]> git.saurik.com Git - bison.git/blob - NEWS
* src/getargs.h, src/getargs.c: Swap --report and --trace handling
[bison.git] / NEWS
1 Bison News
2 ----------
3
4 Changes in version 2.3+:
5
6 * Locations columns and lines start at 1.
7 In accordance with the GNU Coding Standards and Emacs.
8
9 * Except for LALR(1) parsers in C with POSIX Yacc emulation enabled (with `-y',
10 `--yacc', or `%yacc'), Bison no longer generates #define statements for
11 associating token numbers with token names. Removing the #define statements
12 helps to sanitize the global namespace during preprocessing, but POSIX Yacc
13 requires them. Bison still generates an enum for token names in all cases.
14
15 * Handling of traditional Yacc prologue blocks is now more consistent but
16 potentially incompatible with previous releases of Bison.
17
18 As before, you declare prologue blocks in your grammar file with the
19 `%{ ... %}' syntax. To generate the pre-prologue, Bison concatenates all
20 prologue blocks that you've declared before the first %union. To generate
21 the post-prologue, Bison concatenates all prologue blocks that you've
22 declared after the first %union.
23
24 Previous releases of Bison inserted the pre-prologue into both the header
25 file and the code file in all cases except for LALR(1) parsers in C. In the
26 latter case, Bison inserted it only into the code file. For parsers in C++,
27 the point of insertion was before any token definitions (which associate
28 token numbers with names). For parsers in C, the point of insertion was
29 after the token definitions.
30
31 Now, Bison never inserts the pre-prologue into the header file. In the code
32 file, it always inserts it before the token definitions.
33
34 * Bison now provides a more flexible alternative to the traditional Yacc
35 prologue blocks: %before-header, %start-header, %end-header, and
36 %after-header.
37
38 For example, the following declaration order in the grammar file reflects the
39 order in which Bison will output these code blocks. However, you are free to
40 declare these code blocks in your grammar file in whatever order is most
41 convenient for you:
42
43 %before-header {
44 /* Bison treats this block like a pre-prologue block: it inserts it into
45 * the code file before the contents of the header file. It does *not*
46 * insert it into the header file. This is a good place to put
47 * #include's that you want at the top of your code file. A common
48 * example is `#include "system.h"'. */
49 }
50 %start-header {
51 /* Bison inserts this block into both the header file and the code file.
52 * In both files, the point of insertion is before any Bison-generated
53 * token, semantic type, location type, and class definitions. This is a
54 * good place to define %union dependencies, for example. */
55 }
56 %union {
57 /* Unlike the traditional Yacc prologue blocks, the output order for the
58 * new %*-header blocks is not affected by their declaration position
59 * relative to any %union in the grammar file. */
60 }
61 %end-header {
62 /* Bison inserts this block into both the header file and the code file.
63 * In both files, the point of insertion is after the Bison-generated
64 * definitions. This is a good place to declare or define public
65 * functions or data structures that depend on the Bison-generated
66 * definitions. */
67 }
68 %after-header {
69 /* Bison treats this block like a post-prologue block: it inserts it into
70 * the code file after the contents of the header file. It does *not*
71 * insert it into the header file. This is a good place to declare or
72 * define internal functions or data structures that depend on the
73 * Bison-generated definitions. */
74 }
75
76 If you have multiple occurrences of any one of the above declarations, Bison
77 will concatenate the contents in declaration order.
78
79 * The option `--report=look-ahead' has been changed to `--report=lookahead'.
80 The old spelling still works, but is not documented and may be removed
81 in a future release.
82
83 Changes in version 2.3, 2006-06-05:
84
85 * GLR grammars should now use `YYRECOVERING ()' instead of `YYRECOVERING',
86 for compatibility with LALR(1) grammars.
87
88 * It is now documented that any definition of YYSTYPE or YYLTYPE should
89 be to a type name that does not contain parentheses or brackets.
90
91 Changes in version 2.2, 2006-05-19:
92
93 * The distribution terms for all Bison-generated parsers now permit
94 using the parsers in nonfree programs. Previously, this permission
95 was granted only for Bison-generated LALR(1) parsers in C.
