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