1 Ragel 6.5 - May 18, 2009
2 ========================
3 -Fixed a bug in graphviz generation. Ragel crashed when using -V and -M and
4 the specified machine referenced another machine that wasn't included in the
6 -The name "CS" is in use on OpenSolaris, changed to vCS to ease compiling
8 -Converted to automake.
9 -REALLY fixed a bug that was intended to be fixed in 6.4:
10 Fixed a problem reading hex numbers that have the high bit set when the
11 alphabet is signed and we are on 64 bit. This was reported by _why. The
12 fix was provided by Wialliam Morgan. The literal 0xffffffff was used for
13 a fully set long when -1L should be used instead.
14 A null patch (whitespace changes) must have gotten checked after I was
15 testing with and without the critical one-line patch and I forgot to enable
16 make sure it was enabled in the final checkin version.
18 Ragel 6.4 - Mar 22, 2009
19 ========================
20 -Moved back to a single executable. The old intermediate format can still be
21 generated using the -x option. Ragel was split into frontend and backend
22 programs in version 5.0. This was done to encourage interoperability with
23 other tools. Since then, ragel has been made to work with qfsm, with ragel
24 producing the intermediate format and qfsm consuming it. However, there has
25 been no use of Ragel as a consumer of state machine data, with Ragel used as
26 a code generator for DFAs. This is not surprising given that much of the
27 complexity of Ragel is in the frontend, where the regular language to DFA
28 compilation happens. Since the full benefits of the split have not
29 materialized, and the split increases the complexity for users, Ragel has
30 been made once again into a single executable.
31 -Applied a fix to the documentation Makefile from John D. Mitchell.
32 -Use CXXFLAGS instead of CFLAGS for C++ compiling. Patch from Diego
34 -Added support for DESTDIR variable. Patch from Diego 'Flameeyes' Pettenò.
35 -Added a script called unicode2ragel.rb for generating unicode machines to
36 the examples directory. From Rakan El-Khalil.
37 -Fixed a copy-paste error in the documentation that was reported by Jose
39 -Added three new write commands:
43 These generate a reference to the start, first final and error state. When
44 there are many different machine specifications in one file it is easy to
45 get the prefix for these wrong (especially when you do a lot of copy-pasting
46 of boilerplate). The problem can be avoided by using write commands.
47 -Fixed a problem reading hex numbers that have the high bit set when the
48 alphabet is signed and we are on 64 bit. This was reported by _why. The fix
49 was provided by Wialliam Morgan. The literal 0xffffffff was used for a fully
50 set long when -1L should be used instead.
52 Ragel 6.3 - Aug 29, 2008
53 ========================
54 -Fixed an assertion that is too strong. In the condition code we need to copy
55 transitions that have non-empty lmActionTable arrays so we don't assert
56 emptiness in the constructor. Lift out the assertion and copy the array in
58 -Fixed and improved multiple include prevention. We now track the entire
59 include history of a parser state to prevent duplicates.
60 -Fixed crash on failed lookup of goto/call/etc target.
62 Ragel 6.2 - May 9, 2008
63 =======================
64 -Bug fix: The lm_switch actions needs to set p from tokend when there is no
66 -Bug fix: when not using indicies we can't use a transitions's id to identify
67 the eof transition to take. Instead add the transition to the end of the
68 transition list and store its position in a new var called pos. The pos var
69 is then used as the index.
70 -Bug fix: an fnext followed by an fbreak in -G2 was not working. The fbreak
71 was not aware that the fnext causes the cs variable to be forced active. In
72 this case fbreak does not need to save cs because it is already current.
73 -Bug fix: need to compute the low and high character-space keys from the
74 condition-trans overlap when computing an expansion. Can't use the range
75 supplied from the condition overlap since they may not match. An incorrect
76 machine that accepted 1(!cond1, !cond2) was generated for the following
77 grammar. This bug was reported by Tim Chklovski.
79 sc1 = 1..2 when cond1;
80 sc2 = 1..2 when cond2;
81 main := sc1 | c | sc2;
82 -Bug fix: error messages in start label analysis of join operations were
83 causing assertion failures because location info was not set. Fixed by
85 -Include and import file searching now searches for the file name given based
86 on the location of the current file, not ragel's current path.
87 Additional search locations can be given using the -I option.
88 -Rubinius code generation was updated to the latest Rubinius. Patch from Evan
90 -Switched from strcasecmp to strcmp for testing long arguments.
91 -Applied a patch from Andrei Polushin for setting the error message format.
92 --error-format=gnu (default)
94 -Now using the _WIN32 define instead of _WIN32. Other MSVC compilation
95 improvments from Andrei Polushin.
96 -Added the hyperref package to the manual.
98 Ragel 6.1 - Mar 26, 2008
99 ========================
100 -Scanners now ensure that any leaving actions at the end of a pattern are
101 executed. They are always executed before the pattern action.
102 -Added an option -d for turning off the removal of duplicate actions from
104 -Need to unset the final state status of the start state in kleene star if it
105 is set. It is possible to crash ragel when the warning is ignored.
106 -In the dot file generation we need to print any actions that are in
107 State::eofTrans. These come from scanners only.
108 -Use @docdir@ for the docdir Makefile variable.
109 -Check for ar and ranlib in the configure script.
111 Ragel 6.0 - Jan 12, 2008
112 ========================
113 -Removed the 'noend' write option from examples/atoi.rl. This example is
114 referenced a lot as a first example and as such it shouldn't contain a
115 special purpose write option like 'noend'.
116 -Introcuded the "eof" variable for indicating the end of file. The p variable
117 is checked against eof when the processing loop reaches the end of a block.
118 If p == eof at this time then the EOF actions are executed. The variable is
119 required only when EOF actions have been emebedded.
120 -The "write eof" command is no longer needed and was removed.
121 -Scanners now use EOF actions to generate tokens. This eliminates the need to
122 flush the last token.
123 -Restructured the Java driver; a switch statement with fallthrough cases are
124 now used to emulate gotos.
125 -Ruby code generation was also restructured. Gotos are elmulated using a
127 -Went back to 3.X semantics for >, % and error actions. The > operator also
128 embeds a leaving action/priority into the start state if it is final. If EOF
129 happens in a state with a leaving operator then the leaving action is
130 executed. If EOF happens in a non-final state that has an error action, the
131 error action is executed.
132 -The "ragel" program now executes frontend and backend processes separately,
133 connecting them with a temporary file in the current directory. Without the
134 -x option the "ragel" program marshals arguments and calls the frontend and
135 backend. With the -x option the "ragel" program acts as the frontend only.
136 -Added name finding for executables. If any forward slash is found in argv0
137 then it is assumed that the path is explicit and the path to the backend
138 executable should be derived from that. Whe check that location and also go
139 up one then inside a directory of the same name in case we are executing
140 from the source tree. If no forward slash is found it is assumed the file is
141 being run from the installed location. The PREFIX supplied during
142 configuration is used.
143 -On windows GetModuleFileNameEx is used to find out where the the current
144 process's binary is. That location is searched first. If that fails then we
145 go up one directory and look for the executable inside a directory of the
146 same name in case we are executing from the source tree.
147 -Changed the -l option in rlgen-cd to -L because it is covered in the
148 frontend. Added a passthrough in the frontend for the backend options.
149 -Dot file generation can now be invoked using the -V option to ragel. We
150 now require an input file. If standard in is used then we don't have a file
151 name on which to base the output.
152 -Able to build native windows executables using Cygwin+MinGW.
153 -Patch from David Waite: Large arrays are now created by copying in the data
154 from smaller arrays using System.arraycopy(). This eliminates the debug data
155 associated with explicit initialization statements. It is also much easier
156 on the java compiler which can run out of memory compiling very large
157 machines. The downside is that it takes slightly longer to initialize static
159 -The fbreak statement now advances p.
160 -In the :> :>> and <: operators it was possible for the priority assignment
161 to be bypassed via the zero length string. In :> this was fixed
162 automatically with the semantics change to the entering priority operator.
163 If the start state is final it now embeds a leaving action into it,
164 preventing persistance through the zero length string. In :>> and <: this
165 was fixed explicitly. With <: the entering priority operator was used and
166 with :> a special exception was added. Since it uses the finishing
167 transition operator it also adds a leaving priority to the start state if it
169 -Ranlib is now run on the archives. Patch from Kenny MacDermid.
170 -The case statement syntax in ruby code generation used a form depreciated in
171 Ruby 1.9. Updated it.
172 -Made a number of fixes that eliminate warnings in GCC 4.3. Mostly concern
173 the now depreciate automatic conversion of string contsants to "char*" type.
174 Other fixes include adding parenthesis around && within ||.
175 -The "tokstart" and "tokend" variables were changed to "ts" and "te".
177 Ragel 5.25 - Dec 24, 2007
178 =========================
179 -Fixed segfault reported by Ryan Phelps. Affected Java and Ruby code
180 generation. The dataExpr variable was not initialized.
181 -Fixed incorrect case label in test/runtests. Caused Objective-C tests to be
183 -Added missing include to common.cpp.
185 Ragel 5.24 - Sep 16, 2007
186 =========================
187 -Applied patch from Victor Hugo Borja <vic@rubyforge.org>. This patch
188 implements -T1 -F0 -F1 and -G0 in the ruby code generator. Goto-driven code
189 generation is experimental and requires rubinius asm directives (specify
190 with --rbx option). These code generators pass all the ruby tests.
191 -If the condition embedding code runs out of available characters in the
192 keyspace an error message is emitted.
