]> git.saurik.com Git - bison.git/blame_incremental - NEWS
* data/yacc.c: Reformat m4 a little.
[bison.git] / NEWS
... / ...
CommitLineData
1Bison News
2----------
3
4* Changes in version ?.? (????-??-??):
5
6** Additional yylex/yyparse arguments
7
8 The new directive %param declare additional argument to both yylex
9 and yyparse. The %lex-param, %parse-param, and %param directive
10 support one or more arguments. Instead of
11
12 %lex-param {arg1_type *arg1}
13 %lex-param {arg2_type *arg2}
14 %parse-param {arg1_type *arg1}
15 %parse-param {arg2_type *arg2}
16
17 one may now declare
18
19 %param {arg1_type *arg1} {arg2_type *arg2}
20
21** Java skeleton improvements
22
23 The constants for token names were moved to the Lexer interface.
24 Also, it is possible to add code to the parser's constructors using
25 "%code init" and "%define init_throws".
26
27** Variable api.tokens.prefix
28
29 The variable api.tokens.prefix changes the way tokens are identified in
30 the generated files. This is especially useful to avoid collisions
31 with identifiers in the target language. For instance
32
33 %token FILE for ERROR
34 %define api.tokens.prefix "TOK_"
35 %%
36 start: FILE for ERROR;
37
38 will generate the definition of the symbols TOK_FILE, TOK_for, and
39 TOK_ERROR in the generated sources. In particular, the scanner must
40 use these prefixed token names, although the grammar itself still
41 uses the short names (as in the sample rule given above).
42
43** Variable api.namespace
44
45 The "namespace" variable is renamed "api.namespace". Backward
46 compatibility is ensured, but upgrading is recommended.
47
48** Variable parse.error
49
50 The variable error controls the verbosity of error messages. The
51 use of the %error-verbose directive is deprecated in favor of
52 %define parse.error "verbose".
53
54* Changes in version 2.5 (????-??-??):
55
56** Named References Support
57
58 Historically, Yacc and Bison have supported positional references
59 ($n, $$) to allow access to symbol values from inside of semantic
60 actions code.
61
62 Starting from this version, Bison can also accept named references.
63 When no ambiguity is possible, original symbol names may be used
64 as named references:
65
66 if_stmt : 'if' cond_expr 'then' then_stmt ';'
67 { $if_stmt = mk_if_stmt($cond_expr, $then_stmt); }
68
69 In the more common case, explicit names may be declared:
70
71 stmt[res] : 'if' expr[cond] 'then' stmt[then] 'else' stmt[else] ';'
72 { $res = mk_if_stmt($cond, $then, $else); }
73
74 Location information is also accessible using @name syntax. When
75 accessing symbol names containing dots or dashes, explicit bracketing
76 ($[sym.1]) must be used.
77
78 These features are experimental in this version. More user feedback
79 will help to stabilize them.
80
81** IELR(1) and Canonical LR(1) Support
82
83 IELR(1) is a minimal LR(1) parser table generation algorithm. That
84 is, given any context-free grammar, IELR(1) generates parser tables
85 with the full language recognition power of canonical LR(1) but with
86 nearly the same number of parser states as LALR(1). This reduction in
87 parser states is often an order of magnitude. More importantly,
88 because canonical LR(1)'s extra parser states may contain duplicate
89 conflicts in the case of non-LR(1) grammars, the number of conflicts
90 for IELR(1) is often an order of magnitude less as well. This can
91 significantly reduce the complexity of developing of a grammar.
92
93 Bison can now generate IELR(1) and canonical LR(1) parser tables in
94 place of its traditional LALR(1) parser tables, which remain the
95 default. You can specify the type of parser tables in the grammar
96 file with these directives:
97
98 %define lr.type lalr
99 %define lr.type ielr
100 %define lr.type canonical-lr
101
102 The default reduction optimization in the parser tables can also be
103 adjusted using `%define lr.default-reductions'. See the documentation
104 for `%define lr.type' and `%define lr.default-reductions' in the
105 section `Bison Declaration Summary' in the Bison manual for the
106 details.
107
108 These features are experimental. More user feedback will help to
109 stabilize them.
110
111** Unrecognized %code qualifiers are now an error not a warning.
112
113** %define improvements.
114
115*** Unrecognized variables are now an error not a warning.
116
117*** Multiple invocations for any variable is now an error not a warning.
118
119*** Can now be invoked via the command line.
120
121 Each of these command-line options
122
123 -D NAME[=VALUE]
124 --define=NAME[=VALUE]
125
126 -F NAME[=VALUE]
127 --force-define=NAME[=VALUE]
128
129 is equivalent to this grammar file declaration
130
131 %define NAME ["VALUE"]
132
133 except that the manner in which Bison processes multiple definitions
134 for the same NAME differs. Most importantly, -F and --force-define
135 quietly override %define, but -D and --define do not. For further
136 details, see the section "Bison Options" in the Bison manual.
137
138*** Variables renamed.
139
140 The following %define variables
141
142 api.push_pull
143 lr.keep_unreachable_states
144
145 have been renamed to
146
147 api.push-pull
148 lr.keep-unreachable-states
149
150 The old names are now deprecated but will be maintained indefinitely
151 for backward compatibility.
152
153*** Values no longer need to be quoted in grammar file.
154
155 If a %define value is an identifier, it no longer needs to be placed
156 within quotations marks. For example,
157
158 %define api.push-pull "push"
159
160 can be rewritten as
161
162 %define api.push-pull push
163
164** Symbol names.
165
166 Consistently with directives (such as %error-verbose) and variables
167 (e.g. push-pull), symbol names may include dashes in any position,
168 similarly to periods and underscores. This is GNU extension over
169 POSIX Yacc whose use is reported by -Wyacc, and rejected in Yacc
170 mode (--yacc).
171
172** Temporary hack for adding a semicolon to the user action.
