CC Mode

CC Mode is a GNU Emacs mode for editing files containing C, C++, Objective-C, Java, CORBA IDL (and the variants PSDL and CIDL), Pike and AWK code. It provides syntax-based indentation, font locking, and has several handy commands and some minor modes to make the editing easier. It does not provide tools to look up and navigate between functions, classes, etc.; there are other packages for that.

This manual is for CC Mode in Emacs.

Copyright © 1995–2014 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 no Invariant Sections, 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.”

Introduction
Overview
Getting Started
Commands
Font Locking
Config Basics
Custom Filling and Breaking
Custom Auto-newlines
Clean-ups
Indentation Engine Basics
Customizing Indentation
Custom Macros
Odds and Ends
Sample Init File
Performance Issues
Limitations and Known Bugs
FAQ
Updating CC Mode
Mailing Lists and Bug Reports
GNU Free Documentation License
Command and Function Index
Variable Index
Concept and Key Index

Detailed Node Listing

Commands
Indentation Commands
Comment Commands
Movement Commands
Filling and Breaking
Minor Modes
Electric Keys
Auto-newlines
Hungry WS Deletion
Subword Movement
Other Commands
Font Locking
Font Locking Preliminaries
Faces
Doc Comments
AWK Mode Font Locking
Configuration Basics
CC Hooks
Style Variables
Styles
Styles
Built-in Styles
Choosing a Style
Adding Styles
Guessing the Style
File Styles
Customizing Auto-newlines
Hanging Braces
Hanging Colons
Hanging Semicolons and Commas
Hanging Braces
Custom Braces
Indentation Engine Basics
Syntactic Analysis
Syntactic Symbols
Indentation Calculation
Syntactic Symbols
Function Symbols
Class Symbols
Conditional Construct Symbols
Switch Statement Symbols
Brace List Symbols
External Scope Symbols
Paren List Symbols
Literal Symbols
Multiline Macro Symbols
Objective-C Method Symbols
Java Symbols
Statement Block Symbols
K&R Symbols
Customizing Indentation
c-offsets-alist
Interactive Customization
Line-Up Functions
Custom Line-Up
Other Indentation
Line-Up Functions
Brace/Paren Line-Up
List Line-Up
Operator Line-Up
Comment Line-Up
Misc Line-Up
Customizing Macros
Macro Backslashes
Macros with ;

Next: , Previous: Top, Up: Top

1 Introduction

Welcome to CC Mode, a GNU Emacs mode for editing files containing C, C++, Objective-C, Java, CORBA IDL (and the variants CORBA PSDL and CIDL), Pike and AWK code. This incarnation of the mode is descended from c-mode.el (also called “Boring Old C Mode” or BOCM :-), c++-mode.el version 2, which Barry Warsaw had been maintaining since 1992, and awk-mode.el, a long neglected mode in the (X)Emacs base.

Late in 1997, Martin Stjernholm joined Barry on the CC Mode Maintainers Team, and implemented the Pike support. In 2000 Martin took over as the sole maintainer. In 2001 Alan Mackenzie joined the team, implementing AWK support in version 5.30. CC Mode did not originally contain the font lock support for its languages; that was added in version 5.30.

This manual describes CC Mode version 5.32.

CC Mode supports the editing of K&R and ANSI C, C++, Objective-C, Java, CORBA's Interface Definition Language, Pike1 and AWK files. In this way, you can easily set up consistent font locking and coding styles for use in editing all of these languages, although AWK is not yet as uniformly integrated as the other languages.

Note that the name of this package is “CC Mode”, but there is no top level cc-mode entry point. All of the variables, commands, and functions in CC Mode are prefixed with c-thing, and c-mode, c++-mode, objc-mode, java-mode, idl-mode, pike-mode, and awk-mode entry points are provided. This package is intended to be a replacement for c-mode.el, c++-mode.el and awk-mode.el.

A special word of thanks goes to Krishna Padmasola for his work in converting the original README file to Texinfo format. I'd also like to thank all the CC Mode victims who help enormously during the early beta stages of CC Mode's development.

Next: , Previous: Introduction, Up: Top

2 Overview of the Manual

The manual starts with several introductory chapters (including this one).

The next chunk of the manual describes the day to day use of CC Mode (as contrasted with how to customize it).

The next chunk of the manual describes how to customize CC Mode. Typically, an overview of a topic is given at the chapter level, then the sections and subsections describe the material in increasing detail.

The manual ends with “this and that”, things that don't fit cleanly into any of the previous chunks.

Finally, there are the customary indices.

Next: , Previous: Overview, Up: Top

3 Getting Started

If you got this version of CC Mode with Emacs or XEmacs, it should work just fine right out of the box. Note however that you might not have the latest CC Mode release and might want to upgrade your copy (see below).

You should probably start by skimming through the entire Commands chapter (see Commands) to get an overview of CC Mode's capabilities.

After trying out some commands, you may dislike some aspects of CC Mode's default configuration. Here is an outline of how to change some of the settings that newcomers to CC Mode most often want to change:

c-basic-offset
This Lisp variable holds an integer, the number of columns CC Mode indents nested code. To set this value to 6, customize c-basic-offset or put this into your .emacs:
          (setq c-basic-offset 6)

The (indentation) style
The basic “shape” of indentation created by CC Mode—by default, this is gnu style (except for Java and AWK buffers). A list of the available styles and their descriptions can be found in Built-in Styles. A complete specification of the CC Mode style system, including how to create your own style, can be found in the chapter Styles. To set your style to linux, either customize c-default-style or put this into your .emacs:
          (setq c-default-style '((java-mode . "java")
                                  (awk-mode . "awk")
                                  (other . "linux")))

Electric Indentation
Normally, when you type “punctuation” characters such as ‘;’ or ‘{’, CC Mode instantly reindents the current line. This can be disconcerting until you get used to it. To disable electric indentation in the current buffer, type C-c C-l. Type the same thing to enable it again. To have electric indentation disabled by default, put the following into your .emacs file2:
          (setq-default c-electric-flag nil)

Details of this and other similar “Minor Modes” appear in the section Minor Modes.

Making the <RET> key indent the new line
The standard Emacs binding for <RET> just adds a new line. If you want it to reindent the new line as well, rebind the key. Note that the action of rebinding would fail if the pertinent keymap didn't yet exist—we thus need to delay the action until after CC Mode has been loaded. Put the following code into your .emacs:
          (defun my-make-CR-do-indent ()
            (define-key c-mode-base-map "\C-m" 'c-context-line-break))
          (add-hook 'c-initialization-hook 'my-make-CR-do-indent)

This example demonstrates the use of a very powerful CC Mode (and Emacs) facility, the hook. The use of CC Mode's hooks is described in CC Hooks.

All these settings should occur in your .emacs before any CC Mode buffers get loaded—in particular, before any call of desktop-read.

As you get to know the mode better, you may want to make more ambitious changes to your configuration. For this, you should start reading the chapter Config Basics.

If you are upgrading an existing CC Mode installation, please see the README file for installation details. In particular, if you are going to be editing AWK files, README describes how to configure your (X)Emacs so that CC Mode will supersede the obsolete awk-mode.el which might have been supplied with your (X)Emacs. CC Mode might not work with older versions of Emacs or XEmacs. See the CC Mode release notes at http://cc-mode.sourceforge.net for the latest information on Emacs version and package compatibility (see Updating CC Mode).

— Command: c-version

You can find out what version of CC Mode you are using by visiting a C file and entering M-x c-version RET. You should see this message in the echo area:

          Using CC Mode version 5.XX

where ‘XX’ is the minor release number.

Next: , Previous: Getting Started, Up: Top

4 Commands

This chapter specifies all of CC Mode's commands, and thus contains nearly everything you need to know to use CC Mode (as contrasted with configuring it). Commands here means both control key sequences and electric keys, these being characters such as ‘;’ which, as well as inserting themselves into the buffer, also do other things.

You might well want to review Moving by Parens, which describes commands for moving around brace and parenthesis structures.

Next: , Previous: Commands, Up: Commands

4.1 Indentation Commands

The following commands reindent C constructs. Note that when you change your coding style, either interactively or through some other means, your file does not automatically get reindented. You will need to execute one of the following commands to see the effects of your changes.

Also, variables like c-hanging-* and c-cleanup-list (see Custom Auto-newlines) only affect how on-the-fly code is formatted. Changing the “hanginess” of a brace and then reindenting, will not move the brace to a different line. For this, you're better off getting an external program like GNU indent, which will rearrange brace location, amongst other things.

Preprocessor directives are handled as syntactic whitespace from other code, i.e., they can be interspersed anywhere without affecting the indentation of the surrounding code, just like comments.

The code inside macro definitions is, by default, still analyzed syntactically so that you get relative indentation there just as you'd get if the same code was outside a macro. However, since there is no hint about the syntactic context, i.e., whether the macro expands to an expression, to some statements, or perhaps to whole functions, the syntactic recognition can be wrong. CC Mode manages to figure it out correctly most of the time, though.

Some macros, when invoked, ”have their own semicolon”. To get the next line indented correctly, rather than as a continuation line, See Macros with ;.

Reindenting large sections of code can take a long time. When CC Mode reindents a region of code, it is essentially equivalent to hitting <TAB> on every line of the region.

These commands indent code:

<TAB> (c-indent-command)
This command indents the current line. That is all you need to know about it for normal use.

c-indent-command does different things, depending on the setting of c-syntactic-indentation (see Indentation Engine Basics):

  • When it's non-nil (which it normally is), the command indents the line according to its syntactic context. With a prefix argument (C-u <TAB>), it will re-indent the entire expression3 that begins at the line's left margin.
  • When it's nil, the command indents the line by an extra c-basic-offset columns. A prefix argument acts as a multiplier. A bare prefix (C-u <TAB>) is equivalent to -1, removing c-basic-offset columns from the indentation.

The precise behavior is modified by several variables: With c-tab-always-indent, you can make <TAB> insert whitespace in some circumstances—c-insert-tab-function then defines precisely what sort of “whitespace” this will be. Set the standard Emacs variable indent-tabs-mode to t if you want real ‘tab’ characters to be used in the indentation, to nil if you want only spaces. See Just Spaces.

— User Option: c-tab-always-indent

This variable modifies how <TAB> operates.

  • When it is t (the default), <TAB> simply indents the current line.
  • When it is nil, <TAB> (re)indents the line only if point is to the left of the first non-whitespace character on the line. Otherwise it inserts some whitespace (a tab or an equivalent number of spaces; see below) at point.
  • With some other value, the line is reindented. Additionally, if point is within a string or comment, some whitespace is inserted.

— User Option: c-insert-tab-function

When “some whitespace” is inserted as described above, what actually happens is that the function stored in c-insert-tab-function is called. Normally, this is insert-tab, which inserts a real tab character or the equivalent number of spaces (depending on indent-tabs-mode). Some people, however, set c-insert-tab-function to tab-to-tab-stop so as to get hard tab stops when indenting.

The kind of indentation the next five commands do depends on the setting of c-syntactic-indentation (see Indentation Engine Basics):

C-M-q (c-indent-exp)
Indents an entire balanced brace or parenthesis expression. Note that point must be on the opening brace or parenthesis of the expression you want to indent.
C-c C-q (c-indent-defun)
Indents the entire top-level function, class or macro definition encompassing point. It leaves point unchanged. This function can't be used to reindent a nested brace construct, such as a nested class or function, or a Java method. The top-level construct being reindented must be complete, i.e., it must have both a beginning brace and an ending brace.
C-M-\ (indent-region)
Indents an arbitrary region of code. This is a standard Emacs command, tailored for C code in a CC Mode buffer. Note, of course, that point and mark must delineate the region you want to indent.
C-M-h (c-mark-function)
While not strictly an indentation command, this is useful for marking the current top-level function or class definition as the current region. As with c-indent-defun, this command operates on top-level constructs, and can't be used to mark say, a Java method.

These variables are also useful when indenting code:

— User Option: indent-tabs-mode

This is a standard Emacs variable that controls how line indentation is composed. When it's non-nil, tabs can be used in a line's indentation, otherwise only spaces are used.

— User Option: c-progress-interval

When indenting large regions of code, this variable controls how often a progress message is displayed. Set this variable to nil to inhibit the progress messages, or set it to an integer which is how often (in seconds) progress messages are to be displayed.

Next: , Previous: Indentation Commands, Up: Commands

4.2 Comment Commands

C-c C-c (comment-region)
This command comments out the lines that start in the region. With a negative argument, it does the opposite: it deletes the comment delimiters from these lines. See Multi-Line Comments, for fuller details. comment-region isn't actually part of CC Mode; it is given a CC Mode binding for convenience.
M-; (comment-dwim or indent-for-comment 4)
Insert a comment at the end of the current line, if none is there already. Then reindent the comment according to comment-column (see Options for Comments) and the variables below. Finally, position the point after the comment starter. C-u M-; kills any comment on the current line, together with any whitespace before it. This is a standard Emacs command, but CC Mode enhances it a bit with two variables:
— User Option: c-indent-comment-alist

This style variable allows you to vary the column that M-; puts the comment at, depending on what sort of code is on the line, and possibly the indentation of any similar comment on the preceding line. It is an association list that maps different types of lines to actions describing how they should be handled. If a certain line type isn't present on the list then the line is indented to the column specified by comment-column.

See the documentation string for a full description of this variable (use C-h v c-indent-comment-alist).

— User Option: c-indent-comments-syntactically-p

Normally, when this style variable is nil, M-; will indent comment-only lines according to c-indent-comment-alist, just as it does with lines where other code precede the comments. However, if you want it to act just like <TAB> for comment-only lines you can get that by setting c-indent-comments-syntactically-p to non-nil.

If c-indent-comments-syntactically-p is non-nil then c-indent-comment-alist won't be consulted at all for comment-only lines.

Next: , Previous: Comment Commands, Up: Commands

4.3 Movement Commands

CC Mode contains some useful commands for moving around in C code.

C-M-a (c-beginning-of-defun)
C-M-e (c-end-of-defun)
Move to the beginning or end of the current or next function. Other constructs (such as a structs or classes) which have a brace block also count as “functions” here. To move over several functions, you can give these commands a repeat count.

The start of a function is at its header. The end of the function is after its closing brace, or after the semicolon of a construct (such as a struct) which doesn't end at the brace. These two commands try to leave point at the beginning of a line near the actual start or end of the function. This occasionally causes point not to move at all.

By default, these commands will recognize functions contained within a declaration scope such as a C++ class or namespace construct, should the point start inside it. If CC Mode fails to find function beginnings or ends inside the current declaration scope, it will search the enclosing scopes. If you want CC Mode to recognize functions only at the top level5, set c-defun-tactic to t.

These functions are analogous to the Emacs built-in commands beginning-of-defun and end-of-defun, except they eliminate the constraint that the top-level opening brace of the defun must be in column zero. See Defuns, for more information.

C-M-a (AWK Mode) (c-awk-beginning-of-defun)
C-M-e (AWK Mode) (c-awk-end-of-defun)
Move to the beginning or end of the current or next AWK defun. These commands can take prefix-arguments, their functionality being entirely equivalent to beginning-of-defun and end-of-defun.

AWK Mode defuns are either pattern/action pairs (either of which might be implicit) or user defined functions. Having the ‘{’ and ‘}’ (if there are any) in column zero, as is suggested for some modes, is neither necessary nor helpful in AWK mode.

M-a (c-beginning-of-statement)
M-e (c-end-of-statement)
Move to the beginning or end of the innermost C statement. If point is already there, move to the next beginning or end of a statement, even if that means moving into a block. (Use C-M-b or C-M-f to move over a balanced block.) A prefix argument n means move over n statements.

If point is within or next to a comment or a string which spans more than one line, these commands move by sentences instead of statements.

When called from a program, these functions take three optional arguments: the repetition count, a buffer position limit which is the farthest back to search for the syntactic context, and a flag saying whether to do sentence motion in or near comments and multiline strings.

C-c C-u (c-up-conditional)
Move back to the containing preprocessor conditional, leaving the mark behind. A prefix argument acts as a repeat count. With a negative argument, move forward to the end of the containing preprocessor conditional.

#elif’ is treated like ‘#else’ followed by ‘#if’, so the function stops at them when going backward, but not when going forward.

This key sequence is not bound in AWK Mode, which doesn't have preprocessor statements.

M-x c-up-conditional-with-else
A variety of c-up-conditional that also stops at ‘#else’ lines. Normally those lines are ignored.
M-x c-down-conditional
Move forward into the next nested preprocessor conditional, leaving the mark behind. A prefix argument acts as a repeat count. With a negative argument, move backward into the previous nested preprocessor conditional.

#elif’ is treated like ‘#else’ followed by ‘#if’, so the function stops at them when going forward, but not when going backward.

M-x c-down-conditional-with-else
A variety of c-down-conditional that also stops at ‘#else’ lines. Normally those lines are ignored.
C-c C-p (c-backward-conditional)
C-c C-n (c-forward-conditional)
Move backward or forward across a preprocessor conditional, leaving the mark behind. A prefix argument acts as a repeat count. With a negative argument, move in the opposite direction.

These key sequences are not bound in AWK Mode, which doesn't have preprocessor statements.

M-x c-backward-into-nomenclature
M-x c-forward-into-nomenclature
A popular programming style, especially for object-oriented languages such as C++ is to write symbols in a mixed case format, where the first letter of each word is capitalized, and not separated by underscores. E.g., ‘SymbolsWithMixedCaseAndNoUnderlines’.

These commands move backward or forward to the beginning of the next capitalized word. With prefix argument n, move n times. If n is negative, move in the opposite direction.

Note that these two commands have been superseded by subword-mode, which you should use instead. See Subword Movement. They might be removed from a future release of CC Mode.

Next: , Previous: Movement Commands, Up: Commands

4.4 Filling and Line Breaking Commands

Since there's a lot of normal text in comments and string literals, CC Mode provides features to edit these like in text mode. The goal is to do it seamlessly, i.e., you can use auto fill mode, sentence and paragraph movement, paragraph filling, adaptive filling etc. wherever there's a piece of normal text without having to think much about it. CC Mode keeps the indentation, fixes suitable comment line prefixes, and so on.

You can configure the exact way comments get filled and broken, and where Emacs does auto-filling (see see Custom Filling and Breaking). Typically, the style system (see Styles) will have set this up for you, so you probably won't have to bother.

Line breaks are by default handled (almost) the same regardless of whether they are made by auto fill mode (see Auto Fill), by paragraph filling (e.g., with M-q), or explicitly with M-j or similar methods. In string literals, the new line gets the same indentation as the previous nonempty line.6.

M-q (c-fill-paragraph)
This command fills multiline string literals and both block and line style comments. In Java buffers, the Javadoc markup words are recognized as paragraph starters. The line oriented Pike autodoc markup words are recognized in the same way in Pike mode.

The formatting of the starters (/*) and enders (*/) of block comments are kept as they were before the filling. I.e., if either the starter or ender were on a line of its own, then it stays on its own line; conversely, if the delimiter has comment text on its line, it keeps at least one word of that text with it on the line.

This command is the replacement for fill-paragraph in CC Mode buffers.

M-j (c-indent-new-comment-line)
This breaks the current line at point and indents the new line. If point was in a comment, the new line gets the proper comment line prefix. If point was inside a macro, a backslash is inserted before the line break. It is the replacement for indent-new-comment-line.
M-x c-context-line-break
Insert a line break suitable to the context: If the point is inside a comment, the new line gets the suitable indentation and comment line prefix like c-indent-new-comment-line. In normal code it's indented like newline-and-indent would do. In macros it acts like newline-and-indent but additionally inserts and optionally aligns the line ending backslash so that the macro remains unbroken. See Custom Macros, for details about the backslash alignment. In a string, a backslash is inserted only if the string is within a macro7.

This function is not bound to a key by default, but it's intended to be used on the RET key. If you like the behavior of newline-and-indent on RET, you should consider switching to this function. See Sample Init File.

M-x c-context-open-line
This is to C-o (M-x open-line) as c-context-line-break is to RET. I.e., it works just like c-context-line-break but leaves the point before the inserted line break.

Next: , Previous: Filling and Breaking, Up: Commands

4.5 Minor Modes

CC Mode contains several minor-mode-like features that you might find useful while writing new code or editing old code:

electric mode
When this is enabled, certain visible characters cause reformatting as they are typed. This is normally helpful, but can be a nuisance when editing chaotically formatted code. It can also be disconcerting, especially for users who are new to CC Mode.
auto-newline mode
This automatically inserts newlines where you'd probably want to type them yourself, e.g., after typing ‘}’s. Its action is suppressed when electric mode is disabled.
hungry-delete mode
This lets you delete a contiguous block of whitespace with a single key: for example, the newline and indentation just inserted by auto-newline when you want to back up and write a comment after the last statement.
subword mode
This mode makes basic word movement commands like M-f (forward-word) and M-b (backward-word) treat the parts of sillycapsed symbols as different words. E.g., ‘NSGraphicsContext’ is treated as three words ‘NS’, ‘Graphics’, and ‘Context’.
syntactic-indentation mode
When this is enabled (which it normally is), indentation commands such as C-j indent lines of code according to their syntactic structure. Otherwise, a line is simply indented to the same level as the previous one and <TAB> adjusts the indentation in steps of `c-basic-offset'.

Full details on how these minor modes work are at Electric Keys, Auto-newlines, Hungry WS Deletion, Subword Movement, and Indentation Engine Basics.

You can toggle each of these minor modes on and off, and you can configure CC Mode so that it starts up with your favorite combination of them (see Sample Init File). By default, when you initialize a buffer, electric mode and syntactic-indentation mode are enabled but the other three modes are disabled.

CC Mode displays the current state of the first four of these minor modes on the modeline by appending letters to the major mode's name, one letter for each enabled minor mode: ‘l’ for electric mode, ‘a’ for auto-newline mode, ‘h’ for hungry delete mode, and ‘w’ for subword mode. If all these modes were enabled, you'd see ‘C/lahw8.