193 -The first example that appeared in the manual used the special-purpose
194 'noend' write option. This caused confusion. Now a basic example appears
196 -Added two new statements: prepush and postpop. These are code blocks that
197 are written out during call and return statements. The prepush code is
198 written immediately before pushing the current state to the state stack
199 during a call. The postpop code is written immediately after popping the
200 current state during return. These can be used to implement a dynamically
203 Ragel 5.23 - Jul 24, 2007
204 =========================
205 -Eliminated the use of callcc as an alternative to goto. Instead, the named
206 breaks implementation used in the Java code generator is imitated using
207 control flow variables.
208 -Improved the error message given when there is a write statement but no
209 machine instantiations and hence no state machine.
210 -Documentation improvements: updates to "Machine Instantiation", "Write Init"
211 and "Write Exports" sectons. Added the "Variables Used by Ragel" section.
212 -Renamed "Entering Actions" to "Starting Actions."
213 -Other documentation updates.
215 Ragel 5.22 - June 14, 2007
216 ==========================
217 -Bug fix: need to isolate the start state of a scanner before setting the
218 to-state and from-state actions which clear and set tokstart. This affected
219 very simple scanners only. Most scanners have an isolated start state due to
220 the pattern structure.
221 -Bug fix: when -S or -M was given the ragel version number was not emitted,
222 causing the backend to reject the intermediate format. From Tim Potter.
223 -The p varialbe is now set up at the beginning of a scanner action, rather
224 than at the end. This leaves scanner actions free to manipulate p and
225 removes the need for the special holdTE and execTE (TE for tokend) versions
226 of hold and exec. It also removes the need to set p = tokend-1 immediately
227 before any control flow. We loose the ability to determine where in the
228 input stream a scanner action is executed, however this information is of
229 little use because it is primarily an artifact of the scanner implementation
230 (sometimes the last char, other times later on). The gains of this change
231 are consistency and simplicity.
232 -The "data" variable (used in Java and Ruby code generation only) can now be
233 overridden using the variable statement.
235 Ragel 5.21 - May 9, 2007
236 ========================
237 -Fixed an inconsistency in the value of p following an error. In the C
238 directly executable code (rlgen-cd -G2) p is left at the character where
239 the error occurred, which is correct. In all other code generators it was
240 left at the following character. This was fixed. Now in all code generators
241 p is left at the character where the error occurred.
242 -Bug fix: when fhold was used in scanner pattern actions which get executed
243 on the last character of the pattern (pattern matches which do not require
244 any lookahead), fhold was modifying p instead of tokend. This was fixed and
245 the patact.rl test was modified to cover the case.
246 -Fixed typos in the guide, improved the state action embedding operator
247 section and added subsections on the variable, import, and export
249 -Implemented a better solution than the pri hack for resolving the '-'
250 ambiguity: force a shortest match of term.
251 -Fixed bugs in the binary searching for condition keys in both the Ruby and
252 Java code generation.
253 -Can now embed the negative sense of a condition. Added a language-
254 independent test case for this feature and the necessary transformation
256 -Added new condition embedding syntax:
257 expr inwhen cond - The transitions into the machine (starting transitions).
258 expr outwhen cond - The pending transitions out of the machine.
259 -The argument to the variable statement which affects the name of the current
260 state variable was changed from "curstate" to "cs" (the default name used
261 for the current state)
262 -Implemented the other variables names in the variable statement. Now all
263 variables (p, pe, cs, top, stack, act, tokstart, tokend) can be renamed.
264 -Parse errors in the intermediate XML file now cause the backend to exit
265 immediately rather then forge on. The recovery infrastructure isn't there
266 and segfaults are likely.
267 -When no input is given to the backend program, it should not print an error
268 message, it should just return a non-zero exit status. The assumption is
269 that the frontend printed an error.
270 -The version number is now included in the intermediate file. An error is
271 emitted if there is a mismatch.
272 -The alphabet type is now communicated from the frontend to the backend using
273 a one-word internal name instead of an array offset.
274 -The Ruby host language types had been just copied from Java. Reduced them to
275 two basic types: char and int, both signed with the usual C sizes.
277 Ragel 5.20 - Apr 7, 2007
278 ========================
279 -The cs variable is now always initialized, unless the "nocs" option is given
280 to the write init command. If there is no main machine, cs is initialized to
281 the entry point defined by the last machine instantiation.
282 -A number of fixes were made to the Ruby code generator.
283 -The frontend now scans ruby comments and regular expressions.
284 -A transformation for Ruby was added to the language-independent test suite.
285 The Ruby code generator passes on all the language-independent tests.
286 -A new Ruby test and two language-independent tests were added.
287 -Some portability fixes were made (Patches from Josef Goettgens and Aaron
289 -Fixed a make dependency bug which caused a problem for parallel building
290 (Patch from Jeremy Hinegardner).
292 Ragel 5.19 - Mar 14, 2007
293 =========================
294 -Added an import statement to ragel. This statement takes a literal string as
295 an argument, interprets it as a file name, then scrapes the file for
296 sequences of tokens that match the following forms. Tokens inside ragel
297 sections are ignored. An example is in test/import1.rl
301 "define" name lit_string
302 -Added an export mechanism which writes defines for single character machines
303 that have been tagged with the export keyword in their definition. Defines
304 are used for C, ints for D, Java and Ruby. Examples of the export feature
305 are in test/export*.rl.
306 -All machine instantiations are now always generated, even if they are not
307 referenced. In the backend, entry points for all instantiations are written
308 out alongside start, error and first final states.
309 -If the main machine is not present then do not emit an error. Generate the
310 machine without a start state and do not initialize cs in the write init
312 -Added an option -l to rlgen-cd which inhibits the writing of #line
314 -Added a new syntax for verbose embeddings. This adds parentheses:
316 Verbose embeddings without parentheses can make code difficult to read
317 because they force a space in the middle of an action embedding. There is a
318 tendency to associtate spaces with concatenation. Without syntax
319 highlighting to make it clear that the embedding type is a keyword, the
320 problem is especially bad. The danger is that a verbose embedding could be
321 read as an embedding of the keyword representing the empbedding type. With
322 parentheses, verbose embeddings read much more clearly.
323 -Conditions now have a forced order when more than one is executed on a
324 single character. Previously ordering relied on pointers, which caused
325 results to vary by compiler. Ordering is now done using conditon action
326 declaration order. This fixes the failure of cond4.rl which occured with
327 g++ 4.1 and other compiler versions.
328 -In the port from flex to ragel, the name separator :: in Ragel code was
330 -In the examples directory switched from rlcodegen to rlgen-cd. Silenced a
331 warning in statechart.rl.
332 -In the root makefile the distclean target was fixed. It was calling clean in
333 the subdirs. In docs, the clean target was not deleting the new manpages for
334 the rlgen-* programs. Fixed.
335 -Portability and other fixes from Josef Goettgens were applied.
336 -The @datadir@ and @mandir@ variables are made use of in doc/Makefile.in for
337 specifying where documentation should be installed. Patch from Marcus
340 Ragel 5.18 - Feb 13, 2007
341 =========================
342 -There is now a 100% correspondence between state id numbers in the
343 intermediate XML file, Graphviz dot files and generated code. This was
344 achieved by moving code which determines if the error state is necessary
345 into the frontend, and then assigning state numbers before writing out the
347 -Backened class structure was reorganized to make it easier to add new code
348 generators without having to also modify the existing code generators.
349 -The C and D code generation executable was changed to rlgen-cd.
350 -The Java code generation was split out into it's own exectuable (rlgen-java)
351 to allow it to freely diverge from the C/D-based code generation.
352 -The graphviz dot file generation was also split out to it's own executable
354 -The Ruby code generation patch from Victor Hugo Borja was added. This is
355 highly experimental code and is not yet completely functional. It is in the
356 executable rlgen-ruby.
357 -The problem with large state machine machines in Java was fixed. This
358 problem was discovered by Colin Fleming, who also contributed a patch.
359 Rather than specify arrays as comma-separated lists of literals, array
360 initialization is now done in a static function. This is the approach used
361 by the Java compiler. Unlike the compiler Ragel is careful split large
362 initilization functions.
363 -The manual was expanded and reorganized somewhat.
364 -Eliminated per-example directories in examples/.
365 -Made some fixes to the pullscan.rl example.
366 -In the frontend CR characters are now treated as whitespace.
367 -Updated to the latest aapl. This completely eliminates the shallowCopy
368 function. With that, a definitive memory leak is fixed.
369 -Control codes with escape sequences are now printable characters (-p
370 option). Also, the space character is now printed as SP.
371 -Fixed the null dereference and consequential segfault which occurred when
372 trying to create empty machines with [] and // and /a[]b/.
373 -Fixed the segfault which occured when a machine reference failed.
374 -Discontinuing ragel.spec. It is more appropriate for this to be written by
375 package maintenance developers.
377 Ragel 5.17 - Jan 28, 2007
378 =========================
379 -The scanners and parsers in both the frontend and backend programs were
380 completely rewritten using Ragel and Kelbt.
381 -The '%when condition' syntax was functioning like '$when condition'. This
383 -In the Vim syntax file fixes to the matching of embedding operators were
384 made. Also, improvements to the sync patterns were made.
385 -Added pullscan.rl to the examples directory. It is an example of doing
386 pull-based scanning. Also, xmlscan.rl in rlcodegen is a pull scanner.
387 -The introduction chapter of the manual was improved. The manually-drawn
388 figures for the examples were replaced with graphviz-drawn figures.
390 Ragel 5.16 - Nov 20, 2006
391 =========================
392 -Policy change: the fhold and fexec directives did not function correctly in
393 scanner pattern actions. In this context manipulations of p may be lost or
394 made invalid. In the previous version of Ragel they were banned because of
395 this. Instead of banning these directives they have been fixed. The fexec
396 and fhold directives now manipulate tokend, which is now always used to
397 update p when the action terminates.