173
174 Previously, Bison appended a semicolon to every user action for
175 reductions when the output language defaulted to C (specifically, when
176 neither %yacc, %language, %skeleton, or equivalent command-line
177 options were specified). This allowed actions such as
178
179 exp: exp "+" exp { $$ = $1 + $3 };
180
181 instead of
182
183 exp: exp "+" exp { $$ = $1 + $3; };
184
185 As a first step in removing this misfeature, Bison now issues a
186 warning when it appends a semicolon. Moreover, in cases where Bison
187 cannot easily determine whether a semicolon is needed (for example, an
188 action ending with a cpp directive or a braced compound initializer),
189 it no longer appends one. Thus, the C compiler might now complain
190 about a missing semicolon where it did not before. Future releases of
191 Bison will cease to append semicolons entirely.
192
193** Character literals not of length one.
194
195 Previously, Bison quietly converted all character literals to length
196 one. For example, without warning, Bison interpreted the operators in
197 the following grammar to be the same token:
198
199 exp: exp '++'
200 | exp '+' exp
201 ;
202
203 Bison now warns when a character literal is not of length one. In
204 some future release, Bison will report an error instead.
205
206** Verbose error messages fixed for nonassociative tokens.
207
208 When %error-verbose is specified, syntax error messages produced by
209 the generated parser include the unexpected token as well as a list of
210 expected tokens. Previously, this list erroneously included tokens
211 that would actually induce a syntax error because conflicts for them
212 were resolved with %nonassoc. Such tokens are now properly omitted
213 from the list.
214
215* Changes in version 2.4.2 (????-??-??):
216
217** Detection of GNU M4 1.4.6 or newer during configure is improved.
218
219** Warnings from gcc's -Wundef option about undefined YYENABLE_NLS and
220 YYLTYPE_IS_TRIVIAL in C/C++ parsers are now avoided.
221
222** %code is now a permanent feature.
223
224 A traditional Yacc prologue directive is written in the form:
225
226 %{CODE%}
227
228 To provide a more flexible alternative, Bison 2.3b introduced the
229 %code directive with the following forms for C/C++:
230
231 %code {CODE}
232 %code requires {CODE}
233 %code provides {CODE}
234 %code top {CODE}
235
236 These forms are now considered permanent features of Bison. See the
237 %code entries in the section "Bison Declaration Summary" in the Bison
238 manual for a summary of their functionality. See the section
239 "Prologue Alternatives" for a detailed discussion including the
240 advantages of %code over the traditional Yacc prologue directive.
241
242 Bison's Java feature as a whole including its current usage of %code
243 is still considered experimental.
244
245** Internationalization.
246
247 Fix a regression introduced in Bison 2.4: Under some circumstances,
248 message translations were not installed although supported by the
249 host system.
250
251* Changes in version 2.4.1 (2008-12-11):
252
253** In the GLR defines file, unexpanded M4 macros in the yylval and yylloc
254 declarations have been fixed.
255
256** Temporary hack for adding a semicolon to the user action.
257
258 Bison used to prepend a trailing semicolon at the end of the user
259 action for reductions. This allowed actions such as
260
261 exp: exp "+" exp { $$ = $1 + $3 };
262
263 instead of
264
265 exp: exp "+" exp { $$ = $1 + $3; };
266
267 Some grammars still depend on this `feature'. Bison 2.4.1 restores
268 the previous behavior in the case of C output (specifically, when
269 neither %language or %skeleton or equivalent command-line options
270 are used) to leave more time for grammars depending on the old
271 behavior to be adjusted. Future releases of Bison will disable this
272 feature.
273
274** A few minor improvements to the Bison manual.
275
276* Changes in version 2.4 (2008-11-02):
277
278** %language is an experimental feature.
279
280 We first introduced this feature in test release 2.3b as a cleaner
281 alternative to %skeleton. Since then, we have discussed the possibility of
282 modifying its effect on Bison's output file names. Thus, in this release,
283 we consider %language to be an experimental feature that will likely evolve
284 in future releases.
285
286** Forward compatibility with GNU M4 has been improved.
287
288** Several bugs in the C++ skeleton and the experimental Java skeleton have been
289 fixed.
290
291* Changes in version 2.3b (2008-05-27):
292
293** The quotes around NAME that used to be required in the following directive
294 are now deprecated:
295
296 %define NAME "VALUE"
297
298** The directive `%pure-parser' is now deprecated in favor of:
299
300 %define api.pure
301
302 which has the same effect except that Bison is more careful to warn about
303 unreasonable usage in the latter case.
304
305** Push Parsing
306
307 Bison can now generate an LALR(1) parser in C with a push interface. That
308 is, instead of invoking `yyparse', which pulls tokens from `yylex', you can
309 push one token at a time to the parser using `yypush_parse', which will
310 return to the caller after processing each token. By default, the push
311 interface is disabled. Either of the following directives will enable it:
312
313 %define api.push_pull "push" // Just push; does not require yylex.
314 %define api.push_pull "both" // Push and pull; requires yylex.
315
316 See the new section `A Push Parser' in the Bison manual for details.
317
318 The current push parsing interface is experimental and may evolve. More user
319 feedback will help to stabilize it.
320
321** The -g and --graph options now output graphs in Graphviz DOT format,
322 not VCG format. Like --graph, -g now also takes an optional FILE argument
323 and thus cannot be bundled with other short options.
324
325** Java
326
327 Bison can now generate an LALR(1) parser in Java. The skeleton is
328 `data/lalr1.java'. Consider using the new %language directive instead of
329 %skeleton to select it.
330
331 See the new section `Java Parsers' in the Bison manual for details.
332
333 The current Java interface is experimental and may evolve. More user
334 feedback will help to stabilize it.
335
336** %language
337
338 This new directive specifies the programming language of the generated
339 parser, which can be C (the default), C++, or Java. Besides the skeleton
340 that Bison uses, the directive affects the names of the generated files if
341 the grammar file's name ends in ".y".
