4 Changes in version 2.3a+ (????-??-??):
6 * The -g and --graph options now output graphs in Graphviz DOT format,
9 * Bison now recognizes two separate kinds of default %destructor's and
12 1. Place `<*>' in a %destructor/%printer symbol list to define a default
13 %destructor/%printer for all grammar symbols for which you have formally
14 declared semantic type tags.
16 2. Place `<!>' in a %destructor/%printer symbol list to define a default
17 %destructor/%printer for all grammar symbols without declared semantic
20 Bison no longer supports the `%symbol-default' notation from Bison 2.3a.
21 `<*>' and `<!>' combined achieve the same effect with one exception: Bison no
22 longer applies any %destructor to a mid-rule value if that mid-rule value is
23 not actually ever referenced using either $$ or $n in a semantic action.
25 See the section `Freeing Discarded Symbols' in the Bison manual for further
28 * The Yacc prologue alternatives from Bison 2.3a have been rewritten as the
33 Other than semantic actions, this is probably the most common place you
34 should write verbatim code for the parser implementation. For C/C++, it
35 replaces the traditional Yacc prologue, `%{CODE%}', for most purposes.
36 For Java, it inserts your CODE into the parser class. Compare with:
38 - `%{CODE%}' appearing after the first `%union {CODE}' in a C/C++
39 based grammar file. While Bison will continue to support `%{CODE%}'
40 for backward compatibility, `%code {CODE}' is cleaner as its
41 functionality does not depend on its position in the grammar file
42 relative to any `%union {CODE}'. Specifically, `%code {CODE}'
43 always inserts your CODE into the parser code file after the usual
44 contents of the parser header file.
45 - `%after-header {CODE}', which only Bison 2.3a supported.
49 This is the right place to write dependency code for externally exposed
50 definitions required by Bison. For C/C++, such exposed definitions are
51 those usually appearing in the parser header file. Thus, this is the
52 right place to define types referenced in `%union {CODE}' directives,
53 and it is the right place to override Bison's default YYSTYPE and
54 YYLTYPE definitions. For Java, this is the right place to write import
55 directives. Compare with:
57 - `%{CODE%}' appearing before the first `%union {CODE}' in a C/C++
58 based grammar file. Unlike `%{CODE%}', `%requires {CODE}' inserts
59 your CODE both into the parser code file and into the parser header
60 file since Bison's required definitions should depend on it in both
62 - `%start-header {CODE}', which only Bison 2.3a supported.
66 This is the right place to write additional definitions you would like
67 Bison to expose externally. For C/C++, this directive inserts your CODE
68 both into the parser header file and into the parser code file after
69 Bison's required definitions. For Java, it inserts your CODE into the
70 parser java file after the parser class. Compare with:
72 - `%end-header {CODE}', which only Bison 2.3a supported.
76 Occasionally for C/C++ it is desirable to insert code near the top of
77 the parser code file. For example:
84 For Java, `%code-top {CODE}' is currently unused. Compare with:
86 - `%{CODE%}' appearing before the first `%union {CODE}' in a C/C++
87 based grammar file. `%code-top {CODE}' is cleaner as its
88 functionality does not depend on its position in the grammar file
89 relative to any `%union {CODE}'.
90 - `%before-header {CODE}', which only Bison 2.3a supported.
92 If you have multiple occurrences of any one of the above four directives,
93 Bison will concatenate the contents in the order they appear in the grammar
96 Also see the new section `Prologue Alternatives' in the Bison manual.
98 Changes in version 2.3a, 2006-09-13:
100 * Instead of %union, you can define and use your own union type
101 YYSTYPE if your grammar contains at least one <type> tag.
102 Your YYSTYPE need not be a macro; it can be a typedef.
103 This change is for compatibility with other Yacc implementations,
104 and is required by POSIX.
106 * Locations columns and lines start at 1.
107 In accordance with the GNU Coding Standards and Emacs.
109 * You may now declare per-type and default %destructor's and %printer's:
113 %union { char *string; }
114 %token <string> STRING1
115 %token <string> STRING2
116 %type <string> string1
117 %type <string> string2
118 %union { char character; }
119 %token <character> CHR
120 %type <character> chr
121 %destructor { free ($$); } %symbol-default
122 %destructor { free ($$); printf ("%d", @$.first_line); } STRING1 string1
123 %destructor { } <character>
125 guarantees that, when the parser discards any user-defined symbol that has a
126 semantic type tag other than `<character>', it passes its semantic value to
127 `free'. However, when the parser discards a `STRING1' or a `string1', it
128 also prints its line number to `stdout'. It performs only the second
129 `%destructor' in this case, so it invokes `free' only once.