399 Ragel 5.15 - Oct 31, 2006
400 =========================
401 -A language independent test harness was introduced. Test cases can be
402 written using a custom mini-language in the embedded actions. This
403 mini-language is then translated to C, D and Java when generating the
404 language-specific test cases.
405 -Several existing tests have been ported to the language-independent format
406 and a number of new language-independent test cases have been added.
407 -The state-based embedding operators which access states that are not the
408 start state and are not final (the 'middle' states) have changed. They
410 <@/ eof action into middle states
411 <@! error action into middle states
412 <@^ local error action into middle states
413 <@~ to-state action into middle states
414 <@* from-state action into middle states
416 <>/ eof action into middle states
417 <>! error action into middle states
418 <>^ local error action into middle states
419 <>~ to-state action into middle states
420 <>* from-state action into middle states
421 -The verbose form of embeddings using the <- operator have been removed.
422 This syntax was difficult to remember.
423 -A new verbose form of state-based embedding operators have been added.
424 These are like the symbol versions, except they replace the symbols:
426 with literal keywords:
428 -The following words have been promoted to keywords:
429 when eof err lerr to from
430 -The write statment now gets its own lexical scope in the scanner to ensure
431 that commands are passed through as is (not affected by keywords).
432 -Bug fix: in the code generation of fret in scanner actions the adjustment to
433 p that is needed in some cases (dependent on content of patterns) was not
435 -The fhold directive, which decrements p, cannot be permitted in the pattern
436 action of a scanner item because it will not behave consistently. At the end
437 of a pattern action p could be decremented, set to a new value or left
438 alone. This depends on the contents of the scanner's patterns. The user
439 cannot be expected to predict what will happen to p.
440 -Conditions in D code require a cast to the widec type when computing widec.
441 -Like Java, D code also needs if (true) branches for control flow in actions
442 in order to fool the unreachable code detector. This is now abstracted in
443 all code generators using the CTRL_FLOW() function.
444 -The NULL_ITEM value in java code should be -1. This is needed for
445 maintaining tokstart.
447 Ragel 5.14 - Oct 1, 2006
448 ========================
449 -Fixed the check for use of fcall in actions embedded within longest match
450 items. It was emitting an error if an item's longest-match action had an
451 fcall, which is allowed. This bug was introduced while fixing a segfault in
453 -A new minimization option was added: MinimizeMostOps (-l). This option
454 minimizes at every operation except on chains of expressions and chains of
455 terms (eg, union and concat). On these chains it minimizes only at the last
456 operation. This makes test cases with many states compile faster, without
457 killing the performance on grammars like strings2.rl.
458 -The -l minimiziation option was made the default.
459 -Fixes to Java code: Use of the fc value did not work, now fixed. Static data
460 is now declared with the final keyword. Patch from Colin Fleming. Conditions
461 now work when generating Java code.
462 -The option -p was added to rlcodegen which causes printable characters to be
463 printed in GraphViz output. Patch from Colin Fleming.
464 -The "element" keyword no longer exists, removed from vim syntax file.
465 Updated keyword highlighting.
466 -The host language selection is now made in the frontend.
467 -Native host language types are now used when specifying the alphtype.
468 Previously all languages used the set defined by C, and these were mapped to
469 the appropriate type in the backend.
471 Ragel 5.13 - Sep 7, 2006
472 ========================
473 -Fixed a careless error which broke Java code generation.
475 Ragel 5.12 - Sep 7, 2006
476 ========================
477 -The -o flag did not work in combination with -V. This was fixed.
478 -The split code generation format uses only the required number of digits
479 when writing out the number in the file name of each part.
480 -The -T0, -F0 and -G0 codegens should write out the action list iteration
481 variables only when there are regular, to state or from state actions. The
482 code gens should not use anyActions().
483 -If two states have the same EOF actions, they are written out in the finish
485 -The split and in-place goto formats would sometimes generate _out when it is
486 not needed. This was fixed.
487 -Improved the basic partitioning in the split code gen. The last partition
488 would sometimes be empty. This was fixed.
489 -Use of 'fcall *' was not causing top to be initialized. Fixed.
490 -Implemented a Java backend, specified with -J. Only the table-based format
492 -Implemented range compression in the frontend. This has no effect on the
493 generated code, however it reduces the work of the backend and any programs
494 that read the intermediate format.
496 Ragel 5.11 - Aug 10, 2006
497 =========================
498 -Added a variable to the configure.in script which allows the building of
499 the parsers to be turned off (BUILD_PARSERS). Parser building is off by
500 default for released versions.
501 -Removed configure tests for bison defines header file. Use --defines=file
503 -Configure script doesn't test for bison, flex and gperf when building of the
504 parsers is turned off.
505 -Removed check for YYLTYPE structure from configure script. Since shipped
506 code will not build parsers by default, we don't need to be as accomodating
507 of other versions of bison.
508 -Added a missing include that showed up with g++ 2.95.3.
509 -Failed configure test for Objective-C compiler is now silent.
511 Ragel 5.10 - Jul 31, 2006
512 =========================
513 -Moved the check for error state higher in the table-based processing loop.
514 -Replaced naive implementations of condition searching with proper ones. In
515 the table-based formats the searching is also table-based. In the directly
516 executed formats the searching is also directly executable.
517 -The minimization process was made aware of conditions.
518 -A problem with the condition implementation was fixed. Previously we were
519 taking pointers to transitions and then using them after a call to
520 outTransCopy, which was a bad idea because they may be changed by the call.
521 -Added test mailbox3.rl which is based on mailbox2.rl but includes conditions
522 for restricting header and message body lengths.
523 -Eliminated the initial one-character backup of p just before resuming
525 -Added the -s option to the frontend for printing statistics. This currently
526 includes just the number of states.
527 -Sped up the generation of the in-place goto-driven (-G2) code style.
528 -Implemented a split version of in-place goto-driven code style. This code
529 generation style is suitable for producing fast implementations of very
530 large machines. Partitioning is currently naive. In the future a
531 high-quality partitioning program will be employed. The flag for accessing
532 this feature is -Pn, where n is the number of partitions.
533 -Converted mailbox1.rl, strings2.rl and cppscan1.rl tests to support the
534 split code generation.
535 -Fixes and updates were made to the runtests script: added -c for compiling
536 only, changed the -me option to -e, and added support for testing the split
539 Ragel 5.9 - Jul 19, 2006
540 ========================
541 -Fixed a bug in the include system which caused malformed output from the
542 frontend when the include was made from a multi-line machine spec and the
543 included file ended in a single line spec (or vice versa).
544 -Static data is now const.
545 -Actions which referenced states but were not embedded caused the frontend to
547 -Manual now built with pdflatex.
548 -The manual was reorganized and expanded. Chapter sequence is now:
549 Introduction, Constructing Machines, Embedding Actions, Controlling
550 Nondeterminism and Interfacing to the Host program.
552 Ragel 5.8 - Jun 17, 2006
553 ========================
554 -The internal representation of the alphabet type has been encapsulated
555 into a class and all operations on it have been defined as C++ operators.
556 -The condition implementation now supports range transitions. This allows
557 conditions to be embedded into arbitrary machines. Conditions are still
559 -More condition embedding operators were added
560 1. Isolate the start state and embed a condition into all transitions
563 2. Embed a condition into all transitions:
564 when cond OR $when cond OR $?cond
565 3. Embed a condition into pending out transitions:
567 -Improvements were made to the determinization process to support pending out
569 -The Vim sytax file was fixed so that :> doesn't cause the match of a label.
570 -The test suite was converted to a single-file format which uses less disk
571 space than the old directory-per-test format.
573 Ragel 5.7 - May 14, 2006
574 ========================
575 -Conditions will not be embedded like actions because they involve a
576 manipulation of the state machine they are specified in. They have therefore
577 been taken out of the verbose action embedding form (using the <- compound
578 symbol). A new syntax for specifying conditions has been created:
579 m = '\n' when {i==4};
580 -Fixed a bug which prevented state machine commands like fcurs, fcall, fret,
581 etc, from being accounted for in from-state actions and to-state actions.
582 This prevented some necessary support code from being generated.
583 -Implemented condition testing in remaining code generators.
584 -Configure script now checks for gperf, which is required for building.
585 -Added support for case-insensitive literal strings (in addition to regexes).
586 A case-insensitive string is made by appending an 'i' to the literal, as in
588 -Fixed a bug which caused all or expressions inside of all regular
589 expressions to be case-insensitive. For example /[fo]o bar/ would make the
590 [fo] part case-insensitive even though no 'i' was given following the
593 Ragel 5.6 - Apr 1, 2006
594 =======================
595 -Added a left-guarded concatenation operator. This operator <: is equivalent
596 to ( expr1 $1 . expr2 >0 ). It is useful if you want to prefix a sequence
597 with a sequence of a subset of the characters it matches. For example, one
598 can consume leading whitespace before tokenizing a sequence of whitespace
599 separated words: ( ' '* <: ( ' '+ | [a-z]+ )** )
600 -Removed context embedding code, which has been dead since 5.0.
602 Ragel 5.5 - Mar 28, 2006
603 ========================
604 -Implemented a case-insensitive option for regular expressions: /get/i.
605 -If no input file is given to the ragel program it reads from standard input.
606 -The label of the start state has been changed from START to IN to save on
607 required screen space.
608 -Bug fix: \0 was not working in literal strings, due to a change that reduced
609 memory usage by concatenating components of literal strings. Token data
610 length is now passed from the scanner to the paser so that we do not need to
611 rely on null termination.
613 Ragel 5.4 - Mar 12, 2006
614 ========================
615 -Eliminated the default transition from the frontend implementation. This
616 default transition was a space-saving optimization that at best could reduce
617 the number of allocated transitions by one half. Unfortunately it
618 complicated the implementation and this stood in the way of introducing
619 conditionals. The default transition may be reintroduced in the future.
