method function should test the values of overriding-local-map and overriding-terminal-local-map; if either of these variables is non-nil, the input method should put its argument into a list and return that list with no further processing.

21.8.5 Quoted Character Input

You can use the function read-quoted-char to ask the user to specify a character, and allow the user to specify a control or meta character conveniently, either literally or as an octal character code. The command quoted-insert uses this function.

21.8.6 Miscellaneous Event Input Features

This section describes how to “peek ahead” at events without using them up, how to check for pending input, and how to discard pending input. See also the function read-passwd (see Reading a Password).

21.9 Special Events

Special events are handled at a very low level—as soon as they are read. The read-event function processes these events itself, and never returns them. Instead, it keeps waiting for the first event that is not special and returns that one.

Events that are handled in this way do not echo, they are never grouped into key sequences, and they never appear in the value of last-command-event or (this-command-keys). They do not discard a numeric argument, they cannot be unread with unread-command-events, they may not appear in a keyboard macro, and they are not recorded in a keyboard macro while you are defining one.

These events do, however, appear in last-input-event immediately after they are read, and this is the way for the event's definition to find the actual event.

The events types iconify-frame, make-frame-visible, delete-frame, drag-n-drop, and user signals like sigusr1 are normally handled in this way. The keymap which defines how to handle special events—and which events are special—is in the variable special-event-map (see Active Keymaps).

21.10 Waiting for Elapsed Time or Input

The wait functions are designed to wait for a certain amount of time to pass or until there is input. For example, you may wish to pause in the middle of a computation to allow the user time to view the display. sit-for pauses and updates the screen, and returns immediately if input comes in, while sleep-for pauses without updating the screen.

See Time of Day, for functions to get the current time.

21.11 Quitting

Typing C-g while a Lisp function is running causes Emacs to quit whatever it is doing. This means that control returns to the innermost active command loop.

Typing C-g while the command loop is waiting for keyboard input does not cause a quit; it acts as an ordinary input character. In the simplest case, you cannot tell the difference, because C-g normally runs the command keyboard-quit, whose effect is to quit. However, when C-g follows a prefix key, they combine to form an undefined key. The effect is to cancel the prefix key as well as any prefix argument.

In the minibuffer, C-g has a different definition: it aborts out of the minibuffer. This means, in effect, that it exits the minibuffer and then quits. (Simply quitting would return to the command loop within the minibuffer.) The reason why C-g does not quit directly when the command reader is reading input is so that its meaning can be redefined in the minibuffer in this way. C-g following a prefix key is not redefined in the minibuffer, and it has its normal effect of canceling the prefix key and prefix argument. This too would not be possible if C-g always quit directly.

When C-g does directly quit, it does so by setting the variable quit-flag to t. Emacs checks this variable at appropriate times and quits if it is not nil. Setting quit-flag non-nil in any way thus causes a quit.

At the level of C code, quitting cannot happen just anywhere; only at the special places that check quit-flag. The reason for this is that quitting at other places might leave an inconsistency in Emacs's internal state. Because quitting is delayed until a safe place, quitting cannot make Emacs crash.

Certain functions such as read-key-sequence or read-quoted-char prevent quitting entirely even though they wait for input. Instead of quitting, C-g serves as the requested input. In the case of read-key-sequence, this serves to bring about the special behavior of C-g in the command loop. In the case of read-quoted-char, this is so that C-q can be used to quote a C-g.

You can prevent quitting for a portion of a Lisp function by binding the variable inhibit-quit to a non-nil value. Then, although C-g still sets quit-flag to t as usual, the usual result of this—a quit—is prevented. Eventually, inhibit-quit will become nil again, such as when its binding is unwound at the end of a let form. At that time, if quit-flag is still non-nil, the requested quit happens immediately. This behavior is ideal when you wish to make sure that quitting does not happen within a “critical section” of the program.

In some functions (such as read-quoted-char), C-g is handled in a special way that does not involve quitting. This is done by reading the input with inhibit-quit bound to t, and setting quit-flag to nil before inhibit-quit becomes nil again. This excerpt from the definition of read-quoted-char shows how this is done; it also shows that normal quitting is permitted after the first character of input.

     (defun read-quoted-char (&optional prompt)
       "...documentation..."
       (let ((message-log-max nil) done (first t) (code 0) char)
         (while (not done)
           (let ((inhibit-quit first)
                 ...)
             (and prompt (message "%s-" prompt))
             (setq char (read-event))
             (if inhibit-quit (setq quit-flag nil)))
           ...set the variable code...)
         code))

You can specify a character other than C-g to use for quitting. See the function set-input-mode in Terminal Input.

21.12 Prefix Command Arguments

Most Emacs commands can use a prefix argument, a number specified before the command itself. (Don't confuse prefix arguments with prefix keys.) The prefix argument is at all times represented by a value, which may be nil, meaning there is currently no prefix argument. Each command may use the prefix argument or ignore it.

There are two representations of the prefix argument: raw and numeric. The editor command loop uses the raw representation internally, and so do the Lisp variables that store the information, but commands can request either representation.

Here are the possible values of a raw prefix argument:

• nil, meaning there is no prefix argument. Its numeric value is 1, but numerous commands make a distinction between nil and the integer 1.
• An integer, which stands for itself.
• A list of one element, which is an integer. This form of prefix argument results from one or a succession of C-u's with no digits. The numeric value is the integer in the list, but some commands make a distinction between such a list and an integer alone.
• The symbol -. This indicates that M-- or C-u - was typed, without following digits. The equivalent numeric value is −1, but some commands make a distinction between the integer −1 and the symbol -.

We illustrate these possibilities by calling the following function with various prefixes:

     (defun display-prefix (arg)
       "Display the value of the raw prefix arg."
       (interactive "P")
       (message "%s" arg))

Here are the results of calling display-prefix with various raw prefix arguments:

             M-x display-prefix  -| nil
     
     C-u     M-x display-prefix  -| (4)
     
     C-u C-u M-x display-prefix  -| (16)
     
     C-u 3   M-x display-prefix  -| 3
     
     M-3     M-x display-prefix  -| 3      ; (Same as C-u 3.)
     
     C-u -   M-x display-prefix  -| -
     
     M--     M-x display-prefix  -| -      ; (Same as C-u -.)
     
     C-u - 7 M-x display-prefix  -| -7
     
     M-- 7   M-x display-prefix  -| -7     ; (Same as C-u -7.)

Emacs uses two variables to store the prefix argument: prefix-arg and current-prefix-arg. Commands such as universal-argument that set up prefix arguments for other commands store them in prefix-arg. In contrast, current-prefix-arg conveys the prefix argument to the current command, so setting it has no effect on the prefix arguments for future commands.

Normally, commands specify which representation to use for the prefix argument, either numeric or raw, in the interactive specification. (See Using Interactive.) Alternatively, functions may look at the value of the prefix argument directly in the variable current-prefix-arg, but this is less clean.

The following commands exist to set up prefix arguments for the following command. Do not call them for any other reason.

21.13 Recursive Editing

The Emacs command loop is entered automatically when Emacs starts up. This top-level invocation of the command loop never exits; it keeps running as long as Emacs does. Lisp programs can also invoke the command loop. Since this makes more than one activation of the command loop, we call it recursive editing. A recursive editing level has the effect of suspending whatever command invoked it and permitting the user to do arbitrary editing before resuming that command.

The commands available during recursive editing are the same ones available in the top-level editing loop and defined in the keymaps. Only a few special commands exit the recursive editing level; the others return to the recursive editing level when they finish. (The special commands for exiting are always available, but they do nothing when recursive editing is not in progress.)

All command loops, including recursive ones, set up all-purpose error handlers so that an error in a command run from the command loop will not exit the loop.

Minibuffer input is a special kind of recursive editing. It has a few special wrinkles, such as enabling display of the minibuffer and the minibuffer window, but fewer than you might suppose. Certain keys behave differently in the minibuffer, but that is only because of the minibuffer's local map; if you switch windows, you get the usual Emacs commands.

To invoke a recursive editing level, call the function recursive-edit. This function contains the command loop; it also contains a call to catch with tag exit, which makes it possible to exit the recursive editing level by throwing to exit (see Catch and Throw). If you throw a value other than t, then recursive-edit returns normally to the function that called it. The command C-M-c (exit-recursive-edit) does this. Throwing a t value causes recursive-edit to quit, so that control returns to the command loop one level up. This is called aborting, and is done by C-] (abort-recursive-edit).

Most applications should not use recursive editing, except as part of using the minibuffer. Usually it is more convenient for the user if you change the major mode of the current buffer temporarily to a special major mode, which should have a command to go back to the previous mode. (The e command in Rmail uses this technique.) Or, if you wish to give the user different text to edit “recursively,” create and select a new buffer in a special mode. In this mode, define a command to complete the processing and go back to the previous buffer. (The m command in Rmail does this.)

Recursive edits are useful in debugging. You can insert a call to debug into a function definition as a sort of breakpoint, so that you can look around when the function gets there. debug invokes a recursive edit but also provides the other features of the debugger.

Recursive editing levels are also used when you type C-r in query-replace or use C-x q (kbd-macro-query).

21.14 Disabling Commands

Disabling a command marks the command as requiring user confirmation before it can be executed. Disabling is used for commands which might be confusing to beginning users, to prevent them from using the commands by accident.

The low-level mechanism for disabling a command is to put a non-nil disabled property on the Lisp symbol for the command. These properties are normally set up by the user's init file (see Init File) with Lisp expressions such as this:

     (put 'upcase-region 'disabled t)

For a few commands, these properties are present by default (you can remove them in your init file if you wish).

If the value of the disabled property is a string, the message saying the command is disabled includes that string. For example:

     (put 'delete-region 'disabled
          "Text deleted this way cannot be yanked back!\n")

See Disabling, for the details on what happens when a disabled command is invoked interactively. Disabling a command has no effect on calling it as a function from Lisp programs.

21.15 Command History

The command loop keeps a history of the complex commands that have been executed, to make it convenient to repeat these commands. A complex command is one for which the interactive argument reading uses the minibuffer. This includes any M-x command, any M-: command, and any command whose interactive specification reads an argument from the minibuffer. Explicit use of the minibuffer during the execution of the command itself does not cause the command to be considered complex.

This history list is actually a special case of minibuffer history (see Minibuffer History), with one special twist: the elements are expressions rather than strings.

There are a number of commands devoted to the editing and recall of previous commands. The commands repeat-complex-command, and list-command-history are described in the user manual (see Repetition). Within the minibuffer, the usual minibuffer history commands are available.

21.16 Keyboard Macros

A keyboard macro is a canned sequence of input events that can be considered a command and made the definition of a key. The Lisp representation of a keyboard macro is a string or vector containing the events. Don't confuse keyboard macros with Lisp macros (see Macros).

22 Keymaps

The command bindings of input events are recorded in data structures called keymaps. Each entry in a keymap associates (or binds) an individual event type, either to another keymap or to a command. When an event type is bound to a keymap, that keymap is used to look up the next input event; this continues until a command is found. The whole process is called key lookup.

• Key Sequences: Key sequences as Lisp objects.
• Keymap Basics: Basic concepts of keymaps.
• Format of Keymaps: What a keymap looks like as a Lisp object.
• Creating Keymaps: Functions to create and copy keymaps.
• Inheritance and Keymaps: How one keymap can inherit the bindings of another keymap.
• Prefix Keys: Defining a key with a keymap as its definition.
• Active Keymaps: How Emacs searches the active keymaps for a key binding.
• Searching Keymaps: A pseudo-Lisp summary of searching active maps.
• Controlling Active Maps: Each buffer has a local keymap to override the standard (global) bindings. A minor mode can also override them.
• Key Lookup: Finding a key's binding in one keymap.
• Functions for Key Lookup: How to request key lookup.
• Changing Key Bindings: Redefining a key in a keymap.
• Remapping Commands: A keymap can translate one command to another.
• Translation Keymaps: Keymaps for translating sequences of events.
• Key Binding Commands: Interactive interfaces for redefining keys.
• Scanning Keymaps: Looking through all keymaps, for printing help.
• Menu Keymaps: Defining a menu as a keymap.

22.1 Key Sequences

A key sequence, or key for short, is a sequence of one or more input events that form a unit. Input events include characters, function keys, and mouse actions (see Input Events). The Emacs Lisp representation for a key sequence is a string or vector. Unless otherwise stated, any Emacs Lisp function that accepts a key sequence as an argument can handle both representations.

In the string representation, alphanumeric characters ordinarily stand for themselves; for example, "a" represents a and "2" represents 2. Control character events are prefixed by the substring "\C-", and meta characters by "\M-"; for example, "\C-x" represents the key C-x. In addition, the <TAB>, <RET>, <ESC>, and <DEL> events are represented by "\t", "\r", "\e", and "\d" respectively. The string representation of a complete key sequence is the concatenation of the string representations of the constituent events; thus, "\C-xl" represents the key sequence C-x l.

Key sequences containing function keys, mouse button events, or non-ASCII characters such as C-= or H-a cannot be represented as strings; they have to be represented as vectors.

In the vector representation, each element of the vector represents an input event, in its Lisp form. See Input Events. For example, the vector [?\C-x ?l] represents the key sequence C-x l.

For examples of key sequences written in string and vector representations, Init Rebinding.

22.2 Keymap Basics

A keymap is a Lisp data structure that specifies key bindings for various key sequences.

A single keymap directly specifies definitions for individual events. When a key sequence consists of a single event, its binding in a keymap is the keymap's definition for that event. The binding of a longer key sequence is found by an iterative process: first find the definition of the first event (which must itself be a keymap); then find the second event's definition in that keymap, and so on until all the events in the key sequence have been processed.

If the binding of a key sequence is a keymap, we call the key sequence a prefix key. Otherwise, we call it a complete key (because no more events can be added to it). If the binding is nil, we call the key undefined. Examples of prefix keys are C-c, C-x, and C-x 4. Examples of defined complete keys are X, <RET>, and C-x 4 C-f. Examples of undefined complete keys are C-x C-g, and C-c 3. See Prefix Keys, for more details.

The rule for finding the binding of a key sequence assumes that the intermediate bindings (found for the events before the last) are all keymaps; if this is not so, the sequence of events does not form a unit—it is not really one key sequence. In other words, removing one or more events from the end of any valid key sequence must always yield a prefix key. For example, C-f C-n is not a key sequence; C-f is not a prefix key, so a longer sequence starting with C-f cannot be a key sequence.

The set of possible multi-event key sequences depends on the bindings for prefix keys; therefore, it can be different for different keymaps, and can change when bindings are changed. However, a one-event sequence is always a key sequence, because it does not depend on any prefix keys for its well-formedness.

At any time, several primary keymaps are active—that is, in use for finding key bindings. These are the global map, which is shared by all buffers; the local keymap, which is usually associated with a specific major mode; and zero or more minor mode keymaps, which belong to currently enabled minor modes. (Not all minor modes have keymaps.) The local keymap bindings shadow (i.e., take precedence over) the corresponding global bindings. The minor mode keymaps shadow both local and global keymaps. See Active Keymaps, for details.

22.3 Format of Keymaps

Each keymap is a list whose car is the symbol keymap. The remaining elements of the list define the key bindings of the keymap. A symbol whose function definition is a keymap is also a keymap. Use the function keymapp (see below) to test whether an object is a keymap.

Several kinds of elements may appear in a keymap, after the symbol keymap that begins it:

(type . binding)

This specifies one binding, for events of type type. Each ordinary binding applies to events of a particular event type, which is always a character or a symbol. See Classifying Events. In this kind of binding, binding is a command.

(type item-name [cache] . binding)

This specifies a binding which is also a simple menu item that displays as item-name in the menu. cache, if present, caches certain information for display in the menu. See Simple Menu Items.

(type item-name help-string [cache] . binding)

This is a simple menu item with help string help-string.

(type menu-item . details)

This specifies a binding which is also an extended menu item. This allows use of other features. See Extended Menu Items.

(t . binding)

This specifies a default key binding; any event not bound by other elements of the keymap is given binding as its binding. Default bindings allow a keymap to bind all possible event types without having to enumerate all of them. A keymap that has a default binding completely masks any lower-precedence keymap, except for events explicitly bound to nil (see below).

char-table

If an element of a keymap is a char-table, it counts as holding bindings for all character events with no modifier bits (see modifier bits): element n is the binding for the character with code n. This is a compact way to record lots of bindings. A keymap with such a char-table is called a full keymap. Other keymaps are called sparse keymaps.

string

Aside from elements that specify bindings for keys, a keymap can also have a string as an element. This is called the overall prompt string and makes it possible to use the keymap as a menu. See Defining Menus.

When the binding is nil, it doesn't constitute a definition but it does take precedence over a default binding or a binding in the parent keymap. On the other hand, a binding of nil does not override lower-precedence keymaps; thus, if the local map gives a binding of nil, Emacs uses the binding from the global map.

Keymaps do not directly record bindings for the meta characters. Instead, meta characters are regarded for purposes of key lookup as sequences of two characters, the first of which is <ESC> (or whatever is currently the value of meta-prefix-char). Thus, the key M-a is internally represented as <ESC> a, and its global binding is found at the slot for a in esc-map (see Prefix Keys).

This conversion applies only to characters, not to function keys or other input events; thus, M-<end> has nothing to do with <ESC> <end>.

Here as an example is the local keymap for Lisp mode, a sparse keymap. It defines bindings for <DEL> and <TAB>, plus C-c C-l, M-C-q, and M-C-x.

     lisp-mode-map
     ⇒
     (keymap
      (3 keymap
         ;; C-c C-z
         (26 . run-lisp))
      (27 keymap
          ;; M-C-x, treated as <ESC> C-x
          (24 . lisp-send-defun)
          keymap
          ;; M-C-q, treated as <ESC> C-q
          (17 . indent-sexp))
      ;; This part is inherited from lisp-mode-shared-map.
      keymap
      ;; <DEL>
      (127 . backward-delete-char-untabify)
      (27 keymap
          ;; M-C-q, treated as <ESC> C-q
          (17 . indent-sexp))
      (9 . lisp-indent-line))

22.4 Creating Keymaps

Here we describe the functions for creating keymaps.

22.5 Inheritance and Keymaps

A keymap can inherit the bindings of another keymap, which we call the parent keymap. Such a keymap looks like this:

     (keymap elements... . parent-keymap)

The effect is that this keymap inherits all the bindings of parent-keymap, whatever they may be at the time a key is looked up, but can add to them or override them with elements.

If you change the bindings in parent-keymap using define-key or other key-binding functions, these changed bindings are visible in the inheriting keymap, unless shadowed by the bindings made by elements. The converse is not true: if you use define-key to change bindings in the inheriting keymap, these changes are recorded in elements, but have no effect on parent-keymap.

The proper way to construct a keymap with a parent is to use set-keymap-parent; if you have code that directly constructs a keymap with a parent, please convert the program to use set-keymap-parent instead.

Here is an example showing how to make a keymap that inherits from text-mode-map:

     (let ((map (make-sparse-keymap)))
       (set-keymap-parent map text-mode-map)
       map)

A non-sparse keymap can have a parent too, but this is not very useful. A non-sparse keymap always specifies something as the binding for every numeric character code without modifier bits, even if it is nil, so these character's bindings are never inherited from the parent keymap.

22.6 Prefix Keys

A prefix key is a key sequence whose binding is a keymap. The keymap defines what to do with key sequences that extend the prefix key. For example, C-x is a prefix key, and it uses a keymap that is also stored in the variable ctl-x-map. This keymap defines bindings for key sequences starting with C-x.

Some of the standard Emacs prefix keys use keymaps that are also found in Lisp variables:

• esc-map is the global keymap for the <ESC> prefix key. Thus, the global definitions of all meta characters are actually found here. This map is also the function definition of ESC-prefix.
• help-map is the global keymap for the C-h prefix key.
• mode-specific-map is the global keymap for the prefix key C-c. This map is actually global, not mode-specific, but its name provides useful information about C-c in the output of C-h b (display-bindings), since the main use of this prefix key is for mode-specific bindings.
• ctl-x-map is the global keymap used for the C-x prefix key. This map is found via the function cell of the symbol Control-X-prefix.
• mule-keymap is the global keymap used for the C-x <RET> prefix key.
• ctl-x-4-map is the global keymap used for the C-x 4 prefix key.
• ctl-x-5-map is the global keymap used for the C-x 5 prefix key.
• 2C-mode-map is the global keymap used for the C-x 6 prefix key.
• vc-prefix-map is the global keymap used for the C-x v prefix key.
• goto-map is the global keymap used for the M-g prefix key.
• search-map is the global keymap used for the M-s prefix key.
• facemenu-keymap is the global keymap used for the M-o prefix key.
• The other Emacs prefix keys are C-x @, C-x a i, C-x <ESC> and <ESC> <ESC>. They use keymaps that have no special names.

The keymap binding of a prefix key is used for looking up the event that follows the prefix key. (It may instead be a symbol whose function definition is a keymap. The effect is the same, but the symbol serves as a name for the prefix key.) Thus, the binding of C-x is the symbol Control-X-prefix, whose function cell holds the keymap for C-x commands. (The same keymap is also the value of ctl-x-map.)

Prefix key definitions can appear in any active keymap. The definitions of C-c, C-x, C-h and <ESC> as prefix keys appear in the global map, so these prefix keys are always available. Major and minor modes can redefine a key as a prefix by putting a prefix key definition for it in the local map or the minor mode's map. See Active Keymaps.

If a key is defined as a prefix in more than one active map, then its various definitions are in effect merged: the commands defined in the minor mode keymaps come first, followed by those in the local map's prefix definition, and then by those from the global map.

In the following example, we make C-p a prefix key in the local keymap, in such a way that C-p is identical to C-x. Then the binding for C-p C-f is the function find-file, just like C-x C-f. The key sequence C-p 6 is not found in any active keymap.

     (use-local-map (make-sparse-keymap))
         ⇒ nil
     (local-set-key "\C-p" ctl-x-map)
         ⇒ nil
     (key-binding "\C-p\C-f")
         ⇒ find-file
     
     (key-binding "\C-p6")
         ⇒ nil

22.7 Active Keymaps

Emacs normally contains many keymaps; at any given time, just a few of them are active, meaning that they participate in the interpretation of user input. All the active keymaps are used together to determine what command to execute when a key is entered.

Normally the active keymaps are the keymap property keymap, the keymaps of any enabled minor modes, the current buffer's local keymap, and the global keymap, in that order. Emacs searches for each input key sequence in all these keymaps. See Searching Keymaps, for more details of this procedure.

When the key sequence starts with a mouse event (optionally preceded by a symbolic prefix), the active keymaps are determined based on the position in that event. If the event happened on a string embedded with a display, before-string, or after-string property (see Special Properties), the non-nil map properties of the string override those of the buffer (if the underlying buffer text contains map properties in its text properties or overlays, they are ignored).

The global keymap holds the bindings of keys that are defined regardless of the current buffer, such as C-f. The variable global-map holds this keymap, which is always active.

Each buffer may have another keymap, its local keymap, which may contain new or overriding definitions for keys. The current buffer's local keymap is always active except when overriding-local-map overrides it. The local-map text or overlay property can specify an alternative local keymap for certain parts of the buffer; see Special Properties.

