]> git.saurik.com Git - bison.git/blame - NEWS
Merge remote-tracking branch 'origin/maint'
[bison.git] / NEWS
CommitLineData
ed7658fe 1GNU Bison NEWS
3af4feb2 2
52718713 3* Noteworthy changes in release ?.? (????-??-??) [?]
74553c98 4
47db7ed1
AD
5** Incompatible changes
6
7*** Obsolete features
8
de62ae86 9 Support for YYFAIL is removed (deprecated in Bison 2.4.2).
5f819b2f 10 Support for yystype and yyltype (instead of YYSTYPE and YYLTYPE)
de62ae86
AD
11 is removed (deprecated in Bison 1.875).
12 Support for YYPARSE_PARAM is removed (deprecated in Bison 1.875).
47db7ed1 13
73370a9d
VS
14** Warnings
15
1048a1c9 16*** Enhancements of the -Werror option
518e8830 17
1048a1c9
AD
18 The -Werror=CATEGORY option is now recognized, and will treat specified
19 warnings as errors. The warnings need not have been explictly activated
20 using the -W option, this is similar to what gcc 4.7 does.
21
22 For example, given the following command line, Bison will treat both
23 warnings related to POSIX Yacc incompatiblities and S/R conflicts as
24 errors (and only those):
25
26 $ bison -Werror=yacc,error=conflicts-sr input.y
27
28 If no categories are specified, -Werror will make all active warnings into
29 errors. For example, the following line does the same the previous example:
30
31 $ bison -Werror -Wnone -Wyacc -Wconflicts-sr input.y
32
33 (By default -Wconflicts-sr,conflicts-rr,deprecated,other is enabled.)
34
35 Note that the categories in this -Werror option may not be prefixed with
36 "no-". However, -Wno-error[=CATEGORY] is valid.
37
38 Note that -y enables -Werror=yacc. Therefore it is now possible to require
39 Yacc-like behavior (e.g., always generate y.tab.c), but to report
40 incompatibilities as warnings: "-y -Wno-error=yacc".
518e8830 41
46bdb8ec 42*** The display of warnings is now richer
73370a9d 43
46bdb8ec 44 The option that controls a given warning is now displayed:
73370a9d 45
46bdb8ec 46 foo.y:4.6: warning: type clash on default action: <foo> != <bar> [-Wother]
73370a9d 47
46bdb8ec
TR
48 In the case of warnings treated as errors, the prefix is changed from
49 "warning: " to "error: ", and the suffix is displayed, in a manner similar
50 to gcc, as [-Werror=CATEGORY].
1048a1c9 51
46bdb8ec
TR
52 For instance, where the previous version of Bison would report (and exit
53 with failure):
1048a1c9
AD
54
55 bison: warnings being treated as errors
46bdb8ec 56 input.y:1.1: warning: stray ',' treated as white space
1048a1c9 57
46bdb8ec 58 it now reports:
1048a1c9 59
1048a1c9
AD
60 input.y:1.1: error: stray ',' treated as white space [-Werror=other]
61
62*** Deprecated constructs
63
64 The new 'deprecated' warning category flags obsolete constructs whose
65 support will be discontinued. It is enabled by default. These warnings
66 used to be reported as 'other' warnings.
67
73370a9d 68*** Useless semantic types
9641b918
VS
69
70 Bison now warns about useless (uninhabited) semantic types. Since
71 semantic types are not declared to Bison (they are defined in the opaque
72 %union structure), it is %printer/%destructor directives about useless
73 types that trigger the warning:
74
75 %token <type1> term
76 %type <type2> nterm
77 %printer {} <type1> <type3>
78 %destructor {} <type2> <type4>
79 %%
80 nterm: term { $$ = $1; };
81
82 3.28-34: warning: type <type3> is used, but is not associated to any symbol
83 4.28-34: warning: type <type4> is used, but is not associated to any symbol
84
31557b9e 85*** Undefined but unused symbols
b921d92f 86
31557b9e
AD
87 Bison used to raise an error for undefined symbols that are not used in
88 the grammar. This is now only a warning.
b921d92f
VS
89
90 %printer {} symbol1
91 %destructor {} symbol2
31557b9e 92 %type <type> symbol3
b921d92f
VS
93 %%
94 exp: "a";
95
73370a9d 96*** Useless destructors or printers
ea9a35c6
VS
97
98 Bison now warns about useless destructors or printers. In the following
99 example, the printer for <type1>, and the destructor for <type2> are
100 useless: all symbols of <type1> (token1) already have a printer, and all
101 symbols of type <type2> (token2) already have a destructor.