620 -Added entry-guarded concatenation. This operator :>, is syntactic sugar
621 for expr1 $0 . expr >1. This operator terminates the matching of the first
622 machine when a first character of the second machine is matched. For
623 example in any* . ';' we never leave the any* machine. If we use any* :> ';'
624 then the any* machine is terminiated upon matching the semi-colon.
625 -Added finish-guarded concatenation. This operator :>>, is syntactic sugar
626 for expr1 $0 . expr @1. This operator is like entry guarded concatenation
627 except the first machine is terminated when the second machine enters a
628 final state. This is useful for delaying the guard until a full pattern is
629 matched. For example as in '/*' any* :>> '*/'.
630 -Added strong subtraction. Where regular subtraction removes from the first
631 machine any strings that are matched by the second machine, strong
632 subtraction removes any strings from the first that contain any strings of
633 the second as a substring. Strong subtraction is syntactic sugar for
634 expr1 - ( any* expr2 any* ).
635 -Eliminated the use of priorities from the examples. Replaced with
636 subtraction, guarded concatenation and longest-match kleene star.
637 -Did some initial work on supporting conditional transitions. Far from
638 complete and very buggy. This code will only be active when conditionals are
641 Ragel 5.3 - Jan 27, 2006
642 ========================
643 -Added missing semi-colons that cause the build to fail when using older
645 -Fix for D code: if the contents of an fexec is a single word, the generated
646 code will get interpreted as a C-style cast. Adding two brackets prevents
647 this. Can now turn eliminate the "access this.;" in cppscan5 that was used to
648 get around this problem.
649 -Improved some of the tag names in the intermediate format.
650 -Added unsigned long to the list of supported alphabet types.
651 -Added ids of actions and action lists to XML intermediate format. Makes it
653 -Updated to latest Aapl package.
655 Ragel 5.2 - Jan 6, 2006
656 ========================
657 -Ragel emits an error if the target of fentry, fcall, fgoto or fnext is inside
658 a longest match operator, or if an action embedding in a longest match
659 machine uses fcall. The fcall command can still be used in pattern actions.
660 -Made improvements to the clang, rlscan, awkemu and cppscan examples.
661 -Some fixes to generated label names: they should all be prefixed with _.
662 -A fix to the Vim syntax highlighting script was made
663 -Many fixes and updates to the documentation. All important features and
664 concepts are now documented. A second chapter describing Ragel's use
667 Ragel 5.1 - Dec 22, 2005
668 ========================
669 -Fixes to the matching of section delimiters in Vim syntax file.
670 -If there is a longest match machine, the tokend var is now initialized by
671 write init. This is not necessary for correct functionality, however
672 prevents compiler warnings.
673 -The rlscan example was ported to the longest match operator and changed to
675 -Fix to the error handling in the frontend: if there are errors in the lookup
676 of names at machine generation time then do not emit anything.
677 -If not compiling the full machine in the frontend (by using -M), avoid
678 errors and segfaults caused by names that are not part of the compiled
680 -Longest match bug fix: need to init tokstart when returing from fsm calls
681 that are inside longest match actions.
682 -In Graphviz drawing, the arrow into the start state is not a real
683 transition, do not draw to-state actions on the label.
684 -A bug fix to the handling of non-tag data within an XML tag was made.
685 -Backend exit value fixed: since the parser now accepts nothing so as to
686 avoid a redundant parse error when the frontend dies, we must force an
687 error. The backend should now be properly reporting errors.
688 -The longest match machine now has it's start state set final. An LM machine
689 is in a final state when it has not matched anything, when it has matched
690 and accepted a token and is ready for another, and when it has matched a
691 token but is waiting for some lookahead before determining what to do about
692 it (similar to kleene star).
693 -Element statement removed from some tests.
694 -Entry point names are propagated to the backend and used to label the entry
695 point arrows in Graphviz output.
697 Ragel 5.0 - Dec 17, 2005
698 ========================
699 (additional details in V5 release notes)
700 -Ragel has been split into two executables: A frontend which compiles
701 machines and emits them in an XML format, and a backend which generates code
702 or a Graphviz dot file from the XML input. The purpose of this split is to
703 allow Ragel to interface with other tools by means of the XML intermediate
704 format and to reduce complexity by strictly separating the previously
705 entangled phases. The intermediate format will provide a better platform
706 inspecting compiled machines and for extending Ragel to support other host
708 -The host language interface has been reduced significantly. Ragel no longer
709 expects the machine to be implemented as a structure or class and does not
710 generate functions corresponding to initialization, execution and EOF.
711 Instead, Ragel just generates the code of these components, allowing all of
712 them to be placed in a single function if desired. The user specifies a
713 machine in the usual manner, then indicates at which place in the program
714 text the state machine code is to be generated. This is done using the write
715 statement. It is possible to specify to Ragel how it should access the
716 variables it needs (such as the current state) using the access statement.
717 -The host language embedding delimiters have been changed. Single line
718 machines start with '%%' and end at newline. Multiline machines start with
719 '%%{' and end with '}%%'. The machine name is given with the machine
720 statement at the very beginning of the specification. This purpose of this
721 change is to make it easier separate Ragel code from the host language. This
722 will ease the addition of supported host languages.
723 -The structure and class parsing which was previously able to extract a
724 machine's name has been removed since this feature is dependent on the host
725 language and inhibits the move towards a more language-independent frontend.
726 -The init, element and interface statements have been made obsolete by the
727 new host language interface and have been removed.
728 -The fexec action statement has been changed to take only the new position to
729 move to. This statement is more useful for moving backwards and reparsing
730 input than for specifying a whole new buffer entirely and has been shifted
731 to this new use. Giving it only one argument also simplifies the parsing of
732 host code embedded in a Ragel specification. This will ease the addition of
733 supported host languages.
734 -Introduced the fbreak statement, which allows one to stop processing data
735 immediately. The machine ends up in the state that the current transition
736 was to go to. The current character is not changed.
737 -Introduced the noend option for writing the execute code. This inhibits
738 checking if we have reached pe. The machine will run until it goes into the
739 error state or fbreak is hit. This allows one to parse null-terminate
740 strings without first computing the length.
741 -The execute code now breaks out of the processing loop when it moves into
742 the error state. Previously it would run until pe was hit. Breaking out
743 makes the noend option useful when an error is encountered and allows
744 user code to determine where in the input the error occured. It also
745 eliminates needlessly iterating the input buffer.
746 -Introduced the noerror, nofinal and noprefix options for writing the machine
747 data. The first two inhibit the writing of the error state and the
748 first-final state should they not be needed. The noprefix eliminates the
749 prefixing of the data items with the machine name.
750 -Support for the D language has been added. This is specified in the backend
752 -Since the new host language interface has been reduced considerably, Ragel
753 no longer needs to distinguish between C-based languages. Support for C, C++
754 and Objective-C has been folded into one option in the backend: -C
755 -The code generator has been made independent of the languages that it
756 supports by pushing the language dependent apsects down into the lower
757 levels of the code generator.
758 -Many improvements to the longest match construction were made. It is no
759 longer considered experimental. A longest match machine must appear at the
760 top level of a machine instantiation. Since it does not generate a pure
761 state machine (it may need to backtrack), it cannot be used as an operand to
763 -References to the current character and current state are now completely
764 banned in EOF actions.
766 Ragel 4.2 - Sep 16, 2005
767 ========================
768 (additional details in V4 release notes)
769 -Fixed a bug in the longest match operator. In some states it's possible that
770 we either match a token or match nothing at all. In these states we need to
771 consult the LmSwitch on error so it must be prepared to execute an error
772 handler. We therefore need to init act to this error value (which is zero).
773 We can compute if we need to do this and the code generator emits the
774 initialization only if necessary.
775 -Changed the definition of the token end of longest match actions. It now
776 points to one past the last token. This makes computing the token length
777 easier because you don't have to add one. The longest match variables token
778 start, action identifier and token end are now properly initialized in
779 generated code. They don't need to be initialized in the user's code.
780 -Implemented to-state and from-state actions. These actions are executed on
781 transitions into the state (after the in transition's actions) and on
782 transitions out of the state (before the out transition's actions). See V4
783 release notes for more information.
784 -Since there are no longer any action embedding operators that embed both on
785 transitions and on EOF, any actions that exist in both places will be there
786 because the user has explicitly done so. Presuming this case is rare, and
787 with code duplication in the hands of the user, we therefore give the EOF
788 actions their own action switch in the finish() function. This is further
789 motivated by the fact that the best solution is to do the same for to-state
790 and from-state actions in the main loop.
791 -Longest match actions can now be specified using a named action. Since a
792 word following a longest match item conflicts with the concatenation of a
793 named machine, the => symbol must come immediately before a named action.
794 -The longest match operator permits action and machine definitions in the
795 middle of a longest match construction. These are parsed as if they came
796 before the machine definition they are contained in. Permitting action and
797 machine definitions in a longest match construction allows objects to be
798 defined closer to their use.
799 -The longest match operator can now handle longest match items with no
800 action, where previously Ragel segfaulted.
801 -Updated to Aapl post 2.12.
802 -Fixed a bug in epsilon transition name lookups. After doing a name lookup
803 the result was stored in the parse tree. This is wrong because if a machine
804 is used more than once, each time it may resolve to different targets,
805 however it will be stored in the same place. We now store name resolutions
806 in a separated data structure so that each walk of a parse tree uses the
807 name resolved during the corresponding walk in the name lookup pass.
808 -The operators used to embed context and actions into states have been
809 modified. The V4 release notes contain the full details.