Each minor mode can have a keymap; if it does, the keymap is active when the minor mode is enabled. Modes for emulation can specify additional active keymaps through the variable emulation-mode-map-alists.

The highest precedence normal keymap comes from the keymap text or overlay property. If that is non-nil, it is the first keymap to be processed, in normal circumstances.

However, there are also special ways for programs to substitute other keymaps for some of those. The variable overriding-local-map, if non-nil, specifies a keymap that replaces all the usual active keymaps except the global keymap. Another way to do this is with overriding-terminal-local-map; it operates on a per-terminal basis. These variables are documented below.

Since every buffer that uses the same major mode normally uses the same local keymap, you can think of the keymap as local to the mode. A change to the local keymap of a buffer (using local-set-key, for example) is seen also in the other buffers that share that keymap.

The local keymaps that are used for Lisp mode and some other major modes exist even if they have not yet been used. These local keymaps are the values of variables such as lisp-mode-map. For most major modes, which are less frequently used, the local keymap is constructed only when the mode is used for the first time in a session.

The minibuffer has local keymaps, too; they contain various completion and exit commands. See Intro to Minibuffers.

Emacs has other keymaps that are used in a different way—translating events within read-key-sequence. See Translation Keymaps.

See Standard Keymaps, for a list of standard keymaps.

22.8 Searching the Active Keymaps

After translation of event subsequences (see Translation Keymaps) Emacs looks for them in the active keymaps. Here is a pseudo-Lisp description of the order and conditions for searching them:

     (or (if overriding-terminal-local-map
             (find-in overriding-terminal-local-map)
           (if overriding-local-map
               (find-in overriding-local-map)
             (or (find-in (get-char-property (point) 'keymap))
                 (find-in-any emulation-mode-map-alists)
                 (find-in-any minor-mode-overriding-map-alist)
                 (find-in-any minor-mode-map-alist)
                 (if (get-text-property (point) 'local-map)
                     (find-in (get-char-property (point) 'local-map))
                   (find-in (current-local-map))))))
         (find-in (current-global-map)))

The find-in and find-in-any are pseudo functions that search in one keymap and in an alist of keymaps, respectively. (Searching a single keymap for a binding is called key lookup; see Key Lookup.) If the key sequence starts with a mouse event, or a symbolic prefix event followed by a mouse event, that event's position is used instead of point and the current buffer. Mouse events on an embedded string use non-nil text properties from that string instead of the buffer.

1. The function finally found may be remapped (see Remapping Commands).
2. Characters that are bound to self-insert-command are translated according to translation-table-for-input before insertion.
3. current-active-maps returns a list of the currently active keymaps at point.
4. When a match is found (see Key Lookup), if the binding in the keymap is a function, the search is over. However if the keymap entry is a symbol with a value or a string, Emacs replaces the input key sequences with the variable's value or the string, and restarts the search of the active keymaps.

22.9 Controlling the Active Keymaps

current-local-map returns a reference to the local keymap, not a copy of it; if you use define-key or other functions on it you will alter local bindings.

22.10 Key Lookup

Key lookup is the process of finding the binding of a key sequence from a given keymap. The execution or use of the binding is not part of key lookup.

Key lookup uses just the event type of each event in the key sequence; the rest of the event is ignored. In fact, a key sequence used for key lookup may designate a mouse event with just its types (a symbol) instead of the entire event (a list). See Input Events. Such a “key sequence” is insufficient for command-execute to run, but it is sufficient for looking up or rebinding a key.

When the key sequence consists of multiple events, key lookup processes the events sequentially: the binding of the first event is found, and must be a keymap; then the second event's binding is found in that keymap, and so on until all the events in the key sequence are used up. (The binding thus found for the last event may or may not be a keymap.) Thus, the process of key lookup is defined in terms of a simpler process for looking up a single event in a keymap. How that is done depends on the type of object associated with the event in that keymap.

Let's use the term keymap entry to describe the value found by looking up an event type in a keymap. (This doesn't include the item string and other extra elements in a keymap element for a menu item, because lookup-key and other key lookup functions don't include them in the returned value.) While any Lisp object may be stored in a keymap as a keymap entry, not all make sense for key lookup. Here is a table of the meaningful types of keymap entries:

nil

nil means that the events used so far in the lookup form an undefined key. When a keymap fails to mention an event type at all, and has no default binding, that is equivalent to a binding of nil for that event type.

command

The events used so far in the lookup form a complete key, and command is its binding. See What Is a Function.

array

The array (either a string or a vector) is a keyboard macro. The events used so far in the lookup form a complete key, and the array is its binding. See Keyboard Macros, for more information.

keymap

The events used so far in the lookup form a prefix key. The next event of the key sequence is looked up in keymap.

list

The meaning of a list depends on what it contains:

• If the car of list is the symbol keymap, then the list is a keymap, and is treated as a keymap (see above).
• If the car of list is lambda, then the list is a lambda expression. This is presumed to be a function, and is treated as such (see above). In order to execute properly as a key binding, this function must be a command—it must have an interactive specification. See Defining Commands.
• If the car of list is a keymap and the cdr is an event type, then this is an indirect entry:

                 (othermap . othertype)
  

  When key lookup encounters an indirect entry, it looks up instead the binding of othertype in othermap and uses that.

  This feature permits you to define one key as an alias for another key. For example, an entry whose car is the keymap called esc-map and whose cdr is 32 (the code for <SPC>) means, “Use the global binding of Meta-<SPC>, whatever that may be.”

symbol

The function definition of symbol is used in place of symbol. If that too is a symbol, then this process is repeated, any number of times. Ultimately this should lead to an object that is a keymap, a command, or a keyboard macro. A list is allowed if it is a keymap or a command, but indirect entries are not understood when found via symbols.

Note that keymaps and keyboard macros (strings and vectors) are not valid functions, so a symbol with a keymap, string, or vector as its function definition is invalid as a function. It is, however, valid as a key binding. If the definition is a keyboard macro, then the symbol is also valid as an argument to command-execute (see Interactive Call).

The symbol undefined is worth special mention: it means to treat the key as undefined. Strictly speaking, the key is defined, and its binding is the command undefined; but that command does the same thing that is done automatically for an undefined key: it rings the bell (by calling ding) but does not signal an error.

undefined is used in local keymaps to override a global key binding and make the key “undefined” locally. A local binding of nil would fail to do this because it would not override the global binding.

anything else

If any other type of object is found, the events used so far in the lookup form a complete key, and the object is its binding, but the binding is not executable as a command.

In short, a keymap entry may be a keymap, a command, a keyboard macro, a symbol that leads to one of them, or an indirection or nil. Here is an example of a sparse keymap with two characters bound to commands and one bound to another keymap. This map is the normal value of emacs-lisp-mode-map. Note that 9 is the code for <TAB>, 127 for <DEL>, 27 for <ESC>, 17 for C-q and 24 for C-x.

     (keymap (9 . lisp-indent-line)
             (127 . backward-delete-char-untabify)
             (27 keymap (17 . indent-sexp) (24 . eval-defun)))

22.11 Functions for Key Lookup

Here are the functions and variables pertaining to key lookup.

22.12 Changing Key Bindings

The way to rebind a key is to change its entry in a keymap. If you change a binding in the global keymap, the change is effective in all buffers (though it has no direct effect in buffers that shadow the global binding with a local one). If you change the current buffer's local map, that usually affects all buffers using the same major mode. The global-set-key and local-set-key functions are convenient interfaces for these operations (see Key Binding Commands). You can also use define-key, a more general function; then you must specify explicitly the map to change.

When choosing the key sequences for Lisp programs to rebind, please follow the Emacs conventions for use of various keys (see Key Binding Conventions).

In writing the key sequence to rebind, it is good to use the special escape sequences for control and meta characters (see String Type). The syntax ‘\C-’ means that the following character is a control character and ‘\M-’ means that the following character is a meta character. Thus, the string "\M-x" is read as containing a single M-x, "\C-f" is read as containing a single C-f, and "\M-\C-x" and "\C-\M-x" are both read as containing a single C-M-x. You can also use this escape syntax in vectors, as well as others that aren't allowed in strings; one example is ‘[?\C-\H-x home]’. See Character Type.

The key definition and lookup functions accept an alternate syntax for event types in a key sequence that is a vector: you can use a list containing modifier names plus one base event (a character or function key name). For example, (control ?a) is equivalent to ?\C-a and (hyper control left) is equivalent to C-H-left. One advantage of such lists is that the precise numeric codes for the modifier bits don't appear in compiled files.

The functions below signal an error if keymap is not a keymap, or if key is not a string or vector representing a key sequence. You can use event types (symbols) as shorthand for events that are lists. The kbd macro (see Key Sequences) is a convenient way to specify the key sequence.

This example creates a sparse keymap and makes a number of bindings in it:

     (setq map (make-sparse-keymap))
         ⇒ (keymap)
     (define-key map "\C-f" 'forward-char)
         ⇒ forward-char
     map
         ⇒ (keymap (6 . forward-char))
     
     ;; Build sparse submap for C-x and bind f in that.
     (define-key map (kbd "C-x f") 'forward-word)
         ⇒ forward-word
     map
     ⇒ (keymap
         (24 keymap                ; C-x
             (102 . forward-word)) ;      f
         (6 . forward-char))       ; C-f
     
     ;; Bind C-p to the ctl-x-map.
     (define-key map (kbd "C-p") ctl-x-map)
     ;; ctl-x-map
     ⇒ [nil ... find-file ... backward-kill-sentence]
     
     ;; Bind C-f to foo in the ctl-x-map.
     (define-key map (kbd "C-p C-f") 'foo)
     ⇒ 'foo
     map
     ⇒ (keymap     ; Note foo in ctl-x-map.
         (16 keymap [nil ... foo ... backward-kill-sentence])
         (24 keymap
             (102 . forward-word))
         (6 . forward-char))

Note that storing a new binding for C-p C-f actually works by changing an entry in ctl-x-map, and this has the effect of changing the bindings of both C-p C-f and C-x C-f in the default global map.

The function substitute-key-definition scans a keymap for keys that have a certain binding and rebinds them with a different binding. Another feature which is cleaner and can often produce the same results to remap one command into another (see Remapping Commands).

22.13 Remapping Commands

A special kind of key binding, using a special “key sequence” which includes a command name, has the effect of remapping that command into another. Here's how it works. You make a key binding for a key sequence that starts with the dummy event remap, followed by the command name you want to remap. Specify the remapped definition as the definition in this binding. The remapped definition is usually a command name, but it can be any valid definition for a key binding.

Here's an example. Suppose that My mode uses special commands my-kill-line and my-kill-word, which should be invoked instead of kill-line and kill-word. It can establish this by making these two command-remapping bindings in its keymap:

     (define-key my-mode-map [remap kill-line] 'my-kill-line)
     (define-key my-mode-map [remap kill-word] 'my-kill-word)

Whenever my-mode-map is an active keymap, if the user types C-k, Emacs will find the standard global binding of kill-line (assuming nobody has changed it). But my-mode-map remaps kill-line to my-kill-line, so instead of running kill-line, Emacs runs my-kill-line.

Remapping only works through a single level. In other words,

     (define-key my-mode-map [remap kill-line] 'my-kill-line)
     (define-key my-mode-map [remap my-kill-line] 'my-other-kill-line)

does not have the effect of remapping kill-line into my-other-kill-line. If an ordinary key binding specifies kill-line, this keymap will remap it to my-kill-line; if an ordinary binding specifies my-kill-line, this keymap will remap it to my-other-kill-line.

To undo the remapping of a command, remap it to nil; e.g.

     (define-key my-mode-map [remap kill-line] nil)

22.14 Keymaps for Translating Sequences of Events

This section describes keymaps that are used during reading a key sequence, to translate certain event sequences into others. read-key-sequence checks every subsequence of the key sequence being read, as it is read, against input-decode-map, then local-function-key-map, and then against key-translation-map.

You can use input-decode-map, local-function-key-map, or key-translation-map for more than simple aliases, by using a function, instead of a key sequence, as the “translation” of a key. Then this function is called to compute the translation of that key.

The key translation function receives one argument, which is the prompt that was specified in read-key-sequence—or nil if the key sequence is being read by the editor command loop. In most cases you can ignore the prompt value.

If the function reads input itself, it can have the effect of altering the event that follows. For example, here's how to define C-c h to turn the character that follows into a Hyper character:

     (defun hyperify (prompt)
       (let ((e (read-event)))
         (vector (if (numberp e)
                     (logior (lsh 1 24) e)
                   (if (memq 'hyper (event-modifiers e))
                       e
                     (add-event-modifier "H-" e))))))
     
     (defun add-event-modifier (string e)
       (let ((symbol (if (symbolp e) e (car e))))
         (setq symbol (intern (concat string
                                      (symbol-name symbol))))
         (if (symbolp e)
             symbol
           (cons symbol (cdr e)))))
     
     (define-key local-function-key-map "\C-ch" 'hyperify)

If you have enabled keyboard character set decoding using set-keyboard-coding-system, decoding is done after the translations listed above. See Terminal I/O Encoding. However, in future Emacs versions, character set decoding may be done at an earlier stage.

22.15 Commands for Binding Keys

This section describes some convenient interactive interfaces for changing key bindings. They work by calling define-key.

People often use global-set-key in their init files (see Init File) for simple customization. For example,

     (global-set-key (kbd "C-x C-\\") 'next-line)

or

     (global-set-key [?\C-x ?\C-\\] 'next-line)

or

     (global-set-key [(control ?x) (control ?\\)] 'next-line)

redefines C-x C-\ to move down a line.

     (global-set-key [M-mouse-1] 'mouse-set-point)

redefines the first (leftmost) mouse button, entered with the Meta key, to set point where you click.

Be careful when using non-ASCII text characters in Lisp specifications of keys to bind. If these are read as multibyte text, as they usually will be in a Lisp file (see Loading Non-ASCII), you must type the keys as multibyte too. For instance, if you use this:

     (global-set-key "ö" 'my-function) ; bind o-umlaut

or

     (global-set-key ?ö 'my-function) ; bind o-umlaut

and your language environment is multibyte Latin-1, these commands actually bind the multibyte character with code 2294, not the unibyte Latin-1 character with code 246 (M-v). In order to use this binding, you need to enter the multibyte Latin-1 character as keyboard input. One way to do this is by using an appropriate input method (see Input Methods).

If you want to use a unibyte character in the key binding, you can construct the key sequence string using multibyte-char-to-unibyte or string-make-unibyte (see Converting Representations).

22.16 Scanning Keymaps

This section describes functions used to scan all the current keymaps for the sake of printing help information.

22.17 Menu Keymaps

A keymap can operate as a menu as well as defining bindings for keyboard keys and mouse buttons. Menus are usually actuated with the mouse, but they can function with the keyboard also. If a menu keymap is active for the next input event, that activates the keyboard menu feature.

• Defining Menus: How to make a keymap that defines a menu.
• Mouse Menus: How users actuate the menu with the mouse.
• Keyboard Menus: How users actuate the menu with the keyboard.
• Menu Example: Making a simple menu.
• Menu Bar: How to customize the menu bar.
• Tool Bar: A tool bar is a row of images.
• Modifying Menus: How to add new items to a menu.

22.17.1 Defining Menus

A keymap acts as a menu if it has an overall prompt string, which is a string that appears as an element of the keymap. (See Format of Keymaps.) The string should describe the purpose of the menu's commands. Emacs displays the overall prompt string as the menu title in some cases, depending on the toolkit (if any) used for displaying menus.¹⁰ Keyboard menus also display the overall prompt string.

The easiest way to construct a keymap with a prompt string is to specify the string as an argument when you call make-keymap, make-sparse-keymap (see Creating Keymaps), or define-prefix-command (see Definition of define-prefix-command). If you do not want the keymap to operate as a menu, don't specify a prompt string for it.

The menu's items are the bindings in the keymap. Each binding associates an event type to a definition, but the event types have no significance for the menu appearance. (Usually we use pseudo-events, symbols that the keyboard cannot generate, as the event types for menu item bindings.) The menu is generated entirely from the bindings that correspond in the keymap to these events.

The order of items in the menu is the same as the order of bindings in the keymap. Since define-key puts new bindings at the front, you should define the menu items starting at the bottom of the menu and moving to the top, if you care about the order. When you add an item to an existing menu, you can specify its position in the menu using define-key-after (see Modifying Menus).

• Simple Menu Items: A simple kind of menu key binding, limited in capabilities.
• Extended Menu Items: More powerful menu item definitions let you specify keywords to enable various features.
• Menu Separators: Drawing a horizontal line through a menu.
• Alias Menu Items: Using command aliases in menu items.

22.17.1.1 Simple Menu Items

The simpler (and original) way to define a menu item is to bind some event type (it doesn't matter what event type) to a binding like this:

     (item-string . real-binding)

The car, item-string, is the string to be displayed in the menu. It should be short—preferably one to three words. It should describe the action of the command it corresponds to. Note that it is not generally possible to display non-ASCII text in menus. It will work for keyboard menus and will work to a large extent when Emacs is built with the Gtk+ toolkit.¹¹

You can also supply a second string, called the help string, as follows:

     (item-string help . real-binding)

help specifies a “help-echo” string to display while the mouse is on that item in the same way as help-echo text properties (see Help display).

As far as define-key is concerned, item-string and help-string are part of the event's binding. However, lookup-key returns just real-binding, and only real-binding is used for executing the key.

If real-binding is nil, then item-string appears in the menu but cannot be selected.

If real-binding is a symbol and has a non-nil menu-enable property, that property is an expression that controls whether the menu item is enabled. Every time the keymap is used to display a menu, Emacs evaluates the expression, and it enables the menu item only if the expression's value is non-nil. When a menu item is disabled, it is displayed in a “fuzzy” fashion, and cannot be selected.

The menu bar does not recalculate which items are enabled every time you look at a menu. This is because the X toolkit requires the whole tree of menus in advance. To force recalculation of the menu bar, call force-mode-line-update (see Mode Line Format).

You've probably noticed that menu items show the equivalent keyboard key sequence (if any) to invoke the same command. To save time on recalculation, menu display caches this information in a sublist in the binding, like this:

     (item-string [help] (key-binding-data) . real-binding)

Don't put these sublists in the menu item yourself; menu display calculates them automatically. Don't mention keyboard equivalents in the item strings themselves, since that is redundant.

22.17.1.2 Extended Menu Items

An extended-format menu item is a more flexible and also cleaner alternative to the simple format. You define an event type with a binding that's a list starting with the symbol menu-item. For a non-selectable string, the binding looks like this:

     (menu-item item-name)

A string starting with two or more dashes specifies a separator line; see Menu Separators.

To define a real menu item which can be selected, the extended format binding looks like this:

     (menu-item item-name real-binding
         . item-property-list)

Here, item-name is an expression which evaluates to the menu item string. Thus, the string need not be a constant. The third element, real-binding, is the command to execute. The tail of the list, item-property-list, has the form of a property list which contains other information.

When an equivalent keyboard key binding is cached, the extended menu item binding looks like this:

     (menu-item item-name real-binding (key-binding-data)
         . item-property-list)

Here is a table of the properties that are supported:

:enable form

The result of evaluating form determines whether the item is enabled (non-nil means yes). If the item is not enabled, you can't really click on it.

:visible form

The result of evaluating form determines whether the item should actually appear in the menu (non-nil means yes). If the item does not appear, then the menu is displayed as if this item were not defined at all.

:help help

The value of this property, help, specifies a “help-echo” string to display while the mouse is on that item. This is displayed in the same way as help-echo text properties (see Help display). Note that this must be a constant string, unlike the help-echo property for text and overlays.

:button (type . selected)

This property provides a way to define radio buttons and toggle buttons. The car, type, says which: it should be :toggle or :radio. The cdr, selected, should be a form; the result of evaluating it says whether this button is currently selected.

A toggle is a menu item which is labeled as either “on” or “off” according to the value of selected. The command itself should toggle selected, setting it to t if it is nil, and to nil if it is t. Here is how the menu item to toggle the debug-on-error flag is defined:

          (menu-item "Debug on Error" toggle-debug-on-error
                     :button (:toggle
                              . (and (boundp 'debug-on-error)
                                     debug-on-error)))

This works because toggle-debug-on-error is defined as a command which toggles the variable debug-on-error.

Radio buttons are a group of menu items, in which at any time one and only one is “selected.” There should be a variable whose value says which one is selected at any time. The selected form for each radio button in the group should check whether the variable has the right value for selecting that button. Clicking on the button should set the variable so that the button you clicked on becomes selected.

:key-sequence key-sequence

This property specifies which key sequence is likely to be bound to the same command invoked by this menu item. If you specify the right key sequence, that makes preparing the menu for display run much faster.

If you specify the wrong key sequence, it has no effect; before Emacs displays key-sequence in the menu, it verifies that key-sequence is really equivalent to this menu item.

:key-sequence nil

This property indicates that there is normally no key binding which is equivalent to this menu item. Using this property saves time in preparing the menu for display, because Emacs does not need to search the keymaps for a keyboard equivalent for this menu item.

However, if the user has rebound this item's definition to a key sequence, Emacs ignores the :keys property and finds the keyboard equivalent anyway.

:keys string

This property specifies that string is the string to display as the keyboard equivalent for this menu item. You can use the ‘\\[...]’ documentation construct in string.

:filter filter-fn

This property provides a way to compute the menu item dynamically. The property value filter-fn should be a function of one argument; when it is called, its argument will be real-binding. The function should return the binding to use instead.

Emacs can call this function at any time that it does redisplay or operates on menu data structures, so you should write it so it can safely be called at any time.

22.17.1.3 Menu Separators

A menu separator is a kind of menu item that doesn't display any text—instead, it divides the menu into subparts with a horizontal line. A separator looks like this in the menu keymap:

     (menu-item separator-type)

where separator-type is a string starting with two or more dashes.

In the simplest case, separator-type consists of only dashes. That specifies the default kind of separator. (For compatibility, "" and - also count as separators.)

Certain other values of separator-type specify a different style of separator. Here is a table of them:

"--no-line"

"--space"

An extra vertical space, with no actual line.

"--single-line"

A single line in the menu's foreground color.

"--double-line"

A double line in the menu's foreground color.

"--single-dashed-line"

A single dashed line in the menu's foreground color.

"--double-dashed-line"

A double dashed line in the menu's foreground color.

"--shadow-etched-in"

A single line with a 3D sunken appearance. This is the default, used separators consisting of dashes only.

"--shadow-etched-out"

A single line with a 3D raised appearance.

"--shadow-etched-in-dash"

A single dashed line with a 3D sunken appearance.

"--shadow-etched-out-dash"

A single dashed line with a 3D raised appearance.

"--shadow-double-etched-in"

Two lines with a 3D sunken appearance.

"--shadow-double-etched-out"

Two lines with a 3D raised appearance.

"--shadow-double-etched-in-dash"

Two dashed lines with a 3D sunken appearance.

"--shadow-double-etched-out-dash"

Two dashed lines with a 3D raised appearance.

You can also give these names in another style, adding a colon after the double-dash and replacing each single dash with capitalization of the following word. Thus, "--:singleLine", is equivalent to "--single-line".

Some systems and display toolkits don't really handle all of these separator types. If you use a type that isn't supported, the menu displays a similar kind of separator that is supported.