Here are the commands to toggle these modes:

C-c C-l (c-toggle-electric-state)
Toggle electric minor mode. When the command turns the mode off, it also suppresses auto-newline mode.
C-c C-a (c-toggle-auto-newline)
Toggle auto-newline minor mode. When the command turns the mode on, it also enables electric minor mode.
M-x c-toggle-hungry-state9
Toggle hungry-delete minor mode.
M-x c-toggle-auto-hungry-state10
Toggle both auto-newline and hungry delete minor modes.
C-c C-w (M-x subword-mode)
Toggle subword mode.
M-x c-toggle-syntactic-indentation
Toggle syntactic-indentation mode.

Common to all the toggle functions above is that if they are called programmatically, they take an optional numerical argument. A positive value will turn on the minor mode (or both of them in the case of c-toggle-auto-hungry-state) and a negative value will turn it (or them) off.

Next: , Previous: Minor Modes, Up: Commands

4.6 Electric Keys and Keywords

Most punctuation keys provide electric behavior: as well as inserting themselves they perform some other action, such as reindenting the line. This reindentation saves you from having to reindent a line manually after typing, say, a ‘}’. A few keywords, such as else, also trigger electric action.

You can inhibit the electric behavior described here by disabling electric minor mode (see Minor Modes).

Common to all these keys is that they only behave electrically when used in normal code (as contrasted with getting typed in a string literal or comment). Those which cause re-indentation do so only when c-syntactic-indentation has a non-nil value (which it does by default).

These keys and keywords are:

#
Pound (bound to c-electric-pound) is electric when typed as the first non-whitespace character on a line and not within a macro definition. In this case, the variable c-electric-pound-behavior is consulted for the electric behavior. This variable takes a list value, although the only element currently defined is alignleft, which tells this command to force the ‘#’ character into column zero. This is useful for entering preprocessor macro definitions.

Pound is not electric in AWK buffers, where ‘#’ starts a comment, and is bound to self-insert-command like any typical printable character.

*
/
A star (bound to c-electric-star) or a slash (c-electric-slash) causes reindentation when you type it as the second component of a C style block comment opener (‘/*’) or a C++ line comment opener (‘//’) respectively, but only if the comment opener is the first thing on the line (i.e., there's only whitespace before it).

Additionally, you can configure CC Mode so that typing a slash at the start of a line within a block comment will terminate the comment. You don't need to have electric minor mode enabled to get this behavior. See Clean-ups.

In AWK mode, ‘*’ and ‘/’ do not delimit comments and are not electric.

<
>
A less-than or greater-than sign (bound to c-electric-lt-gt) is electric in two circumstances: when it is an angle bracket in a C++ ‘template’ declaration (and similar constructs in other languages) and when it is the second of two < or > characters in a C++ style stream operator. In either case, the line is reindented. Angle brackets in C ‘#include’ directives are not electric.
(
)
The normal parenthesis characters ‘(’ and ‘)’ (bound to c-electric-paren) reindent the current line. This is useful for getting the closing parenthesis of an argument list aligned automatically.

You can also configure CC Mode to insert a space automatically between a function name and the ‘(’ you've just typed, and to remove it automatically after typing ‘)’, should the argument list be empty. You don't need to have electric minor mode enabled to get these actions. See Clean-ups.

{
}
Typing a brace (bound to c-electric-brace) reindents the current line. Also, one or more newlines might be inserted if auto-newline minor mode is enabled. See Auto-newlines. Additionally, you can configure CC Mode to compact excess whitespace inserted by auto-newline mode in certain circumstances. See Clean-ups.
:
Typing a colon (bound to c-electric-colon) reindents the current line. Additionally, one or more newlines might be inserted if auto-newline minor mode is enabled. See Auto-newlines. If you type a second colon immediately after such an auto-newline, by default the whitespace between the two colons is removed, leaving a C++ scope operator. See Clean-ups.

If you prefer, you can insert ‘::’ in a single operation, avoiding all these spurious reindentations, newlines, and clean-ups. See Other Commands.

;
,
Typing a semicolon or comma (bound to c-electric-semi&comma) reindents the current line. Also, a newline might be inserted if auto-newline minor mode is enabled. See Auto-newlines. Additionally, you can configure CC Mode so that when auto-newline has inserted whitespace after a ‘}’, it will be removed again when you type a semicolon or comma just after it. See Clean-ups.
— Command: c-electric-continued-statement

Certain keywords are electric, causing reindentation when they are preceded only by whitespace on the line. The keywords are those that continue an earlier statement instead of starting a new one: else, while, catch (only in C++ and Java) and finally (only in Java).

An example:

          for (i = 0; i < 17; i++)
            if (a[i])
              res += a[i]->offset;
          else

Here, the else should be indented like the preceding if, since it continues that statement. CC Mode will automatically reindent it after the else has been typed in full, since only then is it possible to decide whether it's a new statement or a continuation of the preceding if.

CC Mode uses Abbrev mode (see Abbrevs) to accomplish this. It's therefore turned on by default in all language modes except IDL mode, since CORBA IDL doesn't have any statements.

Next: , Previous: Electric Keys, Up: Commands

4.7 Auto-newline Insertion

When you have Auto-newline minor mode enabled (see Minor Modes), CC Mode inserts newlines for you automatically (in certain syntactic contexts) when you type a left or right brace, a colon, a semicolon, or a comma. Sometimes a newline appears before the character you type, sometimes after it, sometimes both.

Auto-newline only triggers when the following conditions hold:

You can configure the precise circumstances in which newlines get inserted (see see Custom Auto-newlines). Typically, the style system (see Styles) will have set this up for you, so you probably won't have to bother.

Sometimes CC Mode inserts an auto-newline where you don't want one, such as after a ‘}’ when you're about to type a ‘;’. Hungry deletion can help here (see Hungry WS Deletion), or you can activate an appropriate clean-up, which will remove the excess whitespace after you've typed the ‘;’. See Clean-ups for a full description. See also Electric Keys for a summary of clean-ups listed by key.

Next: , Previous: Auto-newlines, Up: Commands

4.8 Hungry Deletion of Whitespace

If you want to delete an entire block of whitespace at point, you can use hungry deletion. This deletes all the contiguous whitespace either before point or after point in a single operation. “Whitespace” here includes tabs and newlines, but not comments or preprocessor commands. Hungry deletion can markedly cut down on the number of times you have to hit deletion keys when, for example, you've made a mistake on the preceding line and have already pressed C-j.

Hungry deletion is a simple feature that some people find extremely useful. In fact, you might find yourself wanting it in all your editing modes!

Loosely speaking, in what follows, <DEL> means “the backspace key” and <DELETE> means “the forward delete key”. This is discussed in more detail below.

There are two different ways you can use hungry deletion:

Using Hungry Delete Mode with <DEL> and C-d
Here you toggle Hungry Delete minor mode with M-x c-toggle-hungry-state12 (see Minor Modes.) This makes <DEL> and C-d do backwards and forward hungry deletion.
<DEL> (c-electric-backspace)
This command is run by default when you hit the DEL key. When hungry delete mode is enabled, it deletes any amount of whitespace in the backwards direction. Otherwise, or when used with a prefix argument or in a literal (see Auto-newlines), the command just deletes backwards in the usual way. (More precisely, it calls the function contained in the variable c-backspace-function, passing it the prefix argument, if any.)
c-backspace-function
Hook that gets called by c-electric-backspace when it doesn't do an “electric” deletion of the preceding whitespace. The default value is backward-delete-char-untabify (see Deletion, the function which deletes a single character.
C-d (c-electric-delete-forward)
This function, which is bound to C-d by default, works just like c-electric-backspace but in the forward direction. When it doesn't do an “electric” deletion of the following whitespace, it just does delete-char, more or less. (Strictly speaking, it calls the function in c-delete-function with the prefix argument.)
c-delete-function
Hook that gets called by c-electric-delete-forward when it doesn't do an “electric” deletion of the following whitespace. The default value is delete-char.

Using Distinct Bindings
The other (newer and recommended) way to use hungry deletion is to perform c-hungry-delete-backwards and c-hungry-delete-forward directly through their key sequences rather than using the minor mode toggling.
C-c C-<DEL>, or C-c <DEL> (c-hungry-delete-backwards)13
Delete any amount of whitespace in the backwards direction (regardless whether hungry-delete mode is enabled or not). This command is bound to both C-c C-<DEL> and C-c <DEL>, since the more natural one, C-c C-<DEL>, is sometimes difficult to type at a character terminal.
C-c C-d, C-c C-<DELETE>, or C-c <DELETE> (c-hungry-delete-forward)
Delete any amount of whitespace in the forward direction (regardless whether hungry-delete mode is enabled or not). This command is bound to both C-c C-<DELETE> and C-c <DELETE> for the same reason as for <DEL> above.

When we talk about <DEL>, and <DELETE> above, we actually do so without connecting them to the physical keys commonly known as <Backspace> and <Delete>. The default bindings to those two keys depends on the flavor of (X)Emacs you are using.

In XEmacs 20.3 and beyond, the <Backspace> key is bound to c-electric-backspace and the <Delete> key is bound to c-electric-delete. You control the direction it deletes in by setting the variable delete-key-deletes-forward, a standard XEmacs variable. When this variable is non-nil, c-electric-delete will do forward deletion with c-electric-delete-forward, otherwise it does backward deletion with c-electric-backspace. Similarly, C-c <Delete> and C-c C-<Delete> are bound to c-hungry-delete which is controlled in the same way by delete-key-deletes-forward.

Emacs 21 and later automatically binds <Backspace> and <Delete> to DEL and C-d according to your environment, and CC Mode extends those bindings to C-c C-<Backspace> etc. If you need to change the bindings through normal-erase-is-backspace-mode then CC Mode will also adapt its extended bindings accordingly.

In earlier (X)Emacs versions, CC Mode doesn't bind either <Backspace> or <Delete> directly. Only the key codes DEL and C-d are bound, and it's up to the default bindings to map the physical keys to them. You might need to modify this yourself if the defaults are unsuitable.

Getting your <Backspace> and <Delete> keys properly set up can sometimes be tricky. The information in DEL Does Not Delete, might be helpful if you're having trouble with this in GNU Emacs.

Next: , Previous: Hungry WS Deletion, Up: Commands

4.9 Subword Movement and Editing

In spite of the GNU Coding Standards, it is popular to name a symbol by mixing uppercase and lowercase letters, e.g., ‘GtkWidget’, ‘EmacsFrameClass’, or ‘NSGraphicsContext’. Here we call these mixed case symbols nomenclatures. Also, each capitalized (or completely uppercase) part of a nomenclature is called a subword. Here are some examples:

Nomenclature Subwords
————————————————————————————
GtkWindowGtk’ and ‘Window
EmacsFrameClassEmacs’, ‘Frame’, and ‘Class
NSGraphicsContextNS’, ‘Graphics’, and ‘Context

The subword minor mode replaces the basic word oriented movement and editing commands with variants that recognize subwords in a nomenclature and treat them as separate words:

Key Word oriented command Subword oriented command
—————————————————————————————————————–
M-f forward-word c-forward-subword
M-b backward-word c-backward-subword
M-@ mark-word c-mark-subword
M-d kill-word c-kill-subword
M-DEL backward-kill-word c-backward-kill-subword
M-t transpose-words c-transpose-subwords
M-c capitalize-word c-capitalize-subword
M-u upcase-word c-upcase-subword
M-l downcase-word c-downcase-subword

Note that if you have changed the key bindings for the word oriented commands in your .emacs or a similar place, the keys you have configured are also used for the corresponding subword oriented commands.

Type C-c C-w to toggle subword mode on and off. To make the mode turn on automatically, put the following code in your .emacs:

     (add-hook 'c-mode-common-hook
               (lambda () (subword-mode 1)))

As a bonus, you can also use subword-mode in non-CC Mode buffers by typing M-x subword-mode.

Previous: Subword Movement, Up: Commands

4.10 Other Commands

Here are the various other commands that didn't fit anywhere else:

C-c . (c-set-style)
Switch to the specified style in the current buffer. Use like this:
          C-c . style-name <RET>

You can use the <TAB> in the normal way to do completion on the style name. Note that all style names are case insensitive, even the ones you define yourself.

Setting a style in this way does not automatically reindent your file. For commands that you can use to view the effect of your changes, see Indentation Commands and Filling and Breaking.

For details of the CC Mode style system, see Styles.

C-c : (c-scope-operator)
In C++, it is also sometimes desirable to insert the double-colon scope operator without performing the electric behavior of colon insertion. C-c : does just this.
C-c C-\ (c-backslash-region)
This function inserts and aligns or deletes end-of-line backslashes in the current region. These are typically used in multi-line macros.

With no prefix argument, it inserts any missing backslashes and aligns them according to the c-backslash-column and c-backslash-max-column variables. With a prefix argument, it deletes any backslashes.

The function does not modify blank lines at the start of the region. If the region ends at the start of a line, it always deletes the backslash (if any) at the end of the previous line.

To customize the precise workings of this command, Custom Macros.

The recommended line breaking function, c-context-line-break (see Filling and Breaking), is especially nice if you edit multiline macros frequently. When used inside a macro, it automatically inserts and adjusts the mandatory backslash at the end of the line to keep the macro together, and it leaves the point at the right indentation column for the code. Thus you can write code inside macros almost exactly as you can elsewhere, without having to bother with the trailing backslashes.

C-c C-e (c-macro-expand)
This command expands C, C++, Objective C or Pike macros in the region, using an appropriate external preprocessor program. Normally it displays its output in a temporary buffer, but if you give it a prefix arg (with C-u C-c C-e) it will overwrite the original region with the expansion.

The command does not work in any of the other modes, and the key sequence is not bound in these other modes.

c-macro-expand isn't actually part of CC Mode, even though it is bound to a CC Mode key sequence. If you need help setting it up or have other problems with it, you can either read its source code or ask for help in the standard (X)Emacs forums.

Next: , Previous: Commands, Up: Top

5 Font Locking

CC Mode provides font locking for its supported languages by supplying patterns for use with Font Lock mode. This means that you get distinct faces on the various syntactic parts such as comments, strings, keywords and types, which is very helpful in telling them apart at a glance and discovering syntactic errors. See Font Lock, for ways to enable font locking in CC Mode buffers.

Please note: The font locking in AWK mode is currently not integrated with the rest of CC Mode. Only the last section of this chapter, AWK Mode Font Locking, applies to AWK. The other sections apply to the other languages.

Next: , Previous: Font Locking, Up: Font Locking

5.1 Font Locking Preliminaries

The font locking for most of the CC Mode languages were provided directly by the Font Lock package prior to version 5.30 of CC Mode. In the transition to CC Mode the patterns have been reworked completely and are applied uniformly across all the languages except AWK mode, just like the indentation rules (although each language still has some peculiarities of its own, of course). Since the languages previously had completely separate font locking patterns, this means that it's a bit different in most languages now.

The main goal for the font locking in CC Mode is accuracy, to provide a dependable aid in recognizing the various constructs. Some, like strings and comments, are easy to recognize while others, like declarations and types, can be very tricky. CC Mode can go to great lengths to recognize declarations and casts correctly, especially when the types aren't recognized by standard patterns. This is a fairly demanding analysis which can be slow on older hardware, and it can therefore be disabled by choosing a lower decoration level with the variable font-lock-maximum-decoration (see Font Lock).

The decoration levels are used as follows:

  1. Minimal font locking: Fontify only comments, strings and preprocessor directives (in the languages that use cpp).
  2. Fast font locking: In addition to level 1, fontify keywords, simple types and declarations that are easy to recognize. The variables *-font-lock-extra-types (where ‘*’ is the name of the language) are used to recognize types (see below). Documentation comments like Javadoc are fontified according to c-doc-comment-style (see Doc Comments).

    Use this if you think the font locking is too slow. It's the closest corresponding level to level 3 in the old font lock patterns.

  3. Accurate font locking: Like level 2 but uses a different approach that can recognize types and declarations much more accurately. The *-font-lock-extra-types variables are still used, but user defined types are recognized correctly anyway in most cases. Therefore those variables should be fairly restrictive and not contain patterns that are uncertain.

    This level is designed for fairly modern hardware and a font lock support mode like Lazy Lock or Just-in-time Lock mode that only fontifies the parts that are actually shown. Fontifying the whole buffer at once can easily get bothersomely slow even on contemporary hardware. See Font Lock.

Since user defined types are hard to recognize you can provide additional regexps to match those you use:

— User Option: c-font-lock-extra-types
— User Option: c++-font-lock-extra-types
— User Option: objc-font-lock-extra-types
— User Option: java-font-lock-extra-types
— User Option: idl-font-lock-extra-types
— User Option: pike-font-lock-extra-types

For each language there's a variable *-font-lock-extra-types, where ‘*’ stands for the language in question. It contains a list of regexps that matches identifiers that should be recognized as types, e.g., ‘\\sw+_t’ to recognize all identifiers ending with ‘_t’ as is customary in C code. Each regexp should not match more than a single identifier.

The default values contain regexps for many types in standard runtime libraries that are otherwise difficult to recognize, and patterns for standard type naming conventions like the ‘_t’ suffix in C and C++. Java, Objective-C and Pike have as a convention to start class names with capitals, so there are patterns for that in those languages.

Despite the names of these variables, they are not only used for fontification but in other places as well where CC Mode needs to recognize types.

Next: , Previous: Font Locking Preliminaries, Up: Font Locking

5.2 Faces

CC Mode attempts to use the standard faces for programming languages in accordance with their intended purposes as far as possible. No extra faces are currently provided, with the exception of a replacement face c-invalid-face for emacsen that don't provide font-lock-warning-face.

Next: , Previous: Faces, Up: Font Locking

5.3 Documentation Comments

There are various tools to supply documentation in the source as specially structured comments, e.g., the standard Javadoc tool in Java. CC Mode provides an extensible mechanism to fontify such comments and the special markup inside them.

— User Option: c-doc-comment-style

This is a style variable that specifies which documentation comment style to recognize, e.g., javadoc for Javadoc comments.

The value may also be a list of styles, in which case all of them are recognized simultaneously (presumably with markup cues that don't conflict).

The value may also be an association list to specify different comment styles for different languages. The symbol for the major mode is then looked up in the alist, and the value of that element is interpreted as above if found. If it isn't found then the symbol `other' is looked up and its value is used instead.

The default value for c-doc-comment-style is ((java-mode . javadoc) (pike-mode . autodoc) (c-mode . gtkdoc)).

Note that CC Mode uses this variable to set other variables that handle fontification etc. That's done at mode initialization or when you switch to a style which sets this variable. Thus, if you change it in some other way, e.g., interactively in a CC Mode buffer, you will need to do M-x java-mode (or whatever mode you're currently using) to reinitialize.

Note also that when CC Mode starts up, the other variables are modified before the mode hooks are run. If you change this variable in a mode hook, you'll have to call c-setup-doc-comment-style afterwards to redo that work.

CC Mode currently provides handing of the following doc comment styles:

javadoc
Javadoc comments, the standard tool in Java.
autodoc
For Pike autodoc markup, the standard in Pike.
gtkdoc
For GtkDoc markup, widely used in the Gnome community.

The above is by no means complete. If you'd like to see support for other doc comment styles, please let us know (see Mailing Lists and Bug Reports).

You can also write your own doc comment fontification support to use with c-doc-comment-style: Supply a variable or function *-font-lock-keywords where ‘*’ is the name you want to use in c-doc-comment-style. If it's a variable, it's prepended to font-lock-keywords. If it's a function, it's called at mode initialization and the result is prepended. For an example, see javadoc-font-lock-keywords in cc-fonts.el.

If you add support for another doc comment style, please consider contributing it: send a note to bug-cc-mode@gnu.org.

Previous: Doc Comments, Up: Font Locking

5.4 AWK Mode Font Locking

The general appearance of font-locking in AWK mode is much like in any other programming mode. See Faces for Font Lock.

The following faces are, however, used in a non-standard fashion in AWK mode:

font-lock-variable-name-face
This face was intended for variable declarations. Since variables are not declared in AWK, this face is used instead for AWK system variables (such as NF) and “Special File Names” (such as "/dev/stderr").
font-lock-builtin-face (Emacs)/font-lock-preprocessor-face (XEmacs)
This face is normally used for preprocessor directives in CC Mode. There are no such things in AWK, so this face is used instead for standard functions (such as match).
font-lock-string-face
As well as being used for strings, including localizable strings, (delimited by ‘"’ and ‘_"’), this face is also used for AWK regular expressions (delimited by ‘/’).
font-lock-warning-face (Emacs)/c-invalid-face (XEmacs)
This face highlights the following syntactically invalid AWK constructs:
  • An unterminated string or regular expression. Here the opening delimiter (‘"’ or ‘/’ or ‘_"’) is displayed in font-lock-warning-face. This is most noticeable when typing in a new string/regular expression into a buffer, when the warning-face serves as a continual reminder to terminate the construct.

    AWK mode fontifies unterminated strings/regular expressions differently from other modes: Only the text up to the end of the line is fontified as a string (escaped newlines being handled correctly), rather than the text up to the next string quote.

  • A space between the function name and opening parenthesis when calling a user function. The last character of the function name and the opening parenthesis are highlighted. This font-locking rule will spuriously highlight a valid concatenation expression where an identifier precedes a parenthesized expression. Unfortunately.
  • Whitespace following the ‘\’ in what otherwise looks like an escaped newline. The ‘\’ is highlighted.

Next: , Previous: Font Locking, Up: Top

6 Configuration Basics

You configure CC Mode by setting Lisp variables and calling (and perhaps writing) Lisp functions14, which is usually done by adding code to an Emacs initialization file. This file might be site-start.el or .emacs or init.el or default.el or perhaps some other file. See Init File. For the sake of conciseness, we just call this file “your .emacs” throughout the rest of the manual.

Several of these variables (currently 16), are known collectively as style variables. CC Mode provides a special mechanism, known as styles to make it easier to set these variables as a group, to “inherit” settings from one style into another, and so on. Style variables remain ordinary Lisp variables, whose values can be read and changed independently of the style system. See Style Variables.

There are several ways you can write the code, depending on the precise effect you want—they are described further down on this page. If you are new to CC Mode, we suggest you begin with the simplest method, “Top-level commands or the customization interface”.