96
97 * %name-prefix changes the namespace name in C++ outputs.
98
99 * The C++ parsers export their token_type.
100
101 * Bison now allows multiple %union declarations, and concatenates
102 their contents together.
103
104 * New warning: unused values
105 Right-hand side symbols whose values are not used are reported,
106 if the symbols have destructors. For instance:
107
108 exp: exp "?" exp ":" exp { $1 ? $1 : $3; }
109 | exp "+" exp
110 ;
111
112 will trigger a warning about $$ and $5 in the first rule, and $3 in
113 the second ($1 is copied to $$ by the default rule). This example
114 most likely contains three errors, and could be rewritten as:
115
116 exp: exp "?" exp ":" exp
117 { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); }
118 | exp "+" exp
119 { $$ = $1 ? $1 : $3; if ($1) free ($3); }
120 ;
121
122 However, if the original actions were really intended, memory leaks
123 and all, the warnings can be suppressed by letting Bison believe the
124 values are used, e.g.:
125
126 exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); }
127 | exp "+" exp { $$ = $1; (void) $3; }
128 ;
129
130 If there are mid-rule actions, the warning is issued if no action
131 uses it. The following triggers no warning: $1 and $3 are used.
132
133 exp: exp { push ($1); } '+' exp { push ($3); sum (); };
134
135 The warning is intended to help catching lost values and memory leaks.
136 If a value is ignored, its associated memory typically is not reclaimed.
137
138 * %destructor vs. YYABORT, YYACCEPT, and YYERROR.
139 Destructors are now called when user code invokes YYABORT, YYACCEPT,
140 and YYERROR, for all objects on the stack, other than objects
141 corresponding to the right-hand side of the current rule.
142
143 * %expect, %expect-rr
144 Incorrect numbers of expected conflicts are now actual errors,
145 instead of warnings.
146
147 * GLR, YACC parsers.
148 The %parse-params are available in the destructors (and the
149 experimental printers) as per the documentation.
150
151 * Bison now warns if it finds a stray `$' or `@' in an action.
152
153 * %require "VERSION"
154 This specifies that the grammar file depends on features implemented
155 in Bison version VERSION or higher.
156
157 * lalr1.cc: The token and value types are now class members.
158 The tokens were defined as free form enums and cpp macros. YYSTYPE
159 was defined as a free form union. They are now class members:
160 tokens are enumerations of the `yy::parser::token' struct, and the
161 semantic values have the `yy::parser::semantic_type' type.
162
163 If you do not want or can update to this scheme, the directive
164 `%define "global_tokens_and_yystype" "1"' triggers the global
165 definition of tokens and YYSTYPE. This change is suitable both
166 for previous releases of Bison, and this one.
167
168 If you wish to update, then make sure older version of Bison will
169 fail using `%require "2.2"'.
170
171 * DJGPP support added.
172 \f
173 Changes in version 2.1, 2005-09-16:
174
175 * The C++ lalr1.cc skeleton supports %lex-param.
176
177 * Bison-generated parsers now support the translation of diagnostics like
178 "syntax error" into languages other than English. The default
179 language is still English. For details, please see the new
180 Internationalization section of the Bison manual. Software
181 distributors should also see the new PACKAGING file. Thanks to
182 Bruno Haible for this new feature.
183
184 * Wording in the Bison-generated parsers has been changed slightly to
185 simplify translation. In particular, the message "memory exhausted"
186 has replaced "parser stack overflow", as the old message was not
187 always accurate for modern Bison-generated parsers.
188
189 * Destructors are now called when the parser aborts, for all symbols left
190 behind on the stack. Also, the start symbol is now destroyed after a
191 successful parse. In both cases, the behavior was formerly inconsistent.
192
193 * When generating verbose diagnostics, Bison-generated parsers no longer
194 quote the literal strings associated with tokens. For example, for
195 a syntax error associated with '%token NUM "number"' they might
196 print 'syntax error, unexpected number' instead of 'syntax error,
197 unexpected "number"'.