810 -Added zlen builtin machine to represent the zero length machine. Eventually
811 the name "null" will be phased out in favour of zlen because it is unclear
812 whether null matches the zero length string or if it does not match any
813 string at all (as does the empty builtin).
814 -Added verbose versions of action, context and priority embedding. See the V4
815 release notes for the full details. A small example:
816 machine <- all exec { foo(); } <- final eof act1
817 -Bugfix for machines with epsilon ops, but no join operations. I had
818 wrongfully assumed that because epsilon ops can only increase connectivity,
819 that no states are ever merged and therefore a call to fillInStates() is not
820 necessary. In reality, epsilon transitions within one machine can induce the
821 merging of states. In the following, state 2 follows two paths on 'i':
822 main := 'h' -> i 'i h' i: 'i';
823 -Changed the license of the guide from a custom "do not propagate modified
824 versions of this document" license to the GPL.
826 Ragel 4.1 - Jun 26, 2005
827 ========================
828 (additional details in V4 release notes)
829 -A bug in include processing was fixed. Surrounding code in an include file
830 was being passed through to the output when it should be ignored. Includes
831 are only for including portions of another machine into he current. This
832 went unnoticed because all tested includes were wrapped in #ifndef ...
833 #endif directives and so did not affect the compilation of the file making
835 -Fixes were made to Vim syntax highlighting file.
836 -Duplicate actions are now removed from action lists.
837 -The character-level negation operator ^ was added. This operator produces a
838 machine that matches single characters that are not matched by the machine
839 it is applied to. This unary prefix operator has the same precedence level
841 -The use of + to specify the a positive literal number was discontinued.
842 -The parser now assigns the subtraction operator a higher precedence than
843 the negation of literal number.
845 Ragel 4.0 - May 26, 2005
846 ========================
847 (additional details in V4 release notes)
848 -Operators now strictly embed into a machine either on a specific class of
849 characters or on EOF, but never both. This gives a cleaner association
850 between the operators and the physical state machine entitites they operate
851 on. This change is made up of several parts:
852 1. '%' operator embeds only into leaving characters.
853 2. All global and local error operators only embed on error character
854 transitions, their action will not be triggerend on EOF in non-final
856 3. EOF action embedding operators have been added for all classes of states
857 to make up for functionality removed from other operators. These are
859 4. Start transition operator '>' no longer implicitly embeds into leaving
860 transtions when start state is final.
861 -Ragel now emits warnings about the improper use of statements and values in
862 action code that is embedded as an EOF action. Warnings are emitted for fpc,
863 fc, fexec, fbuf and fblen.
864 -Added a longest match construction operator |* machine opt-action; ... *|.
865 This is for repetition where an ability to revert to a shorter, previously
866 matched item is required. This is the same behaviour as flex and re2c. The
867 longest match operator is not a pure FSM construction, it introduces
868 transitions that implicitly hold the current character or reset execution to
869 a previous location in the input. Use of this operator requires the caller
870 of the machine to occasionally hold onto data after a call to the exectute
871 routine. Use of machines generated with this operator as the input to other
872 operators may have undefined results. See examples/cppscan for an example.
873 This is very experimental code.
874 -Action ids are only assigned to actions that are referenced in the final
875 constructed machine, preventing gaps in the action id sequence. Previously
876 an action id was assigned if the action was referenced during parsing.
877 -Machine specifications now begin with %% and are followed with an optional
878 name and either a single Ragel statement or a sequence of statements
880 -Ragel no longer generates the FSM's structure or class. It is up to the user
881 to declare the structure and to give it a variable named curs of type
882 integer. If the machine uses the call stack the user must also declare a
883 array of integers named stack and an integer variable named top.
884 -In the case of Objective-C, Ragel no longer generates the interface or
885 implementation directives, allowing the user to declare additional methods.
886 -If a machine specification does not have a name then Ragel tries to find a
887 name for it by first checking if the specification is inside a struct, class
888 or interface. If it is not then it uses the name of the previous machine
889 specification. If still no name is found then this is an error.
890 -Fsm specifications now persist in memory and statements accumulate.
891 -Ragel now has an include statement for including the statements of a machine
892 spec in another file (perhaps because it is the corresponding header file).
893 The include statement can also be used to draw in the statements of another
894 fsm spec in the current file.
895 -The fstack statement is now obsolete and has been removed.
896 -A new statement, simply 'interface;', indicates that ragel should generate
897 the machine's interface. If Ragel sees the main machine it generates the
898 code sections of the machine. Previously, the header portion was generated
899 if the (now removed) struct statement was found and code was generated if
900 any machine definition was found.
901 -Fixed a bug in the resolution of fsm name references in actions. The name
902 resolution code did not recurse into inline code items with children
903 (fgoto*, fcall*, fnext*, and fexec), causing a segfault at code generation
905 -Cleaned up the code generators. FsmCodeGen was made into a virtual base
906 class allowing for the language/output-style specific classes to inherit
907 both a language specific and style-specific base class while retaining only
908 one copy of FsmCodeGen. Language specific output can now be moved into the
909 language specific code generators, requiring less duplication of code in the
910 language/output-style specific leaf classes.
911 -Fixed bugs in fcall* implementation of IpgGoto code generation.
912 -If the element type has not been defined Ragel now uses a constant version
913 of the alphtype, not the exact alphtype. In most cases the data pointer of
914 the execute routine should be const. A non-const element type can still be
915 defined with the element statement.
916 -The fc special value now uses getkey for retrieving the current char rather
917 than *_p, which is wrong if the element type is a structure.
918 -User guide converted to TeX and updated for new 4.0 syntax and semantics.
920 Ragel 3.7 - Oct 31, 2004
921 ========================
922 -Bug fix: unreferenced machine instantiations causing segfault due to name
923 tree and parse tree walk becomming out of syncronization.
924 -Rewrote representation of inline code blocks using a tree data structure.
925 This allows special keywords such as fbuf to be used as the operatands of
927 -Documentation updates.
928 -When deciding whether or not to generate machine instantiations, search the
929 entire name tree beneath the instantiation for references, not just the
931 -Removed stray ';' in keller2.rl
932 -Added fexec for restarting the machine with new buffer data (state stays the
933 same), fbuf for retrieving the the start of the buf, and fblen for
934 retrieving the orig buffer length.
935 -Implemented test/cppscan2 using fexec. This allows token emitting and restart
936 to stay inside the execute routine, instead of leaving and re-entering on
938 -Changed examples/cppscan to use fexec and thereby go much faster.
939 -Implemented flex and re2c versions of examples/cppscan. Ragel version
940 goes faster than flex version but not as fast as re2c version.
941 -Merged in Objective-C patch from Erich Ocean.
942 -Turned off syncing with stdio in C++ tests to make them go faster.
943 -Renamed C++ code generaion classes with the Cpp Prefix instead of CC to make
945 -In the finish function emit fbuf as 0 cast to a pointer to the element type
946 so it's type is not interpreted as an integer.
947 -The number -128 underflows char alphabets on some architectures. Removed
949 -Disabled the keller2 test because it causes problems on many architectures
950 due to its large size and compilation requirements.
952 Ragel 3.6 - Jul 10, 2004
953 ========================
954 -Many documentation updates.
955 -When resolving names, return a set of values so that a reference in an
956 action block that is embedded more than once won't report distinct entry
957 points that are actually the same.
958 -Implemented flat tables. Stores a linear array of indicies into the
959 transition array and only a low and high key value. Faster than binary
960 searching for keys but not usable for large alphabets.
961 -Fixed bug in deleting of transitions leftover from converstion from bst to
962 list implementation of transitions. Other code cleanup.
963 -In table based output calculate the cost of using an index. Don't use if
965 -Changed fstate() value available in init and action code to to fentry() to
966 reflect the fact that the values returned are intended to be used as targets
967 in fgoto, fnext and fcall statements. The returned state is not a unique
968 state representing the label. There can be any number of states representing
970 -Added keller2 test, C++ scanning tests and C++ scanning example.
971 -In table based output split up transitions into targets and actions. This
972 allows actions to be omitted.
973 -Broke the components of the state array into separate arrays. Requires
974 adding some fields where they could previously be omitted, however allows
975 finer grained control over the sizes of items and an overal size reduction.
976 Also means that state numbers are not an offset into the state array but
977 instead a sequence of numbers, meaning the context array does not have any
979 -Action lists and transition also have their types chosen to be the smallest
980 possible for accomodating the contained values.
981 -Changed curs state stored in fsm struct from _cs to curs. Keep fsm->curs ==
982 -1 while in machine. Added tests curs1 and curs2.
983 -Implemented the notion of context. Context can be embedded in states using
984 >:, $:, @: and %: operators. These embed a named context into start states,
985 all states, non-start/non-final and final states. If the context is declared
986 using a context statment
988 then the context can be quered for any state using fsm_name_ctx_name(state)
989 in C code and fsm_name::ctx_name(state) in C++ code. This feature makes it
990 possible to determine what "part" of the machine is currently active.
991 -Fixed crash on machine generation of graphs with no final state. If there
992 is no reference to a final state in a join operation, don't generate one.
993 -Updated Vim sytax: added labels to inline code, added various C++ keywords.
994 Don't highlight name separations as labels. Added switch labels, improved
995 alphtype, element and getkey.
996 -Fixed line info in error reporting of bad epsilon trans.
997 -Fixed fstate() for tab code gen.
998 -Removed references to malloc.h.
1000 Ragel 3.5 - May 29, 2004
1001 ========================
1002 -When parse errors occur, the partially generated output file is deleted and
1003 an non-zero exit status is returned.
1004 -Updated Vim syntax file.
1005 -Implemented the setting of the element type that is passed to the execute
1006 routine as well as method for specifying how ragel should retrive the key
1007 from the element type. This lets ragel process arbitrary structures inside
1008 of which is the key that is parsed.