If you make conflicting settings in several of these ways, the way that takes precedence is the one that appears latest in this list:

Here is a summary of the different ways of writing your configuration settings:

Top-level commands or the “customization interface”
Most simply, you can write setq and similar commands at the top level of your .emacs file. When you load a CC Mode buffer, it initializes its configuration from these global values (at least, for those settings you have given values to), so it makes sense to have these setq commands run before CC Mode is first initialized—in particular, before any call to desktop-read (see Saving Emacs Sessions). For example, you might set c-basic-offset thus:
          (setq c-basic-offset 4)

You can use the more user friendly Customization interface instead, but this manual does not cover in detail how that works. To do this, start by typing M-x customize-group <RET> c <RET>. See Easy Customization. Emacs normally writes the customizations at the end of your .emacs file. If you use desktop-read, you should edit your .emacs to place the call to desktop-read after the customizations.

The first initialization of CC Mode puts a snapshot of the configuration settings into the special style user. See Built-in Styles.

For basic use of Emacs, either of these ways of configuring is adequate. However, the settings are then the same in all CC Mode buffers and it can be clumsy to communicate them between programmers. For more flexibility, you'll want to use one (or both) of CC Mode's more sophisticated facilities, hooks and styles.

Hooks
An Emacs hook is a place to put Lisp functions that you want Emacs to execute later in specific circumstances. See Hooks. CC Mode supplies a main hook and a language-specific hook for each language it supports; any functions you put onto these hooks get executed as the last part of a buffer's initialization. Typically you put most of your customization within the main hook, and use the language-specific hooks to vary the customization settings between language modes. For example, if you wanted different (non-standard) values of c-basic-offset in C Mode and Java Mode buffers, you could do it like this:
          (defun my-c-mode-hook ()
            (setq c-basic-offset 3))
          (add-hook 'c-mode-hook 'my-c-mode-hook)
          
          (defun my-java-mode-hook ()
            (setq c-basic-offset 6))
          (add-hook 'java-mode-hook 'my-java-mode-hook)

See CC Hooks for more details on the use of CC Mode hooks.

Styles
A CC Mode style is a coherent collection of customizations with a name. At any time, exactly one style is active in each CC Mode buffer, either the one you have selected or a default. CC Mode is delivered with several existing styles. Additionally, you can create your own styles, possibly based on these existing styles. If you worked in a programming team called the “Free Group”, which had its own coding standards, you might well have this in your .emacs file:
          (setq c-default-style '((java-mode . "java")
                                  (awk-mode . "awk")
                                  (other . "free-group-style")))

See Styles for fuller details on using CC Mode styles and how to create them.

File Local Variable setting
A file local variable setting is a setting which applies to an individual source file. You put this in a local variables list, a special block at the end of the source file (see Specifying File Variables).
File Styles
A file style is a rarely used variant of the “style” mechanism described above, which applies to an individual source file. See File Styles. You use this by setting certain special variables in a local variables list (see Specifying File Variables).
Hooks with Styles
For ultimate flexibility, you can use hooks and styles together. For example, if your team were developing a product which required a Linux driver, you'd probably want to use the “linux” style for the driver, and your own team's style for the rest of the code. You could achieve this with code like this in your .emacs:
          (defun my-c-mode-hook ()
            (c-set-style
             (if (and (buffer-file-name)
                      (string-match "/usr/src/linux" (buffer-file-name)))
                 "linux"
               "free-group-style")))
          (add-hook 'c-mode-hook 'my-c-mode-hook)

In a programming team, a hook is a also a good place for each member to put his own personal preferences. For example, you might be the only person in your team who likes Auto-newline minor mode. You could have it enabled by default by placing the following in your .emacs:

          (defun my-turn-on-auto-newline ()
            (c-toggle-auto-newline 1))
          (add-hook 'c-mode-common-hook 'my-turn-on-auto-newline)

Next: , Previous: Config Basics, Up: Config Basics

6.1 Hooks

CC Mode provides several hooks that you can use to customize the mode for your coding style. The main hook is c-mode-common-hook; typically, you'll put the bulk of your customizations here. In addition, each language mode has its own hook, allowing you to fine tune your settings individually for the different CC Mode languages, and there is a package initialization hook. Finally, there is c-special-indent-hook, which enables you to solve anomalous indentation problems. It is described in Other Indentation, not here. All these hooks adhere to the standard Emacs conventions.

When you open a buffer, CC Mode first initializes it with the currently active style (see Styles). Then it calls c-mode-common-hook, and finally it calls the language-specific hook. Thus, any style settings done in these hooks will override those set by c-default-style.

— Variable: c-initialization-hook

Hook run only once per Emacs session, when CC Mode is initialized. This is a good place to change key bindings (or add new ones) in any of the CC Mode key maps. See Sample Init File.

— Variable: c-mode-common-hook

Common hook across all languages. It's run immediately before the language specific hook.

— Variable: c-mode-hook
— Variable: c++-mode-hook
— Variable: objc-mode-hook
— Variable: java-mode-hook
— Variable: idl-mode-hook
— Variable: pike-mode-hook
— Variable: awk-mode-hook

The language specific mode hooks. The appropriate one is run as the last thing when you enter that language mode.

Although these hooks are variables defined in CC Mode, you can give them values before CC Mode's code is loaded—indeed, this is the only way to use c-initialization-hook. Their values aren't overwritten when CC Mode gets loaded.

Here's a simplified example of what you can add to your .emacs file to do things whenever any CC Mode language is edited. See the Emacs manuals for more information on customizing Emacs via hooks. See Sample Init File, for a more complete sample .emacs file.

     (defun my-c-mode-common-hook ()
       ;; my customizations for all of c-mode and related modes
       (no-case-fold-search)
       )
     (add-hook 'c-mode-common-hook 'my-c-mode-common-hook)

Next: , Previous: CC Hooks, Up: Config Basics

6.2 Style Variables

The variables that CC Mode's style system control are called style variables. Note that style variables are ordinary Lisp variables, which the style system initializes; you can change their values at any time (e.g., in a hook function). The style system can also set other variables, to some extent. See Styles.

Style variables are handled specially in several ways:

The style variables are:
c-indent-comment-alist, c-indent-comments-syntactically-p (see Indentation Commands);
c-doc-comment-style (see Doc Comments);
c-block-comment-prefix, c-comment-prefix-regexp (see Custom Filling and Breaking);
c-hanging-braces-alist (see Hanging Braces);
c-hanging-colons-alist (see Hanging Colons);
c-hanging-semi&comma-criteria (see Hanging Semicolons and Commas);
c-cleanup-list (see Clean-ups);
c-basic-offset (see Customizing Indentation);
c-offsets-alist (see c-offsets-alist);
c-comment-only-line-offset (see Comment Line-Up);
c-special-indent-hook, c-label-minimum-indentation (see Other Indentation);
c-backslash-column, c-backslash-max-column (see Custom Macros).

Previous: Style Variables, Up: Config Basics

6.3 Styles

By style we mean the layout of the code—things like how many columns to indent a block of code, whether an opening brace gets indented to the level of the code it encloses, or of the construct that introduces it, or “hangs” at the end of a line.

Most people only need to edit code formatted in just a few well-defined and consistent styles. For example, their organization might impose a “blessed” style that all its programmers must conform to. Similarly, people who work on GNU software will have to use the GNU coding style. Some shops are more lenient, allowing a variety of coding styles, and as programmers come and go, there could be a number of styles in use. For this reason, CC Mode makes it convenient for you to set up logical groupings of customizations called styles, associate a single name for any particular style, and pretty easily start editing new or existing code using these styles.

As an alternative to writing a style definition yourself, you can have CC Mode guess (at least part of) your style by looking at an already formatted piece of your code, Guessing the Style.

Next: , Previous: Styles, Up: Styles

6.3.1 Built-in Styles

If you're lucky, one of CC Mode's built-in styles might be just what you're looking for. These are:

gnu
Coding style blessed by the Free Software Foundation for C code in GNU programs.
k&r
The classic Kernighan and Ritchie style for C code.
bsd
Also known as “Allman style” after Eric Allman.
whitesmith
Popularized by the examples that came with Whitesmiths C, an early commercial C compiler.
stroustrup
The classic Stroustrup style for C++ code.
ellemtel
Popular C++ coding standards as defined by “Programming in C++, Rules and Recommendations,” Erik Nyquist and Mats Henricson, Ellemtel18.
linux
C coding standard for Linux (the kernel).
python
C coding standard for Python extension modules19.
java
The style for editing Java code. Note that the default value for c-default-style installs this style when you enter java-mode.
awk
The style for editing AWK code. Note that the default value for c-default-style installs this style when you enter awk-mode.
user
This is a special style created by you. It consists of the factory defaults for all the style variables as modified by the customizations you do either with the Customization interface or by writing setqs and c-set-offsets at the top level of your .emacs file (see Config Basics). The style system creates this style as part of its initialization and doesn't modify it afterwards.

Next: , Previous: Built-in Styles, Up: Styles

6.3.2 Choosing a Style

When you create a new buffer, its style will be set from c-default-style. The factory default is the style gnu, except in Java and AWK modes where it's java and awk.

Remember that if you set a style variable with the Customization interface or at the top level of your .emacs file before the style system is initialized (see Config Basics), this setting will override the one that the style system would have given the variable.

To set a buffer's style interactively, use the command C-c . (see Other Commands). To set it from a file's local variable list, File Styles.

— User Option: c-default-style

This variable specifies which style to install by default in new buffers. It takes either a style name string, or an association list of major mode symbols to style names:

  1. When c-default-style is a string, it must be an existing style name. This style is then used for all modes.
  2. When c-default-style is an association list, the mode language is looked up to find a style name string.
  3. If c-default-style is an association list where the mode language mode isn't found then the special symbol ‘other’ is looked up. If it's found then the associated style is used.
  4. If ‘other’ is not found then the ‘gnu’ style is used.

In all cases, the style described in c-default-style is installed before the language hooks are run, so you can always override this setting by including an explicit call to c-set-style in your language mode hook, or in c-mode-common-hook.

The standard value of c-default-style is ((java-mode . "java") (awk-mode . "awk") (other . "gnu")).

— Variable: c-indentation-style

This variable always contains the buffer's current style name, as a string.

Next: , Previous: Choosing a Style, Up: Styles

6.3.3 Adding and Amending Styles

If none of the built-in styles is appropriate, you'll probably want to create a new style definition, possibly based on an existing style. To do this, put the new style's settings into a list with the following format; the list can then be passed as an argument to the function c-add-style. You can see an example of a style definition in Sample Init File.

Structure of a Style Definition List
([base-style] [(variable . value) ...])

Optional base-style, if present, must be a string which is the name of the base style from which this style inherits. At most one base-style is allowed in a style definition. If base-style is not specified, the style inherits from the table of factory default values20 instead. All styles eventually inherit from this internal table. Style loops generate errors. The list of pre-existing styles can be seen in Built-in Styles.

The dotted pairs (variable . value) each consist of a variable and the value it is to be set to when the style is later activated.21 The variable can be either a CC Mode style variable or an arbitrary Emacs variable. In the latter case, it is not made buffer-local by the CC Mode style system.

Two variables are treated specially in the dotted pair list:

c-offsets-alist
The value is in turn a list of dotted pairs of the form
               (syntactic-symbol . offset)

as described in c-offsets-alist. These are passed to c-set-offset so there is no need to set every syntactic symbol in your style, only those that are different from the inherited style.

c-special-indent-hook
The value is added to c-special-indent-hook using add-hook, so any functions already on it are kept. If the value is a list, each element of the list is added with add-hook.

Styles are kept in the c-style-alist variable, but you should never modify this variable directly. Instead, CC Mode provides the function c-add-style for this purpose.

— Function: c-add-style stylename description &optional set-p

Add or update a style called stylename, a string. description is the new style definition in the form described above. If stylename already exists in c-style-alist then it is replaced by description. (Note, this replacement is total. The old style is not merged into the new one.) Otherwise, a new style is added.

If the optional set-p is non-nil then the new style is applied to the current buffer as well. The use of this facility is deprecated and it might be removed from CC Mode in a future release. You should use c-set-style instead.

The sample .emacs file provides a concrete example of how a new style can be added and automatically set. See Sample Init File.

— Variable: c-style-alist

This is the variable that holds the definitions for the styles. It should not be changed directly; use c-add-style instead.

Next: , Previous: Adding Styles, Up: Styles

6.3.4 Guessing the Style

Instead of specifying a style, you can get CC Mode to guess your style by examining an already formatted code buffer. CC Mode then determines the ”most frequent” offset (see c-offsets-alist) for each of the syntactic symbols (see Indentation Engine Basics) encountered in the buffer, and the ”most frequent” value of c-basic-offset (see Customizing Indentation), then merges the current style with these ”guesses” to form a new style. This combined style is known as the guessed style.

To do this, call c-guess (or one of the other 5 guessing commands) on your sample buffer. The analysis of your code may take some time.

You can then set the guessed style in any CC Mode buffer with c-guess-install. You can display the style with c-guess-view, and preserve it by copying it into your .emacs for future use, preferably after editing it.

M-x c-guess-no-install
M-x c-guess-buffer-no-install
M-x c-guess-region-no-install
These commands analyze a part of the current buffer and guess the style from it.

The part of the buffer examined is either the region (c-guess-region-no-install), the entire buffer (c-guess-buffer-no-install), or the first c-guess-region-max bytes (c-guess-no-install).

Each of these commands can be given an optional prefix argument. This instructs CC Mode to combine the new guesses with the current guesses before forming the guessed style.

M-x c-guess
M-x c-guess-buffer
M-x c-guess-region
These commands analyze a part of the current buffer, guess the style from it, then install the guessed style on the buffer. The guessed style is given a name based on the buffer's absolute file name, and you can then set this style on any CC Mode buffer with C-c ..

The part of the buffer examined is either the region (c-guess-region), the entire buffer (c-guess-buffer), or the first c-guess-region-max bytes (c-guess).

Each of these commands can be given an optional prefix argument. This instructs CC Mode to combine the new guesses with the current guesses before forming the guessed style.

— User Option: c-guess-region-max

This variable, default 50000, is the size in bytes of the buffer portion examined by c-guess and c-guess-no-install. If set to nil, the entire buffer is examined.

— User Option: c-guess-offset-threshold

This variable, default 10, is the maximum offset, either outwards or inwards, which will be taken into account by the analysis process. Any offset bigger than this will be ignored. For no limit, set this variable to a large number.

M-x c-guess-install
Set the current buffer's style to the guessed style. This prompts you to enter an optional new style name to give to the guessed style. By default, this name is based on the buffer's absolute file name. You can then use this style like any other.
M-x c-guess-view
Display the most recently guessed style in a temporary buffer. This display is in the form of a c-add-style form (see Adding Styles) which can be easily copied to your .emacs. You will probably want to edit it first.

The display of the guessed style contains these elements:

Placeholder Name
You should replace this with a style name of your own.
Parent Style
The style current when the guessing began, from which the guessed style inherits (see Config Basics) the settings which weren't guessed.
Guessed Offsets
These are the core result of the guessing process. Each of them is marked by a comment.
Inherited Offsets
These are syntactic offsets which have been taken over from the parent style. To avoid possible future conflicts, you should remove either these offsets or the parent style name.

Previous: Guessing the Style, Up: Styles

6.3.5 File Styles

The Emacs manual describes how you can customize certain variables on a per-file basis by including a file local variable block at the end of the file (see Local Variables in Files).

So far, you've only seen a functional interface for setting styles in CC Mode, and this can't be used here. CC Mode fills the gap by providing two variables for use in a file's local variable list. Don't use them anywhere else! These allow you to customize the style on a per-file basis:

— Variable: c-file-style

Set this variable to a style name string in the Local Variables list. From now on, when you visit the file, CC Mode will automatically set the file's style to this one using c-set-style.

— Variable: c-file-offsets

Set this variable (in the Local Variables list) to an association list of the same format as c-offsets-alist. From now on, when you visit the file, CC Mode will automatically institute these offsets using c-set-offset.

Note that file style settings (i.e., c-file-style) are applied before file offset settings (i.e., c-file-offsets)22.

If you set any variable by the file local variables mechanism, that setting takes priority over all other settings, even those in your mode hooks (see CC Hooks). Any individual setting of a variable will override one made through c-file-style or c-file-offsets.

Next: , Previous: Config Basics, Up: Top

7 Customizing Filling and Line Breaking

Since there's a lot of normal text in comments and string literals, CC Mode provides features to edit these like in text mode. It does this by hooking in on the different line breaking functions and tuning relevant variables as necessary.

To make Emacs recognize comments and treat text in them as normal paragraphs, CC Mode makes several standard variables23 buffer-local and modifies them according to the language syntax and the comment line prefix.

— User Option: c-comment-prefix-regexp

This style variable contains the regexp used to recognize the comment line prefix, which is the line decoration that starts every line in a comment. The variable is either the comment line prefix itself, or (more usually) an association list with different values for different languages. The symbol for the major mode is looked up in the alist to get the regexp for the language, and if it isn't found then the special symbol ‘other’ is looked up instead.

When a comment line gets divided by M-j or the like, CC Mode inserts the comment line prefix from a neighboring line at the start of the new line. The default value of c-comment-prefix-regexp is ‘//+\\|\\**’, which matches C++ style line comments like

          // blah blah

with two or more slashes in front of them, and the second and subsequent lines of C style block comments like

          /*
           * blah blah
           */

with zero or more stars at the beginning of every line. If you change this variable, please make sure it still matches the comment starter (i.e., //) of line comments and the line prefix inside block comments.

Also note that since CC Mode uses the value of c-comment-prefix-regexp to set up several other variables at mode initialization, there won't be any effect if you just change it inside a CC Mode buffer. You need to call the command c-setup-paragraph-variables too, to update those other variables. That's also the case if you modify c-comment-prefix-regexp in a mode hook, since CC Mode will already have set up these variables before calling the hook.

In comments, CC Mode uses c-comment-prefix-regexp to adapt the line prefix from the other lines in the comment.

CC Mode uses adaptive fill mode (see Adaptive Fill) to make Emacs correctly keep the line prefix when filling paragraphs. That also makes Emacs preserve the text indentation inside the comment line prefix. E.g., in the following comment, both paragraphs will be filled with the left margins of the texts kept intact:

     /* Make a balanced b-tree of the nodes in the incoming
      * stream.  But, to quote the famous words of Donald E.
      * Knuth,
      *
      *     Beware of bugs in the above code; I have only
      *     proved it correct, not tried it.
      */

It's also possible to use other adaptive filling packages, notably Kyle E. Jones' Filladapt package24, which handles things like bulleted lists nicely. There's a convenience function c-setup-filladapt that tunes the relevant variables in Filladapt for use in CC Mode. Call it from a mode hook, e.g., with something like this in your .emacs:

     (defun my-c-mode-common-hook ()
       (c-setup-filladapt)
       (filladapt-mode 1))
     (add-hook 'c-mode-common-hook 'my-c-mode-common-hook)
— User Option: c-block-comment-prefix

Normally the comment line prefix inserted for a new line inside a comment is deduced from other lines in it. However there's one situation when there's no hint about what the prefix should look like, namely when a block comment is broken for the first time. This style variable25 is used then as the comment prefix. It defaults to ‘* 26, which makes a comment

          /* Got O(n^2) here, which is a Bad Thing. */

break into

          /* Got O(n^2) here, which
           * is a Bad Thing. */

Note that it won't work to adjust the indentation by putting leading spaces in c-block-comment-prefix, since CC Mode still uses the normal indentation engine to indent the line. Thus, the right way to fix the indentation is by customizing the c syntactic symbol. It defaults to c-lineup-C-comments, which handles the indentation of most common comment styles, see Line-Up Functions.

— User Option: c-ignore-auto-fill

When auto fill mode is enabled, CC Mode can selectively ignore it depending on the context the line break would occur in, e.g., to never break a line automatically inside a string literal. This variable takes a list of symbols for the different contexts where auto-filling never should occur:

string
Inside a string or character literal.
c
Inside a C style block comment.
c++
Inside a C++ style line comment.
cpp
Inside a preprocessor directive.
code
Anywhere else, i.e., in normal code.

By default, c-ignore-auto-fill is set to (string cpp code), which means that when auto-fill mode is activated, auto-filling only occurs in comments. In literals, it's often desirable to have explicit control over newlines. In preprocessor directives, the necessary ‘\’ escape character before the newline is not automatically inserted, so an automatic line break would produce invalid code. In normal code, line breaks are normally dictated by some logical structure in the code rather than the last whitespace character, so automatic line breaks there will produce poor results in the current implementation.

If inside a comment and comment-multi-line (see Auto Fill is non-nil, the indentation and line prefix are preserved. If inside a comment and comment-multi-line is nil, a new comment of the same type is started on the next line and indented as appropriate for comments.

Note that CC Mode sets comment-multi-line to t at startup. The reason is that M-j could otherwise produce sequences of single line block comments for texts that should logically be treated as one comment, and the rest of the paragraph handling code (e.g., M-q and M-a) can't cope with that, which would lead to inconsistent behavior.

Next: , Previous: Custom Filling and Breaking, Up: Top

8 Customizing Auto-newlines

CC Mode determines whether to insert auto-newlines in two basically different ways, depending on the character just typed:

Braces and Colons
CC Mode first determines the syntactic context of the brace or colon (see Syntactic Symbols), then looks for a corresponding element in an alist. This element specifies where to put newlines: this is any combination of before and after the brace or colon. If no alist element is found, newlines are inserted both before and after a brace, but none are inserted around a colon. See Hanging Braces and Hanging Colons.
Semicolons and Commas
The variable c-hanging-semi&comma-criteria contains a list of functions which determine whether to insert a newline after a newly typed semicolon or comma. See Hanging Semicolons and Commas.

The names of these configuration variables contain ‘hanging’ because they let you hang the pertinent characters. A character which introduces a C construct is said to hang on the right when it appears at the end of a line after other code, being separated by a line break from the construct it introduces, like the opening brace in:

     while (i < MAX) {
         total += entry[i];
         entry [i++] = 0;
     }

A character hangs on the left when it appears at the start of the line after the construct it closes off, like the above closing brace.

