Debian Codegraph @

ChangeLog

raw
1 Ragel 6.10 - Mar 24, 2017
2 =========================
3 -C codegen: test P vs PE in goto/call/ret statements in EOF actions, just
4 before re-entering. If at the end of the input block then the EOF check is
5 jumped to. This change prevents overrunning the buffer if control flow is
6 issued in an EOF action without fixing the input pointer first. If a program
7 properly issues an fhold before the control flow the program won't be
8 affected.
9 -Updated action label generation. The previous set of conditions for
10 generating the label didn't cover actions coming from the eofAction pointer
11 (eof trans covered since it points into the set of transitions).
12 -Use separate signed/unsigned values for host type min/max. Using separate
13 values avoids the need to type cast before the data goes into FsmCtx structs.
14 Keep it in native types until it is used.
15 -Optionally do not generate entry point variables. Adds noentry write option
16 for data.
17 -Various warning elimination and build updates.
18
19 Ragel 6.9 - Oct 13, 2014
20 ========================
21 -updated command-line synopsis
22 -ocaml: fix missing semicolon
23 -ocaml: support -G1
24 -ocaml: choose a unique name for type state
25 -ruby: reduce the amount of calls to GET_WIDE_KEY()
26 -union test case: warning fix
27 -omit line directives around expression-oriented write statements
28 -use AS_IF and test command to check if the DIST file is present
29 -added missing std:: using
30 -go: added '//line' directive support
31
32 Ragel 6.8 - Feb 11, 2013
33 ========================
34
35 -The -G2 code generator for Go1 was rewritten. Table, flat and switch-based
36 code generators were added. (Anton Ageev)
37 -The CXXFLAGS variable is not longer set in the configure script.
38
39 Ragel 6.7 - May 22, 2011
40 ========================
41 -The C vim file now supports L,l on the end of literal numbers, other syntax
42 highlighting fixes.
43 -Added an explicit cast when modifying _trans to eliminate compiler warnings
44 on Windows/VC++
45 -Fix for ruby 1.9 compatibility.
46 -Fix for include directories (-I option) on Windows/VC++
47 -Can now rename data variable in C# code generator.
48 -Fix for non-char alphtype in C# code generator.
49 -Fix for signedness of wide char in C code generator. when comparing the wide
50 type against a literal we need to pick the right signedness for the literal.
51 -Fixed arithmetic overflow in generated C# code. The C# _acts and _nacts vars
52 should not be typed based on the size of the array elements they point to.
53 Fixes a bug reported by Attila Sztupák.
54 -Made the -L option work for Ruby.
55 -Enhanced ragel.m4 (from Diego).
56 -Applied GO patch from Justine Tunney.
57 -Applied D2 patch from Johannes Pfau.
58 -Applied Ocaml patch from ygrek.
59
60 Ragel 6.6 - Dec 2, 2009
61 =======================
62 -Applied a number of patches from Diego Elio 'Flameeyes' Pettenò. Should not
63 be modifying the program's arguments. Problem raised by const correctness in
64 gcc 4.4. Other const-correctness and include fixes provided.
65 -Fixed improper allocation of checks in makeIncludePathChecks.
66 -Fixed segfault when there are no machine instantiations.
67 -Fixed wrong line directives. Line directives need to use the fileName stored
68 in the InputLoc stuctures from the parse trees, not the root source file,
69 otherwise actions in included files will have the wrong source file names
70 associated with the text.
71 -Made a number of build system improvements. We locate the DIST file using
72 $srcdir and source it. It contains settings for build_parsers and
73 build_manual. This allows the user of a dist to enable only one.
74 -Added missing files to doc/Makefile.am and examples/Makefile.am.
75 -Added checks for pdflatex and fig2dev is build_manual is on.
76 -Use automake --foreign so we don't need to have INSTALL and NEWS present.
77 -Ragel VIM syntax files should be specialized by host language. Updated the
78 VIM syntax files.
79 -Added examples to the dist. Added unicode2ragel.rb to EXTRA_DIST in contrib.
80 -Moved unicode2ragel.rb to the contrib directory.
81
82 Ragel 6.5 - May 18, 2009
83 ========================
84 -Fixed a bug in graphviz generation. Ragel crashed when using -V and -M and
85 the specified machine referenced another machine that wasn't included in the
86 build.
87 -The name "CS" is in use on OpenSolaris, changed to vCS to ease compiling
88 Ragel there.
89 -Converted to automake.
90 -REALLY fixed a bug that was intended to be fixed in 6.4:
91 Fixed a problem reading hex numbers that have the high bit set when the
92 alphabet is signed and we are on 64 bit. This was reported by _why. The
93 fix was provided by Wialliam Morgan. The literal 0xffffffff was used for
94 a fully set long when -1L should be used instead.
95 A null patch (whitespace changes) must have gotten checked after I was
96 testing with and without the critical one-line patch and I forgot to enable
97 make sure it was enabled in the final checkin version.
98
99 Ragel 6.4 - Mar 22, 2009
100 ========================
101 -Moved back to a single executable. The old intermediate format can still be
102 generated using the -x option. Ragel was split into frontend and backend
103 programs in version 5.0. This was done to encourage interoperability with
104 other tools. Since then, ragel has been made to work with qfsm, with ragel
105 producing the intermediate format and qfsm consuming it. However, there has
106 been no use of Ragel as a consumer of state machine data, with Ragel used as
107 a code generator for DFAs. This is not surprising given that much of the
108 complexity of Ragel is in the frontend, where the regular language to DFA
109 compilation happens. Since the full benefits of the split have not
110 materialized, and the split increases the complexity for users, Ragel has
111 been made once again into a single executable.
112 -Applied a fix to the documentation Makefile from John D. Mitchell.
113 -Use CXXFLAGS instead of CFLAGS for C++ compiling. Patch from Diego
114 'Flameeyes' Pettenò.
115 -Added support for DESTDIR variable. Patch from Diego 'Flameeyes' Pettenò.
116 -Added a script called unicode2ragel.rb for generating unicode machines to
117 the examples directory. From Rakan El-Khalil.
118 -Fixed a copy-paste error in the documentation that was reported by Jose
119 Quinteiro.
120 -Added three new write commands:
121 write start;
122 write first_final;
123 write error;
124 These generate a reference to the start, first final and error state. When
125 there are many different machine specifications in one file it is easy to
126 get the prefix for these wrong (especially when you do a lot of copy-pasting
127 of boilerplate). The problem can be avoided by using write commands.
128 -Fixed a problem reading hex numbers that have the high bit set when the
129 alphabet is signed and we are on 64 bit. This was reported by _why. The fix
130 was provided by Wialliam Morgan. The literal 0xffffffff was used for a fully
131 set long when -1L should be used instead.
132
133 Ragel 6.3 - Aug 29, 2008
134 ========================
135 -Fixed an assertion that is too strong. In the condition code we need to copy
136 transitions that have non-empty lmActionTable arrays so we don't assert
137 emptiness in the constructor. Lift out the assertion and copy the array in
138 the constructor.
139 -Fixed and improved multiple include prevention. We now track the entire
140 include history of a parser state to prevent duplicates.
141 -Fixed crash on failed lookup of goto/call/etc target.
142
143 Ragel 6.2 - May 9, 2008
144 =======================
145 -Bug fix: The lm_switch actions needs to set p from tokend when there is no
146 user action.
147 -Bug fix: when not using indicies we can't use a transitions's id to identify
148 the eof transition to take. Instead add the transition to the end of the
149 transition list and store its position in a new var called pos. The pos var
150 is then used as the index.
151 -Bug fix: an fnext followed by an fbreak in -G2 was not working. The fbreak
152 was not aware that the fnext causes the cs variable to be forced active. In
153 this case fbreak does not need to save cs because it is already current.
154 -Bug fix: need to compute the low and high character-space keys from the
155 condition-trans overlap when computing an expansion. Can't use the range
156 supplied from the condition overlap since they may not match. An incorrect
157 machine that accepted 1(!cond1, !cond2) was generated for the following
158 grammar. This bug was reported by Tim Chklovski.
159 c = 2 @matched_c;
160 sc1 = 1..2 when cond1;
161 sc2 = 1..2 when cond2;
162 main := sc1 | c | sc2;
163 -Bug fix: error messages in start label analysis of join operations were
164 causing assertion failures because location info was not set. Fixed by
165 adding locations.
166 -Include and import file searching now searches for the file name given based
167 on the location of the current file, not ragel's current path.
168 Additional search locations can be given using the -I option.
169 -Rubinius code generation was updated to the latest Rubinius. Patch from Evan
170 Phoenix.
171 -Switched from strcasecmp to strcmp for testing long arguments.
172 -Applied a patch from Andrei Polushin for setting the error message format.
173 --error-format=gnu (default)
174 --error-fromat=msvc
175 -Now using the _WIN32 define instead of _WIN32. Other MSVC compilation
176 improvments from Andrei Polushin.
177 -Added the hyperref package to the manual.
178
179 Ragel 6.1 - Mar 26, 2008
180 ========================
181 -Scanners now ensure that any leaving actions at the end of a pattern are
182 executed. They are always executed before the pattern action.
183 -Added an option -d for turning off the removal of duplicate actions from
184 actions lists.
185 -Need to unset the final state status of the start state in kleene star if it
186 is set. It is possible to crash ragel when the warning is ignored.
187 -In the dot file generation we need to print any actions that are in
188 State::eofTrans. These come from scanners only.
189 -Use @docdir@ for the docdir Makefile variable.
190 -Check for ar and ranlib in the configure script.
191
192 Ragel 6.0 - Jan 12, 2008
193 ========================
194 -Removed the 'noend' write option from examples/atoi.rl. This example is
195 referenced a lot as a first example and as such it shouldn't contain a
196 special purpose write option like 'noend'.
197 -Introcuded the "eof" variable for indicating the end of file. The p variable
198 is checked against eof when the processing loop reaches the end of a block.
199 If p == eof at this time then the EOF actions are executed. The variable is
200 required only when EOF actions have been emebedded.
201 -The "write eof" command is no longer needed and was removed.
202 -Scanners now use EOF actions to generate tokens. This eliminates the need to
203 flush the last token.
204 -Restructured the Java driver; a switch statement with fallthrough cases are
205 now used to emulate gotos.
206 -Ruby code generation was also restructured. Gotos are elmulated using a
207 series of if tests.
208 -Went back to 3.X semantics for >, % and error actions. The > operator also
209 embeds a leaving action/priority into the start state if it is final. If EOF
210 happens in a state with a leaving operator then the leaving action is
211 executed. If EOF happens in a non-final state that has an error action, the
212 error action is executed.