342
343** XML Automaton Report
344
345 Bison can now generate an XML report of the LALR(1) automaton using the new
346 `--xml' option. The current XML schema is experimental and may evolve. More
347 user feedback will help to stabilize it.
348
349** The grammar file may now specify the name of the parser header file using
350 %defines. For example:
351
352 %defines "parser.h"
353
354** When reporting useless rules, useless nonterminals, and unused terminals,
355 Bison now employs the terms "useless in grammar" instead of "useless",
356 "useless in parser" instead of "never reduced", and "unused in grammar"
357 instead of "unused".
358
359** Unreachable State Removal
360
361 Previously, Bison sometimes generated parser tables containing unreachable
362 states. A state can become unreachable during conflict resolution if Bison
363 disables a shift action leading to it from a predecessor state. Bison now:
364
365 1. Removes unreachable states.
366
367 2. Does not report any conflicts that appeared in unreachable states.
368 WARNING: As a result, you may need to update %expect and %expect-rr
369 directives in existing grammar files.
370
371 3. For any rule used only in such states, Bison now reports the rule as
372 "useless in parser due to conflicts".
373
374 This feature can be disabled with the following directive:
375
376 %define lr.keep_unreachable_states
377
378 See the %define entry in the `Bison Declaration Summary' in the Bison manual
379 for further discussion.
380
381** Lookahead Set Correction in the `.output' Report
382
383 When instructed to generate a `.output' file including lookahead sets
384 (using `--report=lookahead', for example), Bison now prints each reduction's
385 lookahead set only next to the associated state's one item that (1) is
386 associated with the same rule as the reduction and (2) has its dot at the end
387 of its RHS. Previously, Bison also erroneously printed the lookahead set
388 next to all of the state's other items associated with the same rule. This
389 bug affected only the `.output' file and not the generated parser source
390 code.
391
392** --report-file=FILE is a new option to override the default `.output' file
393 name.
394
395** The `=' that used to be required in the following directives is now
396 deprecated:
397
398 %file-prefix "parser"
399 %name-prefix "c_"
400 %output "parser.c"
401
402** An Alternative to `%{...%}' -- `%code QUALIFIER {CODE}'
403
404 Bison 2.3a provided a new set of directives as a more flexible alternative to
405 the traditional Yacc prologue blocks. Those have now been consolidated into
406 a single %code directive with an optional qualifier field, which identifies
407 the purpose of the code and thus the location(s) where Bison should generate
408 it:
409
410 1. `%code {CODE}' replaces `%after-header {CODE}'
411 2. `%code requires {CODE}' replaces `%start-header {CODE}'
412 3. `%code provides {CODE}' replaces `%end-header {CODE}'
413 4. `%code top {CODE}' replaces `%before-header {CODE}'
414
415 See the %code entries in section `Bison Declaration Summary' in the Bison
416 manual for a summary of the new functionality. See the new section `Prologue
417 Alternatives' for a detailed discussion including the advantages of %code
418 over the traditional Yacc prologues.
419
420 The prologue alternatives are experimental. More user feedback will help to
421 determine whether they should become permanent features.
422
423** Revised warning: unset or unused mid-rule values
424
425 Since Bison 2.2, Bison has warned about mid-rule values that are set but not
426 used within any of the actions of the parent rule. For example, Bison warns
427 about unused $2 in:
428
429 exp: '1' { $$ = 1; } '+' exp { $$ = $1 + $4; };
430
431 Now, Bison also warns about mid-rule values that are used but not set. For
432 example, Bison warns about unset $$ in the mid-rule action in:
433
434 exp: '1' { $1 = 1; } '+' exp { $$ = $2 + $4; };
435
436 However, Bison now disables both of these warnings by default since they
437 sometimes prove to be false alarms in existing grammars employing the Yacc
438 constructs $0 or $-N (where N is some positive integer).
439
440 To enable these warnings, specify the option `--warnings=midrule-values' or
441 `-W', which is a synonym for `--warnings=all'.
442
443** Default %destructor or %printer with `<*>' or `<>'
444
445 Bison now recognizes two separate kinds of default %destructor's and
446 %printer's:
447
448 1. Place `<*>' in a %destructor/%printer symbol list to define a default
449 %destructor/%printer for all grammar symbols for which you have formally
450 declared semantic type tags.
451
452 2. Place `<>' in a %destructor/%printer symbol list to define a default
453 %destructor/%printer for all grammar symbols without declared semantic
454 type tags.
455
456 Bison no longer supports the `%symbol-default' notation from Bison 2.3a.
457 `<*>' and `<>' combined achieve the same effect with one exception: Bison no
458 longer applies any %destructor to a mid-rule value if that mid-rule value is
459 not actually ever referenced using either $$ or $n in a semantic action.
460
461 The default %destructor's and %printer's are experimental. More user
462 feedback will help to determine whether they should become permanent
463 features.
464
465 See the section `Freeing Discarded Symbols' in the Bison manual for further
466 details.
467
468** %left, %right, and %nonassoc can now declare token numbers. This is required
469 by POSIX. However, see the end of section `Operator Precedence' in the Bison
470 manual for a caveat concerning the treatment of literal strings.
471
472** The nonfunctional --no-parser, -n, and %no-parser options have been
473 completely removed from Bison.
474
475* Changes in version 2.3a, 2006-09-13:
476
477** Instead of %union, you can define and use your own union type
478 YYSTYPE if your grammar contains at least one <type> tag.
479 Your YYSTYPE need not be a macro; it can be a typedef.
480 This change is for compatibility with other Yacc implementations,
481 and is required by POSIX.
482
483** Locations columns and lines start at 1.