The next chapter, “Clean-ups”, describes how to configure CC Mode to remove these automatically added newlines in certain specific circumstances. See Clean-ups.

Next: , Previous: Custom Auto-newlines, Up: Custom Auto-newlines

8.1 Hanging Braces

To specify which kinds of braces you want auto-newlines put around, you set the style variable c-hanging-braces-alist. Its structure and semantics are described in this section. Details of how to set it up, and its relationship to CC Mode's style system are given in Style Variables.

Say you wanted an auto-newline after (but not before) the following ‘{’:

     if (foo < 17) {

First you need to find the syntactic context of the brace—type a <RET> before the brace to get it on a line of its own27, then type C-c C-s. That will tell you something like:

     ((substatement-open 1061))

So here you need to put the entry (substatement-open . (after)) into c-hanging-braces-alist.

If you don't want any auto-newlines for a particular syntactic symbol, put this into c-hanging-braces-alist:

     (brace-entry-open)

If some brace syntactic symbol is not in c-hanging-brace-alist, its entry is taken by default as (before after)—insert a newline both before and after the brace. In place of a “before/after” list you can specify a function in this alist—this is useful when the auto newlines depend on the code around the brace.

— User Option: c-hanging-braces-alist

This variable is an association list which maps syntactic symbols to lists of places to insert a newline. See Association Lists. The key of each element is the syntactic symbol, the associated value is either nil, a list, or a function.

The Key: the syntactic symbol
The syntactic symbols that are useful as keys in this list are brace-list-intro, statement-cont, inexpr-class-open, inexpr-class-close, and all the *-open and *-close symbols. See Syntactic Symbols, for a more detailed description of these syntactic symbols, except for inexpr-class-open and inexpr-class-close, which aren't actual syntactic symbols. Elements with any other value as a key get ignored.

The braces of anonymous inner classes in Java are given the special symbols inexpr-class-open and inexpr-class-close, so that they can be distinguished from the braces of normal classes28.

Note that the aggregate constructs in Pike mode, ‘({’, ‘})’, ‘([’, ‘])’, and ‘(<’, ‘>)’, do not count as brace lists in this regard, even though they do for normal indentation purposes. It's currently not possible to set automatic newlines on these constructs.

The associated value: the “ACTION” list or function
The value associated with each syntactic symbol in this association list is called an action, which can be either a list or a function which returns a list. See Custom Braces, for how to use a function as a brace hanging action.

The list action (or the list returned by action when it's a function) contains some combination of the symbols before and after, directing CC Mode where to put newlines in relationship to the brace being inserted. Thus, if the list contains only the symbol after, then the brace hangs on the right side of the line, as in:

               // here, open braces always `hang'
               void spam( int i ) {
                   if( i == 7 ) {
                       dosomething(i);
                   }
               }

When the list contains both after and before, the braces will appear on a line by themselves, as shown by the close braces in the above example. The list can also be empty, in which case newlines are added neither before nor after the brace.

If a syntactic symbol is missing entirely from c-hanging-braces-alist, it's treated in the same way as an action with a list containing before and after, so that braces by default end up on their own line.

For example, the default value of c-hanging-braces-alist is:

          ((brace-list-open)
           (brace-entry-open)
           (statement-cont)
           (substatement-open after)
           (block-close . c-snug-do-while)
           (extern-lang-open after)
           (namespace-open after)
           (module-open after)
           (composition-open after)
           (inexpr-class-open after)
           (inexpr-class-close before))

which says that brace-list-open, brace-entry-open and statement-cont29 braces should both hang on the right side and allow subsequent text to follow on the same line as the brace. Also, substatement-open, extern-lang-open, and inexpr-class-open braces should hang on the right side, but subsequent text should follow on the next line. The opposite holds for inexpr-class-close braces; they won't hang, but the following text continues on the same line. Here, in the block-close entry, you also see an example of using a function as an action. In all other cases, braces are put on a line by themselves.

Previous: Hanging Braces, Up: Hanging Braces

8.1.1 Custom Brace Hanging

Syntactic symbols aren't the only place where you can customize CC Mode with the lisp equivalent of callback functions. Remember that actions are usually a list containing some combination of the symbols before and after (see Hanging Braces). For more flexibility, you can instead specify brace “hanginess” by giving a syntactic symbol an action function in c-hanging-braces-alist; this function determines the “hanginess” of a brace, usually by looking at the code near it.

An action function is called with two arguments: the syntactic symbol for the brace (e.g., substatement-open), and the buffer position where the brace has been inserted. Point is undefined on entry to an action function, but the function must preserve it (e.g., by using save-excursion). The return value should be a list containing some combination of before and after, including neither of them (i.e., nil).

— Variable: c-syntactic-context

During the call to the indentation or brace hanging action function, this variable is bound to the full syntactic analysis list. This might be, for example, ‘((block-close 73))’. Don't ever give c-syntactic-context a value yourself—this would disrupt the proper functioning of CC Mode.

This variable is also bound in three other circumstances: (i) when calling a c-hanging-semi&comma-criteria function (see Hanging Semicolons and Commas); (ii) when calling a line-up function (see Custom Line-Up); (iii) when calling a c-special-indent-hook function (see Other Indentation).

As an example, CC Mode itself uses this feature to dynamically determine the hanginess of braces which close “do-while” constructs:

     void do_list( int count, char** atleast_one_string )
     {
         int i=0;
         do {
             handle_string( atleast_one_string[i] );
             i++;
         } while( i < count );
     }

CC Mode assigns the block-close syntactic symbol to the brace that closes the do construct, and normally we'd like the line that follows a block-close brace to begin on a separate line. However, with “do-while” constructs, we want the while clause to follow the closing brace. To do this, we associate the block-close symbol with the action function c-snug-do-while:

     (defun c-snug-do-while (syntax pos)
       "Dynamically calculate brace hanginess for do-while statements."
       (save-excursion
         (let (langelem)
           (if (and (eq syntax 'block-close)
                    (setq langelem (assq 'block-close c-syntactic-context))
                    (progn (goto-char (cdr langelem))
                           (if (= (following-char) ?{)
                               (forward-sexp -1))
                           (looking-at "\\<do\\>[^_]")))
               '(before)
             '(before after)))))

This function simply looks to see if the brace closes a “do-while” clause and if so, returns the list ‘(before)’ indicating that a newline should be inserted before the brace, but not after it. In all other cases, it returns the list ‘(before after)’ so that the brace appears on a line by itself.

Next: , Previous: Hanging Braces, Up: Custom Auto-newlines

8.2 Hanging Colons

Using a mechanism similar to brace hanging (see Hanging Braces), colons can also be made to hang using the style variable c-hanging-colons-alist: when a colon is typed, CC Mode determines its syntactic context, looks this up in the alist c-changing-colons-alist and inserts up to two newlines accordingly. Here, however, If CC Mode fails to find an entry for a syntactic symbol in the alist, no newlines are inserted around the newly typed colon.

— User Option: c-hanging-colons-alist

The Key: the syntactic symbol
The syntactic symbols appropriate as keys in this association list are: case-label, label, access-label, member-init-intro, and inher-intro. See Syntactic Symbols. Elements with any other value as a key get ignored.
The associated value: the “ACTION” list
The action here is simply a list containing a combination of the symbols before and after. Unlike in c-hanging-braces-alist, functions as actions are not supported; there doesn't seem to be any need for them.

In C++, double-colons are used as a scope operator but because these colons always appear right next to each other, newlines before and after them are controlled by a different mechanism, called clean-ups in CC Mode. See Clean-ups, for details.

Previous: Hanging Colons, Up: Custom Auto-newlines

8.3 Hanging Semicolons and Commas

— User Option: c-hanging-semi&comma-criteria

This style variable takes a list of functions; these get called when you type a semicolon or comma. The functions are called in order without arguments. When these functions are entered, point is just after the newly inserted ‘;’ or ‘,’ and they must preserve point (e.g., by using save-excursion). During the call, the variable c-syntactic-context is bound to the syntactic context of the current line30 see Custom Braces. These functions don't insert newlines themselves, rather they direct CC Mode whether or not to do so. They should return one of the following values:

t
A newline is to be inserted after the ‘;’ or ‘,’, and no more functions from the list are to be called.
stop
No more functions from the list are to be called, and no newline is to be inserted.
nil
No determination has been made, and the next function in the list is to be called.

Note that auto-newlines are never inserted before a semicolon or comma. If every function in the list is called without a determination being made, then no newline is added.

In AWK mode, this variable is set by default to nil. In the other modes, the default value is a list containing a single function, c-semi&comma-inside-parenlist. This inserts newlines after all semicolons, apart from those separating for-clause statements.

— Function: c-semi&comma-no-newlines-before-nonblanks

This is an example of a criteria function, provided by CC Mode. It prevents newlines from being inserted after semicolons when there is a non-blank following line. Otherwise, it makes no determination. To use, add this function to the front of the c-hanging-semi&comma-criteria list.

          (defun c-semi&comma-no-newlines-before-nonblanks ()
            (save-excursion
              (if (and (eq last-command-char ?\;)
                       (zerop (forward-line 1))
                       (not (looking-at "^[ \t]*$")))
                  'stop
                nil)))
— Function: c-semi&comma-inside-parenlist

— Function: c-semi&comma-no-newlines-for-oneline-inliners

The function c-semi&comma-inside-parenlist is what prevents newlines from being inserted inside the parenthesis list of for statements. In addition to c-semi&comma-no-newlines-before-nonblanks described above, CC Mode also comes with the criteria function c-semi&comma-no-newlines-for-oneline-inliners, which suppresses newlines after semicolons inside one-line inline method definitions (e.g., in C++ or Java).

Next: , Previous: Custom Auto-newlines, Up: Top

9 Clean-ups

Clean-ups are mechanisms which remove (or exceptionally, add) whitespace in specific circumstances and are complementary to colon and brace hanging. You enable a clean-up by adding its symbol into c-cleanup-list, e.g., like this:

     (add-to-list 'c-cleanup-list 'space-before-funcall)

On the surface, it would seem that clean-ups overlap the functionality provided by the c-hanging-*-alist variables. Clean-ups, however, are used to adjust code “after-the-fact”, i.e., to adjust the whitespace in constructs later than when they were typed.

Most of the clean-ups remove automatically inserted newlines, and are only active when auto-newline minor mode is turned on. Others will work all the time. Note that clean-ups are only performed when there is nothing but whitespace appearing between the individual components of the construct, and (apart from comment-close-slash) when the construct does not occur within a literal (see Auto-newlines).

— User Option: c-cleanup-list

You configure CC Mode's clean-ups by setting the style variable c-cleanup-list, which is a list of clean-up symbols. By default, CC Mode cleans up only the scope-operator construct, which is necessary for proper C++ support.

These are the clean-ups that are only active when electric and auto-newline minor modes are enabled:

brace-else-brace
Clean up ‘} else {’ constructs by placing the entire construct on a single line. Clean up occurs when the open brace after the ‘else’ is typed. So for example, this:
          void spam(int i)
          {
              if( i==7 ) {
                  dosomething();
              }
              else
              {

appears like this after the last open brace is typed:

          void spam(int i)
          {
              if( i==7 ) {
                  dosomething();
              } else {

brace-elseif-brace
Similar to the brace-else-brace clean-up, but this cleans up ‘} else if (...) {’ constructs. For example:
          void spam(int i)
          {
              if( i==7 ) {
                  dosomething();
              }
              else if( i==3 )
              {

appears like this after the last open parenthesis is typed:

          void spam(int i)
          {
              if( i==7 ) {
                  dosomething();
              } else if(

and like this after the last open brace is typed:

          void spam(int i)
          {
              if( i==7 ) {
                  dosomething();
              } else if( i==3 ) {

brace-catch-brace
Analogous to brace-elseif-brace, but cleans up ‘} catch (...) {’ in C++ and Java mode.
empty-defun-braces
Clean up braces following a top-level function or class definition that contains no body. Clean up occurs when the closing brace is typed. Thus the following:
          class Spam
          {
          }

is transformed into this when the close brace is typed:

          class Spam
          {}

defun-close-semi
Clean up the terminating semicolon on top-level function or class definitions when they follow a close brace. Clean up occurs when the semicolon is typed. So for example, the following:
          class Spam
          {
          ...
          }
          ;

is transformed into this when the semicolon is typed:

          class Spam
          {
          ...
          };

list-close-comma
Clean up commas following braces in array and aggregate initializers. Clean up occurs when the comma is typed. The space before the comma is zapped just like the space before the semicolon in defun-close-semi.
scope-operator
Clean up double colons which might designate a C++ scope operator split across multiple lines31. Clean up occurs when the second colon is typed. You will always want scope-operator in the c-cleanup-list when you are editing C++ code.
one-liner-defun
Clean up a single line of code enclosed by defun braces by removing the whitespace before and after the code. The clean-up happens when the closing brace is typed. If the variable c-max-one-liner-length is set, the cleanup is only done if the resulting line would be no longer than the value of that variable.

For example, consider this AWK code:

          BEGIN {
              FS = "\t" # use <TAB> as a field separator
          }

It gets compacted to the following when the closing brace is typed:

          BEGIN {FS = "\t"} # use <TAB> as a field separator
— User Option: c-max-one-liner-length

The maximum length of the resulting line for which the clean-up one-liner-defun will be triggered. This length is that of the entire line, including any leading whitespace and any trailing comment. Its default value is 80. If the value is zero or nil, no limit applies.

The following clean-ups are always active when they occur on c-cleanup-list, regardless of whether Electric minor mode or Auto-newline minor mode are enabled:

space-before-funcall
Insert a space between the function name and the opening parenthesis of a function call. This produces function calls in the style mandated by the GNU coding standards, e.g., ‘signal (SIGINT, SIG_IGN)’ and ‘abort ()’. Clean up occurs when the opening parenthesis is typed. This clean-up should never be active in AWK Mode, since such a space is syntactically invalid for user defined functions.
compact-empty-funcall
Clean up any space between the function name and the opening parenthesis of a function call that has no arguments. This is typically used together with space-before-funcall if you prefer the GNU function call style for functions with arguments but think it looks ugly when it's only an empty parenthesis pair. I.e., you will get ‘signal (SIGINT, SIG_IGN)’, but ‘abort()’. Clean up occurs when the closing parenthesis is typed.
comment-close-slash
When inside a block comment, terminate the comment when you type a slash at the beginning of a line (i.e., immediately after the comment prefix). This clean-up removes whitespace preceding the slash and if needed, inserts a star to complete the token ‘*/’. Type C-q / in this situation if you just want a literal ‘/’ inserted.

Next: , Previous: Clean-ups, Up: Top

10 Indentation Engine Basics

This chapter will briefly cover how CC Mode indents lines of code. It is helpful to understand the indentation model being used so that you will know how to customize CC Mode for your personal coding style. All the details are in Customizing Indentation.

CC Mode has an indentation engine that provides a flexible and general mechanism for customizing indentation. When CC Mode indents a line of code, it separates its calculations into two steps:

  1. It analyzes the line to determine its syntactic symbol(s) (the kind of language construct it's looking at) and its anchor position (the position earlier in the file that CC Mode will indent the line relative to). The anchor position might be the location of an opening brace in the previous line, for example. See Syntactic Analysis.
  2. It looks up the syntactic symbol(s) in the configuration to get the corresponding offset(s). The symbol +, which means “indent this line one more level” is a typical offset. CC Mode then applies these offset(s) to the anchor position, giving the indentation for the line. The different sorts of offsets are described in c-offsets-alist.

In exceptional circumstances, the syntax directed indentation described here may be a nuisance rather than a help. You can disable it by setting c-syntactic-indentation to nil. (To set the variable interactively, Minor Modes).

— User Option: c-syntactic-indentation

When this is non-nil (which it is by default), the indentation of code is done according to its syntactic structure. When it's nil, every line is just indented to the same level as the previous one, and TAB (c-indent-command) adjusts the indentation in steps of c-basic-offset. The current style (see Config Basics) then has no effect on indentation, nor do any of the variables associated with indentation, not even c-special-indent-hook.

Next: , Previous: Indentation Engine Basics, Up: Indentation Engine Basics

10.1 Syntactic Analysis

The first thing CC Mode does when indenting a line of code, is to analyze the line, determining the syntactic context of the (first) construct on that line. It's a list of syntactic elements, where each syntactic element in turn is a list32 Here is a brief and typical example:

     ((defun-block-intro 1959))

The first thing inside each syntactic element is always a syntactic symbol. It describes the kind of construct that was recognized, e.g., statement, substatement, class-open, class-close, etc. See Syntactic Symbols, for a complete list of currently recognized syntactic symbols and their semantics. The remaining entries are various data associated with the recognized construct; there might be zero or more.

Conceptually, a line of code is always indented relative to some position higher up in the buffer (typically the indentation of the previous line). That position is the anchor position in the syntactic element. If there is an entry after the syntactic symbol in the syntactic element list then it's either nil or that anchor position.

Here is an example. Suppose we had the following code as the only thing in a C++ buffer 33:

      1: void swap( int& a, int& b )
      2: {
      3:     int tmp = a;
      4:     a = b;
      5:     b = tmp;
      6: }

We can use C-c C-s (c-show-syntactic-information) to report what the syntactic analysis is for the current line:

C-c C-s (c-show-syntactic-information)
This command calculates the syntactic analysis of the current line and displays it in the minibuffer. The command also highlights the anchor position(s).

Running this command on line 4 of this example, we'd see in the echo area34:

     ((statement 35))

and the ‘i’ of int on line 3 would be highlighted. This tells us that the line is a statement and it is indented relative to buffer position 35, the highlighted position. If you were to move point to line 3 and hit C-c C-s, you would see:

     ((defun-block-intro 29))

This indicates that the ‘int’ line is the first statement in a top level function block, and is indented relative to buffer position 29, which is the brace just after the function header.

Here's another example:

      1: int add( int val, int incr, int doit )
      2: {
      3:     if( doit )
      4:         {
      5:             return( val + incr );
      6:         }
      7:     return( val );
      8: }

Hitting C-c C-s on line 4 gives us:

     ((substatement-open 46))

which tells us that this is a brace that opens a substatement block. 35

Syntactic contexts can contain more than one element, and syntactic elements need not have anchor positions. The most common example of this is a comment-only line:

      1: void draw_list( List<Drawables>& drawables )
      2: {
      3:         // call the virtual draw() method on each element in list
      4:     for( int i=0; i < drawables.count(), ++i )
      5:     {
      6:         drawables[i].draw();
      7:     }
      8: }

Hitting C-c C-s on line 3 of this example gives:

     ((comment-intro) (defun-block-intro 46))

and you can see that the syntactic context contains two syntactic elements. Notice that the first element, ‘(comment-intro)’, has no anchor position.

Next: , Previous: Syntactic Analysis, Up: Indentation Engine Basics

10.2 Syntactic Symbols

This section is a complete list of the syntactic symbols which appear in the c-offsets-alist style variable, along with brief descriptions. The previous section (see Syntactic Analysis) states what syntactic symbols are and how the indentation engine uses them.

More detailed descriptions of these symbols, together with snippets of source code to which they apply, appear in the examples in the subsections below. Note that, in the interests of brevity, the anchor position associated with most syntactic symbols is not specified. In cases of doubt, type C-c C-s on a pertinent line—this highlights the anchor position.

The syntactic symbols which indicate brace constructs follow a general naming convention. When a line begins with an open or close brace, its syntactic symbol will contain the suffix -open or -close respectively. The first line within the brace block construct will contain the suffix -block-intro.

In constructs which can span several lines, a distinction is usually made between the first line that introduces the construct and the lines that continue it. The syntactic symbols that indicate these lines will contain the suffixes -intro or -cont respectively.