198 \f
199 Changes in version 2.0, 2004-12-25:
200
201 * Possibly-incompatible changes
202
203 - Bison-generated parsers no longer default to using the alloca function
204 (when available) to extend the parser stack, due to widespread
205 problems in unchecked stack-overflow detection. You can "#define
206 YYSTACK_USE_ALLOCA 1" to require the use of alloca, but please read
207 the manual to determine safe values for YYMAXDEPTH in that case.
208
209 - Error token location.
210 During error recovery, the location of the syntax error is updated
211 to cover the whole sequence covered by the error token: it includes
212 the shifted symbols thrown away during the first part of the error
213 recovery, and the lookahead rejected during the second part.
214
215 - Semicolon changes:
216 . Stray semicolons are no longer allowed at the start of a grammar.
217 . Semicolons are now required after in-grammar declarations.
218
219 - Unescaped newlines are no longer allowed in character constants or
220 string literals. They were never portable, and GCC 3.4.0 has
221 dropped support for them. Better diagnostics are now generated if
222 forget a closing quote.
223
224 - NUL bytes are no longer allowed in Bison string literals, unfortunately.
225
226 * New features
227
228 - GLR grammars now support locations.
229
230 - New directive: %initial-action.
231 This directive allows the user to run arbitrary code (including
232 initializing @$) from yyparse before parsing starts.
233
234 - A new directive "%expect-rr N" specifies the expected number of
235 reduce/reduce conflicts in GLR parsers.
236
237 - %token numbers can now be hexadecimal integers, e.g., `%token FOO 0x12d'.
238 This is a GNU extension.
239
240 - The option `--report=lookahead' was changed to `--report=look-ahead'.
241 [However, this was changed back after 2.3.]
242
243 - Experimental %destructor support has been added to lalr1.cc.
244
245 - New configure option --disable-yacc, to disable installation of the
246 yacc command and -ly library introduced in 1.875 for POSIX conformance.
247
248 * Bug fixes
249
250 - For now, %expect-count violations are now just warnings, not errors.
251 This is for compatibility with Bison 1.75 and earlier (when there are
252 reduce/reduce conflicts) and with Bison 1.30 and earlier (when there
253 are too many or too few shift/reduce conflicts). However, in future
254 versions of Bison we plan to improve the %expect machinery so that
255 these violations will become errors again.
256
257 - Within Bison itself, numbers (e.g., goto numbers) are no longer
258 arbitrarily limited to 16-bit counts.
259
260 - Semicolons are now allowed before "|" in grammar rules, as POSIX requires.
261 \f
262 Changes in version 1.875, 2003-01-01:
263
264 * The documentation license has been upgraded to version 1.2
265 of the GNU Free Documentation License.
266
267 * syntax error processing
268
269 - In Yacc-style parsers YYLLOC_DEFAULT is now used to compute error
270 locations too. This fixes bugs in error-location computation.
271
272 - %destructor
273 It is now possible to reclaim the memory associated to symbols
274 discarded during error recovery. This feature is still experimental.
275
276 - %error-verbose
277 This new directive is preferred over YYERROR_VERBOSE.
278
279 - #defining yyerror to steal internal variables is discouraged.
280 It is not guaranteed to work forever.
281
282 * POSIX conformance
283
284 - Semicolons are once again optional at the end of grammar rules.
285 This reverts to the behavior of Bison 1.33 and earlier, and improves
286 compatibility with Yacc.
287
288 - `parse error' -> `syntax error'
289 Bison now uniformly uses the term `syntax error'; formerly, the code
290 and manual sometimes used the term `parse error' instead. POSIX
291 requires `syntax error' in diagnostics, and it was thought better to
292 be consistent.
293
294 - The documentation now emphasizes that yylex and yyerror must be
295 declared before use. C99 requires this.
296
297 - Bison now parses C99 lexical constructs like UCNs and
298 backslash-newline within C escape sequences, as POSIX 1003.1-2001 requires.
299
300 - File names are properly escaped in C output. E.g., foo\bar.y is
301 output as "foo\\bar.y".
302
303 - Yacc command and library now available
304 The Bison distribution now installs a `yacc' command, as POSIX requires.
305 Also, Bison now installs a small library liby.a containing
306 implementations of Yacc-compatible yyerror and main functions.