22.17.1.4 Alias Menu Items

Sometimes it is useful to make menu items that use the “same” command but with different enable conditions. The best way to do this in Emacs now is with extended menu items; before that feature existed, it could be done by defining alias commands and using them in menu items. Here's an example that makes two aliases for toggle-read-only and gives them different enable conditions:

     (defalias 'make-read-only 'toggle-read-only)
     (put 'make-read-only 'menu-enable '(not buffer-read-only))
     (defalias 'make-writable 'toggle-read-only)
     (put 'make-writable 'menu-enable 'buffer-read-only)

When using aliases in menus, often it is useful to display the equivalent key bindings for the “real” command name, not the aliases (which typically don't have any key bindings except for the menu itself). To request this, give the alias symbol a non-nil menu-alias property. Thus,

     (put 'make-read-only 'menu-alias t)
     (put 'make-writable 'menu-alias t)

causes menu items for make-read-only and make-writable to show the keyboard bindings for toggle-read-only.

22.17.2 Menus and the Mouse

The usual way to make a menu keymap produce a menu is to make it the definition of a prefix key. (A Lisp program can explicitly pop up a menu and receive the user's choice—see Pop-Up Menus.)

If the prefix key ends with a mouse event, Emacs handles the menu keymap by popping up a visible menu, so that the user can select a choice with the mouse. When the user clicks on a menu item, the event generated is whatever character or symbol has the binding that brought about that menu item. (A menu item may generate a series of events if the menu has multiple levels or comes from the menu bar.)

It's often best to use a button-down event to trigger the menu. Then the user can select a menu item by releasing the button.

A single keymap can appear as multiple menu panes, if you explicitly arrange for this. The way to do this is to make a keymap for each pane, then create a binding for each of those maps in the main keymap of the menu. Give each of these bindings an item string that starts with ‘@’. The rest of the item string becomes the name of the pane. See the file lisp/mouse.el for an example of this. Any ordinary bindings with ‘@’-less item strings are grouped into one pane, which appears along with the other panes explicitly created for the submaps.

X toolkit menus don't have panes; instead, they can have submenus. Every nested keymap becomes a submenu, whether the item string starts with ‘@’ or not. In a toolkit version of Emacs, the only thing special about ‘@’ at the beginning of an item string is that the ‘@’ doesn't appear in the menu item.

Multiple keymaps that define the same menu prefix key produce separate panes or separate submenus.

22.17.3 Menus and the Keyboard

When a prefix key ending with a keyboard event (a character or function key) has a definition that is a menu keymap, the keymap operates as a keyboard menu; the user specifies the next event by choosing a menu item with the keyboard.

Emacs displays the keyboard menu with the map's overall prompt string, followed by the alternatives (the item strings of the map's bindings), in the echo area. If the bindings don't all fit at once, the user can type <SPC> to see the next line of alternatives. Successive uses of <SPC> eventually get to the end of the menu and then cycle around to the beginning. (The variable menu-prompt-more-char specifies which character is used for this; <SPC> is the default.)

When the user has found the desired alternative from the menu, he or she should type the corresponding character—the one whose binding is that alternative.

This way of using menus in an Emacs-like editor was inspired by the Hierarkey system.

22.17.4 Menu Example

Here is a complete example of defining a menu keymap. It is the definition of the ‘Replace’ submenu in the ‘Edit’ menu in the menu bar, and it uses the extended menu item format (see Extended Menu Items). First we create the keymap, and give it a name:

     (defvar menu-bar-replace-menu (make-sparse-keymap "Replace"))

Next we define the menu items:

     (define-key menu-bar-replace-menu [tags-repl-continue]
       '(menu-item "Continue Replace" tags-loop-continue
                   :help "Continue last tags replace operation"))
     (define-key menu-bar-replace-menu [tags-repl]
       '(menu-item "Replace in tagged files" tags-query-replace
                   :help "Interactively replace a regexp in all tagged files"))
     (define-key menu-bar-replace-menu [separator-replace-tags]
       '(menu-item "--"))
     ;; ...

Note the symbols which the bindings are “made for”; these appear inside square brackets, in the key sequence being defined. In some cases, this symbol is the same as the command name; sometimes it is different. These symbols are treated as “function keys,” but they are not real function keys on the keyboard. They do not affect the functioning of the menu itself, but they are “echoed” in the echo area when the user selects from the menu, and they appear in the output of where-is and apropos.

The menu in this example is intended for use with the mouse. If a menu is intended for use with the keyboard, that is, if it is bound to a key sequence ending with a keyboard event, then the menu items should be bound to characters or “real” function keys, that can be typed with the keyboard.

The binding whose definition is ("--") is a separator line. Like a real menu item, the separator has a key symbol, in this case separator-replace-tags. If one menu has two separators, they must have two different key symbols.

Here is how we make this menu appear as an item in the parent menu:

     (define-key menu-bar-edit-menu [replace]
       (list 'menu-item "Replace" menu-bar-replace-menu))

Note that this incorporates the submenu keymap, which is the value of the variable menu-bar-replace-menu, rather than the symbol menu-bar-replace-menu itself. Using that symbol in the parent menu item would be meaningless because menu-bar-replace-menu is not a command.

If you wanted to attach the same replace menu to a mouse click, you can do it this way:

     (define-key global-map [C-S-down-mouse-1]
        menu-bar-replace-menu)

22.17.5 The Menu Bar

Most window systems allow each frame to have a menu bar—a permanently displayed menu stretching horizontally across the top of the frame. (In order for a frame to display a menu bar, its menu-bar-lines parameter must be greater than zero. See Layout Parameters.)

The items of the menu bar are the subcommands of the fake “function key” menu-bar, as defined in the active keymaps.

To add an item to the menu bar, invent a fake “function key” of your own (let's call it key), and make a binding for the key sequence [menu-bar key]. Most often, the binding is a menu keymap, so that pressing a button on the menu bar item leads to another menu.

When more than one active keymap defines the same fake function key for the menu bar, the item appears just once. If the user clicks on that menu bar item, it brings up a single, combined menu containing all the subcommands of that item—the global subcommands, the local subcommands, and the minor mode subcommands.

The variable overriding-local-map is normally ignored when determining the menu bar contents. That is, the menu bar is computed from the keymaps that would be active if overriding-local-map were nil. See Active Keymaps.

Here's an example of setting up a menu bar item:

     (modify-frame-parameters (selected-frame)
                              '((menu-bar-lines . 2)))
     
     ;; Make a menu keymap (with a prompt string)
     ;; and make it the menu bar item's definition.
     (define-key global-map [menu-bar words]
       (cons "Words" (make-sparse-keymap "Words")))
     
     ;; Define specific subcommands in this menu.
     (define-key global-map
       [menu-bar words forward]
       '("Forward word" . forward-word))
     (define-key global-map
       [menu-bar words backward]
       '("Backward word" . backward-word))

A local keymap can cancel a menu bar item made by the global keymap by rebinding the same fake function key with undefined as the binding. For example, this is how Dired suppresses the ‘Edit’ menu bar item:

     (define-key dired-mode-map [menu-bar edit] 'undefined)

Here, edit is the fake function key used by the global map for the ‘Edit’ menu bar item. The main reason to suppress a global menu bar item is to regain space for mode-specific items.

Next to every menu bar item, Emacs displays a key binding that runs the same command (if such a key binding exists). This serves as a convenient hint for users who do not know the key binding. If a command has multiple bindings, Emacs normally displays the first one it finds. You can specify one particular key binding by assigning an :advertised-binding symbol property to the command. For instance, the following tells Emacs to show C-/ for the undo menu item:

     (put 'undo :advertised-binding [?\C-/])

If the :advertised-binding property specifies a key binding that the command does not actually have, it is ignored.

22.17.6 Tool bars

A tool bar is a row of icons at the top of a frame, that execute commands when you click on them—in effect, a kind of graphical menu bar.

The frame parameter tool-bar-lines (X resource ‘toolBar’) controls how many lines' worth of height to reserve for the tool bar. A zero value suppresses the tool bar. If the value is nonzero, and auto-resize-tool-bars is non-nil, the tool bar expands and contracts automatically as needed to hold the specified contents.

If the value of auto-resize-tool-bars is grow-only, the tool bar expands automatically, but does not contract automatically. To contract the tool bar, the user has to redraw the frame by entering C-l.

The tool bar contents are controlled by a menu keymap attached to a fake “function key” called tool-bar (much like the way the menu bar is controlled). So you define a tool bar item using define-key, like this:

     (define-key global-map [tool-bar key] item)

where key is a fake “function key” to distinguish this item from other items, and item is a menu item key binding (see Extended Menu Items), which says how to display this item and how it behaves.

The usual menu keymap item properties, :visible, :enable, :button, and :filter, are useful in tool bar bindings and have their normal meanings. The real-binding in the item must be a command, not a keymap; in other words, it does not work to define a tool bar icon as a prefix key.

The :help property specifies a “help-echo” string to display while the mouse is on that item. This is displayed in the same way as help-echo text properties (see Help display).

In addition, you should use the :image property; this is how you specify the image to display in the tool bar:

:image image

images is either a single image specification or a vector of four image specifications. If you use a vector of four, one of them is used, depending on circumstances:

item 0

Used when the item is enabled and selected.

item 1

Used when the item is enabled and deselected.

item 2

Used when the item is disabled and selected.

item 3

Used when the item is disabled and deselected.

If image is a single image specification, Emacs draws the tool bar button in disabled state by applying an edge-detection algorithm to the image.

The :rtl property specifies an alternative image to use for right-to-left languages. Only the Gtk+ version of Emacs supports this at present.

The default tool bar is defined so that items specific to editing do not appear for major modes whose command symbol has a mode-class property of special (see Major Mode Conventions). Major modes may add items to the global bar by binding [tool-bar foo] in their local map. It makes sense for some major modes to replace the default tool bar items completely, since not many can be accommodated conveniently, and the default bindings make this easy by using an indirection through tool-bar-map.

There are two convenience functions for defining tool bar items, as follows.

You can define a special meaning for clicking on a tool bar item with the shift, control, meta, etc., modifiers. You do this by setting up additional items that relate to the original item through the fake function keys. Specifically, the additional items should use the modified versions of the same fake function key used to name the original item.

Thus, if the original item was defined this way,

     (define-key global-map [tool-bar shell]
       '(menu-item "Shell" shell
                   :image (image :type xpm :file "shell.xpm")))

then here is how you can define clicking on the same tool bar image with the shift modifier:

     (define-key global-map [tool-bar S-shell] 'some-command)

See Function Keys, for more information about how to add modifiers to function keys.

22.17.7 Modifying Menus

When you insert a new item in an existing menu, you probably want to put it in a particular place among the menu's existing items. If you use define-key to add the item, it normally goes at the front of the menu. To put it elsewhere in the menu, use define-key-after:

23 Major and Minor Modes

A mode is a set of definitions that customize Emacs and can be turned on and off while you edit. There are two varieties of modes: major modes, which are mutually exclusive and used for editing particular kinds of text, and minor modes, which provide features that users can enable individually.

This chapter describes how to write both major and minor modes, how to indicate them in the mode line, and how they run hooks supplied by the user. For related topics such as keymaps and syntax tables, see Keymaps, and Syntax Tables.

• Hooks: How to use hooks; how to write code that provides hooks.
• Major Modes: Defining major modes.
• Minor Modes: Defining minor modes.
• Mode Line Format: Customizing the text that appears in the mode line.
• Imenu: How a mode can provide a menu of definitions in the buffer.
• Font Lock Mode: How modes can highlight text according to syntax.
• Auto-Indentation: How to teach Emacs to indent for a major mode.
• Desktop Save Mode: How modes can have buffer state saved between Emacs sessions.

23.1 Hooks

A hook is a variable where you can store a function or functions to be called on a particular occasion by an existing program. Emacs provides hooks for the sake of customization. Most often, hooks are set up in the init file (see Init File), but Lisp programs can set them also. See Standard Hooks, for a list of standard hook variables.

Most of the hooks in Emacs are normal hooks. These variables contain lists of functions to be called with no arguments. By convention, whenever the hook name ends in ‘-hook’, that tells you it is normal. We try to make all hooks normal, as much as possible, so that you can use them in a uniform way.

Every major mode function is supposed to run a normal hook called the mode hook as the one of the last steps of initialization. This makes it easy for a user to customize the behavior of the mode, by overriding the buffer-local variable assignments already made by the mode. Most minor mode functions also run a mode hook at the end. But hooks are used in other contexts too. For example, the hook suspend-hook runs just before Emacs suspends itself (see Suspending Emacs).

The recommended way to add a hook function to a normal hook is by calling add-hook (see below). The hook functions may be any of the valid kinds of functions that funcall accepts (see What Is a Function). Most normal hook variables are initially void; add-hook knows how to deal with this. You can add hooks either globally or buffer-locally with add-hook.

If the hook variable's name does not end with ‘-hook’, that indicates it is probably an abnormal hook. That means the hook functions are called with arguments, or their return values are used in some way. The hook's documentation says how the functions are called. You can use add-hook to add a function to an abnormal hook, but you must write the function to follow the hook's calling convention.

By convention, abnormal hook names end in ‘-functions’ or ‘-hooks’. If the variable's name ends in ‘-function’, then its value is just a single function, not a list of functions.

• Running Hooks: How to run a hook.
• Setting Hooks: How to put functions on a hook, or remove them.

23.1.1 Running Hooks

At the appropriate times, Emacs uses the run-hooks function and the other functions below to run particular hooks.

23.1.2 Setting Hooks

Here's an example that uses a mode hook to turn on Auto Fill mode when in Lisp Interaction mode:

     (add-hook 'lisp-interaction-mode-hook 'turn-on-auto-fill)

23.2 Major Modes

Major modes specialize Emacs for editing particular kinds of text. Each buffer has only one major mode at a time. For each major mode there is a function to switch to that mode in the current buffer; its name should end in ‘-mode’. These functions work by setting buffer-local variable bindings and other data associated with the buffer, such as a local keymap. The effect lasts until you switch to another major mode in the same buffer.

• Major Mode Basics
• Major Mode Conventions: Coding conventions for keymaps, etc.
• Auto Major Mode: How Emacs chooses the major mode automatically.
• Mode Help: Finding out how to use a mode.
• Derived Modes: Defining a new major mode based on another major mode.
• Generic Modes: Defining a simple major mode that supports comment syntax and Font Lock mode.
• Mode Hooks: Hooks run at the end of major mode functions.
• Example Major Modes: Text mode and Lisp modes.

23.2.1 Major Mode Basics

The least specialized major mode is called Fundamental mode. This mode has no mode-specific definitions or variable settings, so each Emacs command behaves in its default manner, and each option is in its default state. All other major modes redefine various keys and options. For example, Lisp Interaction mode provides special key bindings for C-j (eval-print-last-sexp), <TAB> (lisp-indent-line), and other keys.

When you need to write several editing commands to help you perform a specialized editing task, creating a new major mode is usually a good idea. In practice, writing a major mode is easy (in contrast to writing a minor mode, which is often difficult).

If the new mode is similar to an old one, it is often unwise to modify the old one to serve two purposes, since it may become harder to use and maintain. Instead, copy and rename an existing major mode definition and alter the copy—or use the define-derived-mode macro to define a derived mode (see Derived Modes). For example, Rmail Edit mode is a major mode that is very similar to Text mode except that it provides two additional commands. Its definition is distinct from that of Text mode, but uses that of Text mode.

Even if the new mode is not an obvious derivative of any other mode, we recommend to use define-derived-mode, since it automatically enforces the most important coding conventions for you.

For a very simple programming language major mode that handles comments and fontification, you can use define-generic-mode. See Generic Modes.

Rmail Edit mode offers an example of changing the major mode temporarily for a buffer, so it can be edited in a different way (with ordinary Emacs commands rather than Rmail commands). In such cases, the temporary major mode usually provides a command to switch back to the buffer's usual mode (Rmail mode, in this case). You might be tempted to present the temporary redefinitions inside a recursive edit and restore the usual ones when the user exits; but this is a bad idea because it constrains the user's options when it is done in more than one buffer: recursive edits must be exited most-recently-entered first. Using an alternative major mode avoids this limitation. See Recursive Editing.

The standard GNU Emacs Lisp library directory tree contains the code for several major modes, in files such as text-mode.el, texinfo.el, lisp-mode.el, c-mode.el, and rmail.el. They are found in various subdirectories of the lisp directory. You can study these libraries to see how modes are written. Text mode is perhaps the simplest major mode aside from Fundamental mode. Rmail mode is a complicated and specialized mode.

23.2.2 Major Mode Conventions

The code for existing major modes follows various coding conventions, including conventions for local keymap and syntax table initialization, global names, and hooks. Please follow these conventions when you define a new major mode. (Fundamental mode is an exception to many of these conventions, because its definition is to present the global state of Emacs.)

This list of conventions is only partial, because each major mode should aim for consistency in general with other Emacs major modes. This makes Emacs as a whole more coherent. It is impossible to list here all the possible points where this issue might come up; if the Emacs developers point out an area where your major mode deviates from the usual conventions, please make it compatible.

• Define a command whose name ends in ‘-mode’, with no arguments, that switches to the new mode in the current buffer. This command should set up the keymap, syntax table, and buffer-local variables in an existing buffer, without changing the buffer's contents.
• Write a documentation string for this command that describes the special commands available in this mode. C-h m (describe-mode) in your mode will display this string.

  The documentation string may include the special documentation substrings, ‘\[command]’, ‘\{keymap}’, and ‘\<keymap>’, which enable the documentation to adapt automatically to the user's own key bindings. See Keys in Documentation.

• The major mode command should start by calling kill-all-local-variables. This runs the normal hook change-major-mode-hook, then gets rid of the buffer-local variables of the major mode previously in effect. See Creating Buffer-Local.
• The major mode command should set the variable major-mode to the major mode command symbol. This is how describe-mode discovers which documentation to print.
• The major mode command should set the variable mode-name to the “pretty” name of the mode, usually a string (but see Mode Line Data, for other possible forms). The name of the mode appears in the mode line.
• Since all global names are in the same name space, all the global variables, constants, and functions that are part of the mode should have names that start with the major mode name (or with an abbreviation of it if the name is long). See Coding Conventions.
• In a major mode for editing some kind of structured text, such as a programming language, indentation of text according to structure is probably useful. So the mode should set indent-line-function to a suitable function, and probably customize other variables for indentation. See Auto-Indentation.
• The major mode should usually have its own keymap, which is used as the local keymap in all buffers in that mode. The major mode command should call use-local-map to install this local map. See Active Keymaps, for more information.

  This keymap should be stored permanently in a global variable named modename-mode-map. Normally the library that defines the mode sets this variable.

  See Tips for Defining, for advice about how to write the code to set up the mode's keymap variable.

• The key sequences bound in a major mode keymap should usually start with C-c, followed by a control character, a digit, or {, }, <, >, : or ;. The other punctuation characters are reserved for minor modes, and ordinary letters are reserved for users.

  A major mode can also rebind the keys M-n, M-p and M-s. The bindings for M-n and M-p should normally be some kind of “moving forward and backward,” but this does not necessarily mean cursor motion.

  It is legitimate for a major mode to rebind a standard key sequence if it provides a command that does “the same job” in a way better suited to the text this mode is used for. For example, a major mode for editing a programming language might redefine C-M-a to “move to the beginning of a function” in a way that works better for that language.

  It is also legitimate for a major mode to rebind a standard key sequence whose standard meaning is rarely useful in that mode. For instance, minibuffer modes rebind M-r, whose standard meaning is rarely of any use in the minibuffer. Major modes such as Dired or Rmail that do not allow self-insertion of text can reasonably redefine letters and other printing characters as special commands.

• Major modes for editing text should not define <RET> to do anything other than insert a newline. However, it is ok for specialized modes for text that users don't directly edit, such as Dired and Info modes, to redefine <RET> to do something entirely different.
• Major modes should not alter options that are primarily a matter of user preference, such as whether Auto-Fill mode is enabled. Leave this to each user to decide. However, a major mode should customize other variables so that Auto-Fill mode will work usefully if the user decides to use it.
• The mode may have its own syntax table or may share one with other related modes. If it has its own syntax table, it should store this in a variable named modename-mode-syntax-table. See Syntax Tables.
• If the mode handles a language that has a syntax for comments, it should set the variables that define the comment syntax. See Options Controlling Comments.
• The mode may have its own abbrev table or may share one with other related modes. If it has its own abbrev table, it should store this in a variable named modename-mode-abbrev-table. If the major mode command defines any abbrevs itself, it should pass t for the system-flag argument to define-abbrev. See Defining Abbrevs.
• The mode should specify how to do highlighting for Font Lock mode, by setting up a buffer-local value for the variable font-lock-defaults (see Font Lock Mode).
• The mode should specify how Imenu should find the definitions or sections of a buffer, by setting up a buffer-local value for the variable imenu-generic-expression, for the two variables imenu-prev-index-position-function and imenu-extract-index-name-function, or for the variable imenu-create-index-function (see Imenu).
• The mode can specify a local value for eldoc-documentation-function to tell ElDoc mode how to handle this mode.
• The mode can specify how to complete various keywords by adding to the special hook completion-at-point-functions.
• Use defvar or defcustom to set mode-related variables, so that they are not reinitialized if they already have a value. (Such reinitialization could discard customizations made by the user.)
• To make a buffer-local binding for an Emacs customization variable, use make-local-variable in the major mode command, not make-variable-buffer-local. The latter function would make the variable local to every buffer in which it is subsequently set, which would affect buffers that do not use this mode. It is undesirable for a mode to have such global effects. See Buffer-Local Variables.

  With rare exceptions, the only reasonable way to use make-variable-buffer-local in a Lisp package is for a variable which is used only within that package. Using it on a variable used by other packages would interfere with them.

• Each major mode should have a normal mode hook named modename-mode-hook. The very last thing the major mode command should do is to call run-mode-hooks. This runs the mode hook, and then runs the normal hook after-change-major-mode-hook. See Mode Hooks.
• The major mode command may start by calling some other major mode command (called the parent mode) and then alter some of its settings. A mode that does this is called a derived mode. The recommended way to define one is to use the define-derived-mode macro, but this is not required. Such a mode should call the parent mode command inside a delay-mode-hooks form. (Using define-derived-mode does this automatically.) See Derived Modes, and Mode Hooks.
• If something special should be done if the user switches a buffer from this mode to any other major mode, this mode can set up a buffer-local value for change-major-mode-hook (see Creating Buffer-Local).
• If this mode is appropriate only for specially-prepared text, then the major mode command symbol should have a property named mode-class with value special, put on as follows:

            (put 'funny-mode 'mode-class 'special)
  

  This tells Emacs that new buffers created while the current buffer is in Funny mode should not inherit Funny mode, in case the default value of major-mode is nil. Modes such as Dired, Rmail, and Buffer List use this feature.

  The define-derived-mode macro automatically marks the derived mode as special if the parent mode is special. The special mode special-mode provides a convenient parent for other special modes to inherit from; it sets buffer-read-only to t, and does little else.

• If you want to make the new mode the default for files with certain recognizable names, add an element to auto-mode-alist to select the mode for those file names (see Auto Major Mode). If you define the mode command to autoload, you should add this element in the same file that calls autoload. If you use an autoload cookie for the mode command, you can also use an autoload cookie for the form that adds the element (see autoload cookie). If you do not autoload the mode command, it is sufficient to add the element in the file that contains the mode definition.
• In the comments that document the file, you should provide a sample autoload form and an example of how to add to auto-mode-alist, that users can include in their init files (see Init File).
• The top-level forms in the file defining the mode should be written so that they may be evaluated more than once without adverse consequences. Even if you never load the file more than once, someone else will.