The best way to understand how all this works is by looking at some examples. Remember that you can see the syntax of any source code line by using C-c C-s.

string
Inside a multiline string. Literal Symbols.
c
Inside a multiline C style block comment. Literal Symbols.
defun-open
Brace that opens a top-level function definition. Function Symbols.
defun-close
Brace that closes a top-level function definition. Function Symbols.
defun-block-intro
The first line in a top-level defun. Function Symbols.
class-open
Brace that opens a class definition. Class Symbols.
class-close
Brace that closes a class definition. Class Symbols.
inline-open
Brace that opens an in-class inline method. Class Symbols.
inline-close
Brace that closes an in-class inline method. Class Symbols.
func-decl-cont
The region between a function definition's argument list and the function opening brace (excluding K&R argument declarations). In C, you cannot put anything but whitespace and comments in this region, however in C++ and Java, throws declarations and other things can appear here. Literal Symbols.
knr-argdecl-intro
First line of a K&R C argument declaration. K&R Symbols.
knr-argdecl
Subsequent lines in a K&R C argument declaration. K&R Symbols.
topmost-intro
The first line in a “topmost” definition. Function Symbols.
topmost-intro-cont
Topmost definition continuation lines. This is only used in the parts that aren't covered by other symbols such as func-decl-cont and knr-argdecl. Function Symbols.
annotation-top-cont
Topmost definition continuation lines where all previous items are annotations. Java Symbols.
member-init-intro
First line in a member initialization list. Class Symbols.
member-init-cont
Subsequent member initialization list lines. Class Symbols.
inher-intro
First line of a multiple inheritance list. Class Symbols.
inher-cont
Subsequent multiple inheritance lines. Class Symbols.
block-open
Statement block open brace. Literal Symbols.
block-close
Statement block close brace. Conditional Construct Symbols.
brace-list-open
Open brace of an enum or static array list. Brace List Symbols.
brace-list-close
Close brace of an enum or static array list. Brace List Symbols.
brace-list-intro
First line in an enum or static array list. Brace List Symbols.
brace-list-entry
Subsequent lines in an enum or static array list. Brace List Symbols.
brace-entry-open
Subsequent lines in an enum or static array list where the line begins with an open brace. Brace List Symbols.
statement
A statement. Function Symbols.
statement-cont
A continuation of a statement. Function Symbols.
annotation-var-cont
A continuation of a statement where all previous items are annotations. Java Symbols.
statement-block-intro
The first line in a new statement block. Conditional Construct Symbols.
statement-case-intro
The first line in a case block. Switch Statement Symbols.
statement-case-open
The first line in a case block that starts with a brace. Switch Statement Symbols.
substatement
The first line after a conditional or loop construct. Conditional Construct Symbols.
substatement-open
The brace that opens a substatement block. Conditional Construct Symbols.
substatement-label
The first line after a conditional or loop construct if it's a label. Conditional Construct Symbols.
case-label
A label in a switch block. Switch Statement Symbols.
access-label
C++ access control label. Class Symbols.
label
Any other label. Literal Symbols.
do-while-closure
The while line that ends a do-while construct. Conditional Construct Symbols.
else-clause
The else line of an if-else construct. Conditional Construct Symbols.
catch-clause
The catch or finally (in Java) line of a try-catch construct. Conditional Construct Symbols.
comment-intro
A line containing only a comment introduction. Literal Symbols.
arglist-intro
The first line in an argument list. Paren List Symbols.
arglist-cont
Subsequent argument list lines when no arguments follow on the same line as the arglist opening paren. Paren List Symbols.
arglist-cont-nonempty
Subsequent argument list lines when at least one argument follows on the same line as the arglist opening paren. Paren List Symbols.
arglist-close
The solo close paren of an argument list. Paren List Symbols.
stream-op
Lines continuing a stream operator (C++ only). Literal Symbols.
inclass
The line is nested inside a class definition. Class Symbols.
cpp-macro
The start of a preprocessor macro definition. Literal Symbols.
cpp-define-intro
The first line inside a multiline preprocessor macro if c-syntactic-indentation-in-macros is set. Multiline Macro Symbols.
cpp-macro-cont
All lines inside multiline preprocessor macros if c-syntactic-indentation-in-macros is nil. Multiline Macro Symbols.
friend
A C++ friend declaration. Class Symbols.
objc-method-intro
The first line of an Objective-C method definition. Objective-C Method Symbols.
objc-method-args-cont
Lines continuing an Objective-C method definition. Objective-C Method Symbols.
objc-method-call-cont
Lines continuing an Objective-C method call. Objective-C Method Symbols.
extern-lang-open
Brace that opens an extern block (e.g., extern "C" {...}). External Scope Symbols.
extern-lang-close
Brace that closes an extern block. External Scope Symbols.
inextern-lang
Analogous to inclass syntactic symbol, but used inside extern blocks. External Scope Symbols.
namespace-open
namespace-close
innamespace
These are analogous to the three extern-lang symbols above, but are returned for C++ namespace blocks. External Scope Symbols.
module-open
module-close
inmodule
Analogous to the above, but for CORBA IDL module blocks. External Scope Symbols.
composition-open
composition-close
incomposition
Analogous to the above, but for CORBA CIDL composition blocks. External Scope Symbols.
template-args-cont
C++ template argument list continuations. Class Symbols.
inlambda
Analogous to inclass syntactic symbol, but used inside lambda (i.e., anonymous) functions. Only used in Pike mode. Statement Block Symbols.
lambda-intro-cont
Lines continuing the header of a lambda function, i.e., between the lambda keyword and the function body. Only used in Pike mode. Statement Block Symbols.
inexpr-statement
A statement block inside an expression. The gcc C and C++ extension for this is recognized. It's also used for the special functions that take a statement block as an argument in Pike. Statement Block Symbols.
inexpr-class
A class definition inside an expression. This is used for anonymous classes in Java. It's also used for anonymous array initializers in Java. Java Symbols.

Next: , Previous: Syntactic Symbols, Up: Syntactic Symbols

10.2.1 Function Symbols

This example shows a typical function declaration.

      1: void
      2: swap( int& a, int& b )
      3: {
      4:     int tmp = a;
      5:     a = b;
      6:     b = tmp;
      7:     int ignored =
      8:         a + b;
      9: }

Line 1 shows a topmost-intro since it is the first line that introduces a top-level construct. Line 2 is a continuation of the top-level construct introduction so it has the syntax topmost-intro-cont. Line 3 shows a defun-open since it is the brace that opens a top-level function definition. Line 9 is the corresponding defun-close since it contains the brace that closes the top-level function definition. Line 4 is a defun-block-intro, i.e., it is the first line of a brace-block, enclosed in a top-level function definition.

Lines 5, 6, and 7 are all given statement syntax since there isn't much special about them. Note however that line 8 is given statement-cont syntax since it continues the statement begun on the previous line.

Next: , Previous: Function Symbols, Up: Syntactic Symbols

10.2.2 Class related Symbols

Here's an example which illustrates some C++ class syntactic symbols:

      1: class Bass
      2:     : public Guitar,
      3:       public Amplifiable
      4: {
      5: public:
      6:     Bass()
      7:         : eString( new BassString( 0.105 )),
      8:           aString( new BassString( 0.085 )),
      9:           dString( new BassString( 0.065 )),
     10:           gString( new BassString( 0.045 ))
     11:     {
     12:         eString.tune( 'E' );
     13:         aString.tune( 'A' );
     14:         dString.tune( 'D' );
     15:         gString.tune( 'G' );
     16:     }
     17:     friend class Luthier;
     18: };

As in the previous example, line 1 has the topmost-intro syntax. Here however, the brace that opens a C++ class definition on line 4 is assigned the class-open syntax. Note that in C++, classes, structs, and unions are essentially equivalent syntactically (and are very similar semantically), so replacing the class keyword in the example above with struct or union would still result in a syntax of class-open for line 4 36. Similarly, line 18 is assigned class-close syntax.

Line 2 introduces the inheritance list for the class so it is assigned the inher-intro syntax, and line 3, which continues the inheritance list is given inher-cont syntax.

Hitting C-c C-s on line 5 shows the following analysis:

     ((inclass 58) (access-label 58))

The primary syntactic symbol for this line is access-label as this is a label keyword that specifies access protection in C++. However, because this line is also a top-level construct inside a class definition, the analysis actually shows two syntactic symbols. The other syntactic symbol assigned to this line is inclass. Similarly, line 6 is given both inclass and topmost-intro syntax:

     ((inclass 58) (topmost-intro 60))

Line 7 introduces a C++ member initialization list and as such is given member-init-intro syntax. Note that in this case it is not assigned inclass since this is not considered a top-level construct. Lines 8 through 10 are all assigned member-init-cont since they continue the member initialization list started on line 7.

Line 11's analysis is a bit more complicated:

     ((inclass 58) (inline-open))

This line is assigned a syntax of both inline-open and inclass because it opens an in-class C++ inline method definition. This is distinct from, but related to, the C++ notion of an inline function in that its definition occurs inside an enclosing class definition, which in C++ implies that the function should be inlined. However, if the definition of the Bass constructor appeared outside the class definition, the construct would be given the defun-open syntax, even if the keyword inline appeared before the method name, as in:

      1: class Bass
      2:     : public Guitar,
      3:       public Amplifiable
      4: {
      5: public:
      6:     Bass();
      7: };
      8:
      9: inline
     10: Bass::Bass()
     11:     : eString( new BassString( 0.105 )),
     12:       aString( new BassString( 0.085 )),
     13:       dString( new BassString( 0.065 )),
     14:       gString( new BassString( 0.045 ))
     15: {
     16:     eString.tune( 'E' );
     17:     aString.tune( 'A' );
     18:     dString.tune( 'D' );
     19:     gString.tune( 'G' );
     20: }

Returning to the previous example, line 16 is given inline-close syntax, while line 12 is given defun-block-open syntax, and lines 13 through 15 are all given statement syntax. Line 17 is interesting in that its syntactic analysis list contains three elements:

     ((inclass 58) (topmost-intro 380) (friend))

The friend and inline-open syntactic symbols are modifiers that do not have anchor positions.

Template definitions introduce yet another syntactic symbol:

      1: ThingManager <int,
      2:    Framework::Callback *,
      3:    Mutex> framework_callbacks;

Here, line 1 is analyzed as a topmost-intro, but lines 2 and 3 are both analyzed as template-args-cont lines.

Next: , Previous: Class Symbols, Up: Syntactic Symbols

10.2.3 Conditional Construct Symbols

Here is a (totally contrived) example which illustrates how syntax is assigned to various conditional constructs:

      1: void spam( int index )
      2: {
      3:     for( int i=0; i<index; i++ )
      4:     {
      5:         if( i == 10 )
      6:             do_something_special();
      7:         else
      8:           silly_label:
      9:             do_something( i );
     10:     }
     11:     do {
     12:         another_thing( i-- );
     13:     }
     14:     while( i > 0 );
     15: }

Only the lines that illustrate new syntactic symbols will be discussed.

Line 4 has a brace which opens a conditional's substatement block. It is thus assigned substatement-open syntax, and since line 5 is the first line in the substatement block, it is assigned statement-block-intro syntax. Line 10 contains the brace that closes the inner substatement block, and is therefore given the syntax block-close37. Line 13 is treated the same way.

Lines 6 and 9 are also substatements of conditionals, but since they don't start blocks they are given substatement syntax instead of substatement-open.

Line 8 contains a label, which is normally given label syntax. This one is however a bit special since it's between a conditional and its substatement. It's analyzed as substatement-label to let you handle this rather odd case differently from normal labels.

Line 7 start with an else that matches the if statement on line 5. It is therefore given the else-clause syntax and is anchored on the matching if. The try-catch constructs in C++ and Java are treated this way too, except that catch and (in Java) finally, are marked with catch-clause.

The while construct on line 14 that closes a do conditional is given the special syntax do-while-closure if it appears on a line by itself. Note that if the while appeared on the same line as the preceding close brace, that line would still have block-close syntax.

Next: , Previous: Conditional Construct Symbols, Up: Syntactic Symbols

10.2.4 Switch Statement Symbols

Switch statements have their own set of syntactic symbols. Here's an example:

      1: void spam( enum Ingredient i )
      2: {
      3:     switch( i ) {
      4:     case Ham:
      5:         be_a_pig();
      6:         break;
      7:     case Salt:
      8:         drink_some_water();
      9:         break;
     10:     default:
     11:         {
     12:             what_is_it();
     13:             break;
     14:         }
     15:     }
     14: }

Here, lines 4, 7, and 10 are all assigned case-label syntax, while lines 5 and 8 are assigned statement-case-intro. Line 11 is treated slightly differently since it contains a brace that opens a block; it is given statement-case-open syntax.

Next: , Previous: Switch Statement Symbols, Up: Syntactic Symbols

10.2.5 Brace List Symbols

There are a set of syntactic symbols that are used to recognize constructs inside of brace lists. A brace list is defined as an enum or aggregate initializer list, such as might statically initialize an array of structs. The three special aggregate constructs in Pike, ({ }), ([ ]) and (< >), are treated as brace lists too. An example:

      1: static char* ingredients[] =
      2: {
      3:     "Ham",
      4:     "Salt",
      5:     NULL
      6: };

Following convention, line 2 in this example is assigned brace-list-open syntax, and line 3 is assigned brace-list-intro syntax. Likewise, line 6 is assigned brace-list-close syntax. Lines 4 and 5 however, are assigned brace-list-entry syntax, as would all subsequent lines in this initializer list.

Your static initializer might be initializing nested structures, for example:

      1: struct intpairs[] =
      2: {
      3:     { 1, 2 },
      4:     {
      5:         3,
      6:         4
      7:     }
      8:     { 1,
      9:       2 },
     10:     { 3, 4 }
     11: };

Here, you've already seen the analysis of lines 1, 2, 3, and 11. On line 4, things get interesting; this line is assigned brace-entry-open syntactic symbol because it's a bracelist entry line that starts with an open brace. Lines 5 and 6 (and line 9) are pretty standard, and line 7 is a brace-list-close as you'd expect. Once again, line 8 is assigned as brace-entry-open as is line 10.

Next: , Previous: Brace List Symbols, Up: Syntactic Symbols

10.2.6 External Scope Symbols

External language definition blocks also have their own syntactic symbols. In this example:

      1: extern "C"
      2: {
      3:     int thing_one( int );
      4:     int thing_two( double );
      5: }

line 2 is given the extern-lang-open syntax, while line 5 is given the extern-lang-close syntax. The analysis for line 3 yields:

     ((inextern-lang) (topmost-intro 14))

where inextern-lang is a modifier similar in purpose to inclass.

There are various other top level blocks like extern, and they are all treated in the same way except that the symbols are named after the keyword that introduces the block. E.g., C++ namespace blocks get the three symbols namespace-open, namespace-close and innamespace. The currently recognized top level blocks are:

extern-lang-open, extern-lang-close, inextern-lang
extern blocks in C and C++.38
namespace-open, namespace-close, innamespace
namespace blocks in C++.
module-open, module-close, inmodule
module blocks in CORBA IDL.
composition-open, composition-close, incomposition
composition blocks in CORBA CIDL.

Next: , Previous: External Scope Symbols, Up: Syntactic Symbols

10.2.7 Parenthesis (Argument) List Symbols

A number of syntactic symbols are associated with parenthesis lists, a.k.a argument lists, as found in function declarations and function calls. This example illustrates these:

      1: void a_function( int line1,
      2:                  int line2 );
      3:
      4: void a_longer_function(
      5:     int line1,
      6:     int line2
      7:     );
      8:
      9: void call_them( int line1, int line2 )
     10: {
     11:     a_function(
     12:         line1,
     13:         line2
     14:         );
     15:
     16:     a_longer_function( line1,
     17:                        line2 );
     18: }

Lines 5 and 12 are assigned arglist-intro syntax since they are the first line following the open parenthesis, and lines 7 and 14 are assigned arglist-close syntax since they contain the parenthesis that closes the argument list.

Lines that continue argument lists can be assigned one of two syntactic symbols. For example, Lines 2 and 17 are assigned arglist-cont-nonempty syntax. What this means is that they continue an argument list, but that the line containing the parenthesis that opens the list is not empty following the open parenthesis. Contrast this against lines 6 and 13 which are assigned arglist-cont syntax. This is because the parenthesis that opens their argument lists is the last character on that line.

Syntactic elements with arglist-intro, arglist-cont-nonempty, and arglist-close contain two buffer positions: the anchor position (the beginning of the declaration or statement) and the position of the open parenthesis. The latter position can be used in a line-up function (see Line-Up Functions).

Note that there is no arglist-open syntax. This is because any parenthesis that opens an argument list, appearing on a separate line, is assigned the statement-cont syntax instead.

Next: , Previous: Paren List Symbols, Up: Syntactic Symbols

10.2.8 Comment String Label and Macro Symbols

A few miscellaneous syntactic symbols that haven't been previously covered are illustrated by this C++ example:

      1: void Bass::play( int volume )
      2: const
      3: {
      4:     /* this line starts a multiline
      5:      * comment.  This line should get `c' syntax */
      6:
      7:     char* a_multiline_string = "This line starts a multiline \
      8: string.  This line should get `string' syntax.";
      9:
     10:   note:
     11:     {
     12: #ifdef LOCK
     13:         Lock acquire();
     14: #endif // LOCK
     15:         slap_pop();
     16:         cout << "I played "
     17:              << "a note\n";
     18:     }
     19: }

The lines to note in this example include:

Next: , Previous: Literal Symbols, Up: Syntactic Symbols

10.2.9 Multiline Macro Symbols

Multiline preprocessor macro definitions are normally handled just like other code, i.e., the lines inside them are indented according to the syntactic analysis of the preceding lines inside the macro. The first line inside a macro definition (i.e., the line after the starting line of the cpp directive itself) gets cpp-define-intro. In this example:

      1: #define LIST_LOOP(cons, listp)                         \
      2:   for (cons = listp; !NILP (cons); cons = XCDR (cons)) \
      3:     if (!CONSP (cons))                                 \
      4:       signal_error ("Invalid list format", listp);     \
      5:     else

line 1 is given the syntactic symbol cpp-macro. The first line of a cpp directive is always given that symbol. Line 2 is given cpp-define-intro, so that you can give the macro body as a whole some extra indentation. Lines 3 through 5 are then analyzed as normal code, i.e., substatement on lines 3 and 4, and else-clause on line 5.

The syntactic analysis inside macros can be turned off with c-syntactic-indentation-in-macros (see Custom Macros). In that case, lines 2 through 5 would all be given cpp-macro-cont with an anchor position pointing to the # which starts the cpp directive39.

See Custom Macros, for more info about the treatment of macros.

Next: , Previous: Multiline Macro Symbols, Up: Syntactic Symbols

10.2.10 Objective-C Method Symbols

In Objective-C buffers, there are three additional syntactic symbols assigned to various message calling constructs. Here's an example illustrating these:

      1: - (void)setDelegate:anObject
      2:           withStuff:stuff
      3: {
      4:     [delegate masterWillRebind:self
      5:               toDelegate:anObject
      6:               withExtraStuff:stuff];
      7: }

Here, line 1 is assigned objc-method-intro syntax, and line 2 is assigned objc-method-args-cont syntax. Lines 5 and 6 are both assigned objc-method-call-cont syntax.

Next: , Previous: Objective-C Method Symbols, Up: Syntactic Symbols

10.2.11 Java Symbols

Java has a concept of anonymous classes which can look something like this:

      1:  @Test
      2:  public void watch(Observable o) {
      3:      @NonNull
      4:      Observer obs = new Observer() {
      5:          public void update(Observable o, Object arg) {
      6:              history.addElement(arg);
      7:          }
      8:      };
      9:      o.addObserver(obs);
      10: }

The brace following the new operator opens the anonymous class. Lines 5 and 8 are assigned the inexpr-class syntax, besides the inclass symbol used in normal classes. Thus, the class will be indented just like a normal class, with the added indentation given to inexpr-class. An inexpr-class syntactic element doesn't have an anchor position.

Line 2 is assigned the annotation-top-cont syntax, due to it being a continuation of a topmost introduction with an annotation symbol preceding the current line. Similarly, line 4 is assigned the annotation-var-cont syntax due to it being a continuation of a variable declaration where preceding the declaration is an annotation.

Next: , Previous: Java Symbols, Up: Syntactic Symbols

10.2.12 Statement Block Symbols

There are a few occasions where a statement block might be used inside an expression. One is in C or C++ code using the gcc extension for this, e.g.:

      1: int res = ({
      2:         int y = foo (); int z;
      3:         if (y > 0) z = y; else z = - y;
      4:         z;
      5:     });

Lines 2 and 5 get the inexpr-statement syntax, besides the symbols they'd get in a normal block. Therefore, the indentation put on inexpr-statement is added to the normal statement block indentation. An inexpr-statement syntactic element doesn't contain an anchor position.

In Pike code, there are a few other situations where blocks occur inside statements, as illustrated here:

      1: array itgob()
      2: {
      3:     string s = map (backtrace()[-2][3..],
      4:                     lambda
      5:                         (mixed arg)
      6:                     {
      7:                         return sprintf ("%t", arg);
      8:                     }) * ", " + "\n";
      9:     return catch {
     10:             write (s + "\n");
     11:         };
     12: }

Lines 4 through 8 contain a lambda function, which CC Mode recognizes by the lambda keyword. If the function argument list is put on a line of its own, as in line 5, it gets the lambda-intro-cont syntax. The function body is handled as an inline method body, with the addition of the inlambda syntactic symbol. This means that line 6 gets inlambda and inline-open, and line 8 gets inline-close40.

On line 9, catch is a special function taking a statement block as its argument. The block is handled as an in-expression statement with the inexpr-statement syntax, just like the gcc extended C example above. The other similar special function, gauge, is handled like this too.

Previous: Statement Block Symbols, Up: Syntactic Symbols

10.2.13 K&R Symbols

Two other syntactic symbols can appear in old style, non-prototyped C code 41:

      1: int add_three_integers(a, b, c)
      2:      int a;
      3:      int b;
      4:      int c;
      5: {
      6:     return a + b + c;
      7: }

Here, line 2 is the first line in an argument declaration list and so is given the knr-argdecl-intro syntactic symbol. Subsequent lines (i.e., lines 3 and 4 in this example), are given knr-argdecl syntax.

Previous: Syntactic Symbols, Up: Indentation Engine Basics

10.3 Indentation Calculation

Indentation for a line is calculated from the syntactic context (see Syntactic Analysis).

First, a buffer position is found whose column will be the base for the indentation calculation. It's the anchor position in the first syntactic element that provides one that is used. If no syntactic element has an anchor position then column zero is used.

Second, the syntactic symbols in each syntactic element are looked up in the c-offsets-alist style variable (see c-offsets-alist), which is an association list of syntactic symbols and the offsets to apply for those symbols. These offsets are added together with the base column to produce the new indentation column.

Let's use our two code examples above to see how this works. Here is our first example again:

      1: void swap( int& a, int& b )
      2: {
      3:     int tmp = a;
      4:     a = b;
      5:     b = tmp;
      6: }

Let's say point is on line 3 and we hit the <TAB> key to reindent the line. The syntactic context for that line is:

     ((defun-block-intro 29))

Since buffer position 29 is the first and only anchor position in the list, CC Mode goes there and asks for the current column. This brace is in column zero, so CC Mode uses ‘0’ as the base column.

Next, CC Mode looks up defun-block-intro in the c-offsets-alist style variable. Let's say it finds the value ‘4’; it adds this to the base column ‘0’, yielding a running total indentation of 4 spaces.

Since there is only one syntactic element on the list for this line, indentation calculation is complete, and the total indentation for the line is 4 spaces.

Here's another example:

      1: int add( int val, int incr, int doit )
      2: {
      3:     if( doit )
      4:         {
      5:             return( val + incr );
      6:         }
      7:     return( val );
      8: }

If we were to hit TAB on line 4 in the above example, the same basic process is performed, despite the differences in the syntactic context. The context for this line is:

     ((substatement-open 46))

Here, CC Mode goes to buffer position 46, which is the ‘i’ in if on line 3. This character is in the fourth column on that line so the base column is ‘4’. Then CC Mode looks up the substatement-open symbol in c-offsets-alist. Let's say it finds the value ‘4’. It's added with the base column and yields an indentation for the line of 8 spaces.

Simple, huh?

Actually, it's a bit more complicated than that since the entries on c-offsets-alist can be much more than plain offsets. See c-offsets-alist, for the full story.

Anyway, the mode usually just does The Right Thing without you having to think about it in this much detail. But when customizing indentation, it's helpful to understand the general indentation model being used.