1009 element struct Element;
1010 getkey fpc->character;
1011 -The current state is now implemented with an int across all machines. This
1012 simplifies working with current state variables. For example this allows a
1013 call stack to be implemented in user code.
1014 -Implemented a method for retrieving the current state, the target state, and
1016 fcurs -retrieve the current state
1017 ftargs -retrieve the target state
1018 fstate(name) -retrieve a named state.
1019 -Implemented a mechanism for jumping to and calling to a state stored in a
1021 fgoto *<expr>; -goto the state returned by the C/C++ expression.
1022 fcall *<expr>; -call the state returned by the C/C++ expression.
1023 -Implemented a mechanism for specifying the next state without immediately
1024 transfering control there (any code following statement is executed).
1025 fnext label; -set the state pointed to by label as the next state.
1026 fnext *<expr>; -set the state returned by the C/C++ expression as the
1028 -Action references are determined from the final machine instead of during
1029 the parse tree walk. Some actions can be referenced in the parse tree but not
1030 show up in the final machine. Machine analysis is now done based on this new
1032 -Named state lookup now employs a breadth-first search in the lookup and
1033 allows the user to fully qualify names, making it possible to specify
1034 jumps/calls into parts of the machine deep in the name hierarchy. Each part
1035 of name (separated by ::) employs a breadth first search from it's starting
1037 -Name references now must always refer to a single state. Since references to
1038 multiple states is not normally intended, it no longer happens
1039 automatically. This frees the programmer from thinking about whether or not
1040 a state reference is unique. It also avoids the added complexity of
1041 determining when to merge the targets of multiple references. The effect of
1042 references to multiple states can be explicitly created using the join
1043 operator and epsilon transitions.
1044 -M option was split into -S and -M. -S specifies the machine spec to generate
1045 for graphviz output and dumping. -M specifies the machine definition or
1047 -Machine function parameters are now prefixed with and underscore to
1048 avoid the hiding of class members.
1050 Ragel 3.4 - May 8, 2004
1051 =======================
1052 -Added the longest match kleene star operator **, which is synonymous
1053 with ( ( <machine> ) $0 %1 ) *.
1054 -Epsilon operators distinguish between leaving transitions (going to an
1055 another expression in a comma separated list) and non-leaving transitions.
1056 Leaving actions and priorities are appropriately transferred.
1057 -Relative priority of following ops changed to:
1061 If label is done first then the isolation of the start state in > operators
1062 will cause the label to point to the old start state that doesn't have the
1063 new action/priority.
1064 -Merged >! and >~, @! and @~, %! and %~, and $! and $~ operators to have one
1065 set of global error action operators (>!, @!, %! and $!) that are invoked on
1066 error by unexpected characters as well as by unexepected EOF.
1067 -Added the fpc keyword for use in action code. This is a pointer to the
1068 current character. *fpc == fc. If an action is invoked on EOF then fpc == 0.
1069 -Added >^, @^, %^, and $^ local error operators. Global error operators (>!,
1070 @!, $!, and %!) cause actions to be invoked if the final machine fails.
1071 Local error actions cause actions to be invoked if if the current machine
1073 -Changed error operators to mean embed global/local error actions in:
1074 >! and !^ -the start state.
1075 @! and @^ -states that are not the start state and are not final.
1076 %! and %^ -final states.
1077 $! and $^ -all states.
1078 -Added >@! which is synonymous >! then @!
1079 -Added >@^ which is synonymous >^ then @^
1080 -Added @%! which is synonymous @! then %!
1081 -Added @%^ which is synonymous >^ then @^
1082 -FsmGraph representation of transition lists was changed from a mapping of
1083 alphabet key -> transition objects using a BST to simply a list of
1084 transition objects. Since the transitions are no longer divided by
1085 single/range, the fast finding of transition objects by key is no longer
1086 required functionality and can be eliminated. This new implementation uses
1087 the same amount of memory however causes less allocations. It also make more
1088 sense for supporting error transitions with actions. Previously an error
1089 transition was represented by a null value in the BST.
1090 -Regular expression ranges are checked to ensure that lower <= upper.
1091 -Added printf-like example.
1092 -Added atoi2, erract2, and gotcallret to the test suite.
1093 -Improved build test to support make -jN and simplified the compiling and
1096 Ragel 3.3 - Mar 7, 2004
1097 =======================
1098 -Portability bug fixes were made. Minimum and maximum integer values are
1099 now taken from the system. An alignment problem on 64bit systems
1102 Ragel 3.2 - Feb 28, 2004
1103 ========================
1104 -Added a Vim syntax file.
1105 -Eliminated length var from generated execute code in favour of an end
1106 pointer. Using length requires two variables be read and written. Using an
1107 end pointer requires one variable read and written and one read. Results in
1108 more optimizable code.
1109 -Minimization is now on by default.
1110 -States are ordered in output by depth first search.
1111 -Bug in minimization fixed. States were not being distinguished based on
1113 -Added null and empty builtin machines.
1114 -Added EOF error action operators. These are >~, >@, $~, and %~. EOF error
1115 operators embed actions to take if the EOF is seen and interpreted as an
1116 error. The operators correspond to the following states:
1118 -any state with a transition to a final state
1119 -any state with a transiion out
1121 -Fixed bug in generation of unreference machine vars using -M. Unreferenced
1122 vars don't have a name tree built underneath when starting from
1123 instantiations. Need to instead build the name tree starting at the var.
1124 -Calls, returns, holds and references to fc in out action code are now
1125 handled for ipgoto output.
1126 -Only actions referenced by an instantiated machine expression are put into
1127 the action index and written out.
1128 -Added rlscan, an example that lexes Ragel input.
1130 Ragel 3.1 - Feb 18, 2004
1131 ========================
1132 -Duplicates in OR literals are removed and no longer cause an assertion
1134 -Duplicate entry points used in goto and call statements are made into
1135 deterministic entry points.
1136 -Base FsmGraph code moved from aapl into ragel, as an increasing amount
1137 of specialization is required. Too much time was spent attempting to
1138 keep it as a general purpose template.
1139 -FsmGraph code de-templatized and heirarchy squashed to a single class.
1140 -Single transitions taken out of FsmGraph code. In the machine construction
1141 stage, transitions are now implemented only with ranges and default
1142 transtions. This reduces memory consumption, simplifies code and prevents
1143 covered transitions. However it requires the automated selection of single
1144 transitions to keep goto-driven code lean.
1145 -Machine reduction completely rewritten to be in-place. As duplicate
1146 transitions and actions are found and the machine is converted to a format
1147 suitable for writing as C code or as GraphViz input, the memory allocated
1148 for states and transitions is reused, instead of newly allocated.
1149 -New reduction code consolodates ranges, selects a default transition, and
1150 selects single transitions with the goal of joining ranges that are split by
1151 any number of single characters.
1152 -Line directive changed from "# <num> <file>" to the more common format
1153 "#line <num> <file>".
1154 -Operator :! changed to @!. This should have happened in last release.
1155 -Added params example.
1157 Ragel 3.0 - Jan 22, 2004
1158 ========================
1159 -Ragel now parses the contents of struct statements and action code.
1160 -The keyword fc replaces the use of *p to reference the current character in
1162 -Machine instantiations other than main are allowed.
1163 -Call, jump and return statements are now available in action code. This
1164 facility makes it possible to jump to an error handling machine, call a
1165 sub-machine for parsing a field or to follow paths through a machine as
1166 determined by arbitrary C code.
1167 -Added labels to the language. Labels can be used anywhere in a machine
1168 expression to define an entry point. Also references to machine definitions
1169 cause the implicit creation of a label.
1170 -Added epsilon transitions to the language. Epsilon operators may reference
1171 labels in the current name scope resolved when join operators are evaluated
1172 and at the root of the expression tree of machine assignment/instantiation.
1173 -Added the comma operator, which joins machines together without drawing any
1174 transitions between them. This operator is useful in combination with
1175 labels, the epsilon operator and user code transitions for defining machines
1176 using the named state and transition list paradigm. It is also useful for
1177 invoking transitions based on some analysis of the input or on the
1179 -Added >!, :!, $!, %! operators for specifying actions to take should the
1180 machine fail. These operators embed actions to execute if the machine
1183 -any state with a transition to a final state
1184 -any state with a transiion out
1186 The general rule is that if an action embedding operator embeds an action
1187 into a set of transitions T, then the error-counterpart with a ! embeds an
1188 action into the error transition taken when any transition T is a candidate,
1189 but does not match the input.
1190 -The finishing augmentation operator ':' has been changed to '@'. This
1191 frees the ':' symbol for machine labels and avoids hacks to the parser to
1192 allow the use of ':' for both labels and finishing augmentations. The best
1193 hack required that label names be distinct from machine definition names as
1194 in main := word : word; This restriction is not good because labels are
1195 local to the machine that they are used in whereas machine names are global
1196 entities. Label name choices should not be restricted by the set of names
1197 that are in use for machines.
1198 -Named priority syntax now requires parenthesis surrounding the name and
1199 value pair. This avoids grammar ambiguities now that the ',' operator has
1200 been introduced and makes it more clear that the name and value are an
1202 -Backslashes are escaped in line directive paths.
1204 Ragel 2.2 - Oct 6, 2003
1205 =======================
1206 -Added {n}, {,n}, {n,} {n,m} repetition operators.
1207 <expr> {n} -- exactly n repetitions
1208 <expr> {,n} -- zero to n repetitions
1209 <expr> {n,} -- n or more repetitions
1210 <expr> {n,m} -- n to m repetitions
1211 -Bug in binary search table in Aapl fixed. Fixes crashing on machines that
1212 add to action tables that are implicitly shared among transitions.