131 * Except for LALR(1) parsers in C with POSIX Yacc emulation enabled (with `-y',
132 `--yacc', or `%yacc'), Bison no longer generates #define statements for
133 associating token numbers with token names. Removing the #define statements
134 helps to sanitize the global namespace during preprocessing, but POSIX Yacc
135 requires them. Bison still generates an enum for token names in all cases.
137 * Handling of traditional Yacc prologue blocks is now more consistent but
138 potentially incompatible with previous releases of Bison.
140 As before, you declare prologue blocks in your grammar file with the
141 `%{ ... %}' syntax. To generate the pre-prologue, Bison concatenates all
142 prologue blocks that you've declared before the first %union. To generate
143 the post-prologue, Bison concatenates all prologue blocks that you've
144 declared after the first %union.
146 Previous releases of Bison inserted the pre-prologue into both the header
147 file and the code file in all cases except for LALR(1) parsers in C. In the
148 latter case, Bison inserted it only into the code file. For parsers in C++,
149 the point of insertion was before any token definitions (which associate
150 token numbers with names). For parsers in C, the point of insertion was
151 after the token definitions.
153 Now, Bison never inserts the pre-prologue into the header file. In the code
154 file, it always inserts it before the token definitions.
156 * Bison now provides a more flexible alternative to the traditional Yacc
157 prologue blocks: %before-header, %start-header, %end-header, and
160 For example, the following declaration order in the grammar file reflects the
161 order in which Bison will output these code blocks. However, you are free to
162 declare these code blocks in your grammar file in whatever order is most
166 /* Bison treats this block like a pre-prologue block: it inserts it into
167 * the code file before the contents of the header file. It does *not*
168 * insert it into the header file. This is a good place to put
169 * #include's that you want at the top of your code file. A common
170 * example is `#include "system.h"'. */
173 /* Bison inserts this block into both the header file and the code file.
174 * In both files, the point of insertion is before any Bison-generated
175 * token, semantic type, location type, and class definitions. This is a
176 * good place to define %union dependencies, for example. */
179 /* Unlike the traditional Yacc prologue blocks, the output order for the
180 * new %*-header blocks is not affected by their declaration position
181 * relative to any %union in the grammar file. */
184 /* Bison inserts this block into both the header file and the code file.
185 * In both files, the point of insertion is after the Bison-generated
186 * definitions. This is a good place to declare or define public
187 * functions or data structures that depend on the Bison-generated
191 /* Bison treats this block like a post-prologue block: it inserts it into
192 * the code file after the contents of the header file. It does *not*
193 * insert it into the header file. This is a good place to declare or
194 * define internal functions or data structures that depend on the
195 * Bison-generated definitions. */
198 If you have multiple occurrences of any one of the above declarations, Bison
199 will concatenate the contents in declaration order.
201 * The option `--report=look-ahead' has been changed to `--report=lookahead'.
202 The old spelling still works, but is not documented and may be removed
205 Changes in version 2.3, 2006-06-05:
207 * GLR grammars should now use `YYRECOVERING ()' instead of `YYRECOVERING',
208 for compatibility with LALR(1) grammars.
210 * It is now documented that any definition of YYSTYPE or YYLTYPE should
211 be to a type name that does not contain parentheses or brackets.
213 Changes in version 2.2, 2006-05-19:
215 * The distribution terms for all Bison-generated parsers now permit
216 using the parsers in nonfree programs. Previously, this permission
217 was granted only for Bison-generated LALR(1) parsers in C.
219 * %name-prefix changes the namespace name in C++ outputs.
221 * The C++ parsers export their token_type.
223 * Bison now allows multiple %union declarations, and concatenates
224 their contents together.