484 In accordance with the GNU Coding Standards and Emacs.
485
486** You may now declare per-type and default %destructor's and %printer's:
487
488 For example:
489
490 %union { char *string; }
491 %token <string> STRING1
492 %token <string> STRING2
493 %type <string> string1
494 %type <string> string2
495 %union { char character; }
496 %token <character> CHR
497 %type <character> chr
498 %destructor { free ($$); } %symbol-default
499 %destructor { free ($$); printf ("%d", @$.first_line); } STRING1 string1
500 %destructor { } <character>
501
502 guarantees that, when the parser discards any user-defined symbol that has a
503 semantic type tag other than `<character>', it passes its semantic value to
504 `free'. However, when the parser discards a `STRING1' or a `string1', it
505 also prints its line number to `stdout'. It performs only the second
506 `%destructor' in this case, so it invokes `free' only once.
507
508 [Although we failed to mention this here in the 2.3a release, the default
509 %destructor's and %printer's were experimental, and they were rewritten in
510 future versions.]
511
512** Except for LALR(1) parsers in C with POSIX Yacc emulation enabled (with `-y',
513 `--yacc', or `%yacc'), Bison no longer generates #define statements for
514 associating token numbers with token names. Removing the #define statements
515 helps to sanitize the global namespace during preprocessing, but POSIX Yacc
516 requires them. Bison still generates an enum for token names in all cases.
517
518** Handling of traditional Yacc prologue blocks is now more consistent but
519 potentially incompatible with previous releases of Bison.
520
521 As before, you declare prologue blocks in your grammar file with the
522 `%{ ... %}' syntax. To generate the pre-prologue, Bison concatenates all
523 prologue blocks that you've declared before the first %union. To generate
524 the post-prologue, Bison concatenates all prologue blocks that you've
525 declared after the first %union.
526
527 Previous releases of Bison inserted the pre-prologue into both the header
528 file and the code file in all cases except for LALR(1) parsers in C. In the
529 latter case, Bison inserted it only into the code file. For parsers in C++,
530 the point of insertion was before any token definitions (which associate
531 token numbers with names). For parsers in C, the point of insertion was
532 after the token definitions.
533
534 Now, Bison never inserts the pre-prologue into the header file. In the code
535 file, it always inserts it before the token definitions.
536
537** Bison now provides a more flexible alternative to the traditional Yacc
538 prologue blocks: %before-header, %start-header, %end-header, and
539 %after-header.
540
541 For example, the following declaration order in the grammar file reflects the
542 order in which Bison will output these code blocks. However, you are free to
543 declare these code blocks in your grammar file in whatever order is most
544 convenient for you:
545
546 %before-header {
547 /* Bison treats this block like a pre-prologue block: it inserts it into
548 * the code file before the contents of the header file. It does *not*
549 * insert it into the header file. This is a good place to put
550 * #include's that you want at the top of your code file. A common
551 * example is `#include "system.h"'. */
552 }
553 %start-header {
554 /* Bison inserts this block into both the header file and the code file.
555 * In both files, the point of insertion is before any Bison-generated
556 * token, semantic type, location type, and class definitions. This is a
557 * good place to define %union dependencies, for example. */
558 }
559 %union {
560 /* Unlike the traditional Yacc prologue blocks, the output order for the
561 * new %*-header blocks is not affected by their declaration position
562 * relative to any %union in the grammar file. */
563 }
564 %end-header {
565 /* Bison inserts this block into both the header file and the code file.
566 * In both files, the point of insertion is after the Bison-generated
567 * definitions. This is a good place to declare or define public
568 * functions or data structures that depend on the Bison-generated
569 * definitions. */
570 }
571 %after-header {
572 /* Bison treats this block like a post-prologue block: it inserts it into
573 * the code file after the contents of the header file. It does *not*
574 * insert it into the header file. This is a good place to declare or
575 * define internal functions or data structures that depend on the
576 * Bison-generated definitions. */
577 }
578
579 If you have multiple occurrences of any one of the above declarations, Bison
580 will concatenate the contents in declaration order.
581
582 [Although we failed to mention this here in the 2.3a release, the prologue
583 alternatives were experimental, and they were rewritten in future versions.]
584
585** The option `--report=look-ahead' has been changed to `--report=lookahead'.
586 The old spelling still works, but is not documented and may be removed
587 in a future release.
588
589* Changes in version 2.3, 2006-06-05:
590
591** GLR grammars should now use `YYRECOVERING ()' instead of `YYRECOVERING',
592 for compatibility with LALR(1) grammars.
593
594** It is now documented that any definition of YYSTYPE or YYLTYPE should
595 be to a type name that does not contain parentheses or brackets.
596
597* Changes in version 2.2, 2006-05-19:
598
599** The distribution terms for all Bison-generated parsers now permit
600 using the parsers in nonfree programs. Previously, this permission
601 was granted only for Bison-generated LALR(1) parsers in C.
602
603** %name-prefix changes the namespace name in C++ outputs.
604
605** The C++ parsers export their token_type.
606
607** Bison now allows multiple %union declarations, and concatenates
608 their contents together.
609
610** New warning: unused values
611 Right-hand side symbols whose values are not used are reported,
612 if the symbols have destructors. For instance:
613
614 exp: exp "?" exp ":" exp { $1 ? $1 : $3; }
615 | exp "+" exp
616 ;
617
618 will trigger a warning about $$ and $5 in the first rule, and $3 in
619 the second ($1 is copied to $$ by the default rule). This example
620 most likely contains three errors, and could be rewritten as:
621
622 exp: exp "?" exp ":" exp
623 { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); }
624 | exp "+" exp
625 { $$ = $1 ? $1 : $3; if ($1) free ($3); }
626 ;
627
628 However, if the original actions were really intended, memory leaks
629 and all, the warnings can be suppressed by letting Bison believe the
630 values are used, e.g.:
631
632 exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); }
633 | exp "+" exp { $$ = $1; (void) $3; }
634 ;
635
636 If there are mid-rule actions, the warning is issued if no action
637 uses it. The following triggers no warning: $1 and $3 are used.