1213 -Tests using obsolete minimization algorithms are no longer built and run by
1215 -Added atoi and concurrent from examples to the test suite.
1217 Ragel 2.1 - Sep 22, 2003
1218 ========================
1219 -Bug in priority comparison code fixed. Segfaulted on some input with many
1220 embedded priorities.
1221 -Added two new examples.
1223 Ragel 2.0 - Sep 7, 2003
1224 =======================
1225 -Optional (?), One or More (+) and Kleene Star (*) operators changed from
1226 prefix to postfix. Rationale is that postfix version is far more common in
1227 regular expression implementations and will be more readily understood.
1228 -All priority values attached to transitions are now accompanied by a name.
1229 Transitions no longer have default priority values of zero assigned
1230 to them. Only transitions that have different priority values assigned
1231 to the same name influence the NFA-DFA conversion. This scheme reduces
1232 side-effects of priorities.
1233 -Removed the %! statement for unsetting pending out priorities. With
1234 named priorities, it is not necessary to clear the priorities of a
1235 machine with $0 %! because non-colliding names can be used to avoid
1237 -Removed the clear keyword, which was for removing actions from a machine.
1238 Not required functionality and it is non-intuitive to have a language
1239 feature that undoes previous definitions.
1240 -Removed the ^ modifier to repetition and concatenation operators. This
1241 undocumented feature prevented out transitions and out priorities from being
1242 transfered from final states to transitions leaving machines. Not required
1243 functionality and complicates the language unnecessarily.
1244 -Keyword 'func' changed to 'action' as a part of the phasing out of the term
1245 'function' in favour of 'action'. Rationale is that the term 'function'
1246 implies that the code is called like a C function, which is not necessarily
1247 the case. The term 'action' is far more common in state machine compiler
1249 -Added the instantiation statement, which looks like a standard variable
1250 assignment except := is used instead of =. Instantiations go into the
1251 same graph dictionary as definitions. In the the future, instantiations
1252 will be used as the target for gotos and calls in action code.
1253 -The main graph should now be explicitly instantiated. If it is not,
1254 a warning is issued.
1255 -Or literal basic machines ([] outside of regular expressions) now support
1256 negation and ranges.
1257 -C and C++ interfaces lowercased. In the C interface an underscore now
1258 separates the fsm machine and the function name. Rationale is that lowercased
1259 library and generated routines are more common.
1261 int fsm_init( struct clang *fsm );
1262 int fsm_execute( struct clang *fsm, char *data, int dlen );
1263 int fsm_finish( struct clang *fsm );
1266 int fsm::execute( char *data, int dlen );
1268 -Init, execute and finish all return -1 if the machine is in the error state
1269 and can never accept, 0 if the machine is in a non-accepting state that has a
1270 path to a final state and 1 if the machine is in an accepting state.
1271 -Accept routine eliminated. Determining whether or not the machine accepts is
1272 done by examining the return value of the finish routine.
1273 -In C output, fsm structure is no longer a typedef, so referencing requires
1274 the struct keyword. This is to stay in line with C language conventions.
1275 -In C++ output, constructor is no longer written by ragel. As a consequence,
1276 init routine is not called automatically. Allows constructor to be supplied
1277 by user as well as the return value of init to be examined without calling it
1279 -Static start state and private structures are taken out of C++ classes.
1281 Ragel 1.5.4 - Jul 14, 2003
1282 ==========================
1283 -Workaround for building with bison 1.875, which produces an
1284 optimization that doesn't build with newer version gcc.
1286 Ragel 1.5.3 - Jul 10, 2003
1287 ==========================
1288 -Fixed building with versions of flex that recognize YY_NO_UNPUT.
1289 -Fixed version numbers in ragel.spec file.
1291 Ragel 1.5.2 - Jul 7, 2003
1292 =========================
1293 -Transition actions and out actions displayed in the graphviz output.
1294 -Transitions on negative numbers handled in graphviz output.
1295 -Warning generated when using bison 1.875 now squashed.
1297 Ragel 1.5.1 - Jun 21, 2003
1298 ==========================
1299 -Bugs fixed: Don't delete the output objects when writing to standard out.
1300 Copy mem into parser buffer with memcpy, not strcpy. Fixes buffer mem errror.
1301 -Fixes for compiling with Sun WorkShop 6 compilers.
1303 Ragel 1.5.0 - Jun 10, 2003
1304 ==========================
1305 -Line directives written to the output so that errors in the action code
1306 are properly reported in the ragel input file.
1307 -Simple graphviz dot file output format is supported. Shows states and
1308 transitions. Does not yet show actions.
1309 -Options -p and -f dropped in favour of -d output format.
1310 -Added option -M for specifying the machine to dump with -d or the graph to
1312 -Error recovery implemented.
1313 -Proper line and column number tracking implemented in the scanner.
1314 -All action/function code is now embedded in the main Execute routine. Avoids
1315 duplication of action code in the Finish routine and the need to call
1316 ExecFuncs which resulted in huge code bloat. Will also allow actions to
1317 modify cs when fsm goto, call and return is supported in action code.
1318 -Fsm spec can have no statements, nothing will be generated.
1319 -Bug fix: Don't accept ] as the opening of a .-. range a reg exp.
1320 -Regular expression or set ranges (ie /[0-9]/) are now handled by the parser
1321 and consequently must be well-formed. The following now generates a parser
1322 error: /[+-]/ and must be rewritten as /[+\-]/. Also fixes a bug whereby ]
1323 might be accepted as the opening of a .-. range causing /[0-9]-[0-9]/ to
1325 -\v, \f, and \r are now treated as whitespace in an fsm spec.
1327 Ragel 1.4.1 - Nov 19, 2002
1328 ==========================
1329 -Compile fixes. The last release (integer alphabets) was so exciting
1330 that usual portability checks got bypassed.
1332 Ragel 1.4.0 - Nov 19, 2002
1333 ==========================
1334 -Arbitrary integer alphabets are now fully supported! A new language
1336 'alphtype <type>' added for specifying the type of the alphabet. Default
1337 is 'char'. Possible alphabet types are:
1338 char, unsigned char, short, unsigned short, int, unsigned int
1339 -Literal machines specified in decimal format can now be negative when the
1340 alphabet is a signed type.
1341 -Literal machines (strings, decimal and hex) have their values checked for
1342 overflow/underflow against the size of the alphabet type.
1343 -Table driven and goto driven output redesigned to support ranges. Table
1344 driven uses a binary search for locating single characters and ranges. Goto
1345 driven uses a switch statement for single characters and nested if blocks for
1347 -Switch driven output removed due to a lack of consistent advantages. Most of
1348 the time the switch driven FSM is of no use because the goto FSM makes
1349 smaller and faster code. Under certain circumstances it can produce smaller
1350 code than a goto driven fsm and be almost as fast, but some sporadic case
1351 does not warrant maintaining it.
1352 -Many warnings changed to errors.
1353 -Added option -p for printing the final fsm before minimization. This lets
1354 priorities be seen. Priorties are all reset to 0 before minimization. The
1355 exiting option -f prints the final fsm after minimization.
1356 -Fixed a bug in the clang test and example that resulted in redundant actions
1359 Ragel 1.3.4 - Nov 6, 2002
1360 =========================
1361 -Fixes to Chapter 1 of the guide.
1362 -Brought back the examples and made them current.
1363 -MSVC is no longer supported for compiling windows binaries because its
1364 support for the C++ standard is frustratingly inadequate, it will cost money
1365 to upgrade if it ever gets better, and MinGW is a much better alternative.
1366 -The build system now supports the --host= option for building ragel
1367 for another system (used for cross compiling a windows binary with MinGW).
1368 -Various design changes and fixes towards the goal of arbitrary integer
1369 alphabets and the handling of larger state machines were made.
1370 -The new shared vector class is now used for action lists in transitions and
1371 states to reduce memory allocations.
1372 -An avl tree is now used for the reduction of transitions and functions of an
1373 fsm graph before making the final machine. The tree allows better scalability
1374 and performance by not requiring consecutively larger heap allocations.
1375 -Final stages in the separation of fsm graph code from action embedding and
1376 priority assignment is complete. Makes the base graph leaner and easier to reuse
1377 in other projects (like Keller).
1379 Ragel 1.3.3 - Oct 22, 2002
1380 ==========================
1381 -More diagrams were added to section 1.7.1 of the user guide.
1382 -FSM Graph code was reworked to spearate the regex/nfa/minimizaion graph
1383 algorithms from the manipulation of state and transition properties.
1384 -An rpm spec file from Cris Bailiff was added. This allows an rpm for ragel
1385 to be built with the command 'rpm -ta ragel-x.x.x.tar.gz'
1386 -Fixes to the build system and corresponding doc updates in the README.
1387 -Removed autil and included the one needed source file directly in the top
1388 level ragel directory.
1389 -Fixed a bug that nullified the 20 times speedup in large compilations
1390 claimed by the last version.
1391 -Removed awk from the doc build (it was added with the last release -- though
1392 not mentioned in the changelog).
1393 -Install of man page was moved to the doc dir. The install also installs the
1394 user guide to $(PREFIX)/share/doc/ragel/
1396 Ragel 1.3.2 - Oct 16, 2002
1397 ==========================
1398 -Added option -v (or --version) to show version information.
1399 -The subtract operator no longer removes transition data from the machine
1400 being subtracted. This is left up to the user for the purpose of making it
1401 possible to transfer transitions using subtract and also for speeding up the
1402 subtract routine. Note that it is possible to explicitly clear transition
1403 data before a doing a subtract.
1404 -Rather severe typo bug fixed. Bug was related to transitions with higher
1405 priorities taking precedence. A wrong ptr was being returned. It appears to
1406 have worked most of the time becuase the old ptr was deleted and the new one
1407 allocated immediatly after so the old ptr often pointed to the same space.