307 This library is normally not useful, but POSIX requires it.
308
309 - Type clashes now generate warnings, not errors.
310
311 - If the user does not define YYSTYPE as a macro, Bison now declares it
312 using typedef instead of defining it as a macro.
313 For consistency, YYLTYPE is also declared instead of defined.
314
315 * Other compatibility issues
316
317 - %union directives can now have a tag before the `{', e.g., the
318 directive `%union foo {...}' now generates the C code
319 `typedef union foo { ... } YYSTYPE;'; this is for Yacc compatibility.
320 The default union tag is `YYSTYPE', for compatibility with Solaris 9 Yacc.
321 For consistency, YYLTYPE's struct tag is now `YYLTYPE' not `yyltype'.
322 This is for compatibility with both Yacc and Bison 1.35.
323
324 - `;' is output before the terminating `}' of an action, for
325 compatibility with Bison 1.35.
326
327 - Bison now uses a Yacc-style format for conflict reports, e.g.,
328 `conflicts: 2 shift/reduce, 1 reduce/reduce'.
329
330 - `yystype' and `yyltype' are now obsolescent macros instead of being
331 typedefs or tags; they are no longer documented and are planned to be
332 withdrawn in a future release.
333
334 * GLR parser notes
335
336 - GLR and inline
337 Users of Bison have to decide how they handle the portability of the
338 C keyword `inline'.
339
340 - `parsing stack overflow...' -> `parser stack overflow'
341 GLR parsers now report `parser stack overflow' as per the Bison manual.
342
343 * Bison now warns if it detects conflicting outputs to the same file,
344 e.g., it generates a warning for `bison -d -o foo.h foo.y' since
345 that command outputs both code and header to foo.h.
346
347 * #line in output files
348 - --no-line works properly.
349
350 * Bison can no longer be built by a K&R C compiler; it requires C89 or
351 later to be built. This change originally took place a few versions
352 ago, but nobody noticed until we recently asked someone to try
353 building Bison with a K&R C compiler.
354 \f
355 Changes in version 1.75, 2002-10-14:
356
357 * Bison should now work on 64-bit hosts.
358
359 * Indonesian translation thanks to Tedi Heriyanto.
360
361 * GLR parsers
362 Fix spurious parse errors.
363
364 * Pure parsers
365 Some people redefine yyerror to steal yyparse' private variables.
366 Reenable this trick until an official feature replaces it.
367
368 * Type Clashes
369 In agreement with POSIX and with other Yaccs, leaving a default
370 action is valid when $$ is untyped, and $1 typed:
371
372 untyped: ... typed;
373
374 but the converse remains an error:
375
376 typed: ... untyped;
377
378 * Values of mid-rule actions
379 The following code:
380
381 foo: { ... } { $$ = $1; } ...
382
383 was incorrectly rejected: $1 is defined in the second mid-rule
384 action, and is equal to the $$ of the first mid-rule action.
385 \f
386 Changes in version 1.50, 2002-10-04:
387
388 * GLR parsing
389 The declaration
390 %glr-parser
391 causes Bison to produce a Generalized LR (GLR) parser, capable of handling
392 almost any context-free grammar, ambiguous or not. The new declarations
393 %dprec and %merge on grammar rules allow parse-time resolution of
394 ambiguities. Contributed by Paul Hilfinger.
395
396 Unfortunately Bison 1.50 does not work properly on 64-bit hosts
397 like the Alpha, so please stick to 32-bit hosts for now.
398
399 * Output Directory
400 When not in Yacc compatibility mode, when the output file was not
401 specified, running `bison foo/bar.y' created `foo/bar.c'. It
402 now creates `bar.c'.
403
404 * Undefined token
405 The undefined token was systematically mapped to 2 which prevented
406 the use of 2 by the user. This is no longer the case.
407
408 * Unknown token numbers
409 If yylex returned an out of range value, yyparse could die. This is
410 no longer the case.
411
412 * Error token
413 According to POSIX, the error token must be 256.
414 Bison extends this requirement by making it a preference: *if* the
415 user specified that one of her tokens is numbered 256, then error
416 will be mapped onto another number.