As you configure CC Mode, you might want to set the variable c-echo-syntactic-information-p to non-nil so that the syntactic context and calculated offset always is echoed in the minibuffer when you hit TAB.

Next: , Previous: Indentation Engine Basics, Up: Top

11 Customizing Indentation

The principal variable for customizing indentation is the style variable c-offsets-alist, which gives an offset (an indentation rule) for each syntactic symbol. Its structure and semantics are completely described in c-offsets-alist. The various ways you can set the variable, including the use of the CC Mode style system, are described in Config Basics and its sections, in particular Style Variables.

The simplest and most used kind of “offset” setting in c-offsets-alist is in terms of multiples of c-basic-offset:

— User Option: c-basic-offset

This style variable holds the basic offset between indentation levels. It's factory default is 4, but all the built-in styles set it themselves, to some value between 2 (for gnu style) and 8 (for bsd, linux, and python styles).

The most flexible “offset” setting you can make in c-offsets-alist is a line-up function (or even a list of them), either one supplied by CC Mode (see Line-Up Functions) or one you write yourself (see Custom Line-Up).

Finally, in Other Indentation you'll find the tool of last resort: a hook which is called after a line has been indented. You can install functions here to make ad-hoc adjustments to any line's indentation.

Next: , Previous: Customizing Indentation, Up: Customizing Indentation

11.1 c-offsets-alist

This section explains the structure and semantics of the style variable c-offsets-alist, the principal variable for configuring indentation. Details of how to set it up, and its relationship to CC Mode's style system are given in Style Variables.

— User Option: c-offsets-alist

This is an alist which associates an offset with each syntactic symbol. This offset is a rule specifying how to indent a line whose syntactic context matches the symbol. See Syntactic Analysis.

Note that the buffer-local binding of this alist in a CC Mode buffer contains an entry for every syntactic symbol. Its global binding and its settings within style specifications usually contain only a few entries. See Style Variables.

The offset specification associated with any particular syntactic symbol can be an integer, a variable name, a vector, a function or lambda expression, a list, or one of the following special symbols: +, -, ++, --, *, or /. The meanings of these values are described in detail below.

Here is an example fragment of a c-offsets-alist, showing some of these kinds of offsets:

          ((statement . 0)
           (substatement . +)
           (cpp-macro . [0])
           (topmost-intro-cont . c-lineup-topmost-intro-cont)
           (statement-block-intro . (add c-lineup-whitesmith-in-block
                                         c-indent-multi-line-block))
           ...
          
)
— Command: c-set-offset (C-c C-o)

This command changes the entry for a syntactic symbol in the current binding of c-offsets-alist, or it inserts a new entry if there isn't already one for that syntactic symbol.

You can use c-set-offsets interactively within a CC Mode buffer to make experimental changes to your indentation settings. C-c C-o prompts you for the syntactic symbol to change (defaulting to that of the current line) and the new offset (defaulting to the current offset).

c-set-offsets takes two arguments when used programmatically: symbol, the syntactic element symbol to change and offset, the new offset for that syntactic element. You can call the command in your .emacs to change the global binding of c-offsets-alist (see Style Variables); you can use it in a hook function to make changes from the current style. CC Mode itself uses this function when initializing styles.

The “offset specifications” in c-offsets-alist can be any of the following:

An integer
The integer specifies a relative offset. All relative offsets42 will be added together and used to calculate the indentation relative to an anchor position earlier in the buffer. See Indentation Calculation, for details. Most of the time, it's probably better to use one of the special symbols like + than an integer (apart from zero).
One of the symbols +, -, ++, --, *, or /
These special symbols describe a relative offset in multiples of c-basic-offset:

By defining a style's indentation in terms of c-basic-offset, you can change the amount of whitespace given to an indentation level while maintaining the same basic shape of your code. Here are the values that the special symbols correspond to:

+
c-basic-offset times 1
-
c-basic-offset times −1
++
c-basic-offset times 2
--
c-basic-offset times −2
*
c-basic-offset times 0.5
/
c-basic-offset times −0.5

A vector
The first element of the vector, an integer, sets the absolute indentation column. This will override any previously calculated indentation, but won't override relative indentation calculated from syntactic elements later on in the syntactic context of the line being indented. See Indentation Calculation. Any elements in the vector beyond the first will be ignored.
A function or lambda expression
The function will be called and its return value will in turn be evaluated as an offset specification. Functions are useful when more context than just the syntactic symbol is needed to get the desired indentation. See Line-Up Functions, and Custom Line-Up, for details about them.
A symbol with a variable binding
If the symbol also has a function binding, the function takes precedence over the variable. Otherwise the value of the variable is used. It must be an integer (which is used as relative offset) or a vector (an absolute offset).
A list
The offset can also be a list containing several offset specifications; these are evaluated recursively and combined. A list is typically only useful when some of the offsets are line-up functions. A common strategy is calling a sequence of functions in turn until one of them recognizes that it is appropriate for the source line and returns a non-nil value.

nil values are always ignored when the offsets are combined. The first element of the list specifies the method of combining the non-nil offsets from the remaining elements:

first
Use the first offset that doesn't evaluate to nil. Subsequent elements of the list don't get evaluated.
min
Use the minimum of all the offsets. All must be either relative or absolute; they can't be mixed.
max
Use the maximum of all the offsets. All must be either relative or absolute; they can't be mixed.
add
Add all the evaluated offsets together. Exactly one of them may be absolute, in which case the result is absolute. Any relative offsets that preceded the absolute one in the list will be ignored in that case.

As a compatibility measure, if the first element is none of the above then it too will be taken as an offset specification and the whole list will be combined according to the method first.

If an offset specification evaluates to nil, then a relative offset of 0 (zero) is used43.

Next: , Previous: c-offsets-alist, Up: Customizing Indentation

11.2 Interactive Customization

As an example of how to customize indentation, let's change the style of this example44:

      1: int add( int val, int incr, int doit )
      2: {
      3:   if( doit )
      4:     {
      5:       return( val + incr );
      6:     }
      7:   return( val );
      8: }

to:

      1: int add( int val, int incr, int doit )
      2: {
      3:   if( doit )
      4:   {
      5:     return( val + incr );
      6:   }
      7:   return( val );
      8: }

In other words, we want to change the indentation of braces that open a block following a condition so that the braces line up under the conditional, instead of being indented. Notice that the construct we want to change starts on line 4. To change the indentation of a line, we need to see which syntactic symbols affect the offset calculations for that line. Hitting C-c C-s on line 4 yields:

     ((substatement-open 44))

so we know that to change the offset of the open brace, we need to change the indentation for the substatement-open syntactic symbol.

To do this interactively, just hit C-c C-o. This prompts you for the syntactic symbol to change, providing a reasonable default. In this case, the default is substatement-open, which is just the syntactic symbol we want to change!

After you hit return, CC Mode will then prompt you for the new offset value, with the old value as the default. The default in this case is ‘+’, but we want no extra indentation so enter ‘0’ and RET. This will associate the offset 0 with the syntactic symbol substatement-open.

To check your changes quickly, just hit C-c C-q (c-indent-defun) to reindent the entire function. The example should now look like:

      1: int add( int val, int incr, int doit )
      2: {
      3:   if( doit )
      4:   {
      5:     return( val + incr );
      6:   }
      7:   return( val );
      8: }

Notice how just changing the open brace offset on line 4 is all we needed to do. Since the other affected lines are indented relative to line 4, they are automatically indented the way you'd expect. For more complicated examples, this might not always work. The general approach to take is to always start adjusting offsets for lines higher up in the file, then reindent and see if any following lines need further adjustments.

— Command: c-set-offset symbol offset

This is the command bound to C-c C-o. It provides a convenient way to set offsets on c-offsets-alist both interactively (see the example above) and from your mode hook.

It takes two arguments when used programmatically: symbol is the syntactic element symbol to change and offset is the new offset for that syntactic element.

Next: , Previous: Interactive Customization, Up: Customizing Indentation

11.3 Line-Up Functions

Often there are cases when a simple offset setting on a syntactic symbol isn't enough to get the desired indentation—for example, you might want to line up a closing parenthesis with the matching opening one rather than indenting relative to its “anchor point”. CC Mode provides this flexibility with line-up functions.

The way you associate a line-up function with a syntactic symbol is described in c-offsets-alist. CC Mode comes with many predefined line-up functions for common situations. If none of these does what you want, you can write your own. See Custom Line-Up. Sometimes, it is easier to tweak the standard indentation by adding a function to c-special-indent-hook (see Other Indentation).

The line-up functions haven't been adapted for AWK buffers or tested with them. Some of them might work serendipitously. There shouldn't be any problems writing custom line-up functions for AWK mode.

The calling convention for line-up functions is described fully in Custom Line-Up. Roughly speaking, the return value is either an offset itself (such as + or [0]) or it's nil, meaning “this function is inappropriate in this case; try a different one”. See c-offsets-alist.

The subsections below describe all the standard line-up functions, categorized by the sort of token the lining-up centers around. For each of these functions there is a “works with” list that indicates which syntactic symbols the function is intended to be used with.

Next: , Previous: Line-Up Functions, Up: Line-Up Functions

11.3.1 Brace and Parenthesis Line-Up Functions

The line-up functions here calculate the indentation for braces, parentheses and statements within brace blocks.

— Function: c-lineup-close-paren

Line up the closing paren under its corresponding open paren if the open paren is followed by code. If the open paren ends its line, no indentation is added. E.g.:

          main (int,
                char **
               )                <- c-lineup-close-paren

and

          main (
              int, char **
          )                     <- c-lineup-close-paren

As a special case, if a brace block is opened at the same line as the open parenthesis of the argument list, the indentation is c-basic-offset instead of the open paren column. See c-lineup-arglist for further discussion of this “DWIM” measure.

Works with: All *-close symbols.

— Function: c-lineup-arglist-close-under-paren

Set your arglist-close syntactic symbol to this line-up function so that parentheses that close argument lists will line up under the parenthesis that opened the argument list. It can also be used with arglist-cont and arglist-cont-nonempty to line up all lines inside a parenthesis under the open paren.

As a special case, if a brace block is opened at the same line as the open parenthesis of the argument list, the indentation is c-basic-offset only. See c-lineup-arglist for further discussion of this “DWIM” measure.

Works with: Almost all symbols, but are typically most useful on arglist-close, brace-list-close, arglist-cont and arglist-cont-nonempty.

— Function: c-indent-one-line-block

Indent a one line block c-basic-offset extra. E.g.:

          if (n > 0)
              {m+=n; n=0;}      <- c-indent-one-line-block
          <--> c-basic-offset

and

          if (n > 0)
          {                     <- c-indent-one-line-block
              m+=n; n=0;
          }

The block may be surrounded by any kind of parenthesis characters. nil is returned if the line doesn't start with a one line block, which makes the function usable in list expressions.

Works with: Almost all syntactic symbols, but most useful on the -open symbols.

— Function: c-indent-multi-line-block

Indent a multiline block c-basic-offset extra. E.g.:

          int *foo[] = {
              NULL,
              {17},             <- c-indent-multi-line-block

and

          int *foo[] = {
              NULL,
                  {             <- c-indent-multi-line-block
                  17
                  },
              <--> c-basic-offset

The block may be surrounded by any kind of parenthesis characters. nil is returned if the line doesn't start with a multiline block, which makes the function usable in list expressions.

Works with: Almost all syntactic symbols, but most useful on the -open symbols.

— Function: c-lineup-runin-statements

Line up statements for coding standards which place the first statement in a block on the same line as the block opening brace45. E.g.:

          int main()
          { puts ("Hello!");
            return 0;           <- c-lineup-runin-statements
          }

If there is no statement after the opening brace to align with, nil is returned. This makes the function usable in list expressions.

Works with: The statement syntactic symbol.

— Function: c-lineup-inexpr-block

This can be used with the in-expression block symbols to indent the whole block to the column where the construct is started. E.g., for Java anonymous classes, this lines up the class under the ‘new’ keyword, and in Pike it lines up the lambda function body under the ‘lambda’ keyword. Returns nil if the block isn't part of such a construct.

Works with: inlambda, inexpr-statement, inexpr-class.

— Function: c-lineup-after-whitesmith-blocks

Compensate for Whitesmith style indentation of blocks. Due to the way CC Mode calculates anchor positions for normal lines inside blocks, this function is necessary for those lines to get correct Whitesmith style indentation. Consider the following examples:

          int foo()
              {
              a;
              x;                 <- c-lineup-after-whitesmith-blocks
          int foo()
              {
                  {
                  a;
                  }
              x;                 <- c-lineup-after-whitesmith-blocks

The fact that the line with x is preceded by a Whitesmith style indented block in the latter case and not the first should not affect its indentation. But since CC Mode in cases like this uses the indentation of the preceding statement as anchor position, the x would in the second case be indented too much if the offset for statement was set simply to zero.

This lineup function corrects for this situation by detecting if the anchor position is at an open paren character. In that case, it instead indents relative to the surrounding block just like c-lineup-whitesmith-in-block.

Works with: brace-list-entry, brace-entry-open, statement, arglist-cont.

— Function: c-lineup-whitesmith-in-block

Line up lines inside a block in Whitesmith style. It's done in a way that works both when the opening brace hangs and when it doesn't. E.g.:

          something
              {
              foo;              <- c-lineup-whitesmith-in-block
              }

and

          something {
              foo;              <- c-lineup-whitesmith-in-block
              }
          <--> c-basic-offset

In the first case the indentation is kept unchanged, in the second c-basic-offset is added.

Works with: defun-close, defun-block-intro, inline-close, block-close, brace-list-close, brace-list-intro, statement-block-intro, arglist-intro, arglist-cont-nonempty, arglist-close, and all in* symbols, e.g., inclass and inextern-lang.

Next: , Previous: Brace/Paren Line-Up, Up: Line-Up Functions

11.3.2 List Line-Up Functions

The line-up functions here calculate the indentation for lines which form lists of items, usually separated by commas.

The function c-lineup-arglist-close-under-paren, which is mainly for indenting a close parenthesis, is also useful for the lines contained within parentheses.

— Function: c-lineup-arglist

Line up the current argument line under the first argument.

As a special case, if an argument on the same line as the open parenthesis starts with a brace block opener, the indentation is c-basic-offset only. This is intended as a “DWIM” measure in cases like macros that contain statement blocks, e.g.:

          A_VERY_LONG_MACRO_NAME ({
                  some (code, with + long, lines * in[it]);
              });
          <--> c-basic-offset

This is motivated partly because it's more in line with how code blocks are handled, and partly since it approximates the behavior of earlier CC Mode versions, which due to inaccurate analysis tended to indent such cases this way.

Works with: arglist-cont-nonempty, arglist-close.

— Function: c-lineup-arglist-intro-after-paren

Line up a line to just after the open paren of the surrounding paren or brace block.

Works with: defun-block-intro, brace-list-intro, statement-block-intro, statement-case-intro, arglist-intro.

— Function: c-lineup-multi-inher

Line up the classes in C++ multiple inheritance clauses and member initializers under each other. E.g.:

          Foo::Foo (int a, int b):
              Cyphr (a),
              Bar (b)           <- c-lineup-multi-inher

and

          class Foo
              : public Cyphr,
                public Bar      <- c-lineup-multi-inher

and

          Foo::Foo (int a, int b)
              : Cyphr (a)
              , Bar (b)         <- c-lineup-multi-inher

Works with: inher-cont, member-init-cont.

— Function: c-lineup-java-inher

Line up Java implements and extends declarations. If class names follow on the same line as the ‘implements’/‘extends’ keyword, they are lined up under each other. Otherwise, they are indented by adding c-basic-offset to the column of the keyword. E.g.:

          class Foo
              extends
                  Bar           <- c-lineup-java-inher
              <--> c-basic-offset

and

          class Foo
              extends Cyphr,
                      Bar       <- c-lineup-java-inher

Works with: inher-cont.

— Function: c-lineup-java-throws

Line up Java throws declarations. If exception names follow on the same line as the throws keyword, they are lined up under each other. Otherwise, they are indented by adding c-basic-offset to the column of the ‘throws’ keyword. The ‘throws’ keyword itself is also indented by c-basic-offset from the function declaration start if it doesn't hang. E.g.:

          int foo()
              throws            <- c-lineup-java-throws
                  Bar           <- c-lineup-java-throws
          <--><--> c-basic-offset

and

          int foo() throws Cyphr,
                           Bar,    <- c-lineup-java-throws
                           Vlod    <- c-lineup-java-throws

Works with: func-decl-cont.

— Function: c-lineup-template-args

Line up the arguments of a template argument list under each other, but only in the case where the first argument is on the same line as the opening ‘<’.

To allow this function to be used in a list expression, nil is returned if there's no template argument on the first line.

Works with: template-args-cont.

— Function: c-lineup-ObjC-method-call

For Objective-C code, line up selector args as Emacs Lisp mode does with function args: go to the position right after the message receiver, and if you are at the end of the line, indent the current line c-basic-offset columns from the opening bracket; otherwise you are looking at the first character of the first method call argument, so lineup the current line with it.

Works with: objc-method-call-cont.

— Function: c-lineup-ObjC-method-args

For Objective-C code, line up the colons that separate args. The colon on the current line is aligned with the one on the first line.

Works with: objc-method-args-cont.

— Function: c-lineup-ObjC-method-args-2

Similar to c-lineup-ObjC-method-args but lines up the colon on the current line with the colon on the previous line.

Works with: objc-method-args-cont.

Next: , Previous: List Line-Up, Up: Line-Up Functions

11.3.3 Operator Line-Up Functions

The line-up functions here calculate the indentation for lines which start with an operator, by lining it up with something on the previous line.

— Function: c-lineup-argcont

Line up a continued argument. E.g.:

          foo (xyz, aaa + bbb + ccc
                    + ddd + eee + fff);  <- c-lineup-argcont

Only continuation lines like this are touched, nil is returned on lines which are the start of an argument.

Within a gcc asm block, : is recognized as an argument separator, but of course only between operand specifications, not in the expressions for the operands.

Works with: arglist-cont, arglist-cont-nonempty.

— Function: c-lineup-arglist-operators

Line up lines starting with an infix operator under the open paren. Return nil on lines that don't start with an operator, to leave those cases to other line-up functions. Example:

          if (  x < 10
             || at_limit (x,     <- c-lineup-arglist-operators
                          list)  <- c-lineup-arglist-operators returns nil
             )

Since this function doesn't do anything for lines without an infix operator you typically want to use it together with some other lineup settings, e.g., as follows (the arglist-close setting is just a suggestion to get a consistent style):

          (c-set-offset 'arglist-cont
                        '(c-lineup-arglist-operators 0))
          (c-set-offset 'arglist-cont-nonempty
                        '(c-lineup-arglist-operators c-lineup-arglist))
          (c-set-offset 'arglist-close
                        '(c-lineup-arglist-close-under-paren))

Works with: arglist-cont, arglist-cont-nonempty.

— Function: c-lineup-assignments

Line up the current line after the assignment operator on the first line in the statement. If there isn't any, return nil to allow stacking with other line-up functions. If the current line contains an assignment operator too, try to align it with the first one.

Works with: topmost-intro-cont, statement-cont, arglist-cont, arglist-cont-nonempty.

— Function: c-lineup-math

Like c-lineup-assignments but indent with c-basic-offset if no assignment operator was found on the first line. I.e., this function is the same as specifying a list (c-lineup-assignments +). It's provided for compatibility with old configurations.

Works with: topmost-intro-cont, statement-cont, arglist-cont, arglist-cont-nonempty.

— Function: c-lineup-cascaded-calls

Line up “cascaded calls” under each other. If the line begins with -> or . and the preceding line ends with one or more function calls preceded by the same token, then the arrow is lined up with the first of those tokens. E.g.:

          r = proc->add(17)->add(18)
                  ->add(19) +         <- c-lineup-cascaded-calls
            offset;                   <- c-lineup-cascaded-calls (inactive)

In any other situation nil is returned to allow use in list expressions.

Works with: topmost-intro-cont, statement-cont, arglist-cont, arglist-cont-nonempty.

— Function: c-lineup-streamop

Line up C++ stream operators (i.e., ‘<<’ and ‘>>’).

Works with: stream-op.

— Function: c-lineup-string-cont

Line up a continued string under the one it continues. A continued string in this sense is where a string literal follows directly after another one. E.g.:

          result = prefix + "A message "
                            "string.";    <- c-lineup-string-cont

nil is returned in other situations, to allow stacking with other lineup functions.

Works with: topmost-intro-cont, statement-cont, arglist-cont, arglist-cont-nonempty.

Next: , Previous: Operator Line-Up, Up: Line-Up Functions

11.3.4 Comment Line-Up Functions

The lineup functions here calculate the indentation for several types of comment structure.

— Function: c-lineup-C-comments

Line up C block comment continuation lines. Various heuristics are used to handle most of the common comment styles. Some examples:

          /*                 /**               /*
           * text             * text             text
           */                 */               */
          /* text            /*                /**
             text            ** text            ** text
          */                 */                 */
          /**************************************************
           * text
           *************************************************/

          /**************************************************
              Free form text comments:
           In comments with a long delimiter line at the
           start, the indentation is kept unchanged for lines
           that start with an empty comment line prefix.  The
           delimiter line is whatever matches the
           comment-start-skip regexp.
          **************************************************/

The style variable c-comment-prefix-regexp is used to recognize the comment line prefix, e.g., the ‘*’ that usually starts every line inside a comment.

Works with: The c syntactic symbol.

— Function: c-lineup-comment

Line up a comment-only line according to the style variable c-comment-only-line-offset. If the comment is lined up with a comment starter on the previous line, that alignment is preserved.

— User Option: c-comment-only-line-offset

This style variable specifies the extra offset for the line. It can contain an integer or a cons cell of the form

               (non-anchored-offset . anchored-offset)

where non-anchored-offset is the amount of offset given to non-column-zero anchored lines, and anchored-offset is the amount of offset to give column-zero anchored lines. Just an integer as value is equivalent to (value . -1000).

Works with: comment-intro.