213 -The "ragel" program now executes frontend and backend processes separately,
214 connecting them with a temporary file in the current directory. Without the
215 -x option the "ragel" program marshals arguments and calls the frontend and
216 backend. With the -x option the "ragel" program acts as the frontend only.
217 -Added name finding for executables. If any forward slash is found in argv0
218 then it is assumed that the path is explicit and the path to the backend
219 executable should be derived from that. Whe check that location and also go
220 up one then inside a directory of the same name in case we are executing
221 from the source tree. If no forward slash is found it is assumed the file is
222 being run from the installed location. The PREFIX supplied during
223 configuration is used.
224 -On windows GetModuleFileNameEx is used to find out where the the current
225 process's binary is. That location is searched first. If that fails then we
226 go up one directory and look for the executable inside a directory of the
227 same name in case we are executing from the source tree.
228 -Changed the -l option in rlgen-cd to -L because it is covered in the
229 frontend. Added a passthrough in the frontend for the backend options.
230 -Dot file generation can now be invoked using the -V option to ragel. We
231 now require an input file. If standard in is used then we don't have a file
232 name on which to base the output.
233 -Able to build native windows executables using Cygwin+MinGW.
234 -Patch from David Waite: Large arrays are now created by copying in the data
235 from smaller arrays using System.arraycopy(). This eliminates the debug data
236 associated with explicit initialization statements. It is also much easier
237 on the java compiler which can run out of memory compiling very large
238 machines. The downside is that it takes slightly longer to initialize static
239 data at run time.
240 -The fbreak statement now advances p.
241 -In the :> :>> and <: operators it was possible for the priority assignment
242 to be bypassed via the zero length string. In :> this was fixed
243 automatically with the semantics change to the entering priority operator.
244 If the start state is final it now embeds a leaving action into it,
245 preventing persistance through the zero length string. In :>> and <: this
246 was fixed explicitly. With <: the entering priority operator was used and
247 with :> a special exception was added. Since it uses the finishing
248 transition operator it also adds a leaving priority to the start state if it
249 is final.
250 -Ranlib is now run on the archives. Patch from Kenny MacDermid.
251 -The case statement syntax in ruby code generation used a form depreciated in
252 Ruby 1.9. Updated it.
253 -Made a number of fixes that eliminate warnings in GCC 4.3. Mostly concern
254 the now depreciate automatic conversion of string contsants to "char*" type.
255 Other fixes include adding parenthesis around && within ||.
256 -The "tokstart" and "tokend" variables were changed to "ts" and "te".
257
258 Ragel 5.25 - Dec 24, 2007
259 =========================
260 -Fixed segfault reported by Ryan Phelps. Affected Java and Ruby code
261 generation. The dataExpr variable was not initialized.
262 -Fixed incorrect case label in test/runtests. Caused Objective-C tests to be
263 ignored.
264 -Added missing include to common.cpp.
265
266 Ragel 5.24 - Sep 16, 2007
267 =========================
268 -Applied patch from Victor Hugo Borja <[email protected]>. This patch
269 implements -T1 -F0 -F1 and -G0 in the ruby code generator. Goto-driven code
270 generation is experimental and requires rubinius asm directives (specify
271 with --rbx option). These code generators pass all the ruby tests.
272 -If the condition embedding code runs out of available characters in the
273 keyspace an error message is emitted.
274 -The first example that appeared in the manual used the special-purpose
275 'noend' write option. This caused confusion. Now a basic example appears
276 first.
277 -Added two new statements: prepush and postpop. These are code blocks that
278 are written out during call and return statements. The prepush code is
279 written immediately before pushing the current state to the state stack
280 during a call. The postpop code is written immediately after popping the
281 current state during return. These can be used to implement a dynamically
282 resizable stack.
283
284 Ragel 5.23 - Jul 24, 2007
285 =========================
286 -Eliminated the use of callcc as an alternative to goto. Instead, the named
287 breaks implementation used in the Java code generator is imitated using
288 control flow variables.
289 -Improved the error message given when there is a write statement but no
290 machine instantiations and hence no state machine.
291 -Documentation improvements: updates to "Machine Instantiation", "Write Init"
292 and "Write Exports" sectons. Added the "Variables Used by Ragel" section.
293 -Renamed "Entering Actions" to "Starting Actions."
294 -Other documentation updates.
295
296 Ragel 5.22 - June 14, 2007
297 ==========================
298 -Bug fix: need to isolate the start state of a scanner before setting the
299 to-state and from-state actions which clear and set tokstart. This affected
300 very simple scanners only. Most scanners have an isolated start state due to
301 the pattern structure.
302 -Bug fix: when -S or -M was given the ragel version number was not emitted,
303 causing the backend to reject the intermediate format. From Tim Potter.
304 -The p varialbe is now set up at the beginning of a scanner action, rather
305 than at the end. This leaves scanner actions free to manipulate p and
306 removes the need for the special holdTE and execTE (TE for tokend) versions
307 of hold and exec. It also removes the need to set p = tokend-1 immediately
308 before any control flow. We loose the ability to determine where in the
309 input stream a scanner action is executed, however this information is of
310 little use because it is primarily an artifact of the scanner implementation
311 (sometimes the last char, other times later on). The gains of this change
312 are consistency and simplicity.
313 -The "data" variable (used in Java and Ruby code generation only) can now be
314 overridden using the variable statement.
315
316 Ragel 5.21 - May 9, 2007
317 ========================
318 -Fixed an inconsistency in the value of p following an error. In the C
319 directly executable code (rlgen-cd -G2) p is left at the character where
320 the error occurred, which is correct. In all other code generators it was
321 left at the following character. This was fixed. Now in all code generators
322 p is left at the character where the error occurred.
323 -Bug fix: when fhold was used in scanner pattern actions which get executed
324 on the last character of the pattern (pattern matches which do not require
325 any lookahead), fhold was modifying p instead of tokend. This was fixed and
326 the patact.rl test was modified to cover the case.
327 -Fixed typos in the guide, improved the state action embedding operator
328 section and added subsections on the variable, import, and export
329 statements.
330 -Implemented a better solution than the pri hack for resolving the '-'
331 ambiguity: force a shortest match of term.
332 -Fixed bugs in the binary searching for condition keys in both the Ruby and
333 Java code generation.
334 -Can now embed the negative sense of a condition. Added a language-
335 independent test case for this feature and the necessary transformation
336 support.
337 -Added new condition embedding syntax:
338 expr inwhen cond - The transitions into the machine (starting transitions).
339 expr outwhen cond - The pending transitions out of the machine.
340 -The argument to the variable statement which affects the name of the current
341 state variable was changed from "curstate" to "cs" (the default name used
342 for the current state)
343 -Implemented the other variables names in the variable statement. Now all
344 variables (p, pe, cs, top, stack, act, tokstart, tokend) can be renamed.
345 -Parse errors in the intermediate XML file now cause the backend to exit
346 immediately rather then forge on. The recovery infrastructure isn't there
347 and segfaults are likely.
348 -When no input is given to the backend program, it should not print an error
349 message, it should just return a non-zero exit status. The assumption is
350 that the frontend printed an error.
351 -The version number is now included in the intermediate file. An error is
352 emitted if there is a mismatch.
353 -The alphabet type is now communicated from the frontend to the backend using
354 a one-word internal name instead of an array offset.
355 -The Ruby host language types had been just copied from Java. Reduced them to
356 two basic types: char and int, both signed with the usual C sizes.
357
358 Ragel 5.20 - Apr 7, 2007
359 ========================
360 -The cs variable is now always initialized, unless the "nocs" option is given
361 to the write init command. If there is no main machine, cs is initialized to
362 the entry point defined by the last machine instantiation.
363 -A number of fixes were made to the Ruby code generator.
364 -The frontend now scans ruby comments and regular expressions.
365 -A transformation for Ruby was added to the language-independent test suite.
366 The Ruby code generator passes on all the language-independent tests.
367 -A new Ruby test and two language-independent tests were added.
368 -Some portability fixes were made (Patches from Josef Goettgens and Aaron
369 Campbell).
370 -Fixed a make dependency bug which caused a problem for parallel building
371 (Patch from Jeremy Hinegardner).
372
373 Ragel 5.19 - Mar 14, 2007
374 =========================
375 -Added an import statement to ragel. This statement takes a literal string as
376 an argument, interprets it as a file name, then scrapes the file for
377 sequences of tokens that match the following forms. Tokens inside ragel
378 sections are ignored. An example is in test/import1.rl
379 name = number
380 name = lit_string
381 "define" name number
382 "define" name lit_string
383 -Added an export mechanism which writes defines for single character machines
384 that have been tagged with the export keyword in their definition. Defines
385 are used for C, ints for D, Java and Ruby. Examples of the export feature
386 are in test/export*.rl.
387 -All machine instantiations are now always generated, even if they are not
388 referenced. In the backend, entry points for all instantiations are written
389 out alongside start, error and first final states.
390 -If the main machine is not present then do not emit an error. Generate the
391 machine without a start state and do not initialize cs in the write init
392 code.
393 -Added an option -l to rlgen-cd which inhibits the writing of #line
394 directives.
395 -Added a new syntax for verbose embeddings. This adds parentheses:
396 $from(action_name);
397 Verbose embeddings without parentheses can make code difficult to read
398 because they force a space in the middle of an action embedding. There is a
399 tendency to associtate spaces with concatenation. Without syntax
400 highlighting to make it clear that the embedding type is a keyword, the
401 problem is especially bad. The danger is that a verbose embedding could be
402 read as an embedding of the keyword representing the empbedding type. With
403 parentheses, verbose embeddings read much more clearly.
404 -Conditions now have a forced order when more than one is executed on a
405 single character. Previously ordering relied on pointers, which caused
406 results to vary by compiler. Ordering is now done using conditon action
407 declaration order. This fixes the failure of cond4.rl which occured with
408 g++ 4.1 and other compiler versions.
409 -In the port from flex to ragel, the name separator :: in Ragel code was
410 lost. Added it back.
411 -In the examples directory switched from rlcodegen to rlgen-cd. Silenced a
412 warning in statechart.rl.
413 -In the root makefile the distclean target was fixed. It was calling clean in
414 the subdirs. In docs, the clean target was not deleting the new manpages for
415 the rlgen-* programs. Fixed.
416 -Portability and other fixes from Josef Goettgens were applied.