417
418 * Verbose error messages
419 They no longer report `..., expecting error or...' for states where
420 error recovery is possible.
421
422 * End token
423 Defaults to `$end' instead of `$'.
424
425 * Error recovery now conforms to documentation and to POSIX
426 When a Bison-generated parser encounters a syntax error, it now pops
427 the stack until it finds a state that allows shifting the error
428 token. Formerly, it popped the stack until it found a state that
429 allowed some non-error action other than a default reduction on the
430 error token. The new behavior has long been the documented behavior,
431 and has long been required by POSIX. For more details, please see
432 Paul Eggert, "Reductions during Bison error handling" (2002-05-20)
433 <http://lists.gnu.org/archive/html/bug-bison/2002-05/msg00038.html>.
434
435 * Traces
436 Popped tokens and nonterminals are now reported.
437
438 * Larger grammars
439 Larger grammars are now supported (larger token numbers, larger grammar
440 size (= sum of the LHS and RHS lengths), larger LALR tables).
441 Formerly, many of these numbers ran afoul of 16-bit limits;
442 now these limits are 32 bits on most hosts.
443
444 * Explicit initial rule
445 Bison used to play hacks with the initial rule, which the user does
446 not write. It is now explicit, and visible in the reports and
447 graphs as rule 0.
448
449 * Useless rules
450 Before, Bison reported the useless rules, but, although not used,
451 included them in the parsers. They are now actually removed.
452
453 * Useless rules, useless nonterminals
454 They are now reported, as a warning, with their locations.
455
456 * Rules never reduced
457 Rules that can never be reduced because of conflicts are now
458 reported.
459
460 * Incorrect `Token not used'
461 On a grammar such as
462
463 %token useless useful
464 %%
465 exp: '0' %prec useful;
466
467 where a token was used to set the precedence of the last rule,
468 bison reported both `useful' and `useless' as useless tokens.
469
470 * Revert the C++ namespace changes introduced in 1.31
471 as they caused too many portability hassles.
472
473 * Default locations
474 By an accident of design, the default computation of @$ was
475 performed after another default computation was performed: @$ = @1.
476 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
477 the computation of @$.
478
479 * Token end-of-file
480 The token end of file may be specified by the user, in which case,
481 the user symbol is used in the reports, the graphs, and the verbose
482 error messages instead of `$end', which remains being the default.
483 For instance
484 %token MYEOF 0
485 or
486 %token MYEOF 0 "end of file"
487
488 * Semantic parser
489 This old option, which has been broken for ages, is removed.
490
491 * New translations
492 Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
493 Croatian, thanks to Denis Lackovic.
494
495 * Incorrect token definitions
496 When given `%token 'a' "A"', Bison used to output `#define 'a' 65'.
497
498 * Token definitions as enums
499 Tokens are output both as the traditional #define's, and, provided
500 the compiler supports ANSI C or is a C++ compiler, as enums.
501 This lets debuggers display names instead of integers.
502
503 * Reports
504 In addition to --verbose, bison supports --report=THINGS, which
505 produces additional information:
506 - itemset
507 complete the core item sets with their closure
508 - lookahead [changed to `look-ahead' in 1.875e through 2.3, but changed back]
509 explicitly associate lookahead tokens to items
510 - solved
511 describe shift/reduce conflicts solving.
512 Bison used to systematically output this information on top of
513 the report. Solved conflicts are now attached to their states.
514
515 * Type clashes
516 Previous versions don't complain when there is a type clash on
517 the default action if the rule has a mid-rule action, such as in:
518
519 %type <foo> bar
520 %%
521 bar: '0' {} '0';
522
523 This is fixed.
524
525 * GNU M4 is now required when using Bison.
526 \f
527 Changes in version 1.35, 2002-03-25:
528
529 * C Skeleton
530 Some projects use Bison's C parser with C++ compilers, and define
531 YYSTYPE as a class. The recent adjustment of C parsers for data
532 alignment and 64 bit architectures made this impossible.
533
534 Because for the time being no real solution for C++ parser
535 generation exists, kludges were implemented in the parser to
536 maintain this use. In the future, when Bison has C++ parsers, this
537 kludge will be disabled.