23.2.3 How Emacs Chooses a Major Mode

Based on information in the file name or in the file itself, Emacs automatically selects a major mode for the new buffer when a file is visited. It also processes local variables specified in the file text.

23.2.4 Getting Help about a Major Mode

The describe-mode function is used to provide information about major modes. It is normally called with C-h m. The describe-mode function uses the value of major-mode, which is why every major mode function needs to set the major-mode variable.

23.2.5 Defining Derived Modes

The recommended way to define a new major mode is to derive it from an existing one using define-derived-mode. If there is no closely related mode, you can inherit from text-mode, special-mode, or in the worst case fundamental-mode.

23.2.6 Generic Modes

Generic modes are simple major modes with basic support for comment syntax and Font Lock mode. To define a generic mode, use the macro define-generic-mode. See the file generic-x.el for some examples of the use of define-generic-mode.

23.2.7 Mode Hooks

Every major mode function should finish by running its mode hook and the mode-independent normal hook after-change-major-mode-hook. It does this by calling run-mode-hooks. If the major mode is a derived mode, that is if it calls another major mode (the parent mode) in its body, it should do this inside delay-mode-hooks so that the parent won't run these hooks itself. Instead, the derived mode's call to run-mode-hooks runs the parent's mode hook too. See Major Mode Conventions.

Emacs versions before Emacs 22 did not have delay-mode-hooks. When user-implemented major modes have not been updated to use it, they won't entirely follow these conventions: they may run the parent's mode hook too early, or fail to run after-change-major-mode-hook. If you encounter such a major mode, please correct it to follow these conventions.

When you defined a major mode using define-derived-mode, it automatically makes sure these conventions are followed. If you define a major mode “by hand,” not using define-derived-mode, use the following functions to handle these conventions automatically.

23.2.8 Major Mode Examples

Text mode is perhaps the simplest mode besides Fundamental mode. Here are excerpts from text-mode.el that illustrate many of the conventions listed above:

     ;; Create the syntax table for this mode.
     (defvar text-mode-syntax-table
       (let ((st (make-syntax-table)))
         (modify-syntax-entry ?\" ".   " st)
         (modify-syntax-entry ?\\ ".   " st)
         ;; Add `p' so M-c on `hello' leads to `Hello', not `hello'.
         (modify-syntax-entry ?' "w p" st)
         st)
       "Syntax table used while in `text-mode'.")
     
     ;; Create the keymap for this mode.
     (defvar text-mode-map
       (let ((map (make-sparse-keymap)))
         (define-key map "\e\t" 'ispell-complete-word)
         (define-key map "\es" 'center-line)
         (define-key map "\eS" 'center-paragraph)
         map)
       "Keymap for `text-mode'.
     Many other modes, such as Mail mode, Outline mode
     and Indented Text mode, inherit all the commands
     defined in this map.")

Here is how the actual mode command is defined now:

     (define-derived-mode text-mode nil "Text"
       "Major mode for editing text written for humans to read.
     In this mode, paragraphs are delimited only by blank or white lines.
     You can thus get the full benefit of adaptive filling
      (see the variable `adaptive-fill-mode').
     \\{text-mode-map}
     Turning on Text mode runs the normal hook `text-mode-hook'."
       (set (make-local-variable 'text-mode-variant) t)
       ;; These two lines are a feature added recently.
       (set (make-local-variable 'require-final-newline)
            mode-require-final-newline)
       (set (make-local-variable 'indent-line-function) 'indent-relative))

(The last line is redundant nowadays, since indent-relative is the default value, and we'll delete it in a future version.)

Here is how it was defined formerly, before define-derived-mode existed:

     ;; This isn't needed nowadays, since define-derived-mode does it.
     (define-abbrev-table 'text-mode-abbrev-table ()
       "Abbrev table used while in text mode.")
     
     (defun text-mode ()
       "Major mode for editing text intended for humans to read...
      Special commands: \\{text-mode-map}
     Turning on text-mode runs the hook `text-mode-hook'."
       (interactive)
       (kill-all-local-variables)
       (use-local-map text-mode-map)
       (setq local-abbrev-table text-mode-abbrev-table)
       (set-syntax-table text-mode-syntax-table)
       ;; These four lines are absent from the current version
       ;; not because this is done some other way, but rather
       ;; because nowadays Text mode uses the normal definition of paragraphs.
       (set (make-local-variable 'paragraph-start)
            (concat "[ \t]*$\\|" page-delimiter))
       (set (make-local-variable 'paragraph-separate) paragraph-start)
       (set (make-local-variable 'indent-line-function) 'indent-relative-maybe)
       (setq mode-name "Text")
       (setq major-mode 'text-mode)
       (run-mode-hooks 'text-mode-hook)) ; Finally, this permits the user to
                                         ;   customize the mode with a hook.

The three Lisp modes (Lisp mode, Emacs Lisp mode, and Lisp Interaction mode) have more features than Text mode and the code is correspondingly more complicated. Here are excerpts from lisp-mode.el that illustrate how these modes are written.

     ;; Create mode-specific table variables.
     (defvar lisp-mode-syntax-table nil "")
     (defvar lisp-mode-abbrev-table nil "")
     
     (defvar emacs-lisp-mode-syntax-table
       (let ((table (make-syntax-table)))
         (let ((i 0))
     
           ;; Set syntax of chars up to ‘0’ to say they are
           ;;   part of symbol names but not words.
           ;;   (The digit ‘0’ is 48 in the ASCII character set.)
           (while (< i ?0)
             (modify-syntax-entry i "_   " table)
             (setq i (1+ i)))
           ;; ... similar code follows for other character ranges.
           ;; Then set the syntax codes for characters that are special in Lisp.
           (modify-syntax-entry ?  "    " table)
           (modify-syntax-entry ?\t "    " table)
           (modify-syntax-entry ?\f "    " table)
           (modify-syntax-entry ?\n ">   " table)
           ;; Give CR the same syntax as newline, for selective-display.
           (modify-syntax-entry ?\^m ">   " table)
           (modify-syntax-entry ?\; "<   " table)
           (modify-syntax-entry ?` "'   " table)
           (modify-syntax-entry ?' "'   " table)
           (modify-syntax-entry ?, "'   " table)
           ;; ...likewise for many other characters...
           (modify-syntax-entry ?\( "()  " table)
           (modify-syntax-entry ?\) ")(  " table)
           (modify-syntax-entry ?\[ "(]  " table)
           (modify-syntax-entry ?\] ")[  " table))
         table))
     ;; Create an abbrev table for lisp-mode.
     (define-abbrev-table 'lisp-mode-abbrev-table ())

The three modes for Lisp share much of their code. For instance, each calls the following function to set various variables:

     (defun lisp-mode-variables (lisp-syntax)
       (when lisp-syntax
         (set-syntax-table lisp-mode-syntax-table))
       (setq local-abbrev-table lisp-mode-abbrev-table)
       ...

In Lisp and most programming languages, we want the paragraph commands to treat only blank lines as paragraph separators. And the modes should understand the Lisp conventions for comments. The rest of lisp-mode-variables sets this up:

       (set (make-local-variable 'paragraph-start) (concat page-delimiter "\\|$" ))
       (set (make-local-variable 'paragraph-separate) paragraph-start)
       ...
       (set (make-local-variable 'comment-indent-function) 'lisp-comment-indent))
       ...

Each of the different Lisp modes has a slightly different keymap. For example, Lisp mode binds C-c C-z to run-lisp, but the other Lisp modes do not. However, all Lisp modes have some commands in common. The following code sets up the common commands:

     (defvar shared-lisp-mode-map
       (let ((map (make-sparse-keymap)))
         (define-key shared-lisp-mode-map "\e\C-q" 'indent-sexp)
         (define-key shared-lisp-mode-map "\177"
                     'backward-delete-char-untabify)
         map)
       "Keymap for commands shared by all sorts of Lisp modes.")

And here is the code to set up the keymap for Lisp mode:

     (defvar lisp-mode-map
       (let ((map (make-sparse-keymap)))
         (set-keymap-parent map shared-lisp-mode-map)
         (define-key map "\e\C-x" 'lisp-eval-defun)
         (define-key map "\C-c\C-z" 'run-lisp)
         map)
       "Keymap for ordinary Lisp mode...")

Finally, here is the complete major mode function definition for Lisp mode.

     (defun lisp-mode ()
       "Major mode for editing Lisp code for Lisps other than GNU Emacs Lisp.
     Commands:
     Delete converts tabs to spaces as it moves back.
     Blank lines separate paragraphs.  Semicolons start comments.
     \\{lisp-mode-map}
     Note that `run-lisp' may be used either to start an inferior Lisp job
     or to switch back to an existing one.
     
     Entry to this mode calls the value of `lisp-mode-hook'
     if that value is non-nil."
       (interactive)
       (kill-all-local-variables)
       (use-local-map lisp-mode-map)          ; Select the mode's keymap.
       (setq major-mode 'lisp-mode)           ; This is how describe-mode
                                              ;   finds out what to describe.
       (setq mode-name "Lisp")                ; This goes into the mode line.
       (lisp-mode-variables t)                ; This defines various variables.
       (set (make-local-variable 'comment-start-skip)
            "\\(\\(^\\|[^\\\\\n]\\)\\(\\\\\\\\\\)*\\)\\(;+\\|#|\\) *")
       (set (make-local-variable 'font-lock-keywords-case-fold-search) t)
       (setq imenu-case-fold-search t)
       (set-syntax-table lisp-mode-syntax-table)
       (run-mode-hooks 'lisp-mode-hook))      ; This permits the user to use a
                                              ;   hook to customize the mode.

23.3 Minor Modes

A minor mode provides features that users may enable or disable independently of the choice of major mode. Minor modes can be enabled individually or in combination. Minor modes would be better named “generally available, optional feature modes,” except that such a name would be unwieldy.

A minor mode is not usually meant as a variation of a single major mode. Usually they are general and can apply to many major modes. For example, Auto Fill mode works with any major mode that permits text insertion. To be general, a minor mode must be effectively independent of the things major modes do.

A minor mode is often much more difficult to implement than a major mode. One reason is that you should be able to activate and deactivate minor modes in any order. A minor mode should be able to have its desired effect regardless of the major mode and regardless of the other minor modes in effect.

Often the biggest problem in implementing a minor mode is finding a way to insert the necessary hook into the rest of Emacs. Minor mode keymaps make this easier than it used to be.

• Minor Mode Conventions: Tips for writing a minor mode.
• Keymaps and Minor Modes: How a minor mode can have its own keymap.
• Defining Minor Modes: A convenient facility for defining minor modes.

23.3.1 Conventions for Writing Minor Modes

There are conventions for writing minor modes just as there are for major modes. Several of the major mode conventions apply to minor modes as well: those regarding the name of the mode initialization function, the names of global symbols, the use of a hook at the end of the initialization function, and the use of keymaps and other tables.

In addition, there are several conventions that are specific to minor modes. (The easiest way to follow all the conventions is to use the macro define-minor-mode; Defining Minor Modes.)

• Make a variable whose name ends in ‘-mode’ to control the minor mode. We call this the mode variable. The minor mode command should set this variable (nil to disable; anything else to enable).

  If possible, implement the mode so that setting the variable automatically enables or disables the mode. Then the minor mode command does not need to do anything except set the variable.

  This variable is used in conjunction with the minor-mode-alist to display the minor mode name in the mode line. It can also enable or disable a minor mode keymap. Individual commands or hooks can also check the variable's value.

  If you want the minor mode to be enabled separately in each buffer, make the variable buffer-local.

• Define a command whose name is the same as the mode variable. Its job is to enable and disable the mode by setting the variable.

  The command should accept one optional argument. If the argument is nil, it should toggle the mode (turn it on if it is off, and off if it is on). It should turn the mode on if the argument is a positive integer, the symbol t, or a list whose car is one of those. It should turn the mode off if the argument is a negative integer or zero, the symbol -, or a list whose car is a negative integer or zero. The meaning of other arguments is not specified.

  Here is an example taken from the definition of transient-mark-mode. It shows the use of transient-mark-mode as a variable that enables or disables the mode's behavior, and also shows the proper way to toggle, enable or disable the minor mode based on the raw prefix argument value.

            (setq transient-mark-mode
                  (if (null arg) (not transient-mark-mode)
                    (> (prefix-numeric-value arg) 0)))
  

• Add an element to minor-mode-alist for each minor mode (see Definition of minor-mode-alist), if you want to indicate the minor mode in the mode line. This element should be a list of the following form:

            (mode-variable string)
  

  Here mode-variable is the variable that controls enabling of the minor mode, and string is a short string, starting with a space, to represent the mode in the mode line. These strings must be short so that there is room for several of them at once.

  When you add an element to minor-mode-alist, use assq to check for an existing element, to avoid duplication. For example:

            (unless (assq 'leif-mode minor-mode-alist)
              (setq minor-mode-alist
                    (cons '(leif-mode " Leif") minor-mode-alist)))
  

  or like this, using add-to-list (see List Variables):

            (add-to-list 'minor-mode-alist '(leif-mode " Leif"))
  

Global minor modes distributed with Emacs should if possible support enabling and disabling via Custom (see Customization). To do this, the first step is to define the mode variable with defcustom, and specify :type 'boolean.

If just setting the variable is not sufficient to enable the mode, you should also specify a :set method which enables the mode by invoking the mode command. Note in the variable's documentation string that setting the variable other than via Custom may not take effect.

Also mark the definition with an autoload cookie (see autoload cookie), and specify a :require so that customizing the variable will load the library that defines the mode. This will copy suitable definitions into loaddefs.el so that users can use customize-option to enable the mode. For example:

     
     ;;;###autoload
     (defcustom msb-mode nil
       "Toggle msb-mode.
     Setting this variable directly does not take effect;
     use either \\[customize] or the function `msb-mode'."
       :set 'custom-set-minor-mode
       :initialize 'custom-initialize-default
       :version "20.4"
       :type    'boolean
       :group   'msb
       :require 'msb)

23.3.2 Keymaps and Minor Modes

Each minor mode can have its own keymap, which is active when the mode is enabled. To set up a keymap for a minor mode, add an element to the alist minor-mode-map-alist. See Definition of minor-mode-map-alist.

One use of minor mode keymaps is to modify the behavior of certain self-inserting characters so that they do something else as well as self-insert. In general, this is the only way to do that, since the facilities for customizing self-insert-command are limited to special cases (designed for abbrevs and Auto Fill mode). (Do not try substituting your own definition of self-insert-command for the standard one. The editor command loop handles this function specially.)

The key sequences bound in a minor mode should consist of C-c followed by one of .,/?`'"[]\|~!#$%^&*()-_+=. (The other punctuation characters are reserved for major modes.)

23.3.3 Defining Minor Modes

The macro define-minor-mode offers a convenient way of implementing a mode in one self-contained definition.

The initial value must be nil except in cases where (1) the mode is preloaded in Emacs, or (2) it is painless for loading to enable the mode even though the user did not request it. For instance, if the mode has no effect unless something else is enabled, and will always be loaded by that time, enabling it by default is harmless. But these are unusual circumstances. Normally, the initial value must be nil.

The name easy-mmode-define-minor-mode is an alias for this macro.

Here is an example of using define-minor-mode:

     (define-minor-mode hungry-mode
       "Toggle Hungry mode.
     With no argument, this command toggles the mode.
     Non-null prefix argument turns on the mode.
     Null prefix argument turns off the mode.
     
     When Hungry mode is enabled, the control delete key
     gobbles all preceding whitespace except the last.
     See the command \\[hungry-electric-delete]."
      ;; The initial value.
      nil
      ;; The indicator for the mode line.
      " Hungry"
      ;; The minor mode bindings.
      '(([C-backspace] . hungry-electric-delete))
      :group 'hunger)

This defines a minor mode named “Hungry mode,” a command named hungry-mode to toggle it, a variable named hungry-mode which indicates whether the mode is enabled, and a variable named hungry-mode-map which holds the keymap that is active when the mode is enabled. It initializes the keymap with a key binding for C-<DEL>. It puts the variable hungry-mode into custom group hunger. There are no body forms—many minor modes don't need any.

Here's an equivalent way to write it:

     (define-minor-mode hungry-mode
       "Toggle Hungry mode.
     With no argument, this command toggles the mode.
     Non-null prefix argument turns on the mode.
     Null prefix argument turns off the mode.
     
     When Hungry mode is enabled, the control delete key
     gobbles all preceding whitespace except the last.
     See the command \\[hungry-electric-delete]."
      ;; The initial value.
      :init-value nil
      ;; The indicator for the mode line.
      :lighter " Hungry"
      ;; The minor mode bindings.
      :keymap
      '(([C-backspace] . hungry-electric-delete)
        ([C-M-backspace]
         . (lambda ()
             (interactive)
             (hungry-electric-delete t))))
      :group 'hunger)

23.4 Mode-Line Format

Each Emacs window (aside from minibuffer windows) typically has a mode line at the bottom, which displays status information about the buffer displayed in the window. The mode line contains information about the buffer, such as its name, associated file, depth of recursive editing, and major and minor modes. A window can also have a header line, which is much like the mode line but appears at the top of the window.

This section describes how to control the contents of the mode line and header line. We include it in this chapter because much of the information displayed in the mode line relates to the enabled major and minor modes.

• Base: Basic ideas of mode line control.
• Data: The data structure that controls the mode line.
• Top: The top level variable, mode-line-format.
• Mode Line Variables: Variables used in that data structure.
• %-Constructs: Putting information into a mode line.
• Properties in Mode: Using text properties in the mode line.
• Header Lines: Like a mode line, but at the top.
• Emulating Mode Line: Formatting text as the mode line would.

23.4.1 Mode Line Basics

mode-line-format is a buffer-local variable that holds a mode line construct, a kind of template, which controls what is displayed on the mode line of the current buffer. The value of header-line-format specifies the buffer's header line in the same way. All windows for the same buffer use the same mode-line-format and header-line-format.

For efficiency, Emacs does not continuously recompute the mode line and header line of a window. It does so when circumstances appear to call for it—for instance, if you change the window configuration, switch buffers, narrow or widen the buffer, scroll, or change the buffer's modification status. If you modify any of the variables referenced by mode-line-format (see Mode Line Variables), or any other variables and data structures that affect how text is displayed (see Display), you may want to force an update of the mode line so as to display the new information or display it in the new way.

The selected window's mode line is usually displayed in a different color using the face mode-line. Other windows' mode lines appear in the face mode-line-inactive instead. See Faces.

23.4.2 The Data Structure of the Mode Line

The mode-line contents are controlled by a data structure called a mode-line construct, made up of lists, strings, symbols, and numbers kept in buffer-local variables. Each data type has a specific meaning for the mode-line appearance, as described below. The same data structure is used for constructing frame titles (see Frame Titles) and header lines (see Header Lines).

A mode-line construct may be as simple as a fixed string of text, but it usually specifies how to combine fixed strings with variables' values to construct the text. Many of these variables are themselves defined to have mode-line constructs as their values.

Here are the meanings of various data types as mode-line constructs:

string

A string as a mode-line construct appears verbatim except for %-constructs in it. These stand for substitution of other data; see %-Constructs.

If parts of the string have face properties, they control display of the text just as they would text in the buffer. Any characters which have no face properties are displayed, by default, in the face mode-line or mode-line-inactive (see Standard Faces). The help-echo and local-map properties in string have special meanings. See Properties in Mode.

symbol

A symbol as a mode-line construct stands for its value. The value of symbol is used as a mode-line construct, in place of symbol. However, the symbols t and nil are ignored, as is any symbol whose value is void.

There is one exception: if the value of symbol is a string, it is displayed verbatim: the %-constructs are not recognized.

Unless symbol is marked as “risky” (i.e., it has a non-nil risky-local-variable property), all text properties specified in symbol's value are ignored. This includes the text properties of strings in symbol's value, as well as all :eval and :propertize forms in it. (The reason for this is security: non-risky variables could be set automatically from file variables without prompting the user.)

(string rest...)

(list rest...)

A list whose first element is a string or list means to process all the elements recursively and concatenate the results. This is the most common form of mode-line construct.

(:eval form)

A list whose first element is the symbol :eval says to evaluate form, and use the result as a string to display. Make sure this evaluation cannot load any files, as doing so could cause infinite recursion.

(:propertize elt props...)

A list whose first element is the symbol :propertize says to process the mode-line construct elt recursively, then add the text properties specified by props to the result. The argument props should consist of zero or more pairs text-property value. (This feature is new as of Emacs 22.1.)

(symbol then else)

A list whose first element is a symbol that is not a keyword specifies a conditional. Its meaning depends on the value of symbol. If symbol has a non-nil value, the second element, then, is processed recursively as a mode-line element. Otherwise, the third element, else, is processed recursively. You may omit else; then the mode-line element displays nothing if the value of symbol is nil or void.

(width rest...)

A list whose first element is an integer specifies truncation or padding of the results of rest. The remaining elements rest are processed recursively as mode-line constructs and concatenated together. When width is positive, the result is space filled on the right if its width is less than width. When width is negative, the result is truncated on the right to −width columns if its width exceeds −width.

For example, the usual way to show what percentage of a buffer is above the top of the window is to use a list like this: (-3 "%p").

23.4.3 The Top Level of Mode Line Control

The variable in overall control of the mode line is mode-line-format.

The default value of mode-line-format is designed to use the values of other variables such as mode-line-position and mode-line-modes (which in turn incorporates the values of the variables mode-name and minor-mode-alist). Very few modes need to alter mode-line-format itself. For most purposes, it is sufficient to alter some of the variables that mode-line-format either directly or indirectly refers to.

If you do alter mode-line-format itself, the new value should use the same variables that appear in the default value (see Mode Line Variables), rather than duplicating their contents or displaying the information in another fashion. This way, customizations made by the user or by Lisp programs (such as display-time and major modes) via changes to those variables remain effective.

Here is an example of a mode-line-format that might be useful for shell-mode, since it contains the host name and default directory.

     (setq mode-line-format
       (list "-"
        'mode-line-mule-info
        'mode-line-modified
        'mode-line-frame-identification
        "%b--"
        ;; Note that this is evaluated while making the list.
        ;; It makes a mode-line construct which is just a string.
        (getenv "HOST")
        ":"
        'default-directory
        "   "
        'global-mode-string
        "   %[("
        '(:eval (mode-line-mode-name))
        'mode-line-process
        'minor-mode-alist
        "%n"
        ")%]--"
        '(which-func-mode ("" which-func-format "--"))
        '(line-number-mode "L%l--")
        '(column-number-mode "C%c--")
        '(-3 "%p")
        "-%-"))

(The variables line-number-mode, column-number-mode and which-func-mode enable particular minor modes; as usual, these variable names are also the minor mode command names.)

23.4.4 Variables Used in the Mode Line

This section describes variables incorporated by the standard value of mode-line-format into the text of the mode line. There is nothing inherently special about these variables; any other variables could have the same effects on the mode line if mode-line-format's value were changed to use them. However, various parts of Emacs set these variables on the understanding that they will control parts of the mode line; therefore, practically speaking, it is essential for the mode line to use them.