226 * New warning: unused values
227 Right-hand side symbols whose values are not used are reported,
228 if the symbols have destructors. For instance:
230 exp: exp "?" exp ":" exp { $1 ? $1 : $3; }
234 will trigger a warning about $$ and $5 in the first rule, and $3 in
235 the second ($1 is copied to $$ by the default rule). This example
236 most likely contains three errors, and could be rewritten as:
238 exp: exp "?" exp ":" exp
239 { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); }
241 { $$ = $1 ? $1 : $3; if ($1) free ($3); }
244 However, if the original actions were really intended, memory leaks
245 and all, the warnings can be suppressed by letting Bison believe the
246 values are used, e.g.:
248 exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); }
249 | exp "+" exp { $$ = $1; (void) $3; }
252 If there are mid-rule actions, the warning is issued if no action
253 uses it. The following triggers no warning: $1 and $3 are used.
255 exp: exp { push ($1); } '+' exp { push ($3); sum (); };
257 The warning is intended to help catching lost values and memory leaks.
258 If a value is ignored, its associated memory typically is not reclaimed.
260 * %destructor vs. YYABORT, YYACCEPT, and YYERROR.
261 Destructors are now called when user code invokes YYABORT, YYACCEPT,
262 and YYERROR, for all objects on the stack, other than objects
263 corresponding to the right-hand side of the current rule.
265 * %expect, %expect-rr
266 Incorrect numbers of expected conflicts are now actual errors,
270 The %parse-params are available in the destructors (and the
271 experimental printers) as per the documentation.
273 * Bison now warns if it finds a stray `$' or `@' in an action.
276 This specifies that the grammar file depends on features implemented
277 in Bison version VERSION or higher.
279 * lalr1.cc: The token and value types are now class members.
280 The tokens were defined as free form enums and cpp macros. YYSTYPE
281 was defined as a free form union. They are now class members:
282 tokens are enumerations of the `yy::parser::token' struct, and the
283 semantic values have the `yy::parser::semantic_type' type.
285 If you do not want or can update to this scheme, the directive
286 `%define "global_tokens_and_yystype" "1"' triggers the global
287 definition of tokens and YYSTYPE. This change is suitable both
288 for previous releases of Bison, and this one.
290 If you wish to update, then make sure older version of Bison will
291 fail using `%require "2.2"'.
293 * DJGPP support added.
295 Changes in version 2.1, 2005-09-16:
297 * The C++ lalr1.cc skeleton supports %lex-param.
299 * Bison-generated parsers now support the translation of diagnostics like
300 "syntax error" into languages other than English. The default
301 language is still English. For details, please see the new
302 Internationalization section of the Bison manual. Software
303 distributors should also see the new PACKAGING file. Thanks to
304 Bruno Haible for this new feature.
306 * Wording in the Bison-generated parsers has been changed slightly to
307 simplify translation. In particular, the message "memory exhausted"
308 has replaced "parser stack overflow", as the old message was not
309 always accurate for modern Bison-generated parsers.
311 * Destructors are now called when the parser aborts, for all symbols left
312 behind on the stack. Also, the start symbol is now destroyed after a
313 successful parse. In both cases, the behavior was formerly inconsistent.
315 * When generating verbose diagnostics, Bison-generated parsers no longer
316 quote the literal strings associated with tokens. For example, for
317 a syntax error associated with '%token NUM "number"' they might
318 print 'syntax error, unexpected number' instead of 'syntax error,
319 unexpected "number"'.
321 Changes in version 2.0, 2004-12-25:
323 * Possibly-incompatible changes
325 - Bison-generated parsers no longer default to using the alloca function
326 (when available) to extend the parser stack, due to widespread
327 problems in unchecked stack-overflow detection. You can "#define
328 YYSTACK_USE_ALLOCA 1" to require the use of alloca, but please read
329 the manual to determine safe values for YYMAXDEPTH in that case.
331 - Error token location.
332 During error recovery, the location of the syntax error is updated
333 to cover the whole sequence covered by the error token: it includes
334 the shifted symbols thrown away during the first part of the error
335 recovery, and the lookahead rejected during the second part.
338 . Stray semicolons are no longer allowed at the start of a grammar.
339 . Semicolons are now required after in-grammar declarations.
341 - Unescaped newlines are no longer allowed in character constants or
342 string literals. They were never portable, and GCC 3.4.0 has
343 dropped support for them. Better diagnostics are now generated if
344 forget a closing quote.