638
639 exp: exp { push ($1); } '+' exp { push ($3); sum (); };
640
641 The warning is intended to help catching lost values and memory leaks.
642 If a value is ignored, its associated memory typically is not reclaimed.
643
644** %destructor vs. YYABORT, YYACCEPT, and YYERROR.
645 Destructors are now called when user code invokes YYABORT, YYACCEPT,
646 and YYERROR, for all objects on the stack, other than objects
647 corresponding to the right-hand side of the current rule.
648
649** %expect, %expect-rr
650 Incorrect numbers of expected conflicts are now actual errors,
651 instead of warnings.
652
653** GLR, YACC parsers.
654 The %parse-params are available in the destructors (and the
655 experimental printers) as per the documentation.
656
657** Bison now warns if it finds a stray `$' or `@' in an action.
658
659** %require "VERSION"
660 This specifies that the grammar file depends on features implemented
661 in Bison version VERSION or higher.
662
663** lalr1.cc: The token and value types are now class members.
664 The tokens were defined as free form enums and cpp macros. YYSTYPE
665 was defined as a free form union. They are now class members:
666 tokens are enumerations of the `yy::parser::token' struct, and the
667 semantic values have the `yy::parser::semantic_type' type.
668
669 If you do not want or can update to this scheme, the directive
670 `%define "global_tokens_and_yystype" "1"' triggers the global
671 definition of tokens and YYSTYPE. This change is suitable both
672 for previous releases of Bison, and this one.
673
674 If you wish to update, then make sure older version of Bison will
675 fail using `%require "2.2"'.
676
677** DJGPP support added.
678\f
679* Changes in version 2.1, 2005-09-16:
680
681** The C++ lalr1.cc skeleton supports %lex-param.
682
683** Bison-generated parsers now support the translation of diagnostics like
684 "syntax error" into languages other than English. The default
685 language is still English. For details, please see the new
686 Internationalization section of the Bison manual. Software
687 distributors should also see the new PACKAGING file. Thanks to
688 Bruno Haible for this new feature.
689
690** Wording in the Bison-generated parsers has been changed slightly to
691 simplify translation. In particular, the message "memory exhausted"
692 has replaced "parser stack overflow", as the old message was not
693 always accurate for modern Bison-generated parsers.
694
695** Destructors are now called when the parser aborts, for all symbols left
696 behind on the stack. Also, the start symbol is now destroyed after a
697 successful parse. In both cases, the behavior was formerly inconsistent.
698
699** When generating verbose diagnostics, Bison-generated parsers no longer
700 quote the literal strings associated with tokens. For example, for
701 a syntax error associated with '%token NUM "number"' they might
702 print 'syntax error, unexpected number' instead of 'syntax error,
703 unexpected "number"'.
704\f
705* Changes in version 2.0, 2004-12-25:
706
707** Possibly-incompatible changes
708
709 - Bison-generated parsers no longer default to using the alloca function
710 (when available) to extend the parser stack, due to widespread
711 problems in unchecked stack-overflow detection. You can "#define
712 YYSTACK_USE_ALLOCA 1" to require the use of alloca, but please read
713 the manual to determine safe values for YYMAXDEPTH in that case.
714
715 - Error token location.
716 During error recovery, the location of the syntax error is updated
717 to cover the whole sequence covered by the error token: it includes
718 the shifted symbols thrown away during the first part of the error
719 recovery, and the lookahead rejected during the second part.
720
721 - Semicolon changes:
722 . Stray semicolons are no longer allowed at the start of a grammar.
723 . Semicolons are now required after in-grammar declarations.
724
725 - Unescaped newlines are no longer allowed in character constants or
726 string literals. They were never portable, and GCC 3.4.0 has
727 dropped support for them. Better diagnostics are now generated if
728 forget a closing quote.
729
730 - NUL bytes are no longer allowed in Bison string literals, unfortunately.
731
732** New features
733
734 - GLR grammars now support locations.
735
736 - New directive: %initial-action.
737 This directive allows the user to run arbitrary code (including
738 initializing @$) from yyparse before parsing starts.
739
740 - A new directive "%expect-rr N" specifies the expected number of
741 reduce/reduce conflicts in GLR parsers.
742
743 - %token numbers can now be hexadecimal integers, e.g., `%token FOO 0x12d'.
744 This is a GNU extension.
745
746 - The option `--report=lookahead' was changed to `--report=look-ahead'.
747 [However, this was changed back after 2.3.]
748
749 - Experimental %destructor support has been added to lalr1.cc.
750
751 - New configure option --disable-yacc, to disable installation of the
752 yacc command and -ly library introduced in 1.875 for POSIX conformance.
753
754** Bug fixes
755
756 - For now, %expect-count violations are now just warnings, not errors.
757 This is for compatibility with Bison 1.75 and earlier (when there are
758 reduce/reduce conflicts) and with Bison 1.30 and earlier (when there
759 are too many or too few shift/reduce conflicts). However, in future
760 versions of Bison we plan to improve the %expect machinery so that
761 these violations will become errors again.
762
763 - Within Bison itself, numbers (e.g., goto numbers) are no longer
764 arbitrarily limited to 16-bit counts.
765
766 - Semicolons are now allowed before "|" in grammar rules, as POSIX requires.
767\f
768* Changes in version 1.875, 2003-01-01:
769
770** The documentation license has been upgraded to version 1.2
771 of the GNU Free Documentation License.
772
773** syntax error processing
774
775 - In Yacc-style parsers YYLLOC_DEFAULT is now used to compute error
776 locations too. This fixes bugs in error-location computation.
777
778 - %destructor
779 It is now possible to reclaim the memory associated to symbols
780 discarded during error recovery. This feature is still experimental.
781
782 - %error-verbose
783 This new directive is preferred over YYERROR_VERBOSE.
784
785 - #defining yyerror to steal internal variables is discouraged.
786 It is not guaranteed to work forever.
787
788** POSIX conformance
789
790 - Semicolons are once again optional at the end of grammar rules.