417 -The @datadir@ and @mandir@ variables are made use of in doc/Makefile.in for
418 specifying where documentation should be installed. Patch from Marcus
419 Rueckert.
420
421 Ragel 5.18 - Feb 13, 2007
422 =========================
423 -There is now a 100% correspondence between state id numbers in the
424 intermediate XML file, Graphviz dot files and generated code. This was
425 achieved by moving code which determines if the error state is necessary
426 into the frontend, and then assigning state numbers before writing out the
427 intermediate file.
428 -Backened class structure was reorganized to make it easier to add new code
429 generators without having to also modify the existing code generators.
430 -The C and D code generation executable was changed to rlgen-cd.
431 -The Java code generation was split out into it's own exectuable (rlgen-java)
432 to allow it to freely diverge from the C/D-based code generation.
433 -The graphviz dot file generation was also split out to it's own executable
434 (rlgen-dot).
435 -The Ruby code generation patch from Victor Hugo Borja was added. This is
436 highly experimental code and is not yet completely functional. It is in the
437 executable rlgen-ruby.
438 -The problem with large state machine machines in Java was fixed. This
439 problem was discovered by Colin Fleming, who also contributed a patch.
440 Rather than specify arrays as comma-separated lists of literals, array
441 initialization is now done in a static function. This is the approach used
442 by the Java compiler. Unlike the compiler Ragel is careful split large
443 initilization functions.
444 -The manual was expanded and reorganized somewhat.
445 -Eliminated per-example directories in examples/.
446 -Made some fixes to the pullscan.rl example.
447 -In the frontend CR characters are now treated as whitespace.
448 -Updated to the latest aapl. This completely eliminates the shallowCopy
449 function. With that, a definitive memory leak is fixed.
450 -Control codes with escape sequences are now printable characters (-p
451 option). Also, the space character is now printed as SP.
452 -Fixed the null dereference and consequential segfault which occurred when
453 trying to create empty machines with [] and // and /a[]b/.
454 -Fixed the segfault which occured when a machine reference failed.
455 -Discontinuing ragel.spec. It is more appropriate for this to be written by
456 package maintenance developers.
457
458 Ragel 5.17 - Jan 28, 2007
459 =========================
460 -The scanners and parsers in both the frontend and backend programs were
461 completely rewritten using Ragel and Kelbt.
462 -The '%when condition' syntax was functioning like '$when condition'. This
463 was fixed.
464 -In the Vim syntax file fixes to the matching of embedding operators were
465 made. Also, improvements to the sync patterns were made.
466 -Added pullscan.rl to the examples directory. It is an example of doing
467 pull-based scanning. Also, xmlscan.rl in rlcodegen is a pull scanner.
468 -The introduction chapter of the manual was improved. The manually-drawn
469 figures for the examples were replaced with graphviz-drawn figures.
470
471 Ragel 5.16 - Nov 20, 2006
472 =========================
473 -Policy change: the fhold and fexec directives did not function correctly in
474 scanner pattern actions. In this context manipulations of p may be lost or
475 made invalid. In the previous version of Ragel they were banned because of
476 this. Instead of banning these directives they have been fixed. The fexec
477 and fhold directives now manipulate tokend, which is now always used to
478 update p when the action terminates.
479
480 Ragel 5.15 - Oct 31, 2006
481 =========================
482 -A language independent test harness was introduced. Test cases can be
483 written using a custom mini-language in the embedded actions. This
484 mini-language is then translated to C, D and Java when generating the
485 language-specific test cases.
486 -Several existing tests have been ported to the language-independent format
487 and a number of new language-independent test cases have been added.
488 -The state-based embedding operators which access states that are not the
489 start state and are not final (the 'middle' states) have changed. They
490 were:
491 <@/ eof action into middle states
492 <@! error action into middle states
493 <@^ local error action into middle states
494 <@~ to-state action into middle states
495 <@* from-state action into middle states
496 They are now:
497 <>/ eof action into middle states
498 <>! error action into middle states
499 <>^ local error action into middle states
500 <>~ to-state action into middle states
501 <>* from-state action into middle states
502 -The verbose form of embeddings using the <- operator have been removed.
503 This syntax was difficult to remember.
504 -A new verbose form of state-based embedding operators have been added.
505 These are like the symbol versions, except they replace the symbols:
506 / ! ^ ~ *
507 with literal keywords:
508 eof err lerr to from
509 -The following words have been promoted to keywords:
510 when eof err lerr to from
511 -The write statment now gets its own lexical scope in the scanner to ensure
512 that commands are passed through as is (not affected by keywords).
513 -Bug fix: in the code generation of fret in scanner actions the adjustment to
514 p that is needed in some cases (dependent on content of patterns) was not
515 happening.
516 -The fhold directive, which decrements p, cannot be permitted in the pattern
517 action of a scanner item because it will not behave consistently. At the end
518 of a pattern action p could be decremented, set to a new value or left
519 alone. This depends on the contents of the scanner's patterns. The user
520 cannot be expected to predict what will happen to p.
521 -Conditions in D code require a cast to the widec type when computing widec.
522 -Like Java, D code also needs if (true) branches for control flow in actions
523 in order to fool the unreachable code detector. This is now abstracted in
524 all code generators using the CTRL_FLOW() function.
525 -The NULL_ITEM value in java code should be -1. This is needed for
526 maintaining tokstart.
527
528 Ragel 5.14 - Oct 1, 2006
529 ========================
530 -Fixed the check for use of fcall in actions embedded within longest match
531 items. It was emitting an error if an item's longest-match action had an
532 fcall, which is allowed. This bug was introduced while fixing a segfault in
533 version 5.8.
534 -A new minimization option was added: MinimizeMostOps (-l). This option
535 minimizes at every operation except on chains of expressions and chains of
536 terms (eg, union and concat). On these chains it minimizes only at the last
537 operation. This makes test cases with many states compile faster, without
538 killing the performance on grammars like strings2.rl.
539 -The -l minimiziation option was made the default.
540 -Fixes to Java code: Use of the fc value did not work, now fixed. Static data
541 is now declared with the final keyword. Patch from Colin Fleming. Conditions
542 now work when generating Java code.
543 -The option -p was added to rlcodegen which causes printable characters to be
544 printed in GraphViz output. Patch from Colin Fleming.
545 -The "element" keyword no longer exists, removed from vim syntax file.
546 Updated keyword highlighting.
547 -The host language selection is now made in the frontend.
548 -Native host language types are now used when specifying the alphtype.
549 Previously all languages used the set defined by C, and these were mapped to
550 the appropriate type in the backend.
551
552 Ragel 5.13 - Sep 7, 2006
553 ========================
554 -Fixed a careless error which broke Java code generation.
555
556 Ragel 5.12 - Sep 7, 2006
557 ========================
558 -The -o flag did not work in combination with -V. This was fixed.
559 -The split code generation format uses only the required number of digits
560 when writing out the number in the file name of each part.
561 -The -T0, -F0 and -G0 codegens should write out the action list iteration
562 variables only when there are regular, to state or from state actions. The
563 code gens should not use anyActions().
564 -If two states have the same EOF actions, they are written out in the finish
565 routine as one case.
566 -The split and in-place goto formats would sometimes generate _out when it is
567 not needed. This was fixed.
568 -Improved the basic partitioning in the split code gen. The last partition
569 would sometimes be empty. This was fixed.
570 -Use of 'fcall *' was not causing top to be initialized. Fixed.
571 -Implemented a Java backend, specified with -J. Only the table-based format
572 is supported.
573 -Implemented range compression in the frontend. This has no effect on the
574 generated code, however it reduces the work of the backend and any programs
575 that read the intermediate format.
576
577 Ragel 5.11 - Aug 10, 2006
578 =========================
579 -Added a variable to the configure.in script which allows the building of
580 the parsers to be turned off (BUILD_PARSERS). Parser building is off by
581 default for released versions.
582 -Removed configure tests for bison defines header file. Use --defines=file
583 instead.
584 -Configure script doesn't test for bison, flex and gperf when building of the
585 parsers is turned off.
586 -Removed check for YYLTYPE structure from configure script. Since shipped
587 code will not build parsers by default, we don't need to be as accomodating
588 of other versions of bison.
589 -Added a missing include that showed up with g++ 2.95.3.
590 -Failed configure test for Objective-C compiler is now silent.
591
592 Ragel 5.10 - Jul 31, 2006
593 =========================
594 -Moved the check for error state higher in the table-based processing loop.
595 -Replaced naive implementations of condition searching with proper ones. In
596 the table-based formats the searching is also table-based. In the directly
597 executed formats the searching is also directly executable.
598 -The minimization process was made aware of conditions.
599 -A problem with the condition implementation was fixed. Previously we were
600 taking pointers to transitions and then using them after a call to
601 outTransCopy, which was a bad idea because they may be changed by the call.
602 -Added test mailbox3.rl which is based on mailbox2.rl but includes conditions
603 for restricting header and message body lengths.
604 -Eliminated the initial one-character backup of p just before resuming
605 execution.
606 -Added the -s option to the frontend for printing statistics. This currently
607 includes just the number of states.
608 -Sped up the generation of the in-place goto-driven (-G2) code style.
609 -Implemented a split version of in-place goto-driven code style. This code
610 generation style is suitable for producing fast implementations of very
611 large machines. Partitioning is currently naive. In the future a
612 high-quality partitioning program will be employed. The flag for accessing
613 this feature is -Pn, where n is the number of partitions.
614 -Converted mailbox1.rl, strings2.rl and cppscan1.rl tests to support the
615 split code generation.
616 -Fixes and updates were made to the runtests script: added -c for compiling
617 only, changed the -me option to -e, and added support for testing the split
618 code style.
619
620 Ragel 5.9 - Jul 19, 2006
621 ========================
622 -Fixed a bug in the include system which caused malformed output from the
623 frontend when the include was made from a multi-line machine spec and the
624 included file ended in a single line spec (or vice versa).
625 -Static data is now const.
626 -Actions which referenced states but were not embedded caused the frontend to
627 segfault, now fixed.
628 -Manual now built with pdflatex.
629 -The manual was reorganized and expanded. Chapter sequence is now:
630 Introduction, Constructing Machines, Embedding Actions, Controlling
631 Nondeterminism and Interfacing to the Host program.
632
633 Ragel 5.8 - Jun 17, 2006
634 ========================
635 -The internal representation of the alphabet type has been encapsulated
636 into a class and all operations on it have been defined as C++ operators.
637 -The condition implementation now supports range transitions. This allows
638 conditions to be embedded into arbitrary machines. Conditions are still
639 exprimental.
640 -More condition embedding operators were added
641 1. Isolate the start state and embed a condition into all transitions
642 leaving it:
643 >when cond OR >?cond
644 2. Embed a condition into all transitions:
645 when cond OR $when cond OR $?cond
646 3. Embed a condition into pending out transitions:
647 %when cond OR %?cond
648 -Improvements were made to the determinization process to support pending out
649 conditions.