346 - NUL bytes are no longer allowed in Bison string literals, unfortunately.
350 - GLR grammars now support locations.
352 - New directive: %initial-action.
353 This directive allows the user to run arbitrary code (including
354 initializing @$) from yyparse before parsing starts.
356 - A new directive "%expect-rr N" specifies the expected number of
357 reduce/reduce conflicts in GLR parsers.
359 - %token numbers can now be hexadecimal integers, e.g., `%token FOO 0x12d'.
360 This is a GNU extension.
362 - The option `--report=lookahead' was changed to `--report=look-ahead'.
363 [However, this was changed back after 2.3.]
365 - Experimental %destructor support has been added to lalr1.cc.
367 - New configure option --disable-yacc, to disable installation of the
368 yacc command and -ly library introduced in 1.875 for POSIX conformance.
372 - For now, %expect-count violations are now just warnings, not errors.
373 This is for compatibility with Bison 1.75 and earlier (when there are
374 reduce/reduce conflicts) and with Bison 1.30 and earlier (when there
375 are too many or too few shift/reduce conflicts). However, in future
376 versions of Bison we plan to improve the %expect machinery so that
377 these violations will become errors again.
379 - Within Bison itself, numbers (e.g., goto numbers) are no longer
380 arbitrarily limited to 16-bit counts.
382 - Semicolons are now allowed before "|" in grammar rules, as POSIX requires.
384 Changes in version 1.875, 2003-01-01:
386 * The documentation license has been upgraded to version 1.2
387 of the GNU Free Documentation License.
389 * syntax error processing
391 - In Yacc-style parsers YYLLOC_DEFAULT is now used to compute error
392 locations too. This fixes bugs in error-location computation.
395 It is now possible to reclaim the memory associated to symbols
396 discarded during error recovery. This feature is still experimental.
399 This new directive is preferred over YYERROR_VERBOSE.
401 - #defining yyerror to steal internal variables is discouraged.
402 It is not guaranteed to work forever.
406 - Semicolons are once again optional at the end of grammar rules.
407 This reverts to the behavior of Bison 1.33 and earlier, and improves
408 compatibility with Yacc.
410 - `parse error' -> `syntax error'
411 Bison now uniformly uses the term `syntax error'; formerly, the code
412 and manual sometimes used the term `parse error' instead. POSIX
413 requires `syntax error' in diagnostics, and it was thought better to
416 - The documentation now emphasizes that yylex and yyerror must be
417 declared before use. C99 requires this.
419 - Bison now parses C99 lexical constructs like UCNs and
420 backslash-newline within C escape sequences, as POSIX 1003.1-2001 requires.
422 - File names are properly escaped in C output. E.g., foo\bar.y is
423 output as "foo\\bar.y".
425 - Yacc command and library now available
426 The Bison distribution now installs a `yacc' command, as POSIX requires.
427 Also, Bison now installs a small library liby.a containing
428 implementations of Yacc-compatible yyerror and main functions.
429 This library is normally not useful, but POSIX requires it.
431 - Type clashes now generate warnings, not errors.
433 - If the user does not define YYSTYPE as a macro, Bison now declares it
434 using typedef instead of defining it as a macro.
435 For consistency, YYLTYPE is also declared instead of defined.
437 * Other compatibility issues
439 - %union directives can now have a tag before the `{', e.g., the
440 directive `%union foo {...}' now generates the C code
441 `typedef union foo { ... } YYSTYPE;'; this is for Yacc compatibility.
442 The default union tag is `YYSTYPE', for compatibility with Solaris 9 Yacc.
443 For consistency, YYLTYPE's struct tag is now `YYLTYPE' not `yyltype'.
444 This is for compatibility with both Yacc and Bison 1.35.
446 - `;' is output before the terminating `}' of an action, for
447 compatibility with Bison 1.35.
449 - Bison now uses a Yacc-style format for conflict reports, e.g.,
450 `conflicts: 2 shift/reduce, 1 reduce/reduce'.
452 - `yystype' and `yyltype' are now obsolescent macros instead of being
453 typedefs or tags; they are no longer documented and are planned to be
454 withdrawn in a future release.