791 This reverts to the behavior of Bison 1.33 and earlier, and improves
792 compatibility with Yacc.
793
794 - `parse error' -> `syntax error'
795 Bison now uniformly uses the term `syntax error'; formerly, the code
796 and manual sometimes used the term `parse error' instead. POSIX
797 requires `syntax error' in diagnostics, and it was thought better to
798 be consistent.
799
800 - The documentation now emphasizes that yylex and yyerror must be
801 declared before use. C99 requires this.
802
803 - Bison now parses C99 lexical constructs like UCNs and
804 backslash-newline within C escape sequences, as POSIX 1003.1-2001 requires.
805
806 - File names are properly escaped in C output. E.g., foo\bar.y is
807 output as "foo\\bar.y".
808
809 - Yacc command and library now available
810 The Bison distribution now installs a `yacc' command, as POSIX requires.
811 Also, Bison now installs a small library liby.a containing
812 implementations of Yacc-compatible yyerror and main functions.
813 This library is normally not useful, but POSIX requires it.
814
815 - Type clashes now generate warnings, not errors.
816
817 - If the user does not define YYSTYPE as a macro, Bison now declares it
818 using typedef instead of defining it as a macro.
819 For consistency, YYLTYPE is also declared instead of defined.
820
821** Other compatibility issues
822
823 - %union directives can now have a tag before the `{', e.g., the
824 directive `%union foo {...}' now generates the C code
825 `typedef union foo { ... } YYSTYPE;'; this is for Yacc compatibility.
826 The default union tag is `YYSTYPE', for compatibility with Solaris 9 Yacc.
827 For consistency, YYLTYPE's struct tag is now `YYLTYPE' not `yyltype'.
828 This is for compatibility with both Yacc and Bison 1.35.
829
830 - `;' is output before the terminating `}' of an action, for
831 compatibility with Bison 1.35.
832
833 - Bison now uses a Yacc-style format for conflict reports, e.g.,
834 `conflicts: 2 shift/reduce, 1 reduce/reduce'.
835
836 - `yystype' and `yyltype' are now obsolescent macros instead of being
837 typedefs or tags; they are no longer documented and are planned to be
838 withdrawn in a future release.
839
840** GLR parser notes
841
842 - GLR and inline
843 Users of Bison have to decide how they handle the portability of the
844 C keyword `inline'.
845
846 - `parsing stack overflow...' -> `parser stack overflow'
847 GLR parsers now report `parser stack overflow' as per the Bison manual.
848
849** Bison now warns if it detects conflicting outputs to the same file,
850 e.g., it generates a warning for `bison -d -o foo.h foo.y' since
851 that command outputs both code and header to foo.h.
852
853** #line in output files
854 - --no-line works properly.
855
856** Bison can no longer be built by a K&R C compiler; it requires C89 or
857 later to be built. This change originally took place a few versions
858 ago, but nobody noticed until we recently asked someone to try
859 building Bison with a K&R C compiler.
860\f
861* Changes in version 1.75, 2002-10-14:
862
863** Bison should now work on 64-bit hosts.
864
865** Indonesian translation thanks to Tedi Heriyanto.
866
867** GLR parsers
868 Fix spurious parse errors.
869
870** Pure parsers
871 Some people redefine yyerror to steal yyparse' private variables.
872 Reenable this trick until an official feature replaces it.
873
874** Type Clashes
875 In agreement with POSIX and with other Yaccs, leaving a default
876 action is valid when $$ is untyped, and $1 typed:
877
878 untyped: ... typed;
879
880 but the converse remains an error:
881
882 typed: ... untyped;
883
884** Values of mid-rule actions
885 The following code:
886
887 foo: { ... } { $$ = $1; } ...
888
889 was incorrectly rejected: $1 is defined in the second mid-rule
890 action, and is equal to the $$ of the first mid-rule action.
891\f
892* Changes in version 1.50, 2002-10-04:
893
894** GLR parsing
895 The declaration
896 %glr-parser
897 causes Bison to produce a Generalized LR (GLR) parser, capable of handling
898 almost any context-free grammar, ambiguous or not. The new declarations
899 %dprec and %merge on grammar rules allow parse-time resolution of
900 ambiguities. Contributed by Paul Hilfinger.
901
902 Unfortunately Bison 1.50 does not work properly on 64-bit hosts
903 like the Alpha, so please stick to 32-bit hosts for now.
904
905** Output Directory
906 When not in Yacc compatibility mode, when the output file was not
907 specified, running `bison foo/bar.y' created `foo/bar.c'. It
908 now creates `bar.c'.
909
910** Undefined token
911 The undefined token was systematically mapped to 2 which prevented
912 the use of 2 by the user. This is no longer the case.
913
914** Unknown token numbers
915 If yylex returned an out of range value, yyparse could die. This is
916 no longer the case.
917
918** Error token
919 According to POSIX, the error token must be 256.
920 Bison extends this requirement by making it a preference: *if* the
921 user specified that one of her tokens is numbered 256, then error
922 will be mapped onto another number.
923
924** Verbose error messages
925 They no longer report `..., expecting error or...' for states where
926 error recovery is possible.
927
928** End token
929 Defaults to `$end' instead of `$'.
930
931** Error recovery now conforms to documentation and to POSIX
932 When a Bison-generated parser encounters a syntax error, it now pops
933 the stack until it finds a state that allows shifting the error
934 token. Formerly, it popped the stack until it found a state that
935 allowed some non-error action other than a default reduction on the
936 error token. The new behavior has long been the documented behavior,
937 and has long been required by POSIX. For more details, please see
938 Paul Eggert, "Reductions during Bison error handling" (2002-05-20)
939 <http://lists.gnu.org/archive/html/bug-bison/2002-05/msg00038.html>.
940
941** Traces
942 Popped tokens and nonterminals are now reported.
943
944** Larger grammars
945 Larger grammars are now supported (larger token numbers, larger grammar
946 size (= sum of the LHS and RHS lengths), larger LALR tables).
