AutoFSM is an add-on package to AutoGen.
AutoGen actually includes AutoFSM because it uses its functionality itself.
When you install AutoGen, the Finite State Machine templates
will be installed in
Finite state machines can be constructed in a variety of ways.
Some of them can be assisted with generated text and others cannot.
Or, at least, they cannot be generated very well. What can be
generated very well are the transition tables derived from
the machine definition.
Here is the list of attributes that various flavors of software finite
state machines may have.
- Transition Event Detection
There is no way to select among these detection methods.
``state-independent'' is assumed, but you have ``adjustments''
available to you when your transition handling code is invoked.
- If the transition event can be recognized without reference
to the current machine state, it is state-independent.
If you cannot determine the next transition event without knowing the state,
then it is state-dependent.
- Generally, transition tables are not
very helpful for a state-dependent machine.
- Very often, you can mostly determine the transition event
independently of the current state. For these occasions, this
Finite State Machine generator will allow you to ``adjust'' the
state during transition processing.
- FSM Type
- The finite state machine may be either the master, driving the
other parts of the program, or act as a subroutine keeping track
of state between calls. Consequently, the type attribute
may be set to:
- If the machine processes transitions until
it reaches a terminal state (error or done).
- If the FSM code must only process a single transition
and then return.
- If the application must run the FSM from multiple independent
threads and the FSM code must only process a single transition
and then return.
- implementation method
- AutoFSM gives you a choice of implementations.
You can specify which with the method attribute. If you
omit the attribute or specify ``none'', then only a header file
containing a transition table will be produced.
- This method will use a callout table instead of a switch
statement to implement the machine.
- This is the alternate implementation method.
- Choosing this method suppresses the prototype implementation.
AutoFSM supports all four variations of the state-independent
or nearly-state-independent transition machines, plus the
transition table only output.
To use this, you must set up a Finite State Machine AutoGen definitions file that defines:
- a list of
- a list of
- a list of
states are simply names you
provide for the transition types and processing states. The
transition entries are entries that specify an initial state, the
result state and a transition type. By default, the transition type is
named after the original state and the transition event. However, you may
also specify a name in order to allow several different transitions to be
handled with the same code.
The output is currently a set of defines, a transition table
suitable for compiling into a C or C++ program, and, optionally, a
prototype implementation. There is nothing to restrict output in
another language, there just happens not to be any templates to
produce other languages at the moment. Watch this space if you have
interest in finite state machines in other languages.
Here is an example of a
method machine, and a usage page for a description
of the FSM attributes. AutoGen itself uses AutoFSM for three parsing
functions. Only the two .c files are linked to here as the
generated .h header files are not archived under CVS:
|agen5/cgi.def || ||produces cgi-fsm.c and cgi-fsm.h. These are used to scan CGI script values for conversion into AutoGen definitions.|
|pseudo.def || ||produces pseudo-fsm.h, which is used in the file loadPseudo.c. That FSM does not produce any code. The table produced is used to parse the pseudo macro at the start of a template.|
|defParse.def || ||produces defParse-fsm.c and defParse-fsm.h. This code drives the parsing of the definition files.|