650 -The Vim sytax file was fixed so that :> doesn't cause the match of a label.
651 -The test suite was converted to a single-file format which uses less disk
652 space than the old directory-per-test format.
653
654 Ragel 5.7 - May 14, 2006
655 ========================
656 -Conditions will not be embedded like actions because they involve a
657 manipulation of the state machine they are specified in. They have therefore
658 been taken out of the verbose action embedding form (using the <- compound
659 symbol). A new syntax for specifying conditions has been created:
660 m = '\n' when {i==4};
661 -Fixed a bug which prevented state machine commands like fcurs, fcall, fret,
662 etc, from being accounted for in from-state actions and to-state actions.
663 This prevented some necessary support code from being generated.
664 -Implemented condition testing in remaining code generators.
665 -Configure script now checks for gperf, which is required for building.
666 -Added support for case-insensitive literal strings (in addition to regexes).
667 A case-insensitive string is made by appending an 'i' to the literal, as in
668 'cmd'i or "cmd"i.
669 -Fixed a bug which caused all or expressions inside of all regular
670 expressions to be case-insensitive. For example /[fo]o bar/ would make the
671 [fo] part case-insensitive even though no 'i' was given following the
672 regular expression.
673
674 Ragel 5.6 - Apr 1, 2006
675 =======================
676 -Added a left-guarded concatenation operator. This operator <: is equivalent
677 to ( expr1 $1 . expr2 >0 ). It is useful if you want to prefix a sequence
678 with a sequence of a subset of the characters it matches. For example, one
679 can consume leading whitespace before tokenizing a sequence of whitespace
680 separated words: ( ' '* <: ( ' '+ | [a-z]+ )** )
681 -Removed context embedding code, which has been dead since 5.0.
682
683 Ragel 5.5 - Mar 28, 2006
684 ========================
685 -Implemented a case-insensitive option for regular expressions: /get/i.
686 -If no input file is given to the ragel program it reads from standard input.
687 -The label of the start state has been changed from START to IN to save on
688 required screen space.
689 -Bug fix: \0 was not working in literal strings, due to a change that reduced
690 memory usage by concatenating components of literal strings. Token data
691 length is now passed from the scanner to the paser so that we do not need to
692 rely on null termination.
693
694 Ragel 5.4 - Mar 12, 2006
695 ========================
696 -Eliminated the default transition from the frontend implementation. This
697 default transition was a space-saving optimization that at best could reduce
698 the number of allocated transitions by one half. Unfortunately it
699 complicated the implementation and this stood in the way of introducing
700 conditionals. The default transition may be reintroduced in the future.
701 -Added entry-guarded concatenation. This operator :>, is syntactic sugar
702 for expr1 $0 . expr >1. This operator terminates the matching of the first
703 machine when a first character of the second machine is matched. For
704 example in any* . ';' we never leave the any* machine. If we use any* :> ';'
705 then the any* machine is terminiated upon matching the semi-colon.
706 -Added finish-guarded concatenation. This operator :>>, is syntactic sugar
707 for expr1 $0 . expr @1. This operator is like entry guarded concatenation
708 except the first machine is terminated when the second machine enters a
709 final state. This is useful for delaying the guard until a full pattern is
710 matched. For example as in '/*' any* :>> '*/'.
711 -Added strong subtraction. Where regular subtraction removes from the first
712 machine any strings that are matched by the second machine, strong
713 subtraction removes any strings from the first that contain any strings of
714 the second as a substring. Strong subtraction is syntactic sugar for
715 expr1 - ( any* expr2 any* ).
716 -Eliminated the use of priorities from the examples. Replaced with
717 subtraction, guarded concatenation and longest-match kleene star.
718 -Did some initial work on supporting conditional transitions. Far from
719 complete and very buggy. This code will only be active when conditionals are
720 used.
721
722 Ragel 5.3 - Jan 27, 2006
723 ========================
724 -Added missing semi-colons that cause the build to fail when using older
725 versions of Bison.
726 -Fix for D code: if the contents of an fexec is a single word, the generated
727 code will get interpreted as a C-style cast. Adding two brackets prevents
728 this. Can now turn eliminate the "access this.;" in cppscan5 that was used to
729 get around this problem.
730 -Improved some of the tag names in the intermediate format.
731 -Added unsigned long to the list of supported alphabet types.
732 -Added ids of actions and action lists to XML intermediate format. Makes it
733 more human readable.
734 -Updated to latest Aapl package.
735
736 Ragel 5.2 - Jan 6, 2006
737 ========================
738 -Ragel emits an error if the target of fentry, fcall, fgoto or fnext is inside
739 a longest match operator, or if an action embedding in a longest match
740 machine uses fcall. The fcall command can still be used in pattern actions.
741 -Made improvements to the clang, rlscan, awkemu and cppscan examples.
742 -Some fixes to generated label names: they should all be prefixed with _.
743 -A fix to the Vim syntax highlighting script was made
744 -Many fixes and updates to the documentation. All important features and
745 concepts are now documented. A second chapter describing Ragel's use
746 was added.
747
748 Ragel 5.1 - Dec 22, 2005
749 ========================
750 -Fixes to the matching of section delimiters in Vim syntax file.
751 -If there is a longest match machine, the tokend var is now initialized by
752 write init. This is not necessary for correct functionality, however
753 prevents compiler warnings.
754 -The rlscan example was ported to the longest match operator and changed to
755 emit XML data.
756 -Fix to the error handling in the frontend: if there are errors in the lookup
757 of names at machine generation time then do not emit anything.
758 -If not compiling the full machine in the frontend (by using -M), avoid
759 errors and segfaults caused by names that are not part of the compiled
760 machine.
761 -Longest match bug fix: need to init tokstart when returing from fsm calls
762 that are inside longest match actions.
763 -In Graphviz drawing, the arrow into the start state is not a real
764 transition, do not draw to-state actions on the label.
765 -A bug fix to the handling of non-tag data within an XML tag was made.
766 -Backend exit value fixed: since the parser now accepts nothing so as to
767 avoid a redundant parse error when the frontend dies, we must force an
768 error. The backend should now be properly reporting errors.
769 -The longest match machine now has it's start state set final. An LM machine
770 is in a final state when it has not matched anything, when it has matched
771 and accepted a token and is ready for another, and when it has matched a
772 token but is waiting for some lookahead before determining what to do about
773 it (similar to kleene star).
774 -Element statement removed from some tests.
775 -Entry point names are propagated to the backend and used to label the entry
776 point arrows in Graphviz output.
777
778 Ragel 5.0 - Dec 17, 2005
779 ========================
780 (additional details in V5 release notes)
781 -Ragel has been split into two executables: A frontend which compiles
782 machines and emits them in an XML format, and a backend which generates code
783 or a Graphviz dot file from the XML input. The purpose of this split is to
784 allow Ragel to interface with other tools by means of the XML intermediate
785 format and to reduce complexity by strictly separating the previously
786 entangled phases. The intermediate format will provide a better platform
787 inspecting compiled machines and for extending Ragel to support other host
788 languages.
789 -The host language interface has been reduced significantly. Ragel no longer
790 expects the machine to be implemented as a structure or class and does not
791 generate functions corresponding to initialization, execution and EOF.
792 Instead, Ragel just generates the code of these components, allowing all of
793 them to be placed in a single function if desired. The user specifies a
794 machine in the usual manner, then indicates at which place in the program
795 text the state machine code is to be generated. This is done using the write
796 statement. It is possible to specify to Ragel how it should access the
797 variables it needs (such as the current state) using the access statement.
798 -The host language embedding delimiters have been changed. Single line
799 machines start with '%%' and end at newline. Multiline machines start with
800 '%%{' and end with '}%%'. The machine name is given with the machine
801 statement at the very beginning of the specification. This purpose of this
802 change is to make it easier separate Ragel code from the host language. This
803 will ease the addition of supported host languages.
804 -The structure and class parsing which was previously able to extract a
805 machine's name has been removed since this feature is dependent on the host
806 language and inhibits the move towards a more language-independent frontend.
807 -The init, element and interface statements have been made obsolete by the
808 new host language interface and have been removed.
809 -The fexec action statement has been changed to take only the new position to
810 move to. This statement is more useful for moving backwards and reparsing
811 input than for specifying a whole new buffer entirely and has been shifted
812 to this new use. Giving it only one argument also simplifies the parsing of
813 host code embedded in a Ragel specification. This will ease the addition of
814 supported host languages.
815 -Introduced the fbreak statement, which allows one to stop processing data
816 immediately. The machine ends up in the state that the current transition
817 was to go to. The current character is not changed.
818 -Introduced the noend option for writing the execute code. This inhibits
819 checking if we have reached pe. The machine will run until it goes into the
820 error state or fbreak is hit. This allows one to parse null-terminate
821 strings without first computing the length.
822 -The execute code now breaks out of the processing loop when it moves into
823 the error state. Previously it would run until pe was hit. Breaking out
824 makes the noend option useful when an error is encountered and allows
825 user code to determine where in the input the error occured. It also
826 eliminates needlessly iterating the input buffer.
827 -Introduced the noerror, nofinal and noprefix options for writing the machine
828 data. The first two inhibit the writing of the error state and the
829 first-final state should they not be needed. The noprefix eliminates the
830 prefixing of the data items with the machine name.
831 -Support for the D language has been added. This is specified in the backend
832 with the -D switch.
833 -Since the new host language interface has been reduced considerably, Ragel
834 no longer needs to distinguish between C-based languages. Support for C, C++
835 and Objective-C has been folded into one option in the backend: -C
836 -The code generator has been made independent of the languages that it
837 supports by pushing the language dependent apsects down into the lower
838 levels of the code generator.
839 -Many improvements to the longest match construction were made. It is no
840 longer considered experimental. A longest match machine must appear at the
841 top level of a machine instantiation. Since it does not generate a pure
842 state machine (it may need to backtrack), it cannot be used as an operand to
843 other operators.
844 -References to the current character and current state are now completely
845 banned in EOF actions.
846
847 Ragel 4.2 - Sep 16, 2005
848 ========================
849 (additional details in V4 release notes)
850 -Fixed a bug in the longest match operator. In some states it's possible that
851 we either match a token or match nothing at all. In these states we need to
852 consult the LmSwitch on error so it must be prepared to execute an error
853 handler. We therefore need to init act to this error value (which is zero).
854 We can compute if we need to do this and the code generator emits the
855 initialization only if necessary.