947 Formerly, many of these numbers ran afoul of 16-bit limits;
948 now these limits are 32 bits on most hosts.
949
950** Explicit initial rule
951 Bison used to play hacks with the initial rule, which the user does
952 not write. It is now explicit, and visible in the reports and
953 graphs as rule 0.
954
955** Useless rules
956 Before, Bison reported the useless rules, but, although not used,
957 included them in the parsers. They are now actually removed.
958
959** Useless rules, useless nonterminals
960 They are now reported, as a warning, with their locations.
961
962** Rules never reduced
963 Rules that can never be reduced because of conflicts are now
964 reported.
965
966** Incorrect `Token not used'
967 On a grammar such as
968
969 %token useless useful
970 %%
971 exp: '0' %prec useful;
972
973 where a token was used to set the precedence of the last rule,
974 bison reported both `useful' and `useless' as useless tokens.
975
976** Revert the C++ namespace changes introduced in 1.31
977 as they caused too many portability hassles.
978
979** Default locations
980 By an accident of design, the default computation of @$ was
981 performed after another default computation was performed: @$ = @1.
982 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
983 the computation of @$.
984
985** Token end-of-file
986 The token end of file may be specified by the user, in which case,
987 the user symbol is used in the reports, the graphs, and the verbose
988 error messages instead of `$end', which remains being the default.
989 For instance
990 %token MYEOF 0
991 or
992 %token MYEOF 0 "end of file"
993
994** Semantic parser
995 This old option, which has been broken for ages, is removed.
996
997** New translations
998 Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
999 Croatian, thanks to Denis Lackovic.
1000
1001** Incorrect token definitions
1002 When given `%token 'a' "A"', Bison used to output `#define 'a' 65'.
1003
1004** Token definitions as enums
1005 Tokens are output both as the traditional #define's, and, provided
1006 the compiler supports ANSI C or is a C++ compiler, as enums.
1007 This lets debuggers display names instead of integers.
1008
1009** Reports
1010 In addition to --verbose, bison supports --report=THINGS, which
1011 produces additional information:
1012 - itemset
1013 complete the core item sets with their closure
1014 - lookahead [changed to `look-ahead' in 1.875e through 2.3, but changed back]
1015 explicitly associate lookahead tokens to items
1016 - solved
1017 describe shift/reduce conflicts solving.
1018 Bison used to systematically output this information on top of
1019 the report. Solved conflicts are now attached to their states.
1020
1021** Type clashes
1022 Previous versions don't complain when there is a type clash on
1023 the default action if the rule has a mid-rule action, such as in:
1024
1025 %type <foo> bar
1026 %%
1027 bar: '0' {} '0';
1028
1029 This is fixed.
1030
1031** GNU M4 is now required when using Bison.
1032\f
1033* Changes in version 1.35, 2002-03-25:
1034
1035** C Skeleton
1036 Some projects use Bison's C parser with C++ compilers, and define
1037 YYSTYPE as a class. The recent adjustment of C parsers for data
1038 alignment and 64 bit architectures made this impossible.
1039
1040 Because for the time being no real solution for C++ parser
1041 generation exists, kludges were implemented in the parser to
1042 maintain this use. In the future, when Bison has C++ parsers, this
1043 kludge will be disabled.
1044
1045 This kludge also addresses some C++ problems when the stack was
1046 extended.
1047\f
1048* Changes in version 1.34, 2002-03-12:
1049
1050** File name clashes are detected
1051 $ bison foo.y -d -o foo.x
1052 fatal error: header and parser would both be named `foo.x'
1053
1054** A missing `;' at the end of a rule triggers a warning
1055 In accordance with POSIX, and in agreement with other
1056 Yacc implementations, Bison will mandate this semicolon in the near
1057 future. This eases the implementation of a Bison parser of Bison
1058 grammars by making this grammar LALR(1) instead of LR(2). To
1059 facilitate the transition, this release introduces a warning.
1060
1061** Revert the C++ namespace changes introduced in 1.31, as they caused too
1062 many portability hassles.
1063
1064** DJGPP support added.
1065
1066** Fix test suite portability problems.
1067\f
1068* Changes in version 1.33, 2002-02-07:
1069
1070** Fix C++ issues
1071 Groff could not be compiled for the definition of size_t was lacking
1072 under some conditions.
1073
1074** Catch invalid @n
1075 As is done with $n.
1076\f
1077* Changes in version 1.32, 2002-01-23:
1078
1079** Fix Yacc output file names
1080
1081** Portability fixes
1082
1083** Italian, Dutch translations
1084\f
1085* Changes in version 1.31, 2002-01-14:
1086
1087** Many Bug Fixes
1088
1089** GNU Gettext and %expect
1090 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
1091 Bison dies on incorrect %expectations, we fear there will be
1092 too many bug reports for Gettext, so _for the time being_, %expect
1093 does not trigger an error when the input file is named `plural.y'.
1094
1095** Use of alloca in parsers
1096 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
1097 malloc exclusively. Since 1.29, but was not NEWS'ed.
1098
1099 alloca is used only when compiled with GCC, to avoid portability
1100 problems as on AIX.
1101
1102** yyparse now returns 2 if memory is exhausted; formerly it dumped core.
1103
1104** When the generated parser lacks debugging code, YYDEBUG is now 0
1105 (as POSIX requires) instead of being undefined.
1106
1107** User Actions
1108 Bison has always permitted actions such as { $$ = $1 }: it adds the
1109 ending semicolon. Now if in Yacc compatibility mode, the semicolon
1110 is no longer output: one has to write { $$ = $1; }.
1111
1112** Better C++ compliance
1113 The output parsers try to respect C++ namespaces.
1114 [This turned out to be a failed experiment, and it was reverted later.]