459 Users of Bison have to decide how they handle the portability of the
462 - `parsing stack overflow...' -> `parser stack overflow'
463 GLR parsers now report `parser stack overflow' as per the Bison manual.
465 * Bison now warns if it detects conflicting outputs to the same file,
466 e.g., it generates a warning for `bison -d -o foo.h foo.y' since
467 that command outputs both code and header to foo.h.
469 * #line in output files
470 - --no-line works properly.
472 * Bison can no longer be built by a K&R C compiler; it requires C89 or
473 later to be built. This change originally took place a few versions
474 ago, but nobody noticed until we recently asked someone to try
475 building Bison with a K&R C compiler.
477 Changes in version 1.75, 2002-10-14:
479 * Bison should now work on 64-bit hosts.
481 * Indonesian translation thanks to Tedi Heriyanto.
484 Fix spurious parse errors.
487 Some people redefine yyerror to steal yyparse' private variables.
488 Reenable this trick until an official feature replaces it.
491 In agreement with POSIX and with other Yaccs, leaving a default
492 action is valid when $$ is untyped, and $1 typed:
496 but the converse remains an error:
500 * Values of mid-rule actions
503 foo: { ... } { $$ = $1; } ...
505 was incorrectly rejected: $1 is defined in the second mid-rule
506 action, and is equal to the $$ of the first mid-rule action.
508 Changes in version 1.50, 2002-10-04:
513 causes Bison to produce a Generalized LR (GLR) parser, capable of handling
514 almost any context-free grammar, ambiguous or not. The new declarations
515 %dprec and %merge on grammar rules allow parse-time resolution of
516 ambiguities. Contributed by Paul Hilfinger.
518 Unfortunately Bison 1.50 does not work properly on 64-bit hosts
519 like the Alpha, so please stick to 32-bit hosts for now.
522 When not in Yacc compatibility mode, when the output file was not
523 specified, running `bison foo/bar.y' created `foo/bar.c'. It
527 The undefined token was systematically mapped to 2 which prevented
528 the use of 2 by the user. This is no longer the case.
530 * Unknown token numbers
531 If yylex returned an out of range value, yyparse could die. This is
535 According to POSIX, the error token must be 256.
536 Bison extends this requirement by making it a preference: *if* the
537 user specified that one of her tokens is numbered 256, then error
538 will be mapped onto another number.
540 * Verbose error messages
541 They no longer report `..., expecting error or...' for states where
542 error recovery is possible.
545 Defaults to `$end' instead of `$'.
547 * Error recovery now conforms to documentation and to POSIX
548 When a Bison-generated parser encounters a syntax error, it now pops
549 the stack until it finds a state that allows shifting the error
550 token. Formerly, it popped the stack until it found a state that
551 allowed some non-error action other than a default reduction on the
552 error token. The new behavior has long been the documented behavior,
553 and has long been required by POSIX. For more details, please see
554 Paul Eggert, "Reductions during Bison error handling" (2002-05-20)
555 <http://lists.gnu.org/archive/html/bug-bison/2002-05/msg00038.html>.
558 Popped tokens and nonterminals are now reported.
561 Larger grammars are now supported (larger token numbers, larger grammar
562 size (= sum of the LHS and RHS lengths), larger LALR tables).
563 Formerly, many of these numbers ran afoul of 16-bit limits;
564 now these limits are 32 bits on most hosts.
566 * Explicit initial rule
567 Bison used to play hacks with the initial rule, which the user does
568 not write. It is now explicit, and visible in the reports and
572 Before, Bison reported the useless rules, but, although not used,
573 included them in the parsers. They are now actually removed.
575 * Useless rules, useless nonterminals
576 They are now reported, as a warning, with their locations.
578 * Rules never reduced
579 Rules that can never be reduced because of conflicts are now
582 * Incorrect `Token not used'
585 %token useless useful
587 exp: '0' %prec useful;
589 where a token was used to set the precedence of the last rule,
590 bison reported both `useful' and `useless' as useless tokens.
592 * Revert the C++ namespace changes introduced in 1.31
593 as they caused too many portability hassles.
596 By an accident of design, the default computation of @$ was
597 performed after another default computation was performed: @$ = @1.
598 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
599 the computation of @$.
602 The token end of file may be specified by the user, in which case,
603 the user symbol is used in the reports, the graphs, and the verbose
604 error messages instead of `$end', which remains being the default.