The following three variables are used in mode-line-modes:

Here is a simplified version of the default value of mode-line-format. The real default value also specifies addition of text properties.

     ("-"
      mode-line-mule-info
      mode-line-modified
      mode-line-frame-identification
      mode-line-buffer-identification
      "   "
      mode-line-position
      (vc-mode vc-mode)
      "   "
      mode-line-modes
      (which-func-mode ("" which-func-format "--"))
      (global-mode-string ("--" global-mode-string))
      "-%-")

23.4.5 %-Constructs in the Mode Line

Strings used as mode-line constructs can use certain %-constructs to substitute various kinds of data. Here is a list of the defined %-constructs, and what they mean. In any construct except ‘%%’, you can add a decimal integer after the ‘%’ to specify a minimum field width. If the width is less, the field is padded with spaces to the right.

%b

The current buffer name, obtained with the buffer-name function. See Buffer Names.

%c

The current column number of point.

%e

When Emacs is nearly out of memory for Lisp objects, a brief message saying so. Otherwise, this is empty.

%f

The visited file name, obtained with the buffer-file-name function. See Buffer File Name.

%F

The title (only on a window system) or the name of the selected frame. See Basic Parameters.

%i

The size of the accessible part of the current buffer; basically (- (point-max) (point-min)).

%I

Like ‘%i’, but the size is printed in a more readable way by using ‘k’ for 10^3, ‘M’ for 10^6, ‘G’ for 10^9, etc., to abbreviate.

%l

The current line number of point, counting within the accessible portion of the buffer.

%n

‘Narrow’ when narrowing is in effect; nothing otherwise (see narrow-to-region in Narrowing).

%p

The percentage of the buffer text above the top of window, or ‘Top’, ‘Bottom’ or ‘All’. Note that the default mode-line specification truncates this to three characters.

%P

The percentage of the buffer text that is above the bottom of the window (which includes the text visible in the window, as well as the text above the top), plus ‘Top’ if the top of the buffer is visible on screen; or ‘Bottom’ or ‘All’.

%s

The status of the subprocess belonging to the current buffer, obtained with process-status. See Process Information.

%t

Whether the visited file is a text file or a binary file. This is a meaningful distinction only on certain operating systems (see MS-DOS File Types).

%z

The mnemonics of keyboard, terminal, and buffer coding systems.

%Z

Like ‘%z’, but including the end-of-line format.

%*

‘%’ if the buffer is read only (see buffer-read-only);
‘*’ if the buffer is modified (see buffer-modified-p);
‘-’ otherwise. See Buffer Modification.

%+

‘*’ if the buffer is modified (see buffer-modified-p);
‘%’ if the buffer is read only (see buffer-read-only);
‘-’ otherwise. This differs from ‘%*’ only for a modified read-only buffer. See Buffer Modification.

%&

‘*’ if the buffer is modified, and ‘-’ otherwise.

%[

An indication of the depth of recursive editing levels (not counting minibuffer levels): one ‘[’ for each editing level. See Recursive Editing.

%]

One ‘]’ for each recursive editing level (not counting minibuffer levels).

%-

Dashes sufficient to fill the remainder of the mode line.

%%

The character ‘%’—this is how to include a literal ‘%’ in a string in which %-constructs are allowed.

The following two %-constructs are still supported, but they are obsolete, since you can get the same results with the variables mode-name and global-mode-string.

%m

The value of mode-name.

%M

The value of global-mode-string.

23.4.6 Properties in the Mode Line

Certain text properties are meaningful in the mode line. The face property affects the appearance of text; the help-echo property associates help strings with the text, and local-map can make the text mouse-sensitive.

There are four ways to specify text properties for text in the mode line:

1. Put a string with a text property directly into the mode-line data structure.
2. Put a text property on a mode-line %-construct such as ‘%12b’; then the expansion of the %-construct will have that same text property.
3. Use a (:propertize elt props...) construct to give elt a text property specified by props.
4. Use a list containing :eval form in the mode-line data structure, and make form evaluate to a string that has a text property.

You can use the local-map property to specify a keymap. This keymap only takes real effect for mouse clicks; binding character keys and function keys to it has no effect, since it is impossible to move point into the mode line.

When the mode line refers to a variable which does not have a non-nil risky-local-variable property, any text properties given or specified within that variable's values are ignored. This is because such properties could otherwise specify functions to be called, and those functions could come from file local variables.

23.4.7 Window Header Lines

A window can have a header line at the top, just as it can have a mode line at the bottom. The header line feature works just like the mode-line feature, except that it's controlled by different variables.

A window that is just one line tall never displays a header line. A window that is two lines tall cannot display both a mode line and a header line at once; if it has a mode line, then it does not display a header line.

23.4.8 Emulating Mode-Line Formatting

You can use the function format-mode-line to compute the text that would appear in a mode line or header line based on a certain mode-line specification.

23.5 Imenu

Imenu is a feature that lets users select a definition or section in the buffer, from a menu which lists all of them, to go directly to that location in the buffer. Imenu works by constructing a buffer index which lists the names and buffer positions of the definitions, or other named portions of the buffer; then the user can choose one of them and move point to it. Major modes can add a menu bar item to use Imenu using imenu-add-to-menubar.

The user-level commands for using Imenu are described in the Emacs Manual (see Imenu). This section explains how to customize Imenu's method of finding definitions or buffer portions for a particular major mode.

The usual and simplest way is to set the variable imenu-generic-expression:

Another way to customize Imenu for a major mode is to set the variables imenu-prev-index-position-function and imenu-extract-index-name-function:

The last way to customize Imenu for a major mode is to set the variable imenu-create-index-function:

23.6 Font Lock Mode

Font Lock mode is a feature that automatically attaches face properties to certain parts of the buffer based on their syntactic role. How it parses the buffer depends on the major mode; most major modes define syntactic criteria for which faces to use in which contexts. This section explains how to customize Font Lock for a particular major mode.

Font Lock mode finds text to highlight in two ways: through syntactic parsing based on the syntax table, and through searching (usually for regular expressions). Syntactic fontification happens first; it finds comments and string constants and highlights them. Search-based fontification happens second.

• Font Lock Basics: Overview of customizing Font Lock.
• Search-based Fontification: Fontification based on regexps.
• Customizing Keywords: Customizing search-based fontification.
• Other Font Lock Variables: Additional customization facilities.
• Levels of Font Lock: Each mode can define alternative levels so that the user can select more or less.
• Precalculated Fontification: How Lisp programs that produce the buffer contents can also specify how to fontify it.
• Faces for Font Lock: Special faces specifically for Font Lock.
• Syntactic Font Lock: Fontification based on syntax tables.
• Setting Syntax Properties: Defining character syntax based on context using the Font Lock mechanism.
• Multiline Font Lock: How to coerce Font Lock into properly highlighting multiline constructs.

23.6.1 Font Lock Basics

There are several variables that control how Font Lock mode highlights text. But major modes should not set any of these variables directly. Instead, they should set font-lock-defaults as a buffer-local variable. The value assigned to this variable is used, if and when Font Lock mode is enabled, to set all the other variables.

If your mode fontifies text explicitly by adding font-lock-face properties, it can specify (nil t) for font-lock-defaults to turn off all automatic fontification. However, this is not required; it is possible to fontify some things using font-lock-face properties and set up automatic fontification for other parts of the text.

23.6.2 Search-based Fontification

The most important variable for customizing Font Lock mode is font-lock-keywords. It specifies the search criteria for search-based fontification. You should specify the value of this variable with keywords in font-lock-defaults.

Each element of font-lock-keywords specifies how to find certain cases of text, and how to highlight those cases. Font Lock mode processes the elements of font-lock-keywords one by one, and for each element, it finds and handles all matches. Ordinarily, once part of the text has been fontified already, this cannot be overridden by a subsequent match in the same text; but you can specify different behavior using the override element of a subexp-highlighter.

Each element of font-lock-keywords should have one of these forms:

regexp

Highlight all matches for regexp using font-lock-keyword-face. For example,

          ;; Highlight occurrences of the word ‘foo’
          ;; using font-lock-keyword-face.
          "\\<foo\\>"

The function regexp-opt (see Regexp Functions) is useful for calculating optimal regular expressions to match a number of different keywords.

function

Find text by calling function, and highlight the matches it finds using font-lock-keyword-face.

When function is called, it receives one argument, the limit of the search; it should begin searching at point, and not search beyond the limit. It should return non-nil if it succeeds, and set the match data to describe the match that was found. Returning nil indicates failure of the search.

Fontification will call function repeatedly with the same limit, and with point where the previous invocation left it, until function fails. On failure, function need not reset point in any particular way.

(matcher . subexp)

In this kind of element, matcher is either a regular expression or a function, as described above. The cdr, subexp, specifies which subexpression of matcher should be highlighted (instead of the entire text that matcher matched).

          ;; Highlight the ‘bar’ in each occurrence of ‘fubar’,
          ;; using font-lock-keyword-face.
          ("fu\\(bar\\)" . 1)

If you use regexp-opt to produce the regular expression matcher, you can use regexp-opt-depth (see Regexp Functions) to calculate the value for subexp.

(matcher . facespec)

In this kind of element, facespec is an expression whose value specifies the face to use for highlighting. In the simplest case, facespec is a Lisp variable (a symbol) whose value is a face name.

          ;; Highlight occurrences of ‘fubar’,
          ;; using the face which is the value of fubar-face.
          ("fubar" . fubar-face)

However, facespec can also evaluate to a list of this form:

          (face face prop1 val1 prop2 val2...)

to specify the face face and various additional text properties to put on the text that matches. If you do this, be sure to add the other text property names that you set in this way to the value of font-lock-extra-managed-props so that the properties will also be cleared out when they are no longer appropriate. Alternatively, you can set the variable font-lock-unfontify-region-function to a function that clears these properties. See Other Font Lock Variables.

(matcher . subexp-highlighter)

In this kind of element, subexp-highlighter is a list which specifies how to highlight matches found by matcher. It has the form:

          (subexp facespec [override [laxmatch]])

The car, subexp, is an integer specifying which subexpression of the match to fontify (0 means the entire matching text). The second subelement, facespec, is an expression whose value specifies the face, as described above.

The last two values in subexp-highlighter, override and laxmatch, are optional flags. If override is t, this element can override existing fontification made by previous elements of font-lock-keywords. If it is keep, then each character is fontified if it has not been fontified already by some other element. If it is prepend, the face specified by facespec is added to the beginning of the font-lock-face property. If it is append, the face is added to the end of the font-lock-face property.

If laxmatch is non-nil, it means there should be no error if there is no subexpression numbered subexp in matcher. Obviously, fontification of the subexpression numbered subexp will not occur. However, fontification of other subexpressions (and other regexps) will continue. If laxmatch is nil, and the specified subexpression is missing, then an error is signaled which terminates search-based fontification.

Here are some examples of elements of this kind, and what they do:

          ;; Highlight occurrences of either ‘foo’ or ‘bar’, using
          ;; foo-bar-face, even if they have already been highlighted.
          ;; foo-bar-face should be a variable whose value is a face.
          ("foo\\|bar" 0 foo-bar-face t)
          
          ;; Highlight the first subexpression within each occurrence
          ;; that the function fubar-match finds,
          ;; using the face which is the value of fubar-face.
          (fubar-match 1 fubar-face)

(matcher . anchored-highlighter)

In this kind of element, anchored-highlighter specifies how to highlight text that follows a match found by matcher. So a match found by matcher acts as the anchor for further searches specified by anchored-highlighter. anchored-highlighter is a list of the following form:

          (anchored-matcher pre-form post-form
                                  subexp-highlighters...)

Here, anchored-matcher, like matcher, is either a regular expression or a function. After a match of matcher is found, point is at the end of the match. Now, Font Lock evaluates the form pre-form. Then it searches for matches of anchored-matcher and uses subexp-highlighters to highlight these. A subexp-highlighter is as described above. Finally, Font Lock evaluates post-form.

The forms pre-form and post-form can be used to initialize before, and cleanup after, anchored-matcher is used. Typically, pre-form is used to move point to some position relative to the match of matcher, before starting with anchored-matcher. post-form might be used to move back, before resuming with matcher.

After Font Lock evaluates pre-form, it does not search for anchored-matcher beyond the end of the line. However, if pre-form returns a buffer position that is greater than the position of point after pre-form is evaluated, then the position returned by pre-form is used as the limit of the search instead. It is generally a bad idea to return a position greater than the end of the line; in other words, the anchored-matcher search should not span lines.

For example,

          ;; Highlight occurrences of the word ‘item’ following
          ;; an occurrence of the word ‘anchor’ (on the same line)
          ;; in the value of item-face.
          ("\\<anchor\\>" "\\<item\\>" nil nil (0 item-face))

Here, pre-form and post-form are nil. Therefore searching for ‘item’ starts at the end of the match of ‘anchor’, and searching for subsequent instances of ‘anchor’ resumes from where searching for ‘item’ concluded.

(matcher highlighters...)

This sort of element specifies several highlighter lists for a single matcher. A highlighter list can be of the type subexp-highlighter or anchored-highlighter as described above.

For example,

          ;; Highlight occurrences of the word ‘anchor’ in the value
          ;; of anchor-face, and subsequent occurrences of the word
          ;; ‘item’ (on the same line) in the value of item-face.
          ("\\<anchor\\>" (0 anchor-face)
                          ("\\<item\\>" nil nil (0 item-face)))

(eval . form)

Here form is an expression to be evaluated the first time this value of font-lock-keywords is used in a buffer. Its value should have one of the forms described in this table.

Warning: Do not design an element of font-lock-keywords to match text which spans lines; this does not work reliably. For details, see See Multiline Font Lock.

You can use case-fold in font-lock-defaults to specify the value of font-lock-keywords-case-fold-search which says whether search-based fontification should be case-insensitive.

23.6.3 Customizing Search-Based Fontification

You can use font-lock-add-keywords to add additional search-based fontification rules to a major mode, and font-lock-remove-keywords to remove rules.

For example, this code

     (font-lock-add-keywords 'c-mode
      '(("\\<\\(FIXME\\):" 1 font-lock-warning-face prepend)
        ("\\<\\(and\\|or\\|not\\)\\>" . font-lock-keyword-face)))

adds two fontification patterns for C mode: one to fontify the word ‘FIXME’, even in comments, and another to fontify the words ‘and’, ‘or’ and ‘not’ as keywords.

That example affects only C mode proper. To add the same patterns to C mode and all modes derived from it, do this instead:

     (add-hook 'c-mode-hook
      (lambda ()
       (font-lock-add-keywords nil
        '(("\\<\\(FIXME\\):" 1 font-lock-warning-face prepend)
          ("\\<\\(and\\|or\\|not\\)\\>" .
           font-lock-keyword-face)))))

23.6.4 Other Font Lock Variables

This section describes additional variables that a major mode can set by means of other-vars in font-lock-defaults (see Font Lock Basics).

23.6.5 Levels of Font Lock

Many major modes offer three different levels of fontification. You can define multiple levels by using a list of symbols for keywords in font-lock-defaults. Each symbol specifies one level of fontification; it is up to the user to choose one of these levels, normally by setting font-lock-maximum-decoration (see Font Lock). The chosen level's symbol value is used to initialize font-lock-keywords.

Here are the conventions for how to define the levels of fontification:

• Level 1: highlight function declarations, file directives (such as include or import directives), strings and comments. The idea is speed, so only the most important and top-level components are fontified.
• Level 2: in addition to level 1, highlight all language keywords, including type names that act like keywords, as well as named constant values. The idea is that all keywords (either syntactic or semantic) should be fontified appropriately.
• Level 3: in addition to level 2, highlight the symbols being defined in function and variable declarations, and all builtin function names, wherever they appear.

23.6.6 Precalculated Fontification

Some major modes such as list-buffers and occur construct the buffer text programmatically. The easiest way for them to support Font Lock mode is to specify the faces of text when they insert the text in the buffer.

The way to do this is to specify the faces in the text with the special text property font-lock-face (see Special Properties). When Font Lock mode is enabled, this property controls the display, just like the face property. When Font Lock mode is disabled, font-lock-face has no effect on the display.

It is ok for a mode to use font-lock-face for some text and also use the normal Font Lock machinery. But if the mode does not use the normal Font Lock machinery, it should not set the variable font-lock-defaults.

23.6.7 Faces for Font Lock

You can make Font Lock mode use any face, but several faces are defined specifically for Font Lock mode. Each of these symbols is both a face name, and a variable whose default value is the symbol itself. Thus, the default value of font-lock-comment-face is font-lock-comment-face. This means you can write font-lock-comment-face in a context such as font-lock-keywords where a face-name-valued expression is used.

font-lock-comment-face

Used (typically) for comments.

font-lock-comment-delimiter-face

Used (typically) for comments delimiters.

font-lock-doc-face

Used (typically) for documentation strings in the code.

font-lock-string-face

Used (typically) for string constants.

font-lock-keyword-face

Used (typically) for keywords—names that have special syntactic significance, like for and if in C.

font-lock-builtin-face

Used (typically) for built-in function names.

font-lock-function-name-face

Used (typically) for the name of a function being defined or declared, in a function definition or declaration.

font-lock-variable-name-face

Used (typically) for the name of a variable being defined or declared, in a variable definition or declaration.

font-lock-type-face

Used (typically) for names of user-defined data types, where they are defined and where they are used.

font-lock-constant-face

Used (typically) for constant names.

font-lock-preprocessor-face

Used (typically) for preprocessor commands.

font-lock-negation-char-face

Used (typically) for easily-overlooked negation characters.

font-lock-warning-face

Used (typically) for constructs that are peculiar, or that greatly change the meaning of other text. For example, this is used for ‘;;;###autoload’ cookies in Emacs Lisp, and for #error directives in C.

23.6.8 Syntactic Font Lock

Syntactic fontification uses the syntax table to find comments and string constants (see Syntax Tables). It highlights them using font-lock-comment-face and font-lock-string-face (see Faces for Font Lock), or whatever font-lock-syntactic-face-function chooses. There are several variables that affect syntactic fontification; you should set them by means of font-lock-defaults (see Font Lock Basics).

23.6.9 Setting Syntax Properties

Font Lock mode can be used to update syntax-table properties automatically (see Syntax Properties). This is useful in languages for which a single syntax table by itself is not sufficient.

23.6.10 Multiline Font Lock Constructs

Normally, elements of font-lock-keywords should not match across multiple lines; that doesn't work reliably, because Font Lock usually scans just part of the buffer, and it can miss a multi-line construct that crosses the line boundary where the scan starts. (The scan normally starts at the beginning of a line.)

Making elements that match multiline constructs work properly has two aspects: correct identification and correct rehighlighting. The first means that Font Lock finds all multiline constructs. The second means that Font Lock will correctly rehighlight all the relevant text when a multiline construct is changed—for example, if some of the text that was previously part of a multiline construct ceases to be part of it. The two aspects are closely related, and often getting one of them to work will appear to make the other also work. However, for reliable results you must attend explicitly to both aspects.

There are three ways to ensure correct identification of multiline constructs:

• Add a function to font-lock-extend-region-functions that does the identification and extends the scan so that the scanned text never starts or ends in the middle of a multiline construct.
• Use the font-lock-fontify-region-function hook similarly to extend the scan so that the scanned text never starts or ends in the middle of a multiline construct.
• Somehow identify the multiline construct right when it gets inserted into the buffer (or at any point after that but before font-lock tries to highlight it), and mark it with a font-lock-multiline which will instruct font-lock not to start or end the scan in the middle of the construct.

There are three ways to do rehighlighting of multiline constructs:

• Place a font-lock-multiline property on the construct. This will rehighlight the whole construct if any part of it is changed. In some cases you can do this automatically by setting the font-lock-multiline variable, which see.
• Make sure jit-lock-contextually is set and rely on it doing its job. This will only rehighlight the part of the construct that follows the actual change, and will do it after a short delay. This only works if the highlighting of the various parts of your multiline construct never depends on text in subsequent lines. Since jit-lock-contextually is activated by default, this can be an attractive solution.
• Place a jit-lock-defer-multiline property on the construct. This works only if jit-lock-contextually is used, and with the same delay before rehighlighting, but like font-lock-multiline, it also handles the case where highlighting depends on subsequent lines.

• Font Lock Multiline: Marking multiline chunks with a text property.
• Region to Fontify: Controlling which region gets refontified after a buffer change.

23.6.10.1 Font Lock Multiline

One way to ensure reliable rehighlighting of multiline Font Lock constructs is to put on them the text property font-lock-multiline. It should be present and non-nil for text that is part of a multiline construct.

When Font Lock is about to highlight a range of text, it first extends the boundaries of the range as necessary so that they do not fall within text marked with the font-lock-multiline property. Then it removes any font-lock-multiline properties from the range, and highlights it. The highlighting specification (mostly font-lock-keywords) must reinstall this property each time, whenever it is appropriate.

Warning: don't use the font-lock-multiline property on large ranges of text, because that will make rehighlighting slow.

The font-lock-multiline property is meant to ensure proper refontification; it does not automatically identify new multiline constructs. Identifying the requires that Font-Lock operate on large enough chunks at a time. This will happen by accident on many cases, which may give the impression that multiline constructs magically work. If you set the font-lock-multiline variable non-nil, this impression will be even stronger, since the highlighting of those constructs which are found will be properly updated from then on. But that does not work reliably.

To find multiline constructs reliably, you must either manually place the font-lock-multiline property on the text before Font-Lock looks at it, or use font-lock-fontify-region-function.

23.6.10.2 Region to Fontify after a Buffer Change

When a buffer is changed, the region that Font Lock refontifies is by default the smallest sequence of whole lines that spans the change. While this works well most of the time, sometimes it doesn't—for example, when a change alters the syntactic meaning of text on an earlier line.

You can enlarge (or even reduce) the region to fontify by setting one the following variables:

23.7 Auto-indention of code

For programming languages, an important feature of a major mode is to provide automatic indentation. This is controlled in Emacs by indent-line-function (see Mode-Specific Indent). Writing a good indentation function can be difficult and to a large extent it is still a black art.

Many major mode authors will start by writing a simple indentation function that works for simple cases, for example by comparing with the indentation of the previous text line. For most programming languages that are not really line-based, this tends to scale very poorly: improving such a function to let it handle more diverse situations tends to become more and more difficult, resulting in the end with a large, complex, unmaintainable indentation function which nobody dares to touch.

A good indentation function will usually need to actually parse the text, according to the syntax of the language. Luckily, it is not necessary to parse the text in as much detail as would be needed for a compiler, but on the other hand, the parser embedded in the indentation code will want to be somewhat friendly to syntactically incorrect code.

Good maintainable indentation functions usually fall into 2 categories: either parsing forward from some “safe” starting point until the position of interest, or parsing backward from the position of interest. Neither of the two is a clearly better choice than the other: parsing backward is often more difficult than parsing forward because programming languages are designed to be parsed forward, but for the purpose of indentation it has the advantage of not needing to guess a “safe” starting point, and it generally enjoys the property that only a minimum of text will be analyzed to decide the indentation of a line, so indentation will tend to be unaffected by syntax errors in some earlier unrelated piece of code. Parsing forward on the other hand is usually easier and has the advantage of making it possible to reindent efficiently a whole region at a time, with a single parse.