856 -Changed the definition of the token end of longest match actions. It now
857 points to one past the last token. This makes computing the token length
858 easier because you don't have to add one. The longest match variables token
859 start, action identifier and token end are now properly initialized in
860 generated code. They don't need to be initialized in the user's code.
861 -Implemented to-state and from-state actions. These actions are executed on
862 transitions into the state (after the in transition's actions) and on
863 transitions out of the state (before the out transition's actions). See V4
864 release notes for more information.
865 -Since there are no longer any action embedding operators that embed both on
866 transitions and on EOF, any actions that exist in both places will be there
867 because the user has explicitly done so. Presuming this case is rare, and
868 with code duplication in the hands of the user, we therefore give the EOF
869 actions their own action switch in the finish() function. This is further
870 motivated by the fact that the best solution is to do the same for to-state
871 and from-state actions in the main loop.
872 -Longest match actions can now be specified using a named action. Since a
873 word following a longest match item conflicts with the concatenation of a
874 named machine, the => symbol must come immediately before a named action.
875 -The longest match operator permits action and machine definitions in the
876 middle of a longest match construction. These are parsed as if they came
877 before the machine definition they are contained in. Permitting action and
878 machine definitions in a longest match construction allows objects to be
879 defined closer to their use.
880 -The longest match operator can now handle longest match items with no
881 action, where previously Ragel segfaulted.
882 -Updated to Aapl post 2.12.
883 -Fixed a bug in epsilon transition name lookups. After doing a name lookup
884 the result was stored in the parse tree. This is wrong because if a machine
885 is used more than once, each time it may resolve to different targets,
886 however it will be stored in the same place. We now store name resolutions
887 in a separated data structure so that each walk of a parse tree uses the
888 name resolved during the corresponding walk in the name lookup pass.
889 -The operators used to embed context and actions into states have been
890 modified. The V4 release notes contain the full details.
891 -Added zlen builtin machine to represent the zero length machine. Eventually
892 the name "null" will be phased out in favour of zlen because it is unclear
893 whether null matches the zero length string or if it does not match any
894 string at all (as does the empty builtin).
895 -Added verbose versions of action, context and priority embedding. See the V4
896 release notes for the full details. A small example:
897 machine <- all exec { foo(); } <- final eof act1
898 -Bugfix for machines with epsilon ops, but no join operations. I had
899 wrongfully assumed that because epsilon ops can only increase connectivity,
900 that no states are ever merged and therefore a call to fillInStates() is not
901 necessary. In reality, epsilon transitions within one machine can induce the
902 merging of states. In the following, state 2 follows two paths on 'i':
903 main := 'h' -> i 'i h' i: 'i';
904 -Changed the license of the guide from a custom "do not propagate modified
905 versions of this document" license to the GPL.
906
907 Ragel 4.1 - Jun 26, 2005
908 ========================
909 (additional details in V4 release notes)
910 -A bug in include processing was fixed. Surrounding code in an include file
911 was being passed through to the output when it should be ignored. Includes
912 are only for including portions of another machine into he current. This
913 went unnoticed because all tested includes were wrapped in #ifndef ...
914 #endif directives and so did not affect the compilation of the file making
915 the include.
916 -Fixes were made to Vim syntax highlighting file.
917 -Duplicate actions are now removed from action lists.
918 -The character-level negation operator ^ was added. This operator produces a
919 machine that matches single characters that are not matched by the machine
920 it is applied to. This unary prefix operator has the same precedence level
921 as !.
922 -The use of + to specify the a positive literal number was discontinued.
923 -The parser now assigns the subtraction operator a higher precedence than
924 the negation of literal number.
925
926 Ragel 4.0 - May 26, 2005
927 ========================
928 (additional details in V4 release notes)
929 -Operators now strictly embed into a machine either on a specific class of
930 characters or on EOF, but never both. This gives a cleaner association
931 between the operators and the physical state machine entitites they operate
932 on. This change is made up of several parts:
933 1. '%' operator embeds only into leaving characters.
934 2. All global and local error operators only embed on error character
935 transitions, their action will not be triggerend on EOF in non-final
936 states.
937 3. EOF action embedding operators have been added for all classes of states
938 to make up for functionality removed from other operators. These are
939 >/ $/ @/ %/.
940 4. Start transition operator '>' no longer implicitly embeds into leaving
941 transtions when start state is final.
942 -Ragel now emits warnings about the improper use of statements and values in
943 action code that is embedded as an EOF action. Warnings are emitted for fpc,
944 fc, fexec, fbuf and fblen.
945 -Added a longest match construction operator |* machine opt-action; ... *|.
946 This is for repetition where an ability to revert to a shorter, previously
947 matched item is required. This is the same behaviour as flex and re2c. The
948 longest match operator is not a pure FSM construction, it introduces
949 transitions that implicitly hold the current character or reset execution to
950 a previous location in the input. Use of this operator requires the caller
951 of the machine to occasionally hold onto data after a call to the exectute
952 routine. Use of machines generated with this operator as the input to other
953 operators may have undefined results. See examples/cppscan for an example.
954 This is very experimental code.
955 -Action ids are only assigned to actions that are referenced in the final
956 constructed machine, preventing gaps in the action id sequence. Previously
957 an action id was assigned if the action was referenced during parsing.
958 -Machine specifications now begin with %% and are followed with an optional
959 name and either a single Ragel statement or a sequence of statements
960 enclosed in {}.
961 -Ragel no longer generates the FSM's structure or class. It is up to the user
962 to declare the structure and to give it a variable named curs of type
963 integer. If the machine uses the call stack the user must also declare a
964 array of integers named stack and an integer variable named top.
965 -In the case of Objective-C, Ragel no longer generates the interface or
966 implementation directives, allowing the user to declare additional methods.
967 -If a machine specification does not have a name then Ragel tries to find a
968 name for it by first checking if the specification is inside a struct, class
969 or interface. If it is not then it uses the name of the previous machine
970 specification. If still no name is found then this is an error.
971 -Fsm specifications now persist in memory and statements accumulate.
972 -Ragel now has an include statement for including the statements of a machine
973 spec in another file (perhaps because it is the corresponding header file).
974 The include statement can also be used to draw in the statements of another
975 fsm spec in the current file.
976 -The fstack statement is now obsolete and has been removed.
977 -A new statement, simply 'interface;', indicates that ragel should generate
978 the machine's interface. If Ragel sees the main machine it generates the
979 code sections of the machine. Previously, the header portion was generated
980 if the (now removed) struct statement was found and code was generated if
981 any machine definition was found.
982 -Fixed a bug in the resolution of fsm name references in actions. The name
983 resolution code did not recurse into inline code items with children
984 (fgoto*, fcall*, fnext*, and fexec), causing a segfault at code generation
985 time.
986 -Cleaned up the code generators. FsmCodeGen was made into a virtual base
987 class allowing for the language/output-style specific classes to inherit
988 both a language specific and style-specific base class while retaining only
989 one copy of FsmCodeGen. Language specific output can now be moved into the
990 language specific code generators, requiring less duplication of code in the
991 language/output-style specific leaf classes.
992 -Fixed bugs in fcall* implementation of IpgGoto code generation.
993 -If the element type has not been defined Ragel now uses a constant version
994 of the alphtype, not the exact alphtype. In most cases the data pointer of
995 the execute routine should be const. A non-const element type can still be
996 defined with the element statement.
997 -The fc special value now uses getkey for retrieving the current char rather
998 than *_p, which is wrong if the element type is a structure.
999 -User guide converted to TeX and updated for new 4.0 syntax and semantics.
1000
1001 Ragel 3.7 - Oct 31, 2004
1002 ========================
1003 -Bug fix: unreferenced machine instantiations causing segfault due to name
1004 tree and parse tree walk becomming out of syncronization.
1005 -Rewrote representation of inline code blocks using a tree data structure.
1006 This allows special keywords such as fbuf to be used as the operatands of
1007 other fsm commands.
1008 -Documentation updates.
1009 -When deciding whether or not to generate machine instantiations, search the
1010 entire name tree beneath the instantiation for references, not just the
1011 root.
1012 -Removed stray ';' in keller2.rl
1013 -Added fexec for restarting the machine with new buffer data (state stays the
1014 same), fbuf for retrieving the the start of the buf, and fblen for
1015 retrieving the orig buffer length.
1016 -Implemented test/cppscan2 using fexec. This allows token emitting and restart
1017 to stay inside the execute routine, instead of leaving and re-entering on
1018 every token.
1019 -Changed examples/cppscan to use fexec and thereby go much faster.
1020 -Implemented flex and re2c versions of examples/cppscan. Ragel version
1021 goes faster than flex version but not as fast as re2c version.
1022 -Merged in Objective-C patch from Erich Ocean.
1023 -Turned off syncing with stdio in C++ tests to make them go faster.
1024 -Renamed C++ code generaion classes with the Cpp Prefix instead of CC to make
1025 them easier to read.
1026 -In the finish function emit fbuf as 0 cast to a pointer to the element type
1027 so it's type is not interpreted as an integer.
1028 -The number -128 underflows char alphabets on some architectures. Removed
1029 uses of it in tests.
1030 -Disabled the keller2 test because it causes problems on many architectures
1031 due to its large size and compilation requirements.
1032
1033 Ragel 3.6 - Jul 10, 2004
1034 ========================
1035 -Many documentation updates.
1036 -When resolving names, return a set of values so that a reference in an
1037 action block that is embedded more than once won't report distinct entry
1038 points that are actually the same.
1039 -Implemented flat tables. Stores a linear array of indicies into the
1040 transition array and only a low and high key value. Faster than binary
1041 searching for keys but not usable for large alphabets.
1042 -Fixed bug in deleting of transitions leftover from converstion from bst to
1043 list implementation of transitions. Other code cleanup.
1044 -In table based output calculate the cost of using an index. Don't use if
1045 cheaper.
1046 -Changed fstate() value available in init and action code to to fentry() to
1047 reflect the fact that the values returned are intended to be used as targets
1048 in fgoto, fnext and fcall statements. The returned state is not a unique
1049 state representing the label. There can be any number of states representing
1050 a label.
1051 -Added keller2 test, C++ scanning tests and C++ scanning example.
1052 -In table based output split up transitions into targets and actions. This
1053 allows actions to be omitted.
1054 -Broke the components of the state array into separate arrays. Requires
1055 adding some fields where they could previously be omitted, however allows
1056 finer grained control over the sizes of items and an overal size reduction.
1057 Also means that state numbers are not an offset into the state array but
1058 instead a sequence of numbers, meaning the context array does not have any
1059 wasted bits.