608 %token MYEOF 0 "end of file"
611 This old option, which has been broken for ages, is removed.
614 Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
615 Croatian, thanks to Denis Lackovic.
617 * Incorrect token definitions
618 When given `%token 'a' "A"', Bison used to output `#define 'a' 65'.
620 * Token definitions as enums
621 Tokens are output both as the traditional #define's, and, provided
622 the compiler supports ANSI C or is a C++ compiler, as enums.
623 This lets debuggers display names instead of integers.
626 In addition to --verbose, bison supports --report=THINGS, which
627 produces additional information:
629 complete the core item sets with their closure
630 - lookahead [changed to `look-ahead' in 1.875e through 2.3, but changed back]
631 explicitly associate lookahead tokens to items
633 describe shift/reduce conflicts solving.
634 Bison used to systematically output this information on top of
635 the report. Solved conflicts are now attached to their states.
638 Previous versions don't complain when there is a type clash on
639 the default action if the rule has a mid-rule action, such as in:
647 * GNU M4 is now required when using Bison.
649 Changes in version 1.35, 2002-03-25:
652 Some projects use Bison's C parser with C++ compilers, and define
653 YYSTYPE as a class. The recent adjustment of C parsers for data
654 alignment and 64 bit architectures made this impossible.
656 Because for the time being no real solution for C++ parser
657 generation exists, kludges were implemented in the parser to
658 maintain this use. In the future, when Bison has C++ parsers, this
659 kludge will be disabled.
661 This kludge also addresses some C++ problems when the stack was
664 Changes in version 1.34, 2002-03-12:
666 * File name clashes are detected
667 $ bison foo.y -d -o foo.x
668 fatal error: header and parser would both be named `foo.x'
670 * A missing `;' at the end of a rule triggers a warning
671 In accordance with POSIX, and in agreement with other
672 Yacc implementations, Bison will mandate this semicolon in the near
673 future. This eases the implementation of a Bison parser of Bison
674 grammars by making this grammar LALR(1) instead of LR(2). To
675 facilitate the transition, this release introduces a warning.
677 * Revert the C++ namespace changes introduced in 1.31, as they caused too
678 many portability hassles.
680 * DJGPP support added.
682 * Fix test suite portability problems.
684 Changes in version 1.33, 2002-02-07:
687 Groff could not be compiled for the definition of size_t was lacking
688 under some conditions.
693 Changes in version 1.32, 2002-01-23:
695 * Fix Yacc output file names
699 * Italian, Dutch translations
701 Changes in version 1.31, 2002-01-14:
705 * GNU Gettext and %expect
706 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
707 Bison dies on incorrect %expectations, we fear there will be
708 too many bug reports for Gettext, so _for the time being_, %expect
709 does not trigger an error when the input file is named `plural.y'.
711 * Use of alloca in parsers
712 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
713 malloc exclusively. Since 1.29, but was not NEWS'ed.
715 alloca is used only when compiled with GCC, to avoid portability
718 * yyparse now returns 2 if memory is exhausted; formerly it dumped core.
720 * When the generated parser lacks debugging code, YYDEBUG is now 0
721 (as POSIX requires) instead of being undefined.
724 Bison has always permitted actions such as { $$ = $1 }: it adds the
725 ending semicolon. Now if in Yacc compatibility mode, the semicolon
726 is no longer output: one has to write { $$ = $1; }.
728 * Better C++ compliance
729 The output parsers try to respect C++ namespaces.
730 [This turned out to be a failed experiment, and it was reverted later.]
733 Fixed bugs when reporting useless nonterminals.
736 The parsers work properly on 64 bit hosts.
739 Some calls to strerror resulted in scrambled or missing error messages.
742 When the number of shift/reduce conflicts is correct, don't issue
745 * The verbose report includes the rule line numbers.
747 * Rule line numbers are fixed in traces.
749 * Swedish translation
752 Verbose parse error messages from the parsers are better looking.
753 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
754 Now: parse error: unexpected '/', expecting "number" or '-' or '('
756 * Fixed parser memory leaks.
757 When the generated parser was using malloc to extend its stacks, the
758 previous allocations were not freed.
760 * Fixed verbose output file.
761 Some newlines were missing.
762 Some conflicts in state descriptions were missing.
764 * Fixed conflict report.
