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