1060 -Action lists and transition also have their types chosen to be the smallest
1061 possible for accomodating the contained values.
1062 -Changed curs state stored in fsm struct from _cs to curs. Keep fsm->curs ==
1063 -1 while in machine. Added tests curs1 and curs2.
1064 -Implemented the notion of context. Context can be embedded in states using
1065 >:, $:, @: and %: operators. These embed a named context into start states,
1066 all states, non-start/non-final and final states. If the context is declared
1067 using a context statment
1068 context name;
1069 then the context can be quered for any state using fsm_name_ctx_name(state)
1070 in C code and fsm_name::ctx_name(state) in C++ code. This feature makes it
1071 possible to determine what "part" of the machine is currently active.
1072 -Fixed crash on machine generation of graphs with no final state. If there
1073 is no reference to a final state in a join operation, don't generate one.
1074 -Updated Vim sytax: added labels to inline code, added various C++ keywords.
1075 Don't highlight name separations as labels. Added switch labels, improved
1076 alphtype, element and getkey.
1077 -Fixed line info in error reporting of bad epsilon trans.
1078 -Fixed fstate() for tab code gen.
1079 -Removed references to malloc.h.
1080
1081 Ragel 3.5 - May 29, 2004
1082 ========================
1083 -When parse errors occur, the partially generated output file is deleted and
1084 an non-zero exit status is returned.
1085 -Updated Vim syntax file.
1086 -Implemented the setting of the element type that is passed to the execute
1087 routine as well as method for specifying how ragel should retrive the key
1088 from the element type. This lets ragel process arbitrary structures inside
1089 of which is the key that is parsed.
1090 element struct Element;
1091 getkey fpc->character;
1092 -The current state is now implemented with an int across all machines. This
1093 simplifies working with current state variables. For example this allows a
1094 call stack to be implemented in user code.
1095 -Implemented a method for retrieving the current state, the target state, and
1096 any named states.
1097 fcurs -retrieve the current state
1098 ftargs -retrieve the target state
1099 fstate(name) -retrieve a named state.
1100 -Implemented a mechanism for jumping to and calling to a state stored in a
1101 variable.
1102 fgoto *<expr>; -goto the state returned by the C/C++ expression.
1103 fcall *<expr>; -call the state returned by the C/C++ expression.
1104 -Implemented a mechanism for specifying the next state without immediately
1105 transfering control there (any code following statement is executed).
1106 fnext label; -set the state pointed to by label as the next state.
1107 fnext *<expr>; -set the state returned by the C/C++ expression as the
1108 next.
1109 -Action references are determined from the final machine instead of during
1110 the parse tree walk. Some actions can be referenced in the parse tree but not
1111 show up in the final machine. Machine analysis is now done based on this new
1112 computation.
1113 -Named state lookup now employs a breadth-first search in the lookup and
1114 allows the user to fully qualify names, making it possible to specify
1115 jumps/calls into parts of the machine deep in the name hierarchy. Each part
1116 of name (separated by ::) employs a breadth first search from it's starting
1117 point.
1118 -Name references now must always refer to a single state. Since references to
1119 multiple states is not normally intended, it no longer happens
1120 automatically. This frees the programmer from thinking about whether or not
1121 a state reference is unique. It also avoids the added complexity of
1122 determining when to merge the targets of multiple references. The effect of
1123 references to multiple states can be explicitly created using the join
1124 operator and epsilon transitions.
1125 -M option was split into -S and -M. -S specifies the machine spec to generate
1126 for graphviz output and dumping. -M specifies the machine definition or
1127 instantiation.
1128 -Machine function parameters are now prefixed with and underscore to
1129 avoid the hiding of class members.
1130
1131 Ragel 3.4 - May 8, 2004
1132 =======================
1133 -Added the longest match kleene star operator **, which is synonymous
1134 with ( ( <machine> ) $0 %1 ) *.
1135 -Epsilon operators distinguish between leaving transitions (going to an
1136 another expression in a comma separated list) and non-leaving transitions.
1137 Leaving actions and priorities are appropriately transferred.
1138 -Relative priority of following ops changed to:
1139 1. Action/Priority
1140 2. Epsilon
1141 3. Label
1142 If label is done first then the isolation of the start state in > operators
1143 will cause the label to point to the old start state that doesn't have the
1144 new action/priority.
1145 -Merged >! and >~, @! and @~, %! and %~, and $! and $~ operators to have one
1146 set of global error action operators (>!, @!, %! and $!) that are invoked on
1147 error by unexpected characters as well as by unexepected EOF.
1148 -Added the fpc keyword for use in action code. This is a pointer to the
1149 current character. *fpc == fc. If an action is invoked on EOF then fpc == 0.
1150 -Added >^, @^, %^, and $^ local error operators. Global error operators (>!,
1151 @!, $!, and %!) cause actions to be invoked if the final machine fails.
1152 Local error actions cause actions to be invoked if if the current machine
1153 fails.
1154 -Changed error operators to mean embed global/local error actions in:
1155 >! and !^ -the start state.
1156 @! and @^ -states that are not the start state and are not final.
1157 %! and %^ -final states.
1158 $! and $^ -all states.
1159 -Added >@! which is synonymous >! then @!
1160 -Added >@^ which is synonymous >^ then @^
1161 -Added @%! which is synonymous @! then %!
1162 -Added @%^ which is synonymous >^ then @^
1163 -FsmGraph representation of transition lists was changed from a mapping of
1164 alphabet key -> transition objects using a BST to simply a list of
1165 transition objects. Since the transitions are no longer divided by
1166 single/range, the fast finding of transition objects by key is no longer
1167 required functionality and can be eliminated. This new implementation uses
1168 the same amount of memory however causes less allocations. It also make more
1169 sense for supporting error transitions with actions. Previously an error
1170 transition was represented by a null value in the BST.
1171 -Regular expression ranges are checked to ensure that lower <= upper.
1172 -Added printf-like example.
1173 -Added atoi2, erract2, and gotcallret to the test suite.
1174 -Improved build test to support make -jN and simplified the compiling and
1175 running of tests.
1176
1177 Ragel 3.3 - Mar 7, 2004
1178 =======================
1179 -Portability bug fixes were made. Minimum and maximum integer values are
1180 now taken from the system. An alignment problem on 64bit systems
1181 was fixed.
1182
1183 Ragel 3.2 - Feb 28, 2004
1184 ========================
1185 -Added a Vim syntax file.
1186 -Eliminated length var from generated execute code in favour of an end
1187 pointer. Using length requires two variables be read and written. Using an
1188 end pointer requires one variable read and written and one read. Results in
1189 more optimizable code.
1190 -Minimization is now on by default.
1191 -States are ordered in output by depth first search.
1192 -Bug in minimization fixed. States were not being distinguished based on
1193 error actions.
1194 -Added null and empty builtin machines.
1195 -Added EOF error action operators. These are >~, >@, $~, and %~. EOF error
1196 operators embed actions to take if the EOF is seen and interpreted as an
1197 error. The operators correspond to the following states:
1198 -the start state
1199 -any state with a transition to a final state
1200 -any state with a transiion out
1201 -a final state
1202 -Fixed bug in generation of unreference machine vars using -M. Unreferenced
1203 vars don't have a name tree built underneath when starting from
1204 instantiations. Need to instead build the name tree starting at the var.
1205 -Calls, returns, holds and references to fc in out action code are now
1206 handled for ipgoto output.
1207 -Only actions referenced by an instantiated machine expression are put into
1208 the action index and written out.
1209 -Added rlscan, an example that lexes Ragel input.
1210
1211 Ragel 3.1 - Feb 18, 2004
1212 ========================
1213 -Duplicates in OR literals are removed and no longer cause an assertion
1214 failure.
1215 -Duplicate entry points used in goto and call statements are made into
1216 deterministic entry points.
1217 -Base FsmGraph code moved from aapl into ragel, as an increasing amount
1218 of specialization is required. Too much time was spent attempting to
1219 keep it as a general purpose template.
1220 -FsmGraph code de-templatized and heirarchy squashed to a single class.
1221 -Single transitions taken out of FsmGraph code. In the machine construction
1222 stage, transitions are now implemented only with ranges and default
1223 transtions. This reduces memory consumption, simplifies code and prevents
1224 covered transitions. However it requires the automated selection of single
1225 transitions to keep goto-driven code lean.
1226 -Machine reduction completely rewritten to be in-place. As duplicate
1227 transitions and actions are found and the machine is converted to a format
1228 suitable for writing as C code or as GraphViz input, the memory allocated
1229 for states and transitions is reused, instead of newly allocated.
1230 -New reduction code consolodates ranges, selects a default transition, and
1231 selects single transitions with the goal of joining ranges that are split by
1232 any number of single characters.
1233 -Line directive changed from "# <num> <file>" to the more common format
1234 "#line <num> <file>".
1235 -Operator :! changed to @!. This should have happened in last release.
1236 -Added params example.
1237
1238 Ragel 3.0 - Jan 22, 2004
1239 ========================
1240 -Ragel now parses the contents of struct statements and action code.
1241 -The keyword fc replaces the use of *p to reference the current character in
1242 action code.
1243 -Machine instantiations other than main are allowed.
1244 -Call, jump and return statements are now available in action code. This
1245 facility makes it possible to jump to an error handling machine, call a
1246 sub-machine for parsing a field or to follow paths through a machine as
1247 determined by arbitrary C code.
1248 -Added labels to the language. Labels can be used anywhere in a machine
1249 expression to define an entry point. Also references to machine definitions
1250 cause the implicit creation of a label.
1251 -Added epsilon transitions to the language. Epsilon operators may reference
1252 labels in the current name scope resolved when join operators are evaluated
1253 and at the root of the expression tree of machine assignment/instantiation.
1254 -Added the comma operator, which joins machines together without drawing any
1255 transitions between them. This operator is useful in combination with
1256 labels, the epsilon operator and user code transitions for defining machines
1257 using the named state and transition list paradigm. It is also useful for
1258 invoking transitions based on some analysis of the input or on the
1259 environment.
1260 -Added >!, :!, $!, %! operators for specifying actions to take should the
1261 machine fail. These operators embed actions to execute if the machine
1262 fails in
1263 -the start state
1264 -any state with a transition to a final state
1265 -any state with a transiion out
1266 -a final state
1267 The general rule is that if an action embedding operator embeds an action
1268 into a set of transitions T, then the error-counterpart with a ! embeds an
1269 action into the error transition taken when any transition T is a candidate,
1270 but does not match the input.