Rather than write your own indentation function from scratch, it is often preferable to try and reuse some existing ones or to rely on a generic indentation engine. There are sadly few such engines. The CC-mode indentation code (used with C, C++, Java, Awk and a few other such modes) has been made more generic over the years, so if your language seems somewhat similar to one of those languages, you might try to use that engine. Another one is SMIE which takes an approach in the spirit of Lisp sexps and adapts it to non-Lisp languages.

• SMIE: A simple minded indentation engine

23.7.1 Simple Minded Indentation Engine

SMIE is a package that provides a generic navigation and indentation engine. Based on a very simple parser using an “operator precedence grammar”, it lets major modes extend the sexp-based navigation of Lisp to non-Lisp languages as well as provide a simple to use but reliable auto-indentation.

Operator precedence grammar is a very primitive technology for parsing compared to some of the more common techniques used in compilers. It has the following characteristics: its parsing power is very limited, and it is largely unable to detect syntax errors, but it has the advantage of being algorithmically efficient and able to parse forward just as well as backward. In practice that means that SMIE can use it for indentation based on backward parsing, that it can provide both forward-sexp and backward-sexp functionality, and that it will naturally work on syntactically incorrect code without any extra effort. The downside is that it also means that most programming languages cannot be parsed correctly using SMIE, at least not without resorting to some special tricks (see SMIE Tricks).

• SMIE setup: SMIE setup and features
• Operator Precedence Grammars: A very simple parsing technique
• SMIE Grammar: Defining the grammar of a language
• SMIE Lexer: Defining tokens
• SMIE Tricks: Working around the parser's limitations
• SMIE Indentation: Specifying indentation rules
• SMIE Indentation Helpers: Helper functions for indentation rules
• SMIE Indentation Example: Sample indentation rules

23.7.1.1 SMIE Setup and Features

SMIE is meant to be a one-stop shop for structural navigation and various other features which rely on the syntactic structure of code, in particular automatic indentation. The main entry point is smie-setup which is a function typically called while setting up a major mode.

Calling this function is sufficient to make commands such as forward-sexp, backward-sexp, and transpose-sexps be able to properly handle structural elements other than just the paired parentheses already handled by syntax tables. For example, if the provided grammar is precise enough, transpose-sexps can correctly transpose the two arguments of a + operator, taking into account the precedence rules of the language.

Calling `smie-setup' is also sufficient to make TAB indentation work in the expected way, and provides some commands that you can bind in the major mode keymap.

23.7.1.2 Operator Precedence Grammars

SMIE's precedence grammars simply give to each token a pair of precedences: the left-precedence and the right-precedence. We say T1 < T2 if the right-precedence of token T1 is less than the left-precedence of token T2. A good way to read this < is as a kind of parenthesis: if we find ... T1 something T2 ... then that should be parsed as ... T1 (something T2 ... rather than as ... T1 something) T2 .... The latter interpretation would be the case if we had T1 > T2. If we have T1 = T2, it means that token T2 follows token T1 in the same syntactic construction, so typically we have "begin" = "end". Such pairs of precedences are sufficient to express left-associativity or right-associativity of infix operators, nesting of tokens like parentheses and many other cases.

23.7.1.3 Defining the Grammar of a Language

The usual way to define the SMIE grammar of a language is by defining a new global variable that holds the precedence table by giving a set of BNF rules. For example, the grammar definition for a small Pascal-like language could look like:

     (require 'smie)
     (defvar sample-smie-grammar
       (smie-prec2->grammar
        (smie-bnf->prec2
         '((id)
           (inst ("begin" insts "end")
                 ("if" exp "then" inst "else" inst)
                 (id ":=" exp)
                 (exp))
           (insts (insts ";" insts) (inst))
           (exp (exp "+" exp)
                (exp "*" exp)
                ("(" exps ")"))
           (exps (exps "," exps) (exp)))
         '((assoc ";"))
         '((assoc ","))
         '((assoc "+") (assoc "*")))))

A few things to note:

• The above grammar does not explicitly mention the syntax of function calls: SMIE will automatically allow any sequence of sexps, such as identifiers, balanced parentheses, or begin ... end blocks to appear anywhere anyway.
• The grammar category id has no right hand side: this does not mean that it can match only the empty string, since as mentioned any sequence of sexps can appear anywhere anyway.
• Because non terminals cannot appear consecutively in the BNF grammar, it is difficult to correctly handle tokens that act as terminators, so the above grammar treats ";" as a statement separator instead, which SMIE can handle very well.
• Separators used in sequences (such as "," and ";" above) are best defined with BNF rules such as (foo (foo "separator" foo) ...) which generate precedence conflicts which are then resolved by giving them an explicit (assoc "separator").
• The ("(" exps ")") rule was not needed to pair up parens, since SMIE will pair up any characters that are marked as having paren syntax in the syntax table. What this rule does instead (together with the definition of exps) is to make it clear that "," should not appear outside of parentheses.
• Rather than have a single precs table to resolve conflicts, it is preferable to have several tables, so as to let the BNF part of the grammar specify relative precedences where possible.
• Unless there is a very good reason to prefer left or right, it is usually preferable to mark operators as associative, using assoc. For that reason "+" and "*" are defined above as assoc, although the language defines them formally as left associative.

23.7.1.4 Defining Tokens

SMIE comes with a predefined lexical analyzer which uses syntax tables in the following way: any sequence of characters that have word or symbol syntax is considered a token, and so is any sequence of characters that have punctuation syntax. This default lexer is often a good starting point but is rarely actually correct for any given language. For example, it will consider "2,+3" to be composed of 3 tokens: "2", ",+", and "3".

To describe the lexing rules of your language to SMIE, you need 2 functions, one to fetch the next token, and another to fetch the previous token. Those functions will usually first skip whitespace and comments and then look at the next chunk of text to see if it is a special token. If so it should skip the token and return a description of this token. Usually this is simply the string extracted from the buffer, but it can be anything you want. For example:

     (defvar sample-keywords-regexp
       (regexp-opt '("+" "*" "," ";" ">" ">=" "<" "<=" ":=" "=")))
     (defun sample-smie-forward-token ()
       (forward-comment (point-max))
       (cond
        ((looking-at sample-keywords-regexp)
         (goto-char (match-end 0))
         (match-string-no-properties 0))
        (t (buffer-substring-no-properties
            (point)
            (progn (skip-syntax-forward "w_")
                   (point))))))
     (defun sample-smie-backward-token ()
       (forward-comment (- (point)))
       (cond
        ((looking-back sample-keywords-regexp (- (point) 2) t)
         (goto-char (match-beginning 0))
         (match-string-no-properties 0))
        (t (buffer-substring-no-properties
            (point)
            (progn (skip-syntax-backward "w_")
                   (point))))))

Notice how those lexers return the empty string when in front of parentheses. This is because SMIE automatically takes care of the parentheses defined in the syntax table. More specifically if the lexer returns nil or an empty string, SMIE tries to handle the corresponding text as a sexp according to syntax tables.

23.7.1.5 Living With a Weak Parser

The parsing technique used by SMIE does not allow tokens to behave differently in different contexts. For most programming languages, this manifests itself by precedence conflicts when converting the BNF grammar.

Sometimes, those conflicts can be worked around by expressing the grammar slightly differently. For example, for Modula-2 it might seem natural to have a BNF grammar that looks like this:

       ...
       (inst ("IF" exp "THEN" insts "ELSE" insts "END")
             ("CASE" exp "OF" cases "END")
             ...)
       (cases (cases "|" cases) (caselabel ":" insts) ("ELSE" insts))
       ...

But this will create conflicts for "ELSE": on the one hand, the IF rule implies (among many other things) that "ELSE" = "END"; but on the other hand, since "ELSE" appears within cases, which appears left of "END", we also have "ELSE" > "END". We can solve the conflict either by using:

       ...
       (inst ("IF" exp "THEN" insts "ELSE" insts "END")
             ("CASE" exp "OF" cases "END")
             ("CASE" exp "OF" cases "ELSE" insts "END")
             ...)
       (cases (cases "|" cases) (caselabel ":" insts))
       ...

or

       ...
       (inst ("IF" exp "THEN" else "END")
             ("CASE" exp "OF" cases "END")
             ...)
       (else (insts "ELSE" insts))
       (cases (cases "|" cases) (caselabel ":" insts) (else))
       ...

Reworking the grammar to try and solve conflicts has its downsides, tho, because SMIE assumes that the grammar reflects the logical structure of the code, so it is preferable to keep the BNF closer to the intended abstract syntax tree.

Other times, after careful consideration you may conclude that those conflicts are not serious and simply resolve them via the resolvers argument of smie-bnf->prec2. Usually this is because the grammar is simply ambiguous: the conflict does not affect the set of programs described by the grammar, but only the way those programs are parsed. This is typically the case for separators and associative infix operators, where you want to add a resolver like '((assoc "|")). Another case where this can happen is for the classic dangling else problem, where you will use '((assoc "else" "then")). It can also happen for cases where the conflict is real and cannot really be resolved, but it is unlikely to pose a problem in practice.

Finally, in many cases some conflicts will remain despite all efforts to restructure the grammar. Do not despair: while the parser cannot be made more clever, you can make the lexer as smart as you want. So, the solution is then to look at the tokens involved in the conflict and to split one of those tokens into 2 (or more) different tokens. E.g. if the grammar needs to distinguish between two incompatible uses of the token "begin", make the lexer return different tokens (say "begin-fun" and "begin-plain") depending on which kind of "begin" it finds. This pushes the work of distinguishing the different cases to the lexer, which will thus have to look at the surrounding text to find ad-hoc clues.

23.7.1.6 Specifying Indentation Rules

Based on the provided grammar, SMIE will be able to provide automatic indentation without any extra effort. But in practice, this default indentation style will probably not be good enough. You will want to tweak it in many different cases.

SMIE indentation is based on the idea that indentation rules should be as local as possible. To this end, it relies on the idea of virtual indentation, which is the indentation that a particular program point would have if it were at the beginning of a line. Of course, if that program point is indeed at the beginning of a line, its virtual indentation is its current indentation. But if not, then SMIE uses the indentation algorithm to compute the virtual indentation of that point. Now in practice, the virtual indentation of a program point does not have to be identical to the indentation it would have if we inserted a newline before it. To see how this works, the SMIE rule for indentation after a { in C does not care whether the { is standing on a line of its own or is at the end of the preceding line. Instead, these different cases are handled in the indentation rule that decides how to indent before a {.

Another important concept is the notion of parent: The parent of a token, is the head token of the nearest enclosing syntactic construct. For example, the parent of an else is the if to which it belongs, and the parent of an if, in turn, is the lead token of the surrounding construct. The command backward-sexp jumps from a token to its parent, but there are some caveats: for openers (tokens which start a construct, like if), you need to start with point before the token, while for others you need to start with point after the token. backward-sexp stops with point before the parent token if that is the opener of the token of interest, and otherwise it stops with point after the parent token.

SMIE indentation rules are specified using a function that takes two arguments method and arg where the meaning of arg and the expected return value depend on method.

method can be:

• :after, in which case arg is a token and the function should return the offset to use for indentation after arg.
• :before, in which case arg is a token and the function should return the offset to use to indent arg itself.
• :elem, in which case the function should return either the offset to use to indent function arguments (if arg is the symbol arg) or the basic indentation step (if arg is the symbol basic).
• :list-intro, in which case arg is a token and the function should return non-nil if the token is followed by a list of expressions (not separated by any token) rather than an expression.

When arg is a token, the function is called with point just before that token. A return value of nil always means to fallback on the default behavior, so the function should return nil for arguments it does not expect.

offset can be:

• nil: use the default indentation rule.
• (column . column): indent to column column.
• number: offset by number, relative to a base token which is the current token for :after and its parent for :before.

23.7.1.7 Helper Functions for Indentation Rules

SMIE provides various functions designed specifically for use in the indentation rules function (several of those functions break if used in another context). These functions all start with the prefix smie-rule-.

23.7.1.8 Sample Indentation Rules

Here is an example of an indentation function:

     (eval-when-compile (require 'cl))       ;For the `case' macro.
     (defun sample-smie-rules (kind token)
       (case kind
         (:elem (case token
                  (basic sample-indent-basic)))
         (:after
          (cond
           ((equal token ",") (smie-rule-separator kind))
           ((equal token ":=") sample-indent-basic)))
         (:before
          (cond
           ((equal token ",") (smie-rule-separator kind))
           ((member token '("begin" "(" "{"))
            (if (smie-rule-hanging-p) (smie-rule-parent)))
           ((equal token "if")
            (and (not (smie-rule-bolp)) (smie-rule-prev-p "else")
                 (smie-rule-parent)))))))

A few things to note:

• The first case indicates the basic indentation increment to use. If sample-indent-basic is nil, then SMIE uses the global setting smie-indent-basic. The major mode could have set smie-indent-basic buffer-locally instead, but that is discouraged.
• The two (identical) rules for the token "," make SMIE try to be more clever when the comma separator is placed at the beginning of lines. It tries to outdent the separator so as to align the code after the comma; for example:

            x = longfunctionname (
                    arg1
                  , arg2
                );
  

• The rule for indentation after ":=" exists because otherwise SMIE would treat ":=" as an infix operator and would align the right argument with the left one.
• The rule for indentation before "begin" is an example of the use of virtual indentation: This rule is used only when "begin" is hanging, which can happen only when "begin" is not at the beginning of a line. So this is not used when indenting "begin" itself but only when indenting something relative to this "begin". Concretely, this rule changes the indentation from:

                if x > 0 then begin
                        dosomething(x);
                    end
  

  to

                if x > 0 then begin
                    dosomething(x);
                end
  

• The rule for indentation before "if" is similar to the one for "begin", but where the purpose is to treat "else if" as a single unit, so as to align a sequence of tests rather than indent each test further to the right. This function does this only in the case where the "if" is not placed on a separate line, hence the smie-rule-bolp test.

  If we know that the "else" is always aligned with its "if" and is always at the beginning of a line, we can use a more efficient rule:

            ((equal token "if")
             (and (not (smie-rule-bolp)) (smie-rule-prev-p "else")
                  (save-excursion
                    (sample-smie-backward-token)  ;Jump before the "else".
                    (cons 'column (current-column)))))
  

  The advantage of this formulation is that it reuses the indentation of the previous "else", rather than going all the way back to the first "if" of the sequence.

23.8 Desktop Save Mode

Desktop Save Mode is a feature to save the state of Emacs from one session to another. The user-level commands for using Desktop Save Mode are described in the GNU Emacs Manual (see Saving Emacs Sessions). Modes whose buffers visit a file, don't have to do anything to use this feature.

For buffers not visiting a file to have their state saved, the major mode must bind the buffer local variable desktop-save-buffer to a non-nil value.

For buffers not visiting a file to be restored, the major mode must define a function to do the job, and that function must be listed in the alist desktop-buffer-mode-handlers.

24 Documentation

GNU Emacs Lisp has convenient on-line help facilities, most of which derive their information from the documentation strings associated with functions and variables. This chapter describes how to write good documentation strings for your Lisp programs, as well as how to write programs to access documentation.

Note that the documentation strings for Emacs are not the same thing as the Emacs manual. Manuals have their own source files, written in the Texinfo language; documentation strings are specified in the definitions of the functions and variables they apply to. A collection of documentation strings is not sufficient as a manual because a good manual is not organized in that fashion; it is organized in terms of topics of discussion.

For commands to display documentation strings, see Help. For the conventions for writing documentation strings, see Documentation Tips.

• Documentation Basics: Good style for doc strings. Where to put them. How Emacs stores them.
• Accessing Documentation: How Lisp programs can access doc strings.
• Keys in Documentation: Substituting current key bindings.
• Describing Characters: Making printable descriptions of non-printing characters and key sequences.
• Help Functions: Subroutines used by Emacs help facilities.

24.1 Documentation Basics

A documentation string is written using the Lisp syntax for strings, with double-quote characters surrounding the text of the string. This is because it really is a Lisp string object. The string serves as documentation when it is written in the proper place in the definition of a function or variable. In a function definition, the documentation string follows the argument list. In a variable definition, the documentation string follows the initial value of the variable.

When you write a documentation string, make the first line a complete sentence (or two complete sentences) since some commands, such as apropos, show only the first line of a multi-line documentation string. Also, you should not indent the second line of a documentation string, if it has one, because that looks odd when you use C-h f (describe-function) or C-h v (describe-variable) to view the documentation string. There are many other conventions for doc strings; see Documentation Tips.

Documentation strings can contain several special substrings, which stand for key bindings to be looked up in the current keymaps when the documentation is displayed. This allows documentation strings to refer to the keys for related commands and be accurate even when a user rearranges the key bindings. (See Keys in Documentation.)

Emacs Lisp mode fills documentation strings to the width specified by emacs-lisp-docstring-fill-column.

In Emacs Lisp, a documentation string is accessible through the function or variable that it describes:

• The documentation for a function is usually stored in the function definition itself (see Lambda Expressions). The function documentation knows how to extract it. You can also put function documentation in the function-documentation property of the function name. That is useful with definitions such as keyboard macros that can't hold a documentation string.
• The documentation for a variable is stored in the variable's property list under the property name variable-documentation. The function documentation-property knows how to retrieve it.

To save space, the documentation for preloaded functions and variables (including primitive functions and autoloaded functions) is stored in the file emacs/etc/DOC-version—not inside Emacs. The documentation strings for functions and variables loaded during the Emacs session from byte-compiled files are stored in those files (see Docs and Compilation).

The data structure inside Emacs has an integer offset into the file, or a list containing a file name and an integer, in place of the documentation string. The functions documentation and documentation-property use that information to fetch the documentation string from the appropriate file; this is transparent to the user.

The emacs/lib-src directory contains two utilities that you can use to print nice-looking hardcopy for the file emacs/etc/DOC-version. These are sorted-doc and digest-doc.

24.2 Access to Documentation Strings

Here is an example of using the two functions, documentation and documentation-property, to display the documentation strings for several symbols in a ‘*Help*’ buffer.

     (defun describe-symbols (pattern)
       "Describe the Emacs Lisp symbols matching PATTERN.
     All symbols that have PATTERN in their name are described
     in the `*Help*' buffer."
       (interactive "sDescribe symbols matching: ")
       (let ((describe-func
              (function
               (lambda (s)
                 ;; Print description of symbol.
                 (if (fboundp s)             ; It is a function.
                     (princ
                      (format "%s\t%s\n%s\n\n" s
                        (if (commandp s)
                            (let ((keys (where-is-internal s)))
                              (if keys
                                  (concat
                                   "Keys: "
                                   (mapconcat 'key-description
                                              keys " "))
                                "Keys: none"))
                          "Function")
                        (or (documentation s)
                            "not documented"))))
     
                 (if (boundp s)              ; It is a variable.
                     (princ
                      (format "%s\t%s\n%s\n\n" s
                        (if (user-variable-p s)
                            "Option " "Variable")
                        (or (documentation-property
                              s 'variable-documentation)
                            "not documented")))))))
             sym-list)
     
         ;; Build a list of symbols that match pattern.
         (mapatoms (function
                    (lambda (sym)
                      (if (string-match pattern (symbol-name sym))
                          (setq sym-list (cons sym sym-list))))))
     
         ;; Display the data.
         (help-setup-xref (list 'describe-symbols pattern) (interactive-p))
         (with-help-window (help-buffer)
           (mapcar describe-func (sort sym-list 'string<)))))

The describe-symbols function works like apropos, but provides more information.

     (describe-symbols "goal")
     
     ---------- Buffer: *Help* ----------
     goal-column     Option
     Semipermanent goal column for vertical motion, as set by ...
     
     
     
     set-goal-column Keys: C-x C-n
     Set the current horizontal position as a goal for C-n and C-p.
     
     Those commands will move to this position in the line moved to
     rather than trying to keep the same horizontal position.
     With a non-nil argument, clears out the goal column
     so that C-n and C-p resume vertical motion.
     The goal column is stored in the variable `goal-column'.
     
     temporary-goal-column   Variable
     Current goal column for vertical motion.
     It is the column where point was
     at the start of current run of vertical motion commands.
     When the `track-eol' feature is doing its job, the value is 9999.
     ---------- Buffer: *Help* ----------

24.3 Substituting Key Bindings in Documentation

When documentation strings refer to key sequences, they should use the current, actual key bindings. They can do so using certain special text sequences described below. Accessing documentation strings in the usual way substitutes current key binding information for these special sequences. This works by calling substitute-command-keys. You can also call that function yourself.

Here is a list of the special sequences and what they mean:

\[command]

stands for a key sequence that will invoke command, or ‘M-x command’ if command has no key bindings.

\{mapvar}

stands for a summary of the keymap which is the value of the variable mapvar. The summary is made using describe-bindings.

\<mapvar>

stands for no text itself. It is used only for a side effect: it specifies mapvar's value as the keymap for any following ‘\[command]’ sequences in this documentation string.

\=

quotes the following character and is discarded; thus, ‘\=\[’ puts ‘\[’ into the output, and ‘\=\=’ puts ‘\=’ into the output.

Please note: Each ‘\’ must be doubled when written in a string in Emacs Lisp.

Here are examples of the special sequences:

     (substitute-command-keys
        "To abort recursive edit, type: \\[abort-recursive-edit]")
     ⇒ "To abort recursive edit, type: C-]"
     
     (substitute-command-keys
        "The keys that are defined for the minibuffer here are:
       \\{minibuffer-local-must-match-map}")
     ⇒ "The keys that are defined for the minibuffer here are:
     
     ?               minibuffer-completion-help
     SPC             minibuffer-complete-word
     TAB             minibuffer-complete
     C-j             minibuffer-complete-and-exit
     RET             minibuffer-complete-and-exit
     C-g             abort-recursive-edit
     "
     
     (substitute-command-keys
        "To abort a recursive edit from the minibuffer, type\
     \\<minibuffer-local-must-match-map>\\[abort-recursive-edit].")
     ⇒ "To abort a recursive edit from the minibuffer, type C-g."

There are other special conventions for the text in documentation strings—for instance, you can refer to functions, variables, and sections of this manual. See Documentation Tips, for details.

24.4 Describing Characters for Help Messages

These functions convert events, key sequences, or characters to textual descriptions. These descriptions are useful for including arbitrary text characters or key sequences in messages, because they convert non-printing and whitespace characters to sequences of printing characters. The description of a non-whitespace printing character is the character itself.

24.5 Help Functions

Emacs provides a variety of on-line help functions, all accessible to the user as subcommands of the prefix C-h. For more information about them, see Help. Here we describe some program-level interfaces to the same information.

The following two functions are meant for modes that want to provide help without relinquishing control, such as the “electric” modes. Their names begin with ‘Helper’ to distinguish them from the ordinary help functions.

See describe-symbols example, for an example of using help-buffer, with-help-window, and help-setup-xref.

25 Files

In Emacs, you can find, create, view, save, and otherwise work with files and file directories. This chapter describes most of the file-related functions of Emacs Lisp, but a few others are described in Buffers, and those related to backups and auto-saving are described in Backups and Auto-Saving.

Many of the file functions take one or more arguments that are file names. A file name is actually a string. Most of these functions expand file name arguments by calling expand-file-name, so that ~ is handled correctly, as are relative file names (including ‘../’). These functions don't recognize environment variable substitutions such as ‘$HOME’. See File Name Expansion.