1115
1116** Reduced Grammars
1117 Fixed bugs when reporting useless nonterminals.
1118
1119** 64 bit hosts
1120 The parsers work properly on 64 bit hosts.
1121
1122** Error messages
1123 Some calls to strerror resulted in scrambled or missing error messages.
1124
1125** %expect
1126 When the number of shift/reduce conflicts is correct, don't issue
1127 any warning.
1128
1129** The verbose report includes the rule line numbers.
1130
1131** Rule line numbers are fixed in traces.
1132
1133** Swedish translation
1134
1135** Parse errors
1136 Verbose parse error messages from the parsers are better looking.
1137 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
1138 Now: parse error: unexpected '/', expecting "number" or '-' or '('
1139
1140** Fixed parser memory leaks.
1141 When the generated parser was using malloc to extend its stacks, the
1142 previous allocations were not freed.
1143
1144** Fixed verbose output file.
1145 Some newlines were missing.
1146 Some conflicts in state descriptions were missing.
1147
1148** Fixed conflict report.
1149 Option -v was needed to get the result.
1150
1151** %expect
1152 Was not used.
1153 Mismatches are errors, not warnings.
1154
1155** Fixed incorrect processing of some invalid input.
1156
1157** Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
1158
1159** Fixed some typos in the documentation.
1160
1161** %token MY_EOF 0 is supported.
1162 Before, MY_EOF was silently renumbered as 257.
1163
1164** doc/refcard.tex is updated.
1165
1166** %output, %file-prefix, %name-prefix.
1167 New.
1168
1169** --output
1170 New, aliasing `--output-file'.
1171\f
1172* Changes in version 1.30, 2001-10-26:
1173
1174** `--defines' and `--graph' have now an optional argument which is the
1175 output file name. `-d' and `-g' do not change; they do not take any
1176 argument.
1177
1178** `%source_extension' and `%header_extension' are removed, failed
1179 experiment.
1180
1181** Portability fixes.
1182\f
1183* Changes in version 1.29, 2001-09-07:
1184
1185** The output file does not define const, as this caused problems when used
1186 with common autoconfiguration schemes. If you still use ancient compilers
1187 that lack const, compile with the equivalent of the C compiler option
1188 `-Dconst='. autoconf's AC_C_CONST macro provides one way to do this.
1189
1190** Added `-g' and `--graph'.
1191
1192** The Bison manual is now distributed under the terms of the GNU FDL.
1193
1194** The input and the output files has automatically a similar extension.
1195
1196** Russian translation added.
1197
1198** NLS support updated; should hopefully be less troublesome.
1199
1200** Added the old Bison reference card.
1201
1202** Added `--locations' and `%locations'.
1203
1204** Added `-S' and `--skeleton'.
1205
1206** `%raw', `-r', `--raw' is disabled.
1207
1208** Special characters are escaped when output. This solves the problems
1209 of the #line lines with path names including backslashes.
1210
1211** New directives.
1212 `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose',
1213 `%debug', `%source_extension' and `%header_extension'.
1214
1215** @$
1216 Automatic location tracking.
1217\f
1218* Changes in version 1.28, 1999-07-06:
1219
1220** Should compile better now with K&R compilers.
1221
1222** Added NLS.
1223
1224** Fixed a problem with escaping the double quote character.
1225
1226** There is now a FAQ.
1227\f
1228* Changes in version 1.27:
1229
1230** The make rule which prevented bison.simple from being created on
1231 some systems has been fixed.
1232\f
1233* Changes in version 1.26:
1234
1235** Bison now uses automake.
1236
1237** New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
1238
1239** Token numbers now start at 257 as previously documented, not 258.
1240
1241** Bison honors the TMPDIR environment variable.
1242
1243** A couple of buffer overruns have been fixed.
1244
1245** Problems when closing files should now be reported.
1246
1247** Generated parsers should now work even on operating systems which do
1248 not provide alloca().
1249\f
1250* Changes in version 1.25, 1995-10-16:
1251
1252** Errors in the input grammar are not fatal; Bison keeps reading
1253the grammar file, and reports all the errors found in it.
1254
1255** Tokens can now be specified as multiple-character strings: for
1256example, you could use "<=" for a token which looks like <=, instead
1257of chosing a name like LESSEQ.
1258
1259** The %token_table declaration says to write a table of tokens (names
1260and numbers) into the parser file. The yylex function can use this
1261table to recognize multiple-character string tokens, or for other
1262purposes.
1263
1264** The %no_lines declaration says not to generate any #line preprocessor
1265directives in the parser file.
1266
1267** The %raw declaration says to use internal Bison token numbers, not
1268Yacc-compatible token numbers, when token names are defined as macros.
1269
1270** The --no-parser option produces the parser tables without including
1271the parser engine; a project can now use its own parser engine.
1272The actions go into a separate file called NAME.act, in the form of
1273a switch statement body.
1274\f
1275* Changes in version 1.23:
1276
1277The user can define YYPARSE_PARAM as the name of an argument to be
1278passed into yyparse. The argument should have type void *. It should
1279actually point to an object. Grammar actions can access the variable
1280by casting it to the proper pointer type.
1281
1282Line numbers in output file corrected.
1283\f
1284* Changes in version 1.22:
1285
1286--help option added.
1287\f
1288* Changes in version 1.20:
1289
1290Output file does not redefine const for C++.
1291
1292Local Variables:
1293mode: outline
1294End:
1295
1296-----
1297
1298Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
12992004, 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
1300
1301This file is part of Bison, the GNU Parser Generator.
1302
1303This program is free software: you can redistribute it and/or modify
1304it under the terms of the GNU General Public License as published by
1305the Free Software Foundation, either version 3 of the License, or
1306(at your option) any later version.
1307
1308This program is distributed in the hope that it will be useful,
1309but WITHOUT ANY WARRANTY; without even the implied warranty of
1310MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
1311GNU General Public License for more details.
1312
1313You should have received a copy of the GNU General Public License
1314along with this program. If not, see <http://www.gnu.org/licenses/>.