— Function: c-lineup-knr-region-comment

Line up a comment in the “K&R region” with the declaration. That is the region between the function or class header and the beginning of the block. E.g.:

          int main()
          /* Called at startup. */  <- c-lineup-knr-region-comment
          {
            return 0;
          }

Return nil if called in any other situation, to be useful in list expressions.

Works with: comment-intro.

Previous: Comment Line-Up, Up: Line-Up Functions

11.3.5 Miscellaneous Line-Up Functions

The line-up functions here are the odds and ends which didn't fit into any earlier category.

— Function: c-lineup-dont-change

This lineup function makes the line stay at whatever indentation it already has; think of it as an identity function for lineups.

Works with: Any syntactic symbol.

— Function: c-lineup-cpp-define

Line up macro continuation lines according to the indentation of the construct preceding the macro. E.g.:

          const char msg[] =    <- The beginning of the preceding construct.
            \"Some text.\";
          
          #define X(A, B)  \
          do {             \    <- c-lineup-cpp-define
            printf (A, B); \
          } while (0)

and:

          int dribble() {
            if (!running)       <- The beginning of the preceding construct.
              error(\"Not running!\");
          
          #define X(A, B)    \
            do {             \  <- c-lineup-cpp-define
              printf (A, B); \
            } while (0)

If c-syntactic-indentation-in-macros is non-nil, the function returns the relative indentation to the macro start line to allow accumulation with other offsets. E.g., in the following cases, cpp-define-intro is combined with the statement-block-intro that comes from the ‘do {’ that hangs on the ‘#define’ line:

          const char msg[] =
            \"Some text.\";
          
          #define X(A, B) do { \
            printf (A, B);     \  <- c-lineup-cpp-define
            this->refs++;      \
          } while (0)             <- c-lineup-cpp-define

and:

          int dribble() {
            if (!running)
              error(\"Not running!\");
          
          #define X(A, B) do { \
              printf (A, B);   \  <- c-lineup-cpp-define
              this->refs++;    \
            } while (0)           <- c-lineup-cpp-define

The relative indentation returned by c-lineup-cpp-define is zero and two, respectively, on the two lines in each of these examples. They are then added to the two column indentation that statement-block-intro gives in both cases here.

If the relative indentation is zero, then nil is returned instead. That is useful in a list expression to specify the default indentation on the top level.

If c-syntactic-indentation-in-macros is nil then this function keeps the current indentation, except for empty lines (ignoring the ending backslash) where it takes the indentation from the closest preceding nonempty line in the macro. If there's no such line in the macro then the indentation is taken from the construct preceding it, as described above.

Works with: cpp-define-intro.

— Function: c-lineup-gcc-asm-reg

Line up a gcc asm register under one on a previous line.

              asm ("foo %1, %0\n"
                   "bar %0, %1"
                   : "=r" (w),
                     "=r" (x)
                   :  "0" (y),
                      "1" (z));

The ‘x’ line is aligned to the text after the ‘:’ on the ‘w’ line, and similarly ‘z’ under ‘y’.

This is done only in an ‘asm’ or ‘__asm__’ block, and only to those lines mentioned. Anywhere else nil is returned. The usual arrangement is to have this routine as an extra feature at the start of arglist lineups, e.g.:

          (c-lineup-gcc-asm-reg c-lineup-arglist)

Works with: arglist-cont, arglist-cont-nonempty.

— Function: c-lineup-topmost-intro-cont

Line up declaration continuation lines zero or one indentation step46. For lines preceding a definition, zero is used. For other lines, c-basic-offset is added to the indentation. E.g.:

          int
          neg (int i)           <- c-lineup-topmost-intro-cont
          {
              return -i;
          }

and

          struct
          larch                 <- c-lineup-topmost-intro-cont
          {
              double height;
          }
              the_larch,        <- c-lineup-topmost-intro-cont
              another_larch;    <- c-lineup-topmost-intro-cont
          <--> c-basic-offset

and

          struct larch
          the_larch,            <- c-lineup-topmost-intro-cont
              another_larch;    <- c-lineup-topmost-intro-cont

Works with: topmost-intro-cont.

Next: , Previous: Line-Up Functions, Up: Customizing Indentation

11.4 Custom Line-Up Functions

The most flexible way to customize indentation is by writing custom line-up functions, and associating them with specific syntactic symbols (see c-offsets-alist). Depending on the effect you want, it might be better to write a c-special-indent-hook function rather than a line-up function (see Other Indentation).

CC Mode comes with an extensive set of predefined line-up functions, not all of which are used by the default styles. So there's a good chance the function you want already exists. See Line-Up Functions, for a list of them. If you write your own line-up function, it's probably a good idea to start working from one of these predefined functions, which can be found in the file cc-align.el. If you have written a line-up function that you think is generally useful, you're very welcome to contribute it; please contact bug-cc-mode@gnu.org.

Line-up functions are passed a single argument, the syntactic element (see below). At the time of the call, point will be somewhere on the line being indented. The return value is a c-offsets-alist offset specification: for example, an integer, a symbol such as +, a vector, nil47, or even another line-up function. Full details of these are in c-offsets-alist.

Line-up functions must not move point or change the content of the buffer (except temporarily). They are however allowed to do hidden buffer changes, i.e., setting text properties for caching purposes etc. Buffer undo recording is disabled while they run.

The syntactic element passed as the parameter to a line-up function is a cons cell of the form

     (syntactic-symbol . anchor-position)

where syntactic-symbol is the symbol that the function was called for, and anchor-position is the anchor position (if any) for the construct that triggered the syntactic symbol (see Syntactic Analysis). This cons cell is how the syntactic element of a line used to be represented in CC Mode 5.28 and earlier. Line-up functions are still passed this cons cell, so as to preserve compatibility with older configurations. In the future, we may decide to convert to using the full list format—you can prepare your setup for this by using the access functions (c-langelem-sym, etc.) described below.

Some syntactic symbols, e.g., arglist-cont-nonempty, have more info in the syntactic element: typically other positions that can be interesting besides the anchor position. That info can't be accessed through the passed argument, which is a cons cell. Instead, you can get this information from the variable c-syntactic-element, which is dynamically bound to the complete syntactic element. The variable c-syntactic-context might also be useful: it gets dynamically bound to the complete syntactic context. See Custom Braces.

CC Mode provides a few functions to access parts of syntactic elements in a more abstract way. Besides making the code easier to read, they also hide the difference between the old cons cell form used in the line-up function argument and the new list form used in c-syntactic-element and everywhere else. The functions are:

— Function: c-langelem-sym langelem

Return the syntactic symbol in langelem.

— Function: c-langelem-pos langelem

Return the anchor position in langelem, or nil if there is none.

— Function: c-langelem-col langelem &optional preserve-point

Return the column of the anchor position in langelem. Also move the point to that position unless preserve-point is non-nil.

— Function: c-langelem-2nd-pos langelem

Return the secondary position in langelem, or nil if there is none.

Note that the return value of this function is always nil if langelem is in the old cons cell form. Thus this function is only meaningful when used on syntactic elements taken from c-syntactic-element or c-syntactic-context.

Custom line-up functions can be as simple or as complex as you like, and any syntactic symbol that appears in c-offsets-alist can have a custom line-up function associated with it.

Previous: Custom Line-Up, Up: Customizing Indentation

11.5 Other Special Indentations

To configure macros which you invoke without a terminating ‘;’, see See Macros with ;.

Here are the remaining odds and ends regarding indentation:

— User Option: c-label-minimum-indentation

In ‘gnu’ style (see Built-in Styles), a minimum indentation is imposed on lines inside code blocks. This minimum indentation is controlled by this style variable. The default value is 1.

It's the function c-gnu-impose-minimum that enforces this minimum indentation. It must be present on c-special-indent-hook to work.

— User Option: c-special-indent-hook

This style variable is a standard hook variable that is called after every line is indented by CC Mode. It is called only if c-syntactic-indentation is non-nil (which it is by default (see Indentation Engine Basics)). You can put a function on this hook to do any special indentation or ad hoc line adjustments your style dictates, such as adding extra indentation to constructors or destructor declarations in a class definition, etc. Sometimes it is better to write a custom Line-up Function instead (see Custom Line-Up).

When the indentation engine calls this hook, the variable c-syntactic-context is bound to the current syntactic context (i.e., what you would get by typing C-c C-s on the source line. See Custom Braces.). Note that you should not change point or mark inside a c-special-indent-hook function, i.e., you'll probably want to wrap your function in a save-excursion48.

Setting c-special-indent-hook in style definitions is handled slightly differently from other variables—A style can only add functions to this hook, not remove them. See Style Variables.

Next: , Previous: Customizing Indentation, Up: Top

12 Customizing Macros

Preprocessor macros in C, C++, and Objective C (introduced by #define) have a syntax different from the main language—for example, a macro declaration is not terminated by a semicolon, and if it is more than a line long, line breaks in it must be escaped with backslashes. CC Mode has some commands to manipulate these, see Macro Backslashes.

Normally, the lines in a multi-line macro are indented relative to each other as though they were code. You can suppress this behavior by setting the following user option:

— User Option: c-syntactic-indentation-in-macros

Enable syntactic analysis inside macros, which is the default. If this is nil, all lines inside macro definitions are analyzed as cpp-macro-cont.

Because a macro can expand into anything at all, near where one is invoked CC Mode can only indent and fontify code heuristically. Sometimes it gets it wrong. Usually you should try to design your macros so that they ”look like ordinary code” when you invoke them. However, one situation is so common that CC Mode handles it specially: that is when certain macros needn't (or mustn't) be followed by a ‘;’. You need to configure CC Mode to handle these macros properly, see Macros with ;.

Next: , Previous: Custom Macros, Up: Custom Macros

12.1 Customizing Macro Backslashes

CC Mode provides some tools to help keep the line continuation backslashes in macros neat and tidy. Their precise action is customized with these variables:

— User Option: c-backslash-column

— User Option: c-backslash-max-column

These variables control the alignment columns for line continuation backslashes in multiline macros. They are used by the functions that automatically insert or align such backslashes, e.g., c-backslash-region and c-context-line-break.

c-backslash-column specifies the minimum column for the backslashes. If any line in the macro goes past this column, then the next tab stop (i.e., next multiple of tab-width) in that line is used as the alignment column for all the backslashes, so that they remain in a single column. However, if any lines go past c-backslash-max-column then the backslashes in the rest of the macro will be kept at that column, so that the lines which are too long “stick out” instead.

Don't ever set these variables to nil. If you want to disable the automatic alignment of backslashes, use c-auto-align-backslashes.

— User Option: c-auto-align-backslashes

Align automatically inserted line continuation backslashes if non-nil. When line continuation backslashes are inserted automatically for line breaks in multiline macros, e.g., by c-context-line-break, they are aligned with the other backslashes in the same macro if this flag is set.

If c-auto-align-backslashes is nil, automatically inserted backslashes are preceded by a single space, and backslashes get aligned only when you explicitly invoke the command c-backslash-region (C-c C-\).

Previous: Macro Backslashes, Up: Custom Macros

12.2 Macros with semicolons

Macros which needn't (or mustn't) be followed by a semicolon when you invoke them, macros with semicolons, are very common. These can cause CC Mode to parse the next line wrongly as a statement-cont (see Function Symbols) and thus mis-indent it.

You can prevent this by specifying which macros have semicolons. It doesn't matter whether or not such a macro has a parameter list:

— User Option: c-macro-names-with-semicolon

This buffer-local variable specifies which macros have semicolons. After setting its value, you need to call c-make-macro-with-semi-re for it to take effect. It should be set to one of these values:

nil
There are no macros with semicolons.
a list of strings
Each string is the name of a macro with a semicolon. Only valid #define names are allowed here. For example, to set the default value, you could write the following into your .emacs:
               (setq c-macro-names-with-semicolon
                     '("Q_OBJECT" "Q_PROPERTY" "Q_DECLARE" "Q_ENUMS"))

a regular expression
This matches each symbol which is a macro with a semicolon. It must not match any string which isn't a valid #define name. For example:
               (setq c-macro-names-with-semicolon
                     "\\<\\(CLEAN_UP_AND_RETURN\\|Q_[[:upper:]]+\\)\\>")

— Function: c-make-macro-with-semi-re

Call this (non-interactive) function, which sets internal variables, each time you change the value of c-macro-names-with-semicolon. It takes no arguments, and its return value has no meaning. This function is called by CC Mode's initialization code.

Next: , Previous: Custom Macros, Up: Top

13 Odds and Ends

The stuff that didn't fit in anywhere else is documented here.

— User Option: c-require-final-newline

Controls whether a final newline is enforced when the file is saved. The value is an association list that for each language mode specifies the value to give to require-final-newline (see Saving Buffers) at mode initialization. If a language isn't present on the association list, CC Mode won't touch require-final-newline in buffers for that language.

The default is to set require-final-newline to t in the languages that mandate that source files should end with newlines. These are C, C++ and Objective-C.

— User Option: c-echo-syntactic-information-p

If non-nil, the syntactic analysis for the current line is shown in the echo area when it's indented (unless c-syntactic-indentation is nil). That's useful when finding out which syntactic symbols to modify to get the indentation you want.

— User Option: c-report-syntactic-errors

If non-nil, certain syntactic errors are reported with a ding and a message, for example when an else is indented for which there is no corresponding if.

Note however that CC Mode doesn't make any special effort to check for syntactic errors; that's the job of the compiler. The reason it can report cases like the one above is that it can't find the correct anchoring position to indent the line in that case.

Next: , Previous: Odds and Ends, Up: Top

Appendix A Sample Init File

Here's a sample .emacs file fragment that might help you along the way. Just copy this region and paste it into your .emacs file. You might want to change some of the actual values.

;; Make a non-standard key binding.  We can put this in
;; c-mode-base-map because c-mode-map, c++-mode-map, and so on,
;; inherit from it.
(defun my-c-initialization-hook ()
  (define-key c-mode-base-map "\C-m" 'c-context-line-break))
(add-hook 'c-initialization-hook 'my-c-initialization-hook)

;; offset customizations not in my-c-style
;; This will take precedence over any setting of the syntactic symbol
;; made by a style.
(setq c-offsets-alist '((member-init-intro . ++)))

;; Create my personal style.
(defconst my-c-style
  '((c-tab-always-indent        . t)
    (c-comment-only-line-offset . 4)
    (c-hanging-braces-alist     . ((substatement-open after)
                                   (brace-list-open)))
    (c-hanging-colons-alist     . ((member-init-intro before)
                                   (inher-intro)
                                   (case-label after)
                                   (label after)
                                   (access-label after)))
    (c-cleanup-list             . (scope-operator
                                   empty-defun-braces
                                   defun-close-semi))
    (c-offsets-alist            . ((arglist-close . c-lineup-arglist)
                                   (substatement-open . 0)
                                   (case-label        . 4)
                                   (block-open        . 0)
                                   (knr-argdecl-intro . -)))
    (c-echo-syntactic-information-p . t))
  "My C Programming Style")
(c-add-style "PERSONAL" my-c-style)

;; Customizations for all modes in CC Mode.
(defun my-c-mode-common-hook ()
  ;; set my personal style for the current buffer
  (c-set-style "PERSONAL")
  ;; other customizations
  (setq tab-width 8
        ;; this will make sure spaces are used instead of tabs
        indent-tabs-mode nil)
  ;; we like auto-newline, but not hungry-delete
  (c-toggle-auto-newline 1))
(add-hook 'c-mode-common-hook 'my-c-mode-common-hook)

Next: , Previous: Sample Init File, Up: Top

2 Performance Issues

C and its derivative languages are highly complex creatures. Often, ambiguous code situations arise that require CC Mode to scan large portions of the buffer to determine syntactic context. Such pathological code can cause CC Mode to perform fairly badly. This section gives some insight in how CC Mode operates, how that interacts with some coding styles, and what you can use to improve performance.

The overall goal is that CC Mode shouldn't be overly slow (i.e., take more than a fraction of a second) in any interactive operation. I.e., it's tuned to limit the maximum response time in single operations, which is sometimes at the expense of batch-like operations like reindenting whole blocks. If you find that CC Mode gradually gets slower and slower in certain situations, perhaps as the file grows in size or as the macro or comment you're editing gets bigger, then chances are that something isn't working right. You should consider reporting it, unless it's something that's mentioned in this section.

Because CC Mode has to scan the buffer backwards from the current insertion point, and because C's syntax is fairly difficult to parse in the backwards direction, CC Mode often tries to find the nearest position higher up in the buffer from which to begin a forward scan (it's typically an opening or closing parenthesis of some kind). The farther this position is from the current insertion point, the slower it gets.

In earlier versions of CC Mode, we used to recommend putting the opening brace of a top-level construct49 into the leftmost column. Earlier still, this used to be a rigid Emacs constraint, as embodied in the beginning-of-defun function. CC Mode now caches syntactic information much better, so that the delay caused by searching for such a brace when it's not in column 0 is minimal, except perhaps when you've just moved a long way inside the file.

A special note about defun-prompt-regexp in Java mode: The common style is to hang the opening braces of functions and classes on the right side of the line, and that doesn't work well with the Emacs approach. CC Mode comes with a constant c-Java-defun-prompt-regexp which tries to define a regular expression usable for this style, but there are problems with it. In some cases it can cause beginning-of-defun to hang50. For this reason, it is not used by default, but if you feel adventurous, you can set defun-prompt-regexp to it in your mode hook. In any event, setting and relying on defun-prompt-regexp will definitely slow things down because (X)Emacs will be doing regular expression searches a lot, so you'll probably be taking a hit either way!

CC Mode maintains a cache of the opening parentheses of the blocks surrounding the point, and it adapts that cache as the point is moved around. That means that in bad cases it can take noticeable time to indent a line in a new surrounding, but after that it gets fast as long as the point isn't moved far off. The farther the point is moved, the less useful is the cache. Since editing typically is done in “chunks” rather than on single lines far apart from each other, the cache typically gives good performance even when the code doesn't fit the Emacs approach to finding the defun starts.

XEmacs users can set the variable c-enable-xemacs-performance-kludge-p to non-nil. This tells CC Mode to use XEmacs-specific built-in functions which, in some circumstances, can locate the top-most opening brace much more quickly than beginning-of-defun. Preliminary testing has shown that for styles where these braces are hung (e.g., most JDK-derived Java styles), this hack can improve performance of the core syntax parsing routines from 3 to 60 times. However, for styles which do conform to Emacs's recommended style of putting top-level braces in column zero, this hack can degrade performance by about as much. Thus this variable is set to nil by default, since the Emacs-friendly styles should be more common (and encouraged!). Note that this variable has no effect in Emacs since the necessary built-in functions don't exist (in Emacs 22.1 as of this writing in February 2007).

Text properties are used to speed up skipping over syntactic whitespace, i.e., comments and preprocessor directives. Indenting a line after a huge macro definition can be slow the first time, but after that the text properties are in place and it should be fast (even after you've edited other parts of the file and then moved back).

Font locking can be a CPU hog, especially the font locking done on decoration level 3 which tries to be very accurate. Note that that level is designed to be used with a font lock support mode that only fontifies the text that's actually shown, i.e., Lazy Lock or Just-in-time Lock mode, so make sure you use one of them. Fontification of a whole buffer with some thousand lines can often take over a minute. That is a known weakness; the idea is that it never should happen.

The most effective way to speed up font locking is to reduce the decoration level to 2 by setting font-lock-maximum-decoration appropriately. That level is designed to be as pretty as possible without sacrificing performance. See Font Locking Preliminaries, for more info.

Next: , Previous: Performance Issues, Up: Top

3 Limitations and Known Bugs

Next: , Previous: Limitations and Known Bugs, Up: Top

Appendix A Frequently Asked Questions

Next: , Previous: FAQ, Up: Top

Appendix B Getting the Latest CC Mode Release

CC Mode has been standard with all versions of Emacs since 19.34 and of XEmacs since 19.16.

Due to release schedule skew, it is likely that all of these Emacsen have old versions of CC Mode and so should be upgraded. Access to the CC Mode source code, as well as more detailed information on Emacsen compatibility, etc. are all available on the web site:

http://cc-mode.sourceforge.net/

Next: , Previous: Updating CC Mode, Up: Top

Appendix C Mailing Lists and Submitting Bug Reports

To report bugs, use the C-c C-b (bound to c-submit-bug-report) command. This provides vital information we need to reproduce your problem. Make sure you include a concise, but complete code example. Please try to boil your example down to just the essential code needed to reproduce the problem, and include an exact recipe of steps needed to expose the bug. Be especially sure to include any code that appears before your bug example, if you think it might affect our ability to reproduce it.

Please try to produce the problem in an Emacs instance without any customizations loaded (i.e., start it with the ‘-q --no-site-file’ arguments). If it works correctly there, the problem might be caused by faulty customizations in either your own or your site configuration. In that case, we'd appreciate it if you isolate the Emacs Lisp code that triggers the bug and include it in your report.

Bug reports should be sent to bug-cc-mode@gnu.org. You can also send other questions and suggestions (kudos? ;-) to that address. It's a mailing list which you can join or browse an archive of; see the web site at http://cc-mode.sourceforge.net/ for further details.

If you want to get announcements of new CC Mode releases, send the word subscribe in the body of a message to cc-mode-announce-request@lists.sourceforge.net. It's possible to subscribe from the web site too. Announcements will also be posted to the Usenet newsgroups gnu.emacs.sources, comp.emacs, comp.emacs.xemacs, comp.lang.c, comp.lang.c++, comp.lang.objective-c, comp.lang.java.softwaretools, comp.lang.idl, and comp.lang.awk.

Next: , Previous: Mailing Lists and Bug Reports, Up: Top

Appendix D GNU Free Documentation License

Version 1.3, 3 November 2008
     Copyright © 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc.
     http://fsf.org/
     
     Everyone is permitted to copy and distribute verbatim copies
     of this license document, but changing it is not allowed.
  1. PREAMBLE

    The purpose of this License is to make a manual, textbook, or other functional and useful document free in the sense of freedom: to assure everyone the effective freedom to copy and redistribute it, with or without modifying it, either commercially or noncommercially. Secondarily, this License preserves for the author and publisher a way to get credit for their work, while not being considered responsible for modifications made by others.