1271 -The finishing augmentation operator ':' has been changed to '@'. This
1272 frees the ':' symbol for machine labels and avoids hacks to the parser to
1273 allow the use of ':' for both labels and finishing augmentations. The best
1274 hack required that label names be distinct from machine definition names as
1275 in main := word : word; This restriction is not good because labels are
1276 local to the machine that they are used in whereas machine names are global
1277 entities. Label name choices should not be restricted by the set of names
1278 that are in use for machines.
1279 -Named priority syntax now requires parenthesis surrounding the name and
1280 value pair. This avoids grammar ambiguities now that the ',' operator has
1281 been introduced and makes it more clear that the name and value are an
1282 asscociated pair.
1283 -Backslashes are escaped in line directive paths.
1284
1285 Ragel 2.2 - Oct 6, 2003
1286 =======================
1287 -Added {n}, {,n}, {n,} {n,m} repetition operators.
1288 <expr> {n} -- exactly n repetitions
1289 <expr> {,n} -- zero to n repetitions
1290 <expr> {n,} -- n or more repetitions
1291 <expr> {n,m} -- n to m repetitions
1292 -Bug in binary search table in Aapl fixed. Fixes crashing on machines that
1293 add to action tables that are implicitly shared among transitions.
1294 -Tests using obsolete minimization algorithms are no longer built and run by
1295 default.
1296 -Added atoi and concurrent from examples to the test suite.
1297
1298 Ragel 2.1 - Sep 22, 2003
1299 ========================
1300 -Bug in priority comparison code fixed. Segfaulted on some input with many
1301 embedded priorities.
1302 -Added two new examples.
1303
1304 Ragel 2.0 - Sep 7, 2003
1305 =======================
1306 -Optional (?), One or More (+) and Kleene Star (*) operators changed from
1307 prefix to postfix. Rationale is that postfix version is far more common in
1308 regular expression implementations and will be more readily understood.
1309 -All priority values attached to transitions are now accompanied by a name.
1310 Transitions no longer have default priority values of zero assigned
1311 to them. Only transitions that have different priority values assigned
1312 to the same name influence the NFA-DFA conversion. This scheme reduces
1313 side-effects of priorities.
1314 -Removed the %! statement for unsetting pending out priorities. With
1315 named priorities, it is not necessary to clear the priorities of a
1316 machine with $0 %! because non-colliding names can be used to avoid
1317 side-effects.
1318 -Removed the clear keyword, which was for removing actions from a machine.
1319 Not required functionality and it is non-intuitive to have a language
1320 feature that undoes previous definitions.
1321 -Removed the ^ modifier to repetition and concatenation operators. This
1322 undocumented feature prevented out transitions and out priorities from being
1323 transfered from final states to transitions leaving machines. Not required
1324 functionality and complicates the language unnecessarily.
1325 -Keyword 'func' changed to 'action' as a part of the phasing out of the term
1326 'function' in favour of 'action'. Rationale is that the term 'function'
1327 implies that the code is called like a C function, which is not necessarily
1328 the case. The term 'action' is far more common in state machine compiler
1329 implementations.
1330 -Added the instantiation statement, which looks like a standard variable
1331 assignment except := is used instead of =. Instantiations go into the
1332 same graph dictionary as definitions. In the the future, instantiations
1333 will be used as the target for gotos and calls in action code.
1334 -The main graph should now be explicitly instantiated. If it is not,
1335 a warning is issued.
1336 -Or literal basic machines ([] outside of regular expressions) now support
1337 negation and ranges.
1338 -C and C++ interfaces lowercased. In the C interface an underscore now
1339 separates the fsm machine and the function name. Rationale is that lowercased
1340 library and generated routines are more common.
1341 C output:
1342 int fsm_init( struct clang *fsm );
1343 int fsm_execute( struct clang *fsm, char *data, int dlen );
1344 int fsm_finish( struct clang *fsm );
1345 C++ output:
1346 int fsm::init( );
1347 int fsm::execute( char *data, int dlen );
1348 int fsm::finish( );
1349 -Init, execute and finish all return -1 if the machine is in the error state
1350 and can never accept, 0 if the machine is in a non-accepting state that has a
1351 path to a final state and 1 if the machine is in an accepting state.
1352 -Accept routine eliminated. Determining whether or not the machine accepts is
1353 done by examining the return value of the finish routine.
1354 -In C output, fsm structure is no longer a typedef, so referencing requires
1355 the struct keyword. This is to stay in line with C language conventions.
1356 -In C++ output, constructor is no longer written by ragel. As a consequence,
1357 init routine is not called automatically. Allows constructor to be supplied
1358 by user as well as the return value of init to be examined without calling it
1359 twice.
1360 -Static start state and private structures are taken out of C++ classes.
1361
1362 Ragel 1.5.4 - Jul 14, 2003
1363 ==========================
1364 -Workaround for building with bison 1.875, which produces an
1365 optimization that doesn't build with newer version gcc.
1366
1367 Ragel 1.5.3 - Jul 10, 2003
1368 ==========================
1369 -Fixed building with versions of flex that recognize YY_NO_UNPUT.
1370 -Fixed version numbers in ragel.spec file.
1371
1372 Ragel 1.5.2 - Jul 7, 2003
1373 =========================
1374 -Transition actions and out actions displayed in the graphviz output.
1375 -Transitions on negative numbers handled in graphviz output.
1376 -Warning generated when using bison 1.875 now squashed.
1377
1378 Ragel 1.5.1 - Jun 21, 2003
1379 ==========================
1380 -Bugs fixed: Don't delete the output objects when writing to standard out.
1381 Copy mem into parser buffer with memcpy, not strcpy. Fixes buffer mem errror.
1382 -Fixes for compiling with Sun WorkShop 6 compilers.
1383
1384 Ragel 1.5.0 - Jun 10, 2003
1385 ==========================
1386 -Line directives written to the output so that errors in the action code
1387 are properly reported in the ragel input file.
1388 -Simple graphviz dot file output format is supported. Shows states and
1389 transitions. Does not yet show actions.
1390 -Options -p and -f dropped in favour of -d output format.
1391 -Added option -M for specifying the machine to dump with -d or the graph to
1392 generate with -V.
1393 -Error recovery implemented.
1394 -Proper line and column number tracking implemented in the scanner.
1395 -All action/function code is now embedded in the main Execute routine. Avoids
1396 duplication of action code in the Finish routine and the need to call
1397 ExecFuncs which resulted in huge code bloat. Will also allow actions to
1398 modify cs when fsm goto, call and return is supported in action code.
1399 -Fsm spec can have no statements, nothing will be generated.
1400 -Bug fix: Don't accept ] as the opening of a .-. range a reg exp.
1401 -Regular expression or set ranges (ie /[0-9]/) are now handled by the parser
1402 and consequently must be well-formed. The following now generates a parser
1403 error: /[+-]/ and must be rewritten as /[+\-]/. Also fixes a bug whereby ]
1404 might be accepted as the opening of a .-. range causing /[0-9]-[0-9]/ to
1405 parse incorrectly.
1406 -\v, \f, and \r are now treated as whitespace in an fsm spec.
1407
1408 Ragel 1.4.1 - Nov 19, 2002
1409 ==========================
1410 -Compile fixes. The last release (integer alphabets) was so exciting
1411 that usual portability checks got bypassed.
1412
1413 Ragel 1.4.0 - Nov 19, 2002
1414 ==========================
1415 -Arbitrary integer alphabets are now fully supported! A new language
1416 construct:
1417 'alphtype <type>' added for specifying the type of the alphabet. Default
1418 is 'char'. Possible alphabet types are:
1419 char, unsigned char, short, unsigned short, int, unsigned int
1420 -Literal machines specified in decimal format can now be negative when the
1421 alphabet is a signed type.
1422 -Literal machines (strings, decimal and hex) have their values checked for
1423 overflow/underflow against the size of the alphabet type.
1424 -Table driven and goto driven output redesigned to support ranges. Table
1425 driven uses a binary search for locating single characters and ranges. Goto
1426 driven uses a switch statement for single characters and nested if blocks for
1427 ranges.
1428 -Switch driven output removed due to a lack of consistent advantages. Most of
1429 the time the switch driven FSM is of no use because the goto FSM makes
1430 smaller and faster code. Under certain circumstances it can produce smaller
1431 code than a goto driven fsm and be almost as fast, but some sporadic case
1432 does not warrant maintaining it.
1433 -Many warnings changed to errors.
1434 -Added option -p for printing the final fsm before minimization. This lets
1435 priorities be seen. Priorties are all reset to 0 before minimization. The
1436 exiting option -f prints the final fsm after minimization.
1437 -Fixed a bug in the clang test and example that resulted in redundant actions
1438 being executed.
1439
1440 Ragel 1.3.4 - Nov 6, 2002
1441 =========================
1442 -Fixes to Chapter 1 of the guide.
1443 -Brought back the examples and made them current.
1444 -MSVC is no longer supported for compiling windows binaries because its
1445 support for the C++ standard is frustratingly inadequate, it will cost money
1446 to upgrade if it ever gets better, and MinGW is a much better alternative.
1447 -The build system now supports the --host= option for building ragel
1448 for another system (used for cross compiling a windows binary with MinGW).
1449 -Various design changes and fixes towards the goal of arbitrary integer
1450 alphabets and the handling of larger state machines were made.
1451 -The new shared vector class is now used for action lists in transitions and
1452 states to reduce memory allocations.
1453 -An avl tree is now used for the reduction of transitions and functions of an
1454 fsm graph before making the final machine. The tree allows better scalability
1455 and performance by not requiring consecutively larger heap allocations.
1456 -Final stages in the separation of fsm graph code from action embedding and
1457 priority assignment is complete. Makes the base graph leaner and easier to reuse
1458 in other projects (like Keller).
1459
1460 Ragel 1.3.3 - Oct 22, 2002
1461 ==========================
1462 -More diagrams were added to section 1.7.1 of the user guide.
1463 -FSM Graph code was reworked to spearate the regex/nfa/minimizaion graph
1464 algorithms from the manipulation of state and transition properties.
1465 -An rpm spec file from Cris Bailiff was added. This allows an rpm for ragel
1466 to be built with the command 'rpm -ta ragel-x.x.x.tar.gz'
1467 -Fixes to the build system and corresponding doc updates in the README.
1468 -Removed autil and included the one needed source file directly in the top
1469 level ragel directory.
1470 -Fixed a bug that nullified the 20 times speedup in large compilations
1471 claimed by the last version.
1472 -Removed awk from the doc build (it was added with the last release -- though
1473 not mentioned in the changelog).
1474 -Install of man page was moved to the doc dir. The install also installs the
1475 user guide to $(PREFIX)/share/doc/ragel/
1476
1477 Ragel 1.3.2 - Oct 16, 2002
1478 ==========================
1479 -Added option -v (or --version) to show version information.