When file I/O functions signal Lisp errors, they usually use the condition file-error (see Handling Errors). The error message is in most cases obtained from the operating system, according to locale system-message-locale, and decoded using coding system locale-coding-system (see Locales).

• Visiting Files: Reading files into Emacs buffers for editing.
• Saving Buffers: Writing changed buffers back into files.
• Reading from Files: Reading files into buffers without visiting.
• Writing to Files: Writing new files from parts of buffers.
• File Locks: Locking and unlocking files, to prevent simultaneous editing by two people.
• Information about Files: Testing existence, accessibility, size of files.
• Changing Files: Renaming files, changing protection, etc.
• File Names: Decomposing and expanding file names.
• Contents of Directories: Getting a list of the files in a directory.
• Create/Delete Dirs: Creating and Deleting Directories.
• Magic File Names: Defining "magic" special handling for certain file names.
• Format Conversion: Conversion to and from various file formats.

25.1 Visiting Files

Visiting a file means reading a file into a buffer. Once this is done, we say that the buffer is visiting that file, and call the file “the visited file” of the buffer.

A file and a buffer are two different things. A file is information recorded permanently in the computer (unless you delete it). A buffer, on the other hand, is information inside of Emacs that will vanish at the end of the editing session (or when you kill the buffer). Usually, a buffer contains information that you have copied from a file; then we say the buffer is visiting that file. The copy in the buffer is what you modify with editing commands. Such changes to the buffer do not change the file; therefore, to make the changes permanent, you must save the buffer, which means copying the altered buffer contents back into the file.

In spite of the distinction between files and buffers, people often refer to a file when they mean a buffer and vice-versa. Indeed, we say, “I am editing a file,” rather than, “I am editing a buffer that I will soon save as a file of the same name.” Humans do not usually need to make the distinction explicit. When dealing with a computer program, however, it is good to keep the distinction in mind.

• Visiting Functions: The usual interface functions for visiting.
• Subroutines of Visiting: Lower-level subroutines that they use.

25.1.1 Functions for Visiting Files

This section describes the functions normally used to visit files. For historical reasons, these functions have names starting with ‘find-’ rather than ‘visit-’. See Buffer File Name, for functions and variables that access the visited file name of a buffer or that find an existing buffer by its visited file name.

In a Lisp program, if you want to look at the contents of a file but not alter it, the fastest way is to use insert-file-contents in a temporary buffer. Visiting the file is not necessary and takes longer. See Reading from Files.

25.1.2 Subroutines of Visiting

The find-file-noselect function uses two important subroutines which are sometimes useful in user Lisp code: create-file-buffer and after-find-file. This section explains how to use them.

25.2 Saving Buffers

When you edit a file in Emacs, you are actually working on a buffer that is visiting that file—that is, the contents of the file are copied into the buffer and the copy is what you edit. Changes to the buffer do not change the file until you save the buffer, which means copying the contents of the buffer into the file.

Saving a buffer runs several hooks. It also performs format conversion (see Format Conversion).

See also the function set-visited-file-name (see Buffer File Name).

25.3 Reading from Files

You can copy a file from the disk and insert it into a buffer using the insert-file-contents function. Don't use the user-level command insert-file in a Lisp program, as that sets the mark.

If you want to pass a file name to another process so that another program can read the file, use the function file-local-copy; see Magic File Names.

25.4 Writing to Files

You can write the contents of a buffer, or part of a buffer, directly to a file on disk using the append-to-file and write-region functions. Don't use these functions to write to files that are being visited; that could cause confusion in the mechanisms for visiting.

25.5 File Locks

When two users edit the same file at the same time, they are likely to interfere with each other. Emacs tries to prevent this situation from arising by recording a file lock when a file is being modified. (File locks are not implemented on Microsoft systems.) Emacs can then detect the first attempt to modify a buffer visiting a file that is locked by another Emacs job, and ask the user what to do. The file lock is really a file, a symbolic link with a special name, stored in the same directory as the file you are editing.

When you access files using NFS, there may be a small probability that you and another user will both lock the same file “simultaneously.” If this happens, it is possible for the two users to make changes simultaneously, but Emacs will still warn the user who saves second. Also, the detection of modification of a buffer visiting a file changed on disk catches some cases of simultaneous editing; see Modification Time.

File locking is not supported on some systems. On systems that do not support it, the functions lock-buffer, unlock-buffer and file-locked-p do nothing and return nil.

25.6 Information about Files

The functions described in this section all operate on strings that designate file names. With a few exceptions, all the functions have names that begin with the word ‘file’. These functions all return information about actual files or directories, so their arguments must all exist as actual files or directories unless otherwise noted.

• Testing Accessibility: Is a given file readable? Writable?
• Kinds of Files: Is it a directory? A symbolic link?
• Truenames: Eliminating symbolic links from a file name.
• File Attributes: How large is it? Any other names? Etc.
• Locating Files: How to find a file in standard places.

25.6.1 Testing Accessibility

These functions test for permission to access a file in specific ways. Unless explicitly stated otherwise, they recursively follow symbolic links for their file name arguments, at all levels (at the level of the file itself and at all levels of parent directories).

25.6.2 Distinguishing Kinds of Files

This section describes how to distinguish various kinds of files, such as directories, symbolic links, and ordinary files.

The next two functions recursively follow symbolic links at all levels for filename.

25.6.3 Truenames

The truename of a file is the name that you get by following symbolic links at all levels until none remain, then simplifying away ‘.’ and ‘..’ appearing as name components. This results in a sort of canonical name for the file. A file does not always have a unique truename; the number of distinct truenames a file has is equal to the number of hard links to the file. However, truenames are useful because they eliminate symbolic links as a cause of name variation.

To illustrate the difference between file-chase-links and file-truename, suppose that /usr/foo is a symbolic link to the directory /home/foo, and /home/foo/hello is an ordinary file (or at least, not a symbolic link) or nonexistent. Then we would have:

     (file-chase-links "/usr/foo/hello")
          ;; This does not follow the links in the parent directories.
          ⇒ "/usr/foo/hello"
     (file-truename "/usr/foo/hello")
          ;; Assuming that /home is not a symbolic link.
          ⇒ "/home/foo/hello"

See Buffer File Name, for related information.

25.6.4 Other Information about Files

This section describes the functions for getting detailed information about a file, other than its contents. This information includes the mode bits that control access permission, the owner and group numbers, the number of names, the inode number, the size, and the times of access and modification.

If the filename argument to the next two functions is a symbolic link, then these function do not replace it with its target. However, they both recursively follow symbolic links at all levels of parent directories.

On MS-DOS, there is no such thing as an “executable” file mode bit. So Emacs considers a file executable if its name ends in one of the standard executable extensions, such as .com, .bat, .exe, and some others. Files that begin with the Unix-standard ‘#!’ signature, such as shell and Perl scripts, are also considered as executable files. This is reflected in the values returned by file-modes and file-attributes. Directories are also reported with executable bit set, for compatibility with Unix.

25.6.5 How to Locate Files in Standard Places

This section explains how to search for a file in a list of directories (a path). One example is when you need to look for a program's executable file, e.g., to find out whether a given program is installed on the user's system. Another example is the search for Lisp libraries (see Library Search). Such searches generally need to try various possible file name extensions, in addition to various possible directories. Emacs provides a function for such a generalized search for a file.

25.7 Changing File Names and Attributes

The functions in this section rename, copy, delete, link, and set the modes of files.

In the functions that have an argument newname, if a file by the name of newname already exists, the actions taken depend on the value of the argument ok-if-already-exists:

• Signal a file-already-exists error if ok-if-already-exists is nil.
• Request confirmation if ok-if-already-exists is a number.
• Replace the old file without confirmation if ok-if-already-exists is any other value.

The next four commands all recursively follow symbolic links at all levels of parent directories for their first argument, but, if that argument is itself a symbolic link, then only copy-file replaces it with its (recursive) target.

25.8 File Names

Files are generally referred to by their names, in Emacs as elsewhere. File names in Emacs are represented as strings. The functions that operate on a file all expect a file name argument.

In addition to operating on files themselves, Emacs Lisp programs often need to operate on file names; i.e., to take them apart and to use part of a name to construct related file names. This section describes how to manipulate file names.

The functions in this section do not actually access files, so they can operate on file names that do not refer to an existing file or directory.

On MS-DOS and MS-Windows, these functions (like the function that actually operate on files) accept MS-DOS or MS-Windows file-name syntax, where backslashes separate the components, as well as Unix syntax; but they always return Unix syntax. This enables Lisp programs to specify file names in Unix syntax and work properly on all systems without change.

• File Name Components: The directory part of a file name, and the rest.
• Relative File Names: Some file names are relative to a current directory.
• Directory Names: A directory's name as a directory is different from its name as a file.
• File Name Expansion: Converting relative file names to absolute ones.
• Unique File Names: Generating names for temporary files.
• File Name Completion: Finding the completions for a given file name.
• Standard File Names: If your package uses a fixed file name, how to handle various operating systems simply.

25.8.1 File Name Components

The operating system groups files into directories. To specify a file, you must specify the directory and the file's name within that directory. Therefore, Emacs considers a file name as having two main parts: the directory name part, and the nondirectory part (or file name within the directory). Either part may be empty. Concatenating these two parts reproduces the original file name.

On most systems, the directory part is everything up to and including the last slash (backslash is also allowed in input on MS-DOS or MS-Windows); the nondirectory part is the rest.

For some purposes, the nondirectory part is further subdivided into the name proper and the version number. On most systems, only backup files have version numbers in their names.

25.8.2 Absolute and Relative File Names

All the directories in the file system form a tree starting at the root directory. A file name can specify all the directory names starting from the root of the tree; then it is called an absolute file name. Or it can specify the position of the file in the tree relative to a default directory; then it is called a relative file name. On Unix and GNU/Linux, an absolute file name starts with a slash or a tilde (‘~’), and a relative one does not. On MS-DOS and MS-Windows, an absolute file name starts with a slash or a backslash, or with a drive specification ‘x:/’, where x is the drive letter.

Given a possibly relative file name, you can convert it to an absolute name using expand-file-name (see File Name Expansion). This function converts absolute file names to relative names:

25.8.3 Directory Names

A directory name is the name of a directory. A directory is actually a kind of file, so it has a file name, which is related to the directory name but not identical to it. (This is not quite the same as the usual Unix terminology.) These two different names for the same entity are related by a syntactic transformation. On GNU and Unix systems, this is simple: a directory name ends in a slash, whereas the directory's name as a file lacks that slash. On MS-DOS the relationship is more complicated.

The difference between a directory name and its name as a file is subtle but crucial. When an Emacs variable or function argument is described as being a directory name, a file name of a directory is not acceptable. When file-name-directory returns a string, that is always a directory name.

The following two functions convert between directory names and file names. They do nothing special with environment variable substitutions such as ‘$HOME’, and the constructs ‘~’, ‘.’ and ‘..’.

Given a directory name, you can combine it with a relative file name using concat:

     (concat dirname relfile)

Be sure to verify that the file name is relative before doing that. If you use an absolute file name, the results could be syntactically invalid or refer to the wrong file.

If you want to use a directory file name in making such a combination, you must first convert it to a directory name using file-name-as-directory:

     (concat (file-name-as-directory dirfile) relfile)

Don't try concatenating a slash by hand, as in

     ;;; Wrong!
     (concat dirfile "/" relfile)

because this is not portable. Always use file-name-as-directory.

To convert a directory name to its abbreviation, use this function:

25.8.4 Functions that Expand Filenames

Expansion of a file name means converting a relative file name to an absolute one. Since this is done relative to a default directory, you must specify the default directory name as well as the file name to be expanded. Expansion also simplifies file names by eliminating redundancies such as ./ and name/../.

25.8.5 Generating Unique File Names

Some programs need to write temporary files. Here is the usual way to construct a name for such a file:

     (make-temp-file name-of-application)

The job of make-temp-file is to prevent two different users or two different jobs from trying to use the exact same file name.

The default directory for temporary files is controlled by the variable temporary-file-directory. This variable gives the user a uniform way to specify the directory for all temporary files. Some programs use small-temporary-file-directory instead, if that is non-nil. To use it, you should expand the prefix against the proper directory before calling make-temp-file.

In older Emacs versions where make-temp-file does not exist, you should use make-temp-name instead:

     (make-temp-name
      (expand-file-name name-of-application
                        temporary-file-directory))

25.8.6 File Name Completion

This section describes low-level subroutines for completing a file name. For higher level functions, see Reading File Names.

25.8.7 Standard File Names

Most of the file names used in Lisp programs are entered by the user. But occasionally a Lisp program needs to specify a standard file name for a particular use—typically, to hold customization information about each user. For example, abbrev definitions are stored (by default) in the file ~/.abbrev_defs; the completion package stores completions in the file ~/.completions. These are two of the many standard file names used by parts of Emacs for certain purposes.

Various operating systems have their own conventions for valid file names and for which file names to use for user profile data. A Lisp program which reads a file using a standard file name ought to use, on each type of system, a file name suitable for that system. The function convert-standard-filename makes this easy to do.

The recommended way to specify a standard file name in a Lisp program is to choose a name which fits the conventions of GNU and Unix systems, usually with a nondirectory part that starts with a period, and pass it to convert-standard-filename instead of using it directly. Here is an example from the completion package:

     (defvar save-completions-file-name
             (convert-standard-filename "~/.completions")
       "*The file name to save completions to.")

On GNU and Unix systems, and on some other systems as well, convert-standard-filename returns its argument unchanged. On some other systems, it alters the name to fit the system's conventions.

For example, on MS-DOS the alterations made by this function include converting a leading ‘.’ to ‘_’, converting a ‘_’ in the middle of the name to ‘.’ if there is no other ‘.’, inserting a ‘.’ after eight characters if there is none, and truncating to three characters after the ‘.’. (It makes other changes as well.) Thus, .abbrev_defs becomes _abbrev.def, and .completions becomes _complet.ion.

25.9 Contents of Directories

A directory is a kind of file that contains other files entered under various names. Directories are a feature of the file system.

Emacs can list the names of the files in a directory as a Lisp list, or display the names in a buffer using the ls shell command. In the latter case, it can optionally display information about each file, depending on the options passed to the ls command.

25.10 Creating, Copying and Deleting Directories

Most Emacs Lisp file-manipulation functions get errors when used on files that are directories. For example, you cannot delete a directory with delete-file. These special functions exist to create and delete directories.

25.11 Making Certain File Names “Magic”

You can implement special handling for certain file names. This is called making those names magic. The principal use for this feature is in implementing remote file names (see Remote Files).

To define a kind of magic file name, you must supply a regular expression to define the class of names (all those that match the regular expression), plus a handler that implements all the primitive Emacs file operations for file names that do match.

The variable file-name-handler-alist holds a list of handlers, together with regular expressions that determine when to apply each handler. Each element has this form:

     (regexp . handler)

All the Emacs primitives for file access and file name transformation check the given file name against file-name-handler-alist. If the file name matches regexp, the primitives handle that file by calling handler.

The first argument given to handler is the name of the primitive, as a symbol; the remaining arguments are the arguments that were passed to that primitive. (The first of these arguments is most often the file name itself.) For example, if you do this:

     (file-exists-p filename)

and filename has handler handler, then handler is called like this:

     (funcall handler 'file-exists-p filename)

When a function takes two or more arguments that must be file names, it checks each of those names for a handler. For example, if you do this:

     (expand-file-name filename dirname)

then it checks for a handler for filename and then for a handler for dirname. In either case, the handler is called like this:

     (funcall handler 'expand-file-name filename dirname)

The handler then needs to figure out whether to handle filename or dirname.

If the specified file name matches more than one handler, the one whose match starts last in the file name gets precedence. This rule is chosen so that handlers for jobs such as uncompression are handled first, before handlers for jobs such as remote file access.

Here are the operations that a magic file name handler gets to handle:

access-file, add-name-to-file, byte-compiler-base-file-name,
copy-directory, copy-file, delete-directory, delete-file, diff-latest-backup-file, directory-file-name, directory-files, directory-files-and-attributes, dired-compress-file, dired-uncache,
expand-file-name, file-accessible-directory-p, file-attributes, file-directory-p, file-executable-p, file-exists-p, file-local-copy, file-remote-p, file-modes, file-name-all-completions, file-name-as-directory, file-name-completion, file-name-directory, file-name-nondirectory, file-name-sans-versions, file-newer-than-file-p, file-ownership-preserved-p, file-readable-p, file-regular-p, file-symlink-p, file-truename, file-writable-p, find-backup-file-name, get-file-buffer, insert-directory, insert-file-contents,
load, make-auto-save-file-name, make-directory, make-directory-internal, make-symbolic-link,
process-file, rename-file, set-file-modes, set-file-times, set-visited-file-modtime, shell-command, start-file-process, substitute-in-file-name,
unhandled-file-name-directory, vc-registered, verify-visited-file-modtime,
write-region.

Handlers for insert-file-contents typically need to clear the buffer's modified flag, with (set-buffer-modified-p nil), if the visit argument is non-nil. This also has the effect of unlocking the buffer if it is locked.

The handler function must handle all of the above operations, and possibly others to be added in the future. It need not implement all these operations itself—when it has nothing special to do for a certain operation, it can reinvoke the primitive, to handle the operation “in the usual way.” It should always reinvoke the primitive for an operation it does not recognize. Here's one way to do this:

     (defun my-file-handler (operation &rest args)
       ;; First check for the specific operations
       ;; that we have special handling for.
       (cond ((eq operation 'insert-file-contents) ...)
             ((eq operation 'write-region) ...)
             ...
             ;; Handle any operation we don't know about.
             (t (let ((inhibit-file-name-handlers
                       (cons 'my-file-handler
                             (and (eq inhibit-file-name-operation operation)
                                  inhibit-file-name-handlers)))
                      (inhibit-file-name-operation operation))
                  (apply operation args)))))

When a handler function decides to call the ordinary Emacs primitive for the operation at hand, it needs to prevent the primitive from calling the same handler once again, thus leading to an infinite recursion. The example above shows how to do this, with the variables inhibit-file-name-handlers and inhibit-file-name-operation. Be careful to use them exactly as shown above; the details are crucial for proper behavior in the case of multiple handlers, and for operations that have two file names that may each have handlers.

Handlers that don't really do anything special for actual access to the file—such as the ones that implement completion of host names for remote file names—should have a non-nil safe-magic property. For instance, Emacs normally “protects” directory names it finds in PATH from becoming magic, if they look like magic file names, by prefixing them with ‘/:’. But if the handler that would be used for them has a non-nil safe-magic property, the ‘/:’ is not added.

A file name handler can have an operations property to declare which operations it handles in a nontrivial way. If this property has a non-nil value, it should be a list of operations; then only those operations will call the handler. This avoids inefficiency, but its main purpose is for autoloaded handler functions, so that they won't be loaded except when they have real work to do.

Simply deferring all operations to the usual primitives does not work. For instance, if the file name handler applies to file-exists-p, then it must handle load itself, because the usual load code won't work properly in that case. However, if the handler uses the operations property to say it doesn't handle file-exists-p, then it need not handle load nontrivially.

25.12 File Format Conversion

Emacs performs several steps to convert the data in a buffer (text, text properties, and possibly other information) to and from a representation suitable for storing into a file. This section describes the fundamental functions that perform this format conversion, namely insert-file-contents for reading a file into a buffer, and write-region for writing a buffer into a file.

• Overview: insert-file-contents and write-region.
• Round-Trip: Using format-alist.
• Piecemeal: Specifying non-paired conversion.

25.12.1 Overview

The function insert-file-contents:

• initially, inserts bytes from the file into the buffer;
• decodes bytes to characters as appropriate;
• processes formats as defined by entries in format-alist; and
• calls functions in after-insert-file-functions.

The function write-region:

• initially, calls functions in write-region-annotate-functions;
• processes formats as defined by entries in format-alist;
• encodes characters to bytes as appropriate; and
• modifies the file with the bytes.

This shows the symmetry of the lowest-level operations; reading and writing handle things in opposite order. The rest of this section describes the two facilities surrounding the three variables named above, as well as some related functions. Coding Systems, for details on character encoding and decoding.

25.12.2 Round-Trip Specification

The most general of the two facilities is controlled by the variable format-alist, a list of file format specifications, which describe textual representations used in files for the data in an Emacs buffer. The descriptions for reading and writing are paired, which is why we call this “round-trip” specification (see Format Conversion Piecemeal, for non-paired specification).

Here is what the elements in a format definition mean:

name

The name of this format.

doc-string

A documentation string for the format.

regexp

A regular expression which is used to recognize files represented in this format. If nil, the format is never applied automatically.

from-fn

A shell command or function to decode data in this format (to convert file data into the usual Emacs data representation).

A shell command is represented as a string; Emacs runs the command as a filter to perform the conversion.

If from-fn is a function, it is called with two arguments, begin and end, which specify the part of the buffer it should convert. It should convert the text by editing it in place. Since this can change the length of the text, from-fn should return the modified end position.

One responsibility of from-fn is to make sure that the beginning of the file no longer matches regexp. Otherwise it is likely to get called again.

to-fn

A shell command or function to encode data in this format—that is, to convert the usual Emacs data representation into this format.

If to-fn is a string, it is a shell command; Emacs runs the command as a filter to perform the conversion.

If to-fn is a function, it is called with three arguments: begin and end, which specify the part of the buffer it should convert, and buffer, which specifies which buffer. There are two ways it can do the conversion:

• By editing the buffer in place. In this case, to-fn should return the end-position of the range of text, as modified.
• By returning a list of annotations. This is a list of elements of the form (position . string), where position is an integer specifying the relative position in the text to be written, and string is the annotation to add there. The list must be sorted in order of position when to-fn returns it.

  When write-region actually writes the text from the buffer to the file, it intermixes the specified annotations at the corresponding positions. All this takes place without modifying the buffer.

modify

A flag, t if the encoding function modifies the buffer, and nil if it works by returning a list of annotations.

mode-fn

A minor-mode function to call after visiting a file converted from this format. The function is called with one argument, the integer 1; that tells a minor-mode function to enable the mode.

preserve

A flag, t if format-write-file should not remove this format from buffer-file-format.

The function insert-file-contents automatically recognizes file formats when it reads the specified file. It checks the text of the beginning of the file against the regular expressions of the format definitions, and if it finds a match, it calls the decoding function for that format. Then it checks all the known formats over again. It keeps checking them until none of them is applicable.

Visiting a file, with find-file-noselect or the commands that use it, performs conversion likewise (because it calls insert-file-contents); it also calls the mode function for each format that it decodes. It stores a list of the format names in the buffer-local variable buffer-file-format.

When write-region writes data into a file, it first calls the encoding functions for the formats listed in buffer-file-format, in the order of appearance in the list.

25.12.3 Piecemeal Specification

In contrast to the round-trip specification described in the previous subsection (see Format Conversion Round-Trip), you can use the variables after-insert-file-functions and write-region-annotate-functions to separately control the respective reading and writing conversions.

Conversion starts with one representation and produces another representation. When there is only one conversion to do, there is no conflict about what to start with. However, when there are multiple conversions involved, conflict may arise when two conversions need to start with the same data.