538
539 This kludge also addresses some C++ problems when the stack was
540 extended.
541 \f
542 Changes in version 1.34, 2002-03-12:
543
544 * File name clashes are detected
545 $ bison foo.y -d -o foo.x
546 fatal error: header and parser would both be named `foo.x'
547
548 * A missing `;' at the end of a rule triggers a warning
549 In accordance with POSIX, and in agreement with other
550 Yacc implementations, Bison will mandate this semicolon in the near
551 future. This eases the implementation of a Bison parser of Bison
552 grammars by making this grammar LALR(1) instead of LR(2). To
553 facilitate the transition, this release introduces a warning.
554
555 * Revert the C++ namespace changes introduced in 1.31, as they caused too
556 many portability hassles.
557
558 * DJGPP support added.
559
560 * Fix test suite portability problems.
561 \f
562 Changes in version 1.33, 2002-02-07:
563
564 * Fix C++ issues
565 Groff could not be compiled for the definition of size_t was lacking
566 under some conditions.
567
568 * Catch invalid @n
569 As is done with $n.
570 \f
571 Changes in version 1.32, 2002-01-23:
572
573 * Fix Yacc output file names
574
575 * Portability fixes
576
577 * Italian, Dutch translations
578 \f
579 Changes in version 1.31, 2002-01-14:
580
581 * Many Bug Fixes
582
583 * GNU Gettext and %expect
584 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
585 Bison dies on incorrect %expectations, we fear there will be
586 too many bug reports for Gettext, so _for the time being_, %expect
587 does not trigger an error when the input file is named `plural.y'.
588
589 * Use of alloca in parsers
590 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
591 malloc exclusively. Since 1.29, but was not NEWS'ed.
592
593 alloca is used only when compiled with GCC, to avoid portability
594 problems as on AIX.
595
596 * yyparse now returns 2 if memory is exhausted; formerly it dumped core.
597
598 * When the generated parser lacks debugging code, YYDEBUG is now 0
599 (as POSIX requires) instead of being undefined.
600
601 * User Actions
602 Bison has always permitted actions such as { $$ = $1 }: it adds the
603 ending semicolon. Now if in Yacc compatibility mode, the semicolon
604 is no longer output: one has to write { $$ = $1; }.
605
606 * Better C++ compliance
607 The output parsers try to respect C++ namespaces.
608 [This turned out to be a failed experiment, and it was reverted later.]
609
610 * Reduced Grammars
611 Fixed bugs when reporting useless nonterminals.
612
613 * 64 bit hosts
614 The parsers work properly on 64 bit hosts.
615
616 * Error messages
617 Some calls to strerror resulted in scrambled or missing error messages.
618
619 * %expect
620 When the number of shift/reduce conflicts is correct, don't issue
621 any warning.
622
623 * The verbose report includes the rule line numbers.
624
625 * Rule line numbers are fixed in traces.
626
627 * Swedish translation
628
629 * Parse errors
630 Verbose parse error messages from the parsers are better looking.
631 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
632 Now: parse error: unexpected '/', expecting "number" or '-' or '('
633
634 * Fixed parser memory leaks.
635 When the generated parser was using malloc to extend its stacks, the
636 previous allocations were not freed.
637
638 * Fixed verbose output file.
639 Some newlines were missing.
640 Some conflicts in state descriptions were missing.
641
642 * Fixed conflict report.
643 Option -v was needed to get the result.
644
645 * %expect
646 Was not used.
647 Mismatches are errors, not warnings.
648
649 * Fixed incorrect processing of some invalid input.
650
651 * Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
652
653 * Fixed some typos in the documentation.
654
655 * %token MY_EOF 0 is supported.
656 Before, MY_EOF was silently renumbered as 257.
657
658 * doc/refcard.tex is updated.
659
660 * %output, %file-prefix, %name-prefix.
661 New.
662
663 * --output
664 New, aliasing `--output-file'.
665 \f
666 Changes in version 1.30, 2001-10-26:
667
668 * `--defines' and `--graph' have now an optional argument which is the
669 output file name. `-d' and `-g' do not change; they do not take any
670 argument.