1480 -The subtract operator no longer removes transition data from the machine
1481 being subtracted. This is left up to the user for the purpose of making it
1482 possible to transfer transitions using subtract and also for speeding up the
1483 subtract routine. Note that it is possible to explicitly clear transition
1484 data before a doing a subtract.
1485 -Rather severe typo bug fixed. Bug was related to transitions with higher
1486 priorities taking precedence. A wrong ptr was being returned. It appears to
1487 have worked most of the time becuase the old ptr was deleted and the new one
1488 allocated immediatly after so the old ptr often pointed to the same space.
1489 Just luck though.
1490 -Bug in the removing of dead end paths was fixed. If the start state
1491 has in transitions then those paths were not followed when finding states to
1492 keep. Would result in non-dead end states being removed from the graph.
1493 -In lists and in ranges are no longer maintained as a bst with the key as the
1494 alphabet character and the value as a list of transitions coming in on that
1495 char. There is one list for each of inList, inRange and inDefault. Now that
1496 the required functionality of the graph is well known it is safe to remove
1497 these lists to gain in speed and footprint. They shouldn't be needed.
1498 -IsolateStartState() runs on modification of start data only if the start
1499 state is not already isolated, which is now possible with the new in list
1500 representation.
1501 -Concat, Or and Star operators now use an approximation to
1502 removeUnreachableStates that does not require a traversal of the entire
1503 graph. This combined with an 'on-the-fly' management of final bits and final
1504 state status results is a dramatic speed increase when compiling machines
1505 that use those operators heavily. The strings2 test goes 20 times faster.
1506 -Before the final minimization, after all fsm operations are complete,
1507 priority data is reset which enables better minimization in cases where
1508 priorities would otherwise separate similar states.
1509
1510 Ragel 1.3.1 - Oct 2, 2002
1511 =========================
1512 -Range transitions are now used to implement machines made with /[a-z]/ and
1513 the .. operator as well as most of the builtin machines. The ranges are not
1514 yet reflected in the output code, they are expanded as if they came from the
1515 regular single transitions. This is one step closer to arbitrary integer
1516 output.
1517 -The builtin machine 'any' was added. It is equiv to the builtin extend,
1518 matching any characters.
1519 -The builtin machine 'cntrl' now includes newline.
1520 -The builtin machine 'space' now includes newline.
1521 -The builtin machine 'ascii' is now the range 0-127, not all characters.
1522 -A man page was written.
1523 -A proper user guide was started. Chapter 1: Specifying Ragel Programs
1524 was written. It even has some diagrams :)
1525
1526 Ragel 1.3.0 - Sep 4, 2002
1527 =========================
1528 -NULL keyword no longer used in table output.
1529 -Though not yet in use, underlying graph structure changed to support range
1530 transitions. As a result, most of the code that walks transition lists is now
1531 implemented with an iterator that hides the complexity of the transition
1532 lists and ranges. Range transitions will be used to implement /[a-z]/ style
1533 machines and machines made with the .. operator. Previously a single
1534 transition would be used for each char in the range, which is very costly.
1535 Ranges eliminate much of the space complexity and allow for the .. operator
1536 to be used with very large (integer) alphabets.
1537 -New minimization similar to Hopcroft's alg. It does not require n^2 space and
1538 runs close to O(n*log(n)) (an exact analysis of the alg is very hard). It is
1539 much better than the stable and approx minimization and obsoletes them both.
1540 An exact implementation of Hopcroft's alg is desirable but not possible
1541 because the ragel implementation does not assume a finite alphabet, which
1542 Hopcroft's requires. Ragel will support arbitrary integer alphabets which
1543 must be treated as an infinite set for implementation considerations.
1544 -New option -m using above described minimization to replace all previous
1545 minimization options. Old options sill work but are obsolete and not
1546 advertised with -h.
1547 -Bug fixed in goto style output. The error exit set the current state to 0,
1548 which is actually a valid state. If the machine was entered again it would go
1549 into the first state, very wrong. If the first state happened to be final then
1550 an immediate finish would accept when in fact it should fail.
1551 -Slightly better fsm minimization now capable due to clearing of the
1552 transition ordering numbers just prior to minimization.
1553
1554 Ragel 1.2.2 - May 25, 2002
1555 ==========================
1556 -Configuration option --prefix now works when installing.
1557 -cc file extension changed to cpp for better portability.
1558 -Unlink of output file upon error no longer happens, removes dependency on
1559 unlink system command.
1560 -All multiline strings removed: not standard c++.
1561 -Awk build dependency removed.
1562 -MSVC 6.0 added to the list of supported compilers (with some tweaking of
1563 bison and flex output).
1564
1565 Ragel 1.2.1 - May 13, 2002
1566 ==========================
1567 -Automatic dependencies were fixed, they were not working correctly.
1568 -Updated AUTHORS file to reflect contributors.
1569 -Code is more C++ standards compliant: compiles with g++ 3.0
1570 -Fixed bugs that only showed up in g++ 3.0
1571 -Latest (unreleased) Aapl.
1572 -Configuration script bails out if bison++ is installed. Ragel will not
1573 compile with bison++ because it is coded in c++ and bison++ automatically
1574 generates a c++ parser. Ragel uses a c-style bison parser.
1575
1576 Ragel 1.2.0 - May 3, 2002
1577 =========================
1578 -Underlying graph structure now supports default transitions. The result is
1579 that a transition does not need to be made for each char of the alphabet
1580 when making 'extend' or '/./' machines. Ragel compiles machines that
1581 use the aforementioned primitives WAY faster.
1582 -The ugly hacks needed to pick default transitions now go away due to
1583 the graph supporting default transitions directly.
1584 -If -e is given, but minimization is not turned on, print a warning.
1585 -Makefiles use automatic dependencies.
1586
1587 Ragel 1.1.0 - Apr 15, 2002
1588 ==========================
1589 -Added goto fsm: much faster than any other fsm style.
1590 -Default operator (if two machines are side by side with no operator
1591 between them) is concatenation. First showed up in 1.0.4.
1592 -The fsm machine no longer auotmatically builds the flat table for
1593 transition indicies. Instead it keeps the key,ptr pair. In tabcodegen
1594 the flat table is produced. This way very large alphabets with sparse
1595 transitions will not consume large amounts of mem. This is also in prep
1596 for fsm graph getting a default transition.
1597 -Generated code contains a statement explicitly stating that ragel fsms
1598 are NOT covered by the GPL. Technically, Ragel copies part of itself
1599 to the output to make the generic fsm execution routine (for table driven
1600 fsms only) and so the output could be considered under the GPL. But this
1601 code is very trivial and could easlily be rewritten. The actual fsm data
1602 is subject to the copyright of the source. To promote the use of Ragel,
1603 a special exception is made for the part of the output copied from Ragel:
1604 it may be used without restriction.
1605 -Much more elegant code generation scheme is employed. Code generation
1606 class members need only put the 'codegen' keyword after their 'void' type
1607 in order to be automatically registerd to handle macros of the same name.
1608 An awk script recognises this keyword and generates an appropriate driver.
1609 -Ragel gets a test suite.
1610 -Postfunc and prefunc go away because they are not supported by non
1611 loop-driven fsms (goto, switch) and present duplicate functionality.
1612 Universal funcs can be implemented by using $ operator.
1613 -Automatic dependencies used in build system, no more make depend target.
1614 -Code generation section in docs.
1615 -Uses the latests aapl.
1616
1617 Ragel 1.0.5 - Mar 3, 2002
1618 =========================
1619 -Bugfix in SetErrorState that caused an assertion failure when compiling
1620 simple machines that did not have full transition tables (and thus did
1621 not show up on any example machines). Assertion failure did not occur
1622 when using the switch statement code as ragel does not call SetErrorState
1623 in that case.
1624 -Fixed some missing includes, now compiles on redhat.
1625 -Moved the FsmMachTrans Compare class out of FsmMachTrans. Some compilers
1626 don't deal with nested classes in templates too well.
1627 -Removed old unused BASEREF in fsmgraph and ragel now compiles using
1628 egcs-2.91.66 and presumably SUNWspro. The baseref is no longer needed
1629 because states do not support being elements in multiple lists. I would
1630 rather be able to support more compilers than have this feature.
1631 -Started a README with compilation notes. Started an AUTHORS file.
1632 -Started the user documentation. Describes basic machines and operators.
1633
1634 Ragel 1.0.4 - Mar 1, 2002
1635 =========================
1636 -Ported to the version of Aapl just after 2.2.0 release. See
1637 http://www.ragel.ca/aapl/ for details on aapl.
1638 -Fixed a bug in the clang example: the newline machine was not stared.
1639 -Added explanations to the clang and mailbox examples. This should
1640 help people that want to learn the lanuage as the manual is far from
1641 complete.
1642
1643 Ragel 1.0.3 - Feb 2, 2002
1644 =========================
1645 -Added aapl to the ragel tree. No longer requires you to download
1646 and build aapl separately. Should avoid discouraging impatient users
1647 from compiling ragel.
1648 -Added the examples to the ragel tree.
1649 -Added configure script checks for bison and flex.
1650 -Fixed makefile so as not to die with newer versions of bison that
1651 write the header of the parser to a .hh file.
1652 -Started ChangeLog file.
1653
1654 Ragel 1.0.2 - Jan 30, 2002
1655 ==========================
1656 -Bug fix in calculating highIndex for table based code. Was using
1657 the length of out tranisition table rather than the value at the
1658 end.
1659 -If high/low index are at the limits, output a define in their place,
1660 not the high/low values themselves so as not to cause compiler warnings.
1661 -If the resulting machines don't have any indicies or functions, then
1662 omit the empty unrefereced static arrays so as not to cause compiler
1663 warnings about unused static vars.
1664 -Fixed variable sized indicies support. The header cannot have any
1665 reference to INDEX_TYPE as that info is not known at the time the header
1666 data is written. Forces us to use a void * for pointers to indicies. In
1667 the c++ versions we are forced to make much of the data non-member
1668 static data in the code portion for the same reason.
1669
1670 Ragel 1.0.1 - Jan 28, 2002
1671 ==========================
1672 -Exe name change from reglang to ragel.
1673 -Added ftabcodegen output code style which uses a table for states and
1674 transitions but uses a switch statement for the function execution.
1675 -Reformatted options in usage dump to look better.
1676 -Support escape sequences in [] sections of regular expressions.
1677
1678 Ragel 1.0 - Jan 25, 2002
1679 ========================
1680 -Initial release.