This situation is best understood in the context of converting text properties during write-region. For example, the character at position 42 in a buffer is ‘X’ with a text property foo. If the conversion for foo is done by inserting into the buffer, say, ‘FOO:’, then that changes the character at position 42 from ‘X’ to ‘F’. The next conversion will start with the wrong data straight away.

To avoid conflict, cooperative conversions do not modify the buffer, but instead specify annotations, a list of elements of the form (position . string), sorted in order of increasing position.

If there is more than one conversion, write-region merges their annotations destructively into one sorted list. Later, when the text from the buffer is actually written to the file, it intermixes the specified annotations at the corresponding positions. All this takes place without modifying the buffer.

In contrast, when reading, the annotations intermixed with the text are handled immediately. insert-file-contents sets point to the beginning of some text to be converted, then calls the conversion functions with the length of that text. These functions should always return with point at the beginning of the inserted text. This approach makes sense for reading because annotations removed by the first converter can't be mistakenly processed by a later converter. Each conversion function should scan for the annotations it recognizes, remove the annotation, modify the buffer text (to set a text property, for example), and return the updated length of the text, as it stands after those changes. The value returned by one function becomes the argument to the next function.

We invite users to write Lisp programs to store and retrieve text properties in files, using these hooks, and thus to experiment with various data formats and find good ones. Eventually we hope users will produce good, general extensions we can install in Emacs.

We suggest not trying to handle arbitrary Lisp objects as text property names or values—because a program that general is probably difficult to write, and slow. Instead, choose a set of possible data types that are reasonably flexible, and not too hard to encode.

26 Backups and Auto-Saving

Backup files and auto-save files are two methods by which Emacs tries to protect the user from the consequences of crashes or of the user's own errors. Auto-saving preserves the text from earlier in the current editing session; backup files preserve file contents prior to the current session.

• Backup Files: How backup files are made; how their names are chosen.
• Auto-Saving: How auto-save files are made; how their names are chosen.
• Reverting: revert-buffer, and how to customize what it does.

26.1 Backup Files

A backup file is a copy of the old contents of a file you are editing. Emacs makes a backup file the first time you save a buffer into its visited file. Thus, normally, the backup file contains the contents of the file as it was before the current editing session. The contents of the backup file normally remain unchanged once it exists.

Backups are usually made by renaming the visited file to a new name. Optionally, you can specify that backup files should be made by copying the visited file. This choice makes a difference for files with multiple names; it also can affect whether the edited file remains owned by the original owner or becomes owned by the user editing it.

By default, Emacs makes a single backup file for each file edited. You can alternatively request numbered backups; then each new backup file gets a new name. You can delete old numbered backups when you don't want them any more, or Emacs can delete them automatically.

• Making Backups: How Emacs makes backup files, and when.
• Rename or Copy: Two alternatives: renaming the old file or copying it.
• Numbered Backups: Keeping multiple backups for each source file.
• Backup Names: How backup file names are computed; customization.

26.1.1 Making Backup Files

26.1.2 Backup by Renaming or by Copying?

There are two ways that Emacs can make a backup file:

• Emacs can rename the original file so that it becomes a backup file, and then write the buffer being saved into a new file. After this procedure, any other names (i.e., hard links) of the original file now refer to the backup file. The new file is owned by the user doing the editing, and its group is the default for new files written by the user in that directory.
• Emacs can copy the original file into a backup file, and then overwrite the original file with new contents. After this procedure, any other names (i.e., hard links) of the original file continue to refer to the current (updated) version of the file. The file's owner and group will be unchanged.

The first method, renaming, is the default.

The variable backup-by-copying, if non-nil, says to use the second method, which is to copy the original file and overwrite it with the new buffer contents. The variable file-precious-flag, if non-nil, also has this effect (as a sideline of its main significance). See Saving Buffers.

The following three variables, when non-nil, cause the second method to be used in certain special cases. They have no effect on the treatment of files that don't fall into the special cases.

26.1.3 Making and Deleting Numbered Backup Files

If a file's name is foo, the names of its numbered backup versions are foo.~v~, for various integers v, like this: foo.~1~, foo.~2~, foo.~3~, ..., foo.~259~, and so on.

The use of numbered backups ultimately leads to a large number of backup versions, which must then be deleted. Emacs can do this automatically or it can ask the user whether to delete them.

If there are backups numbered 1, 2, 3, 5, and 7, and both of these variables have the value 2, then the backups numbered 1 and 2 are kept as old versions and those numbered 5 and 7 are kept as new versions; backup version 3 is excess. The function find-backup-file-name (see Backup Names) is responsible for determining which backup versions to delete, but does not delete them itself.

26.1.4 Naming Backup Files

The functions in this section are documented mainly because you can customize the naming conventions for backup files by redefining them. If you change one, you probably need to change the rest.

26.2 Auto-Saving

Emacs periodically saves all files that you are visiting; this is called auto-saving. Auto-saving prevents you from losing more than a limited amount of work if the system crashes. By default, auto-saves happen every 300 keystrokes, or after around 30 seconds of idle time. See Auto Save, for information on auto-save for users. Here we describe the functions used to implement auto-saving and the variables that control them.

26.3 Reverting

If you have made extensive changes to a file and then change your mind about them, you can get rid of them by reading in the previous version of the file with the revert-buffer command. See Reverting a Buffer.

You can customize how revert-buffer does its work by setting the variables described in the rest of this section.

Some major modes customize revert-buffer by making buffer-local bindings for these variables:

27 Buffers

A buffer is a Lisp object containing text to be edited. Buffers are used to hold the contents of files that are being visited; there may also be buffers that are not visiting files. While several buffers may exist at one time, only one buffer is designated the current buffer at any time. Most editing commands act on the contents of the current buffer. Each buffer, including the current buffer, may or may not be displayed in any windows.

• Buffer Basics: What is a buffer?
• Current Buffer: Designating a buffer as current so that primitives will access its contents.
• Buffer Names: Accessing and changing buffer names.
• Buffer File Name: The buffer file name indicates which file is visited.
• Buffer Modification: A buffer is modified if it needs to be saved.
• Modification Time: Determining whether the visited file was changed ``behind Emacs's back''.
• Read Only Buffers: Modifying text is not allowed in a read-only buffer.
• The Buffer List: How to look at all the existing buffers.
• Creating Buffers: Functions that create buffers.
• Killing Buffers: Buffers exist until explicitly killed.
• Indirect Buffers: An indirect buffer shares text with some other buffer.
• Swapping Text: Swapping text between two buffers.
• Buffer Gap: The gap in the buffer.

27.1 Buffer Basics

A buffer is a Lisp object containing text to be edited. Buffers are used to hold the contents of files that are being visited; there may also be buffers that are not visiting files. Although several buffers normally exist, only one buffer is designated the current buffer at any time. Most editing commands act on the contents of the current buffer. Each buffer, including the current buffer, may or may not be displayed in any windows.

Buffers in Emacs editing are objects that have distinct names and hold text that can be edited. Buffers appear to Lisp programs as a special data type. You can think of the contents of a buffer as a string that you can extend; insertions and deletions may occur in any part of the buffer. See Text.

A Lisp buffer object contains numerous pieces of information. Some of this information is directly accessible to the programmer through variables, while other information is accessible only through special-purpose functions. For example, the visited file name is directly accessible through a variable, while the value of point is accessible only through a primitive function.

Buffer-specific information that is directly accessible is stored in buffer-local variable bindings, which are variable values that are effective only in a particular buffer. This feature allows each buffer to override the values of certain variables. Most major modes override variables such as fill-column or comment-column in this way. For more information about buffer-local variables and functions related to them, see Buffer-Local Variables.

For functions and variables related to visiting files in buffers, see Visiting Files and Saving Buffers. For functions and variables related to the display of buffers in windows, see Buffers and Windows.

27.2 The Current Buffer

There are, in general, many buffers in an Emacs session. At any time, one of them is designated as the current buffer. This is the buffer in which most editing takes place, because most of the primitives for examining or changing text in a buffer operate implicitly on the current buffer (see Text). Normally the buffer that is displayed on the screen in the selected window is the current buffer, but this is not always so: a Lisp program can temporarily designate any buffer as current in order to operate on its contents, without changing what is displayed on the screen.

The way to designate a current buffer in a Lisp program is by calling set-buffer. The specified buffer remains current until a new one is designated.

When an editing command returns to the editor command loop, the command loop designates the buffer displayed in the selected window as current, to prevent confusion: the buffer that the cursor is in when Emacs reads a command is the buffer that the command will apply to. (See Command Loop.) Therefore, set-buffer is not the way to switch visibly to a different buffer so that the user can edit it. For that, you must use the functions described in Displaying Buffers.

Warning: Lisp functions that change to a different current buffer should not depend on the command loop to set it back afterwards. Editing commands written in Emacs Lisp can be called from other programs as well as from the command loop; it is convenient for the caller if the subroutine does not change which buffer is current (unless, of course, that is the subroutine's purpose). Therefore, you should normally use set-buffer within a save-current-buffer or save-excursion (see Excursions) form that will restore the current buffer when your function is done. Here, as an example, is a simplified version of the command append-to-buffer:

     (defun append-to-buffer (buffer start end)
       "Append to specified buffer the text of the region."
       (interactive "BAppend to buffer: \nr")
       (let ((oldbuf (current-buffer)))
         (save-current-buffer
           (set-buffer (get-buffer-create buffer))
           (insert-buffer-substring oldbuf start end))))

This function binds a local variable to record the current buffer, and then save-current-buffer arranges to make it current again. Next, set-buffer makes the specified buffer current. Finally, insert-buffer-substring copies the string from the original current buffer to the specified (and now current) buffer.

If the buffer appended to happens to be displayed in some window, the next redisplay will show how its text has changed. Otherwise, you will not see the change immediately on the screen. The buffer becomes current temporarily during the execution of the command, but this does not cause it to be displayed.

If you make local bindings (with let or function arguments) for a variable that may also have buffer-local bindings, make sure that the same buffer is current at the beginning and at the end of the local binding's scope. Otherwise you might bind it in one buffer and unbind it in another! There are two ways to do this. In simple cases, you may see that nothing ever changes the current buffer within the scope of the binding. Otherwise, use save-current-buffer or save-excursion to make sure that the buffer current at the beginning is current again whenever the variable is unbound.

Do not rely on using set-buffer to change the current buffer back, because that won't do the job if a quit happens while the wrong buffer is current. For instance, in the previous example, it would have been wrong to do this:

       (let ((oldbuf (current-buffer)))
         (set-buffer (get-buffer-create buffer))
         (insert-buffer-substring oldbuf start end)
         (set-buffer oldbuf))

Using save-current-buffer, as we did, handles quitting, errors, and throw, as well as ordinary evaluation.

27.3 Buffer Names

Each buffer has a unique name, which is a string. Many of the functions that work on buffers accept either a buffer or a buffer name as an argument. Any argument called buffer-or-name is of this sort, and an error is signaled if it is neither a string nor a buffer. Any argument called buffer must be an actual buffer object, not a name.

Buffers that are ephemeral and generally uninteresting to the user have names starting with a space, so that the list-buffers and buffer-menu commands don't mention them (but if such a buffer visits a file, it is mentioned). A name starting with space also initially disables recording undo information; see Undo.

27.4 Buffer File Name

The buffer file name is the name of the file that is visited in that buffer. When a buffer is not visiting a file, its buffer file name is nil. Most of the time, the buffer name is the same as the nondirectory part of the buffer file name, but the buffer file name and the buffer name are distinct and can be set independently. See Visiting Files.

27.5 Buffer Modification

Emacs keeps a flag called the modified flag for each buffer, to record whether you have changed the text of the buffer. This flag is set to t whenever you alter the contents of the buffer, and cleared to nil when you save it. Thus, the flag shows whether there are unsaved changes. The flag value is normally shown in the mode line (see Mode Line Variables), and controls saving (see Saving Buffers) and auto-saving (see Auto-Saving).

Some Lisp programs set the flag explicitly. For example, the function set-visited-file-name sets the flag to t, because the text does not match the newly-visited file, even if it is unchanged from the file formerly visited.

The functions that modify the contents of buffers are described in Text.

27.6 Buffer Modification Time

Suppose that you visit a file and make changes in its buffer, and meanwhile the file itself is changed on disk. At this point, saving the buffer would overwrite the changes in the file. Occasionally this may be what you want, but usually it would lose valuable information. Emacs therefore checks the file's modification time using the functions described below before saving the file. (See File Attributes, for how to examine a file's modification time.)

27.7 Read-Only Buffers

If a buffer is read-only, then you cannot change its contents, although you may change your view of the contents by scrolling and narrowing.

Read-only buffers are used in two kinds of situations:

• A buffer visiting a write-protected file is normally read-only.

  Here, the purpose is to inform the user that editing the buffer with the aim of saving it in the file may be futile or undesirable. The user who wants to change the buffer text despite this can do so after clearing the read-only flag with C-x C-q.

• Modes such as Dired and Rmail make buffers read-only when altering the contents with the usual editing commands would probably be a mistake.

  The special commands of these modes bind buffer-read-only to nil (with let) or bind inhibit-read-only to t around the places where they themselves change the text.

27.8 The Buffer List

The buffer list is a list of all live buffers. The order of the buffers in this list is based primarily on how recently each buffer has been displayed in a window. Several functions, notably other-buffer, use this ordering. A buffer list displayed for the user also follows this order.

Creating a buffer adds it to the end of the buffer list, and killing a buffer removes it from that list. A buffer moves to the front of this list whenever it is chosen for display in a window (see Displaying Buffers) or a window displaying it is selected (see Selecting Windows). A buffer moves to the end of the list when it is buried (see bury-buffer, below). There are no functions available to the Lisp programmer which directly manipulate the buffer list.

In addition to the fundamental buffer list just described, Emacs maintains a local buffer list for each frame, in which the buffers that have been displayed (or had their windows selected) in that frame come first. (This order is recorded in the frame's buffer-list frame parameter; see Buffer Parameters.) Buffers never displayed in that frame come afterward, ordered according to the fundamental buffer list.

The list returned by buffer-list is constructed specifically; it is not an internal Emacs data structure, and modifying it has no effect on the order of buffers. If you want to change the order of buffers in the fundamental buffer list, here is an easy way:

     (defun reorder-buffer-list (new-list)
       (while new-list
         (bury-buffer (car new-list))
         (setq new-list (cdr new-list))))

With this method, you can specify any order for the list, but there is no danger of losing a buffer or adding something that is not a valid live buffer.

To change the order or value of a specific frame's buffer list, set that frame's buffer-list parameter with modify-frame-parameters (see Parameter Access).

27.9 Creating Buffers

This section describes the two primitives for creating buffers. get-buffer-create creates a buffer if it finds no existing buffer with the specified name; generate-new-buffer always creates a new buffer and gives it a unique name.

Other functions you can use to create buffers include with-output-to-temp-buffer (see Temporary Displays) and create-file-buffer (see Visiting Files). Starting a subprocess can also create a buffer (see Processes).

27.10 Killing Buffers

Killing a buffer makes its name unknown to Emacs and makes the memory space it occupied available for other use.

The buffer object for the buffer that has been killed remains in existence as long as anything refers to it, but it is specially marked so that you cannot make it current or display it. Killed buffers retain their identity, however; if you kill two distinct buffers, they remain distinct according to eq although both are dead.

If you kill a buffer that is current or displayed in a window, Emacs automatically selects or displays some other buffer instead. This means that killing a buffer can in general change the current buffer. Therefore, when you kill a buffer, you should also take the precautions associated with changing the current buffer (unless you happen to know that the buffer being killed isn't current). See Current Buffer.

If you kill a buffer that is the base buffer of one or more indirect buffers, the indirect buffers are automatically killed as well.

The buffer-name of a killed buffer is nil. You can use this feature to test whether a buffer has been killed:

     (defun buffer-killed-p (buffer)
       "Return t if BUFFER is killed."
       (not (buffer-name buffer)))

27.11 Indirect Buffers

An indirect buffer shares the text of some other buffer, which is called the base buffer of the indirect buffer. In some ways it is the analogue, for buffers, of a symbolic link among files. The base buffer may not itself be an indirect buffer.

The text of the indirect buffer is always identical to the text of its base buffer; changes made by editing either one are visible immediately in the other. This includes the text properties as well as the characters themselves.

In all other respects, the indirect buffer and its base buffer are completely separate. They have different names, independent values of point, independent narrowing, independent markers and overlays (though inserting or deleting text in either buffer relocates the markers and overlays for both), independent major modes, and independent buffer-local variable bindings.

An indirect buffer cannot visit a file, but its base buffer can. If you try to save the indirect buffer, that actually saves the base buffer.

Killing an indirect buffer has no effect on its base buffer. Killing the base buffer effectively kills the indirect buffer in that it cannot ever again be the current buffer.

27.12 Swapping Text Between Two Buffers

Specialized modes sometimes need to let the user access from the same buffer several vastly different types of text. For example, you may need to display a summary of the buffer text, in addition to letting the user access the text itself.

This could be implemented with multiple buffers (kept in sync when the user edits the text), or with narrowing (see Narrowing). But these alternatives might sometimes become tedious or prohibitively expensive, especially if each type of text requires expensive buffer-global operations in order to provide correct display and editing commands.

Emacs provides another facility for such modes: you can quickly swap buffer text between two buffers with buffer-swap-text. This function is very fast because it doesn't move any text, it only changes the internal data structures of the buffer object to point to a different chunk of text. Using it, you can pretend that a group of two or more buffers are actually a single virtual buffer that holds the contents of all the individual buffers together.

If you use buffer-swap-text on a file-visiting buffer, you should set up a hook to save the buffer's original text rather than what it was swapped with. write-region-annotate-functions works for this purpose. You should probably set buffer-saved-size to −2 in the buffer, so that changes in the text it is swapped with will not interfere with auto-saving.

27.13 The Buffer Gap

Emacs buffers are implemented using an invisible gap to make insertion and deletion faster. Insertion works by filling in part of the gap, and deletion adds to the gap. Of course, this means that the gap must first be moved to the locus of the insertion or deletion. Emacs moves the gap only when you try to insert or delete. This is why your first editing command in one part of a large buffer, after previously editing in another far-away part, sometimes involves a noticeable delay.

This mechanism works invisibly, and Lisp code should never be affected by the gap's current location, but these functions are available for getting information about the gap status.

28 Windows

This chapter describes most of the functions and variables related to Emacs windows. See Frames and Windows, for how windows relate to frames. See Display, for information on how text is displayed in windows.

• Basic Windows: Basic information on using windows.
• Splitting Windows: Splitting one window into two windows.
• Deleting Windows: Deleting a window gives its space to other windows.
• Selecting Windows: The selected window is the one that you edit in.
• Cyclic Window Ordering: Moving around the existing windows.
• Buffers and Windows: Each window displays the contents of a buffer.
• Displaying Buffers: Higher-level functions for displaying a buffer and choosing a window for it.
• Choosing Window: How to choose a window for displaying a buffer.
• Dedicated Windows: How to avoid displaying another buffer in a specific window.
• Window Point: Each window has its own location of point.
• Window Start and End: Buffer positions indicating which text is on-screen in a window.
• Textual Scrolling: Moving text up and down through the window.
• Vertical Scrolling: Moving the contents up and down on the window.
• Horizontal Scrolling: Moving the contents sideways on the window.
• Size of Window: Accessing the size of a window.
• Resizing Windows: Changing the size of a window.
• Coordinates and Windows: Converting coordinates to windows.
• Window Tree: The layout and sizes of all windows in a frame.
• Window Configurations: Saving and restoring the state of the screen.
• Window Parameters: Associating additional information with windows.
• Window Hooks: Hooks for scrolling, window size changes, redisplay going past a certain point, or window configuration changes.

28.1 Basic Concepts of Emacs Windows

A window in Emacs is the physical area of the screen in which a buffer is displayed. The term is also used to refer to a Lisp object that represents that screen area in Emacs Lisp. It should be clear from the context which is meant.

Emacs groups windows into frames; see Frames. A frame represents an area of screen available for Emacs to use. Each frame always contains at least one window, but you can subdivide it vertically or horizontally into multiple, nonoverlapping Emacs windows.

In each frame, at any time, one and only one window is designated as selected within the frame. The frame's cursor appears in that window, but the other windows have “non-selected” cursors, normally less visible. (See Cursor Parameters, for customizing this.) At any time, one frame is the selected frame; and the window selected within that frame is the selected window. The selected window's buffer is usually the current buffer (except when set-buffer has been used); see Current Buffer.

For practical purposes, a window exists only while it is displayed in a frame. Once removed from the frame, the window is effectively deleted and should not be used, even though there may still be references to it from other Lisp objects; see Deleting Windows. Restoring a saved window configuration is the only way for a window no longer on the screen to come back to life; see Window Configurations.

Users create multiple windows so they can look at several buffers at once. Lisp libraries use multiple windows for a variety of reasons, but most often to display related information. In Rmail, for example, you can move through a summary buffer in one window while the other window shows messages one at a time as they are reached.

The meaning of “window” in Emacs is similar to what it means in the context of general-purpose window systems such as X, but not identical. The X Window System places X windows on the screen; Emacs uses one or more X windows as frames, and subdivides them into Emacs windows. When you use Emacs on a character-only terminal, Emacs treats the whole terminal screen as one frame.

Most window systems support arbitrarily located overlapping windows. In contrast, Emacs windows are tiled; they never overlap, and together they fill the whole screen or frame. Because of the way in which Emacs creates new windows (see Splitting Windows) and resizes them (see Resizing Windows), not all conceivable tilings of windows on an Emacs frame are actually possible.

28.2 Splitting Windows

The functions described below are the primitives used to split a window into two windows. They do not accept a buffer as an argument. Rather, the two “halves” of the split window initially display the same buffer previously visible in the window that was split.

28.3 Deleting Windows

A window remains visible on its frame unless you delete it by calling certain functions that delete windows. A deleted window cannot appear on the screen, but continues to exist as a Lisp object until there are no references to it. There is no way to cancel the deletion of a window aside from restoring a saved window configuration (see Window Configurations). Restoring a window configuration also deletes any windows that aren't part of that configuration.

When you delete a window, the space it took up is given to one of its sibling windows adjacent to it.

28.4 Selecting Windows

When a window is selected, the buffer in the window becomes the current buffer, and the cursor will appear in it.

The following functions choose one of the windows on the screen, offering various criteria for the choice.

28.5 Cyclic Ordering of Windows

When you use the command C-x o (other-window) to select some other window, it moves through the windows on the screen in a specific order. For any given configuration of windows, this order never varies. It is called the cyclic ordering of windows.

For a particular frame, this ordering generally goes from top to bottom, and from left to right. But it may go down first or go right first, depending on the order in which windows were split.

If the first split was vertical (into windows one above each other), and then the subwindows were split horizontally, then the ordering is left to right in the top of the frame, and then left to right in the next lower part of the frame, and so on. If the first split was horizontal, the ordering is top to bottom in the left part, and so on. In general, within each set of siblings at any level in the window tree (see Window Tree), the order is left to right, or top to bottom.

28.6 Buffers and Windows

This section describes low-level functions to examine windows or to display buffers in windows in a precisely controlled fashion. See Displaying Buffers, for related functions that find a window to use and specify a buffer for it. The functions described there are easier to use, but they employ heuristics in choosing or creating a window; use the functions described here when you need complete control.