671
672 * `%source_extension' and `%header_extension' are removed, failed
673 experiment.
674
675 * Portability fixes.
676 \f
677 Changes in version 1.29, 2001-09-07:
678
679 * The output file does not define const, as this caused problems when used
680 with common autoconfiguration schemes. If you still use ancient compilers
681 that lack const, compile with the equivalent of the C compiler option
682 `-Dconst='. autoconf's AC_C_CONST macro provides one way to do this.
683
684 * Added `-g' and `--graph'.
685
686 * The Bison manual is now distributed under the terms of the GNU FDL.
687
688 * The input and the output files has automatically a similar extension.
689
690 * Russian translation added.
691
692 * NLS support updated; should hopefully be less troublesome.
693
694 * Added the old Bison reference card.
695
696 * Added `--locations' and `%locations'.
697
698 * Added `-S' and `--skeleton'.
699
700 * `%raw', `-r', `--raw' is disabled.
701
702 * Special characters are escaped when output. This solves the problems
703 of the #line lines with path names including backslashes.
704
705 * New directives.
706 `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose',
707 `%debug', `%source_extension' and `%header_extension'.
708
709 * @$
710 Automatic location tracking.
711 \f
712 Changes in version 1.28, 1999-07-06:
713
714 * Should compile better now with K&R compilers.
715
716 * Added NLS.
717
718 * Fixed a problem with escaping the double quote character.
719
720 * There is now a FAQ.
721 \f
722 Changes in version 1.27:
723
724 * The make rule which prevented bison.simple from being created on
725 some systems has been fixed.
726 \f
727 Changes in version 1.26:
728
729 * Bison now uses automake.
730
731 * New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
732
733 * Token numbers now start at 257 as previously documented, not 258.
734
735 * Bison honors the TMPDIR environment variable.
736
737 * A couple of buffer overruns have been fixed.
738
739 * Problems when closing files should now be reported.
740
741 * Generated parsers should now work even on operating systems which do
742 not provide alloca().
743 \f
744 Changes in version 1.25, 1995-10-16:
745
746 * Errors in the input grammar are not fatal; Bison keeps reading
747 the grammar file, and reports all the errors found in it.
748
749 * Tokens can now be specified as multiple-character strings: for
750 example, you could use "<=" for a token which looks like <=, instead
751 of chosing a name like LESSEQ.
752
753 * The %token_table declaration says to write a table of tokens (names
754 and numbers) into the parser file. The yylex function can use this
755 table to recognize multiple-character string tokens, or for other
756 purposes.
757
758 * The %no_lines declaration says not to generate any #line preprocessor
759 directives in the parser file.
760
761 * The %raw declaration says to use internal Bison token numbers, not
762 Yacc-compatible token numbers, when token names are defined as macros.
763
764 * The --no-parser option produces the parser tables without including
765 the parser engine; a project can now use its own parser engine.
766 The actions go into a separate file called NAME.act, in the form of
767 a switch statement body.
768 \f
769 Changes in version 1.23:
770
771 The user can define YYPARSE_PARAM as the name of an argument to be
772 passed into yyparse. The argument should have type void *. It should
773 actually point to an object. Grammar actions can access the variable
774 by casting it to the proper pointer type.
775
776 Line numbers in output file corrected.
777 \f
778 Changes in version 1.22:
779
780 --help option added.
781 \f
782 Changes in version 1.20:
783
784 Output file does not redefine const for C++.
785
786 Local Variables:
787 mode: outline
788 End:
789
790 -----
791
792 Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
793 2004, 2005, 2006 Free Software Foundation, Inc.
794
795 This file is part of Bison, the GNU Compiler Compiler.
796
797 Bison is free software; you can redistribute it and/or modify
798 it under the terms of the GNU General Public License as published by
799 the Free Software Foundation; either version 2, or (at your option)
800 any later version.
801
802 Bison is distributed in the hope that it will be useful,
803 but WITHOUT ANY WARRANTY; without even the implied warranty of
804 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
805 GNU General Public License for more details.
806
807 You should have received a copy of the GNU General Public License
808 along with autoconf; see the file COPYING. If not, write to
809 the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
810 Boston, MA 02110-1301, USA.