1409 -Bug in the removing of dead end paths was fixed. If the start state
1410 has in transitions then those paths were not followed when finding states to
1411 keep. Would result in non-dead end states being removed from the graph.
1412 -In lists and in ranges are no longer maintained as a bst with the key as the
1413 alphabet character and the value as a list of transitions coming in on that
1414 char. There is one list for each of inList, inRange and inDefault. Now that
1415 the required functionality of the graph is well known it is safe to remove
1416 these lists to gain in speed and footprint. They shouldn't be needed.
1417 -IsolateStartState() runs on modification of start data only if the start
1418 state is not already isolated, which is now possible with the new in list
1420 -Concat, Or and Star operators now use an approximation to
1421 removeUnreachableStates that does not require a traversal of the entire
1422 graph. This combined with an 'on-the-fly' management of final bits and final
1423 state status results is a dramatic speed increase when compiling machines
1424 that use those operators heavily. The strings2 test goes 20 times faster.
1425 -Before the final minimization, after all fsm operations are complete,
1426 priority data is reset which enables better minimization in cases where
1427 priorities would otherwise separate similar states.
1429 Ragel 1.3.1 - Oct 2, 2002
1430 =========================
1431 -Range transitions are now used to implement machines made with /[a-z]/ and
1432 the .. operator as well as most of the builtin machines. The ranges are not
1433 yet reflected in the output code, they are expanded as if they came from the
1434 regular single transitions. This is one step closer to arbitrary integer
1436 -The builtin machine 'any' was added. It is equiv to the builtin extend,
1437 matching any characters.
1438 -The builtin machine 'cntrl' now includes newline.
1439 -The builtin machine 'space' now includes newline.
1440 -The builtin machine 'ascii' is now the range 0-127, not all characters.
1441 -A man page was written.
1442 -A proper user guide was started. Chapter 1: Specifying Ragel Programs
1443 was written. It even has some diagrams :)
1445 Ragel 1.3.0 - Sep 4, 2002
1446 =========================
1447 -NULL keyword no longer used in table output.
1448 -Though not yet in use, underlying graph structure changed to support range
1449 transitions. As a result, most of the code that walks transition lists is now
1450 implemented with an iterator that hides the complexity of the transition
1451 lists and ranges. Range transitions will be used to implement /[a-z]/ style
1452 machines and machines made with the .. operator. Previously a single
1453 transition would be used for each char in the range, which is very costly.
1454 Ranges eliminate much of the space complexity and allow for the .. operator
1455 to be used with very large (integer) alphabets.
1456 -New minimization similar to Hopcroft's alg. It does not require n^2 space and
1457 runs close to O(n*log(n)) (an exact analysis of the alg is very hard). It is
1458 much better than the stable and approx minimization and obsoletes them both.
1459 An exact implementation of Hopcroft's alg is desirable but not possible
1460 because the ragel implementation does not assume a finite alphabet, which
1461 Hopcroft's requires. Ragel will support arbitrary integer alphabets which
1462 must be treated as an infinite set for implementation considerations.
1463 -New option -m using above described minimization to replace all previous
1464 minimization options. Old options sill work but are obsolete and not
1466 -Bug fixed in goto style output. The error exit set the current state to 0,
1467 which is actually a valid state. If the machine was entered again it would go
1468 into the first state, very wrong. If the first state happened to be final then
1469 an immediate finish would accept when in fact it should fail.
1470 -Slightly better fsm minimization now capable due to clearing of the
1471 transition ordering numbers just prior to minimization.
1473 Ragel 1.2.2 - May 25, 2002
1474 ==========================
1475 -Configuration option --prefix now works when installing.
1476 -cc file extension changed to cpp for better portability.
1477 -Unlink of output file upon error no longer happens, removes dependency on
1478 unlink system command.
1479 -All multiline strings removed: not standard c++.
1480 -Awk build dependency removed.
1481 -MSVC 6.0 added to the list of supported compilers (with some tweaking of
1482 bison and flex output).
1484 Ragel 1.2.1 - May 13, 2002
1485 ==========================
1486 -Automatic dependencies were fixed, they were not working correctly.
1487 -Updated AUTHORS file to reflect contributors.
1488 -Code is more C++ standards compliant: compiles with g++ 3.0
1489 -Fixed bugs that only showed up in g++ 3.0
1490 -Latest (unreleased) Aapl.
1491 -Configuration script bails out if bison++ is installed. Ragel will not
1492 compile with bison++ because it is coded in c++ and bison++ automatically
1493 generates a c++ parser. Ragel uses a c-style bison parser.
1495 Ragel 1.2.0 - May 3, 2002
1496 =========================
1497 -Underlying graph structure now supports default transitions. The result is
1498 that a transition does not need to be made for each char of the alphabet
1499 when making 'extend' or '/./' machines. Ragel compiles machines that
1500 use the aforementioned primitives WAY faster.
1501 -The ugly hacks needed to pick default transitions now go away due to
1502 the graph supporting default transitions directly.
1503 -If -e is given, but minimization is not turned on, print a warning.
1504 -Makefiles use automatic dependencies.
1506 Ragel 1.1.0 - Apr 15, 2002
1507 ==========================
1508 -Added goto fsm: much faster than any other fsm style.
1509 -Default operator (if two machines are side by side with no operator
1510 between them) is concatenation. First showed up in 1.0.4.
1511 -The fsm machine no longer auotmatically builds the flat table for
1512 transition indicies. Instead it keeps the key,ptr pair. In tabcodegen
1513 the flat table is produced. This way very large alphabets with sparse
1514 transitions will not consume large amounts of mem. This is also in prep
1515 for fsm graph getting a default transition.
1516 -Generated code contains a statement explicitly stating that ragel fsms
1517 are NOT covered by the GPL. Technically, Ragel copies part of itself
1518 to the output to make the generic fsm execution routine (for table driven
1519 fsms only) and so the output could be considered under the GPL. But this
1520 code is very trivial and could easlily be rewritten. The actual fsm data
1521 is subject to the copyright of the source. To promote the use of Ragel,
1522 a special exception is made for the part of the output copied from Ragel:
1523 it may be used without restriction.
1524 -Much more elegant code generation scheme is employed. Code generation
1525 class members need only put the 'codegen' keyword after their 'void' type
1526 in order to be automatically registerd to handle macros of the same name.
1527 An awk script recognises this keyword and generates an appropriate driver.
1528 -Ragel gets a test suite.
1529 -Postfunc and prefunc go away because they are not supported by non
1530 loop-driven fsms (goto, switch) and present duplicate functionality.
1531 Universal funcs can be implemented by using $ operator.
1532 -Automatic dependencies used in build system, no more make depend target.
1533 -Code generation section in docs.
1534 -Uses the latests aapl.
1536 Ragel 1.0.5 - Mar 3, 2002
1537 =========================
1538 -Bugfix in SetErrorState that caused an assertion failure when compiling
1539 simple machines that did not have full transition tables (and thus did
1540 not show up on any example machines). Assertion failure did not occur
1541 when using the switch statement code as ragel does not call SetErrorState
1543 -Fixed some missing includes, now compiles on redhat.
1544 -Moved the FsmMachTrans Compare class out of FsmMachTrans. Some compilers
1545 don't deal with nested classes in templates too well.
1546 -Removed old unused BASEREF in fsmgraph and ragel now compiles using
1547 egcs-2.91.66 and presumably SUNWspro. The baseref is no longer needed
1548 because states do not support being elements in multiple lists. I would
1549 rather be able to support more compilers than have this feature.
1550 -Started a README with compilation notes. Started an AUTHORS file.
1551 -Started the user documentation. Describes basic machines and operators.
1553 Ragel 1.0.4 - Mar 1, 2002
1554 =========================
1555 -Ported to the version of Aapl just after 2.2.0 release. See
1556 http://www.ragel.ca/aapl/ for details on aapl.
1557 -Fixed a bug in the clang example: the newline machine was not stared.
1558 -Added explanations to the clang and mailbox examples. This should
1559 help people that want to learn the lanuage as the manual is far from
1562 Ragel 1.0.3 - Feb 2, 2002
1563 =========================
1564 -Added aapl to the ragel tree. No longer requires you to download
1565 and build aapl separately. Should avoid discouraging impatient users
1566 from compiling ragel.
1567 -Added the examples to the ragel tree.
1568 -Added configure script checks for bison and flex.
1569 -Fixed makefile so as not to die with newer versions of bison that
1570 write the header of the parser to a .hh file.
1571 -Started ChangeLog file.
1573 Ragel 1.0.2 - Jan 30, 2002
1574 ==========================
1575 -Bug fix in calculating highIndex for table based code. Was using
1576 the length of out tranisition table rather than the value at the
1578 -If high/low index are at the limits, output a define in their place,
1579 not the high/low values themselves so as not to cause compiler warnings.
1580 -If the resulting machines don't have any indicies or functions, then
1581 omit the empty unrefereced static arrays so as not to cause compiler
1582 warnings about unused static vars.
1583 -Fixed variable sized indicies support. The header cannot have any
1584 reference to INDEX_TYPE as that info is not known at the time the header
1585 data is written. Forces us to use a void * for pointers to indicies. In
1586 the c++ versions we are forced to make much of the data non-member
1587 static data in the code portion for the same reason.
1589 Ragel 1.0.1 - Jan 28, 2002
1590 ==========================
1591 -Exe name change from reglang to ragel.
1592 -Added ftabcodegen output code style which uses a table for states and
1593 transitions but uses a switch statement for the function execution.
1594 -Reformatted options in usage dump to look better.
1595 -Support escape sequences in [] sections of regular expressions.
1597 Ragel 1.0 - Jan 25, 2002
1598 ========================