765 Option -v was needed to get the result.
769 Mismatches are errors, not warnings.
771 * Fixed incorrect processing of some invalid input.
773 * Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
775 * Fixed some typos in the documentation.
777 * %token MY_EOF 0 is supported.
778 Before, MY_EOF was silently renumbered as 257.
780 * doc/refcard.tex is updated.
782 * %output, %file-prefix, %name-prefix.
786 New, aliasing `--output-file'.
788 Changes in version 1.30, 2001-10-26:
790 * `--defines' and `--graph' have now an optional argument which is the
791 output file name. `-d' and `-g' do not change; they do not take any
794 * `%source_extension' and `%header_extension' are removed, failed
799 Changes in version 1.29, 2001-09-07:
801 * The output file does not define const, as this caused problems when used
802 with common autoconfiguration schemes. If you still use ancient compilers
803 that lack const, compile with the equivalent of the C compiler option
804 `-Dconst='. autoconf's AC_C_CONST macro provides one way to do this.
806 * Added `-g' and `--graph'.
808 * The Bison manual is now distributed under the terms of the GNU FDL.
810 * The input and the output files has automatically a similar extension.
812 * Russian translation added.
814 * NLS support updated; should hopefully be less troublesome.
816 * Added the old Bison reference card.
818 * Added `--locations' and `%locations'.
820 * Added `-S' and `--skeleton'.
822 * `%raw', `-r', `--raw' is disabled.
824 * Special characters are escaped when output. This solves the problems
825 of the #line lines with path names including backslashes.
828 `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose',
829 `%debug', `%source_extension' and `%header_extension'.
832 Automatic location tracking.
834 Changes in version 1.28, 1999-07-06:
836 * Should compile better now with K&R compilers.
840 * Fixed a problem with escaping the double quote character.
842 * There is now a FAQ.
844 Changes in version 1.27:
846 * The make rule which prevented bison.simple from being created on
847 some systems has been fixed.
849 Changes in version 1.26:
851 * Bison now uses automake.
853 * New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
855 * Token numbers now start at 257 as previously documented, not 258.
857 * Bison honors the TMPDIR environment variable.
859 * A couple of buffer overruns have been fixed.
861 * Problems when closing files should now be reported.
863 * Generated parsers should now work even on operating systems which do
864 not provide alloca().
866 Changes in version 1.25, 1995-10-16:
868 * Errors in the input grammar are not fatal; Bison keeps reading
869 the grammar file, and reports all the errors found in it.
871 * Tokens can now be specified as multiple-character strings: for
872 example, you could use "<=" for a token which looks like <=, instead
873 of chosing a name like LESSEQ.
875 * The %token_table declaration says to write a table of tokens (names
876 and numbers) into the parser file. The yylex function can use this
877 table to recognize multiple-character string tokens, or for other
880 * The %no_lines declaration says not to generate any #line preprocessor
881 directives in the parser file.
883 * The %raw declaration says to use internal Bison token numbers, not
884 Yacc-compatible token numbers, when token names are defined as macros.
886 * The --no-parser option produces the parser tables without including
887 the parser engine; a project can now use its own parser engine.
888 The actions go into a separate file called NAME.act, in the form of
889 a switch statement body.
891 Changes in version 1.23:
893 The user can define YYPARSE_PARAM as the name of an argument to be
894 passed into yyparse. The argument should have type void *. It should
895 actually point to an object. Grammar actions can access the variable
896 by casting it to the proper pointer type.
898 Line numbers in output file corrected.
900 Changes in version 1.22:
904 Changes in version 1.20:
906 Output file does not redefine const for C++.
914 Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
915 2004, 2005, 2006 Free Software Foundation, Inc.
917 This file is part of Bison, the GNU Compiler Compiler.
919 Bison is free software; you can redistribute it and/or modify
920 it under the terms of the GNU General Public License as published by
921 the Free Software Foundation; either version 2, or (at your option)
924 Bison is distributed in the hope that it will be useful,
925 but WITHOUT ANY WARRANTY; without even the implied warranty of
926 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
927 GNU General Public License for more details.
929 You should have received a copy of the GNU General Public License
930 along with autoconf; see the file COPYING. If not, write to
931 the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
932 Boston, MA 02110-1301, USA.