    This License is a kind of “copyleft”, which means that derivative works of the document must themselves be free in the same sense. It complements the GNU General Public License, which is a copyleft license designed for free software.

    We have designed this License in order to use it for manuals for free software, because free software needs free documentation: a free program should come with manuals providing the same freedoms that the software does. But this License is not limited to software manuals; it can be used for any textual work, regardless of subject matter or whether it is published as a printed book. We recommend this License principally for works whose purpose is instruction or reference.

  2. APPLICABILITY AND DEFINITIONS

    This License applies to any manual or other work, in any medium, that contains a notice placed by the copyright holder saying it can be distributed under the terms of this License. Such a notice grants a world-wide, royalty-free license, unlimited in duration, to use that work under the conditions stated herein. The “Document”, below, refers to any such manual or work. Any member of the public is a licensee, and is addressed as “you”. You accept the license if you copy, modify or distribute the work in a way requiring permission under copyright law.

    A “Modified Version” of the Document means any work containing the Document or a portion of it, either copied verbatim, or with modifications and/or translated into another language.

    A “Secondary Section” is a named appendix or a front-matter section of the Document that deals exclusively with the relationship of the publishers or authors of the Document to the Document's overall subject (or to related matters) and contains nothing that could fall directly within that overall subject. (Thus, if the Document is in part a textbook of mathematics, a Secondary Section may not explain any mathematics.) The relationship could be a matter of historical connection with the subject or with related matters, or of legal, commercial, philosophical, ethical or political position regarding them.

    The “Invariant Sections” are certain Secondary Sections whose titles are designated, as being those of Invariant Sections, in the notice that says that the Document is released under this License. If a section does not fit the above definition of Secondary then it is not allowed to be designated as Invariant. The Document may contain zero Invariant Sections. If the Document does not identify any Invariant Sections then there are none.

    The “Cover Texts” are certain short passages of text that are listed, as Front-Cover Texts or Back-Cover Texts, in the notice that says that the Document is released under this License. A Front-Cover Text may be at most 5 words, and a Back-Cover Text may be at most 25 words.

    A “Transparent” copy of the Document means a machine-readable copy, represented in a format whose specification is available to the general public, that is suitable for revising the document straightforwardly with generic text editors or (for images composed of pixels) generic paint programs or (for drawings) some widely available drawing editor, and that is suitable for input to text formatters or for automatic translation to a variety of formats suitable for input to text formatters. A copy made in an otherwise Transparent file format whose markup, or absence of markup, has been arranged to thwart or discourage subsequent modification by readers is not Transparent. An image format is not Transparent if used for any substantial amount of text. A copy that is not “Transparent” is called “Opaque”.

    Examples of suitable formats for Transparent copies include plain ASCII without markup, Texinfo input format, LaTeX input format, SGML or XML using a publicly available DTD, and standard-conforming simple HTML, PostScript or PDF designed for human modification. Examples of transparent image formats include PNG, XCF and JPG. Opaque formats include proprietary formats that can be read and edited only by proprietary word processors, SGML or XML for which the DTD and/or processing tools are not generally available, and the machine-generated HTML, PostScript or PDF produced by some word processors for output purposes only.

    The “Title Page” means, for a printed book, the title page itself, plus such following pages as are needed to hold, legibly, the material this License requires to appear in the title page. For works in formats which do not have any title page as such, “Title Page” means the text near the most prominent appearance of the work's title, preceding the beginning of the body of the text.

    The “publisher” means any person or entity that distributes copies of the Document to the public.

    A section “Entitled XYZ” means a named subunit of the Document whose title either is precisely XYZ or contains XYZ in parentheses following text that translates XYZ in another language. (Here XYZ stands for a specific section name mentioned below, such as “Acknowledgements”, “Dedications”, “Endorsements”, or “History”.) To “Preserve the Title” of such a section when you modify the Document means that it remains a section “Entitled XYZ” according to this definition.

    The Document may include Warranty Disclaimers next to the notice which states that this License applies to the Document. These Warranty Disclaimers are considered to be included by reference in this License, but only as regards disclaiming warranties: any other implication that these Warranty Disclaimers may have is void and has no effect on the meaning of this License.

  3. VERBATIM COPYING

    You may copy and distribute the Document in any medium, either commercially or noncommercially, provided that this License, the copyright notices, and the license notice saying this License applies to the Document are reproduced in all copies, and that you add no other conditions whatsoever to those of this License. You may not use technical measures to obstruct or control the reading or further copying of the copies you make or distribute. However, you may accept compensation in exchange for copies. If you distribute a large enough number of copies you must also follow the conditions in section 3.

    You may also lend copies, under the same conditions stated above, and you may publicly display copies.

  4. COPYING IN QUANTITY

    If you publish printed copies (or copies in media that commonly have printed covers) of the Document, numbering more than 100, and the Document's license notice requires Cover Texts, you must enclose the copies in covers that carry, clearly and legibly, all these Cover Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on the back cover. Both covers must also clearly and legibly identify you as the publisher of these copies. The front cover must present the full title with all words of the title equally prominent and visible. You may add other material on the covers in addition. Copying with changes limited to the covers, as long as they preserve the title of the Document and satisfy these conditions, can be treated as verbatim copying in other respects.

    If the required texts for either cover are too voluminous to fit legibly, you should put the first ones listed (as many as fit reasonably) on the actual cover, and continue the rest onto adjacent pages.

    If you publish or distribute Opaque copies of the Document numbering more than 100, you must either include a machine-readable Transparent copy along with each Opaque copy, or state in or with each Opaque copy a computer-network location from which the general network-using public has access to download using public-standard network protocols a complete Transparent copy of the Document, free of added material. If you use the latter option, you must take reasonably prudent steps, when you begin distribution of Opaque copies in quantity, to ensure that this Transparent copy will remain thus accessible at the stated location until at least one year after the last time you distribute an Opaque copy (directly or through your agents or retailers) of that edition to the public.

    It is requested, but not required, that you contact the authors of the Document well before redistributing any large number of copies, to give them a chance to provide you with an updated version of the Document.

  5. MODIFICATIONS

    You may copy and distribute a Modified Version of the Document under the conditions of sections 2 and 3 above, provided that you release the Modified Version under precisely this License, with the Modified Version filling the role of the Document, thus licensing distribution and modification of the Modified Version to whoever possesses a copy of it. In addition, you must do these things in the Modified Version:

    1. Use in the Title Page (and on the covers, if any) a title distinct from that of the Document, and from those of previous versions (which should, if there were any, be listed in the History section of the Document). You may use the same title as a previous version if the original publisher of that version gives permission.
    2. List on the Title Page, as authors, one or more persons or entities responsible for authorship of the modifications in the Modified Version, together with at least five of the principal authors of the Document (all of its principal authors, if it has fewer than five), unless they release you from this requirement.
    3. State on the Title page the name of the publisher of the Modified Version, as the publisher.
    4. Preserve all the copyright notices of the Document.
    5. Add an appropriate copyright notice for your modifications adjacent to the other copyright notices.
    6. Include, immediately after the copyright notices, a license notice giving the public permission to use the Modified Version under the terms of this License, in the form shown in the Addendum below.
    7. Preserve in that license notice the full lists of Invariant Sections and required Cover Texts given in the Document's license notice.
    8. Include an unaltered copy of this License.
    9. Preserve the section Entitled “History”, Preserve its Title, and add to it an item stating at least the title, year, new authors, and publisher of the Modified Version as given on the Title Page. If there is no section Entitled “History” in the Document, create one stating the title, year, authors, and publisher of the Document as given on its Title Page, then add an item describing the Modified Version as stated in the previous sentence.
    10. Preserve the network location, if any, given in the Document for public access to a Transparent copy of the Document, and likewise the network locations given in the Document for previous versions it was based on. These may be placed in the “History” section. You may omit a network location for a work that was published at least four years before the Document itself, or if the original publisher of the version it refers to gives permission.
    11. For any section Entitled “Acknowledgements” or “Dedications”, Preserve the Title of the section, and preserve in the section all the substance and tone of each of the contributor acknowledgements and/or dedications given therein.
    12. Preserve all the Invariant Sections of the Document, unaltered in their text and in their titles. Section numbers or the equivalent are not considered part of the section titles.
    13. Delete any section Entitled “Endorsements”. Such a section may not be included in the Modified Version.
    14. Do not retitle any existing section to be Entitled “Endorsements” or to conflict in title with any Invariant Section.
    15. Preserve any Warranty Disclaimers.

    If the Modified Version includes new front-matter sections or appendices that qualify as Secondary Sections and contain no material copied from the Document, you may at your option designate some or all of these sections as invariant. To do this, add their titles to the list of Invariant Sections in the Modified Version's license notice. These titles must be distinct from any other section titles.

    You may add a section Entitled “Endorsements”, provided it contains nothing but endorsements of your Modified Version by various parties—for example, statements of peer review or that the text has been approved by an organization as the authoritative definition of a standard.

    You may add a passage of up to five words as a Front-Cover Text, and a passage of up to 25 words as a Back-Cover Text, to the end of the list of Cover Texts in the Modified Version. Only one passage of Front-Cover Text and one of Back-Cover Text may be added by (or through arrangements made by) any one entity. If the Document already includes a cover text for the same cover, previously added by you or by arrangement made by the same entity you are acting on behalf of, you may not add another; but you may replace the old one, on explicit permission from the previous publisher that added the old one.

    The author(s) and publisher(s) of the Document do not by this License give permission to use their names for publicity for or to assert or imply endorsement of any Modified Version.

  6. COMBINING DOCUMENTS

    You may combine the Document with other documents released under this License, under the terms defined in section 4 above for modified versions, provided that you include in the combination all of the Invariant Sections of all of the original documents, unmodified, and list them all as Invariant Sections of your combined work in its license notice, and that you preserve all their Warranty Disclaimers.

    The combined work need only contain one copy of this License, and multiple identical Invariant Sections may be replaced with a single copy. If there are multiple Invariant Sections with the same name but different contents, make the title of each such section unique by adding at the end of it, in parentheses, the name of the original author or publisher of that section if known, or else a unique number. Make the same adjustment to the section titles in the list of Invariant Sections in the license notice of the combined work.

    In the combination, you must combine any sections Entitled “History” in the various original documents, forming one section Entitled “History”; likewise combine any sections Entitled “Acknowledgements”, and any sections Entitled “Dedications”. You must delete all sections Entitled “Endorsements.”

  7. COLLECTIONS OF DOCUMENTS

    You may make a collection consisting of the Document and other documents released under this License, and replace the individual copies of this License in the various documents with a single copy that is included in the collection, provided that you follow the rules of this License for verbatim copying of each of the documents in all other respects.

    You may extract a single document from such a collection, and distribute it individually under this License, provided you insert a copy of this License into the extracted document, and follow this License in all other respects regarding verbatim copying of that document.

  8. AGGREGATION WITH INDEPENDENT WORKS

    A compilation of the Document or its derivatives with other separate and independent documents or works, in or on a volume of a storage or distribution medium, is called an “aggregate” if the copyright resulting from the compilation is not used to limit the legal rights of the compilation's users beyond what the individual works permit. When the Document is included in an aggregate, this License does not apply to the other works in the aggregate which are not themselves derivative works of the Document.

    If the Cover Text requirement of section 3 is applicable to these copies of the Document, then if the Document is less than one half of the entire aggregate, the Document's Cover Texts may be placed on covers that bracket the Document within the aggregate, or the electronic equivalent of covers if the Document is in electronic form. Otherwise they must appear on printed covers that bracket the whole aggregate.

  9. TRANSLATION

    Translation is considered a kind of modification, so you may distribute translations of the Document under the terms of section 4. Replacing Invariant Sections with translations requires special permission from their copyright holders, but you may include translations of some or all Invariant Sections in addition to the original versions of these Invariant Sections. You may include a translation of this License, and all the license notices in the Document, and any Warranty Disclaimers, provided that you also include the original English version of this License and the original versions of those notices and disclaimers. In case of a disagreement between the translation and the original version of this License or a notice or disclaimer, the original version will prevail.

    If a section in the Document is Entitled “Acknowledgements”, “Dedications”, or “History”, the requirement (section 4) to Preserve its Title (section 1) will typically require changing the actual title.

  10. TERMINATION

    You may not copy, modify, sublicense, or distribute the Document except as expressly provided under this License. Any attempt otherwise to copy, modify, sublicense, or distribute it is void, and will automatically terminate your rights under this License.

    However, if you cease all violation of this License, then your license from a particular copyright holder is reinstated (a) provisionally, unless and until the copyright holder explicitly and finally terminates your license, and (b) permanently, if the copyright holder fails to notify you of the violation by some reasonable means prior to 60 days after the cessation.

    Moreover, your license from a particular copyright holder is reinstated permanently if the copyright holder notifies you of the violation by some reasonable means, this is the first time you have received notice of violation of this License (for any work) from that copyright holder, and you cure the violation prior to 30 days after your receipt of the notice.

    Termination of your rights under this section does not terminate the licenses of parties who have received copies or rights from you under this License. If your rights have been terminated and not permanently reinstated, receipt of a copy of some or all of the same material does not give you any rights to use it.

  11. FUTURE REVISIONS OF THIS LICENSE

    The Free Software Foundation may publish new, revised versions of the GNU Free Documentation License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. See http://www.gnu.org/copyleft/.

    Each version of the License is given a distinguishing version number. If the Document specifies that a particular numbered version of this License “or any later version” applies to it, you have the option of following the terms and conditions either of that specified version or of any later version that has been published (not as a draft) by the Free Software Foundation. If the Document does not specify a version number of this License, you may choose any version ever published (not as a draft) by the Free Software Foundation. If the Document specifies that a proxy can decide which future versions of this License can be used, that proxy's public statement of acceptance of a version permanently authorizes you to choose that version for the Document.

  12. RELICENSING

    “Massive Multiauthor Collaboration Site” (or “MMC Site”) means any World Wide Web server that publishes copyrightable works and also provides prominent facilities for anybody to edit those works. A public wiki that anybody can edit is an example of such a server. A “Massive Multiauthor Collaboration” (or “MMC”) contained in the site means any set of copyrightable works thus published on the MMC site.

    “CC-BY-SA” means the Creative Commons Attribution-Share Alike 3.0 license published by Creative Commons Corporation, a not-for-profit corporation with a principal place of business in San Francisco, California, as well as future copyleft versions of that license published by that same organization.

    “Incorporate” means to publish or republish a Document, in whole or in part, as part of another Document.

    An MMC is “eligible for relicensing” if it is licensed under this License, and if all works that were first published under this License somewhere other than this MMC, and subsequently incorporated in whole or in part into the MMC, (1) had no cover texts or invariant sections, and (2) were thus incorporated prior to November 1, 2008.

    The operator of an MMC Site may republish an MMC contained in the site under CC-BY-SA on the same site at any time before August 1, 2009, provided the MMC is eligible for relicensing.

ADDENDUM: How to use this License for your documents

To use this License in a document you have written, include a copy of the License in the document and put the following copyright and license notices just after the title page:

       Copyright (C)  year  your name.
       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 no Invariant Sections, no Front-Cover Texts, and no Back-Cover
       Texts.  A copy of the license is included in the section entitled ``GNU
       Free Documentation License''.

If you have Invariant Sections, Front-Cover Texts and Back-Cover Texts, replace the “with...Texts.” line with this:

         with the Invariant Sections being list their titles, with
         the Front-Cover Texts being list, and with the Back-Cover Texts
         being list.

If you have Invariant Sections without Cover Texts, or some other combination of the three, merge those two alternatives to suit the situation.

If your document contains nontrivial examples of program code, we recommend releasing these examples in parallel under your choice of free software license, such as the GNU General Public License, to permit their use in free software.

Next: , Previous: GNU Free Documentation License, Up: Top

Command and Function Index

Since most CC Mode commands are prepended with the string ‘c-’, each appears under its c-thing name and its thing (c-) name.

Next: , Previous: Command and Function Index, Up: Top

Variable Index

Since most CC Mode variables are prepended with the string ‘c-’, each appears under its c-thing name and its thing (c-) name.

Previous: Variable Index, Up: Top

Concept and Key Index


Footnotes

[1] A C-like scripting language with its roots in the LPC language used in some MUD engines. See http://pike.ida.liu.se/.

[2] There is no “easy customization” facility for making this change.

[3] this is only useful for a line starting with a comment opener or an opening brace, parenthesis, or string quote.

[4] The name of this command varies between (X)Emacs versions.

[5] this was CC Mode's behavior prior to version 5.32.

[6] You can change this default by setting the string syntactic symbol (see Syntactic Symbols and see Customizing Indentation)

[7] In GCC, unescaped line breaks within strings are valid.

[8] The ‘C’ would be replaced with the name of the language in question for the other languages CC Mode supports.

[9] Prior to CC Mode 5.31, this command was bound to C-c C-d.

[10] Prior to CC Mode 5.31, this command was bound to C-c C-t.

[11] A literal is defined as any comment, string, or preprocessor macro definition. These constructs are also known as syntactic whitespace since they are usually ignored when scanning C code.

[12] Prior to CC Mode 5.31, this command was bound to C-c C-d. C-c C-d is now the default binding for c-hungry-delete-forward.

[13] This command was formerly known as c-hungry-backspace.

[14] DON'T PANIC!!! This isn't difficult.

[15] In earlier versions of CC Mode, a File Style setting took precedence over any other setting apart from a File Local Variable setting.

[16] This is a big change from versions of CC Mode earlier than 5.26, where such settings would get overridden by the style system unless special precautions were taken. That was changed since it was counterintuitive and confusing, especially to novice users. If your configuration depends on the old overriding behavior, you can set the variable c-old-style-variable-behavior to non-nil.

[17] This did not change in version 5.26.

[18] This document is available at http://www.doc.ic.ac.uk/lab/cplus/c++.rules/ among other places.

[19] Python is a high level scripting language with a C/C++ foreign function interface. For more information, see http://www.python.org/.

[20] This table is stored internally in the variable c-fallback-style.

[21] Note that if the variable has been given a value by the Customization interface or a setq at the top level of your .emacs, this value will override the one the style system tries to give it. See Config Basics.

[22] Also, if either of these are set in a file's local variable section, all the style variable values are made local to that buffer, even if c-style-variables-are-local-p is nil. Since this variable is virtually always non-nil anyhow, you're unlikely to notice this effect.

[23] comment-start, comment-end, comment-start-skip, paragraph-start, paragraph-separate, paragraph-ignore-fill-prefix, adaptive-fill-mode, adaptive-fill-regexp, and adaptive-fill-first-line-regexp.

[24] It's available from http://www.wonderworks.com/. As of version 2.12, it does however lack a feature that makes it work suboptimally when c-comment-prefix-regexp matches the empty string (which it does by default). A patch for that is available from the CC Mode web site.

[25] In versions before 5.26, this variable was called c-comment-continuation-stars. As a compatibility measure, CC Mode still uses the value on that variable if it's set.

[26] Actually, this default setting of c-block-comment-prefix typically gets overridden by the default style gnu, which sets it to blank. You can see the line splitting effect described here by setting a different style, e.g., k&r See Choosing a Style.

[27] Also insert a ‘\’ at the end of the previous line if you're in AWK Mode.

[28] The braces of anonymous classes produce a combination of inexpr-class, and class-open or class-close in normal indentation analysis.

[29] Brace lists inside statements, such as initializers for static array variables inside functions in C, are recognized as statement-cont. All normal substatement blocks are recognized with other symbols.

[30] This was first introduced in CC Mode 5.31.

[31] Certain C++ constructs introduce ambiguous situations, so scope-operator clean-ups might not always be correct. This usually only occurs when scoped identifiers appear in switch label tags.

[32] In CC Mode 5.28 and earlier, a syntactic element was a dotted pair; the cons was the syntactic symbol and the cdr was the anchor position. For compatibility's sake, the parameter passed to a line-up function still has this dotted pair form (see Custom Line-Up).

[33] The line numbers in this and future examples don't actually appear in the buffer, of course!

[34] With a universal argument (i.e., C-u C-c C-s) the analysis is inserted into the buffer as a comment on the current line.

[35] A substatement is the line after a conditional statement, such as if, else, while, do, switch, etc. A substatement block is a brace block following one of these conditional statements.

[36] This is the case even for C and Objective-C. For consistency, structs in all supported languages are syntactically equivalent to classes. Note however that the keyword class is meaningless in C and Objective-C.

[37] block-open is used only for “free-standing” blocks, and is somewhat rare (see Literal Symbols for an example.)

[38] These should logically be named extern-open, extern-close and inextern, but that isn't the case for historical reasons.

[39] This is how CC Mode 5.28 and earlier analyzed macros.

[40] You might wonder why it doesn't get inlambda too. It's because the closing brace is relative to the opening brace, which stands on its own line in this example. If the opening brace was hanging on the previous line, then the closing brace would get the inlambda syntax too to be indented correctly.

[41] a.k.a. K&R C, or Kernighan & Ritchie C

[42] The syntactic context ((defun-block-intro 2724) (comment-intro)) would likely have two relative offsets.

[43] There is however a variable c-strict-syntax-p that when set to non-nil will cause an error to be signaled in that case. It's now considered obsolete since it doesn't work well with some of the alignment functions that return nil instead of zero. You should therefore leave c-strict-syntax-p set to nil.

[44] In this and subsequent examples, the original code is formatted using the ‘gnu’ style unless otherwise indicated. See Styles.

[45] Run-in style doesn't really work too well. You might need to write your own custom line-up functions to better support this style.

[46] This function is mainly provided to mimic the behavior of CC Mode 5.28 and earlier where this case wasn't handled consistently so that those lines could be analyzed as either topmost-intro-cont or statement-cont. It's used for topmost-intro-cont by default, but you might consider using + instead.

[47] Returning nil is useful when the offset specification for a syntactic element is a list containing the line-up function (see c-offsets-alist).

[48] The numerical value returned by point will change if you change the indentation of the line within a save-excursion form, but point itself will still be over the same piece of text.

[49] E.g., a function in C, or outermost class definition in C++ or Java.

[50] This has been observed in Emacs 19.34 and XEmacs 19.15.

[51] Using the variable open-paren-in-column-0-is-defun-start.