102
103 %token <type1> token1
104 <type2> token2
105 <type3> token3
106 <type4> token4
107 %printer {} token1 <type1> <type3>
108 %destructor {} token2 <type2> <type4>
109
d87ea54c
AD
110*** Conflicts
111
112 The warnings and error messages about shift/reduce and reduce/reduce
113 conflicts have been normalized. For instance on the following foo.y file:
114
115 %glr-parser
116 %%
117 exp: exp '+' exp | '0' | '0';
118
119 compare the previous version of bison:
120
121 $ bison foo.y
122 foo.y: conflicts: 1 shift/reduce, 2 reduce/reduce
123 $ bison -Werror foo.y
124 bison: warnings being treated as errors
125 foo.y: conflicts: 1 shift/reduce, 2 reduce/reduce
126
127 with the new behavior:
128
129 $ bison foo.y
130 foo.y: warning: 1 shift/reduce conflict [-Wconflicts-sr]
131 foo.y: warning: 2 reduce/reduce conflicts [-Wconflicts-rr]
132 $ bison -Werror foo.y
9503b0a4
TR
133 foo.y: error: 1 shift/reduce conflict [-Werror=conflicts-sr]
134 foo.y: error: 2 reduce/reduce conflicts [-Werror=conflicts-rr]
d87ea54c
AD
135
136 When %expect or %expect-rr is used, such as with bar.y:
137
138 %expect 0
139 %glr-parser
140 %%
141 exp: exp '+' exp | '0' | '0';
142
143 Former behavior:
144
145 $ bison bar.y
146 bar.y: conflicts: 1 shift/reduce, 2 reduce/reduce
147 bar.y: expected 0 shift/reduce conflicts
148 bar.y: expected 0 reduce/reduce conflicts
149
150 New one:
151
152 $ bison bar.y
9503b0a4
TR
153 bar.y: error: shift/reduce conflicts: 1 found, 0 expected
154 bar.y: error: reduce/reduce conflicts: 2 found, 0 expected
d87ea54c 155
2055a44e
AD
156** Additional yylex/yyparse arguments
157
6dc4663d
AD
158 The new directive %param declares additional arguments to both yylex and
159 yyparse. The %lex-param, %parse-param, and %param directives support one
160 or more arguments. Instead of
2055a44e 161
e436fa67
AD
162 %lex-param {arg1_type *arg1}
163 %lex-param {arg2_type *arg2}
164 %parse-param {arg1_type *arg1}
165 %parse-param {arg2_type *arg2}
2055a44e
AD
166
167 one may now declare
168
e436fa67 169 %param {arg1_type *arg1} {arg2_type *arg2}
2055a44e
AD
170
171** Java skeleton improvements
0ea583d2 172
1f77b2e0
AD
173 The constants for token names were moved to the Lexer interface. Also, it
174 is possible to add code to the parser's constructors using "%code init"
175 and "%define init_throws".
0ea583d2 176
34904c57
AD
177** C++ skeletons improvements
178
e4c0985b 179*** The parser header is no longer mandatory (lalr1.cc, glr.cc)
93549bcd 180
e4c0985b
AD
181 Using %defines is now optional. Without it, the needed support classes
182 are defined in the generated parser, instead of additional files (such as
183 location.hh, position.hh and stack.hh).
93549bcd 184
e4c0985b 185*** Locations are no longer mandatory (lalr1.cc, glr.cc)
34904c57
AD
186
187 Both lalr1.cc and glr.cc no longer require %location.
188
189*** syntax_error exception (lalr1.cc)
a6552c5d
AD
190
191 The C++ parser features a syntax_error exception, which can be
192 thrown from the scanner or from user rules to raise syntax errors.
193 This facilitates reporting errors caught in sub-functions (e.g.,
194 rejecting too large integral literals from a conversion function
195 used by the scanner, or rejecting invalid combinations from a
196 factory invoked by the user actions).
197
4c6622c2 198** Variable api.tokens.prefix
99c08fb6 199
4c6622c2 200 The variable api.tokens.prefix changes the way tokens are identified in
99c08fb6
AD
201 the generated files. This is especially useful to avoid collisions
202 with identifiers in the target language. For instance
203
e436fa67
AD
204 %token FILE for ERROR
205 %define api.tokens.prefix "TOK_"
206 %%
207 start: FILE for ERROR;
99c08fb6
AD
208
209 will generate the definition of the symbols TOK_FILE, TOK_for, and
210 TOK_ERROR in the generated sources. In particular, the scanner must
211 use these prefixed token names, although the grammar itself still
212 uses the short names (as in the sample rule given above).
213
67501061
AD
214** Variable api.namespace
215
1f77b2e0 216 The 'namespace' variable is renamed 'api.namespace'. Backward
67501061
AD
217 compatibility is ensured, but upgrading is recommended.
218
31b850d2
AD
219** Variable parse.error
220
1f77b2e0
AD
221 This variable controls the verbosity of error messages. The use of the
222 %error-verbose directive is deprecated in favor of "%define parse.error
223 verbose".
31b850d2 224
ca2a6d15
PH
225** Semantic predicates
226
1f77b2e0
AD
227 The new, experimental, semantic-predicate feature allows actions of the
228 form "%?{ BOOLEAN-EXPRESSION }", which cause syntax errors (as for
ca2a6d15 229 YYERROR) if the expression evaluates to 0, and are evaluated immediately
1f77b2e0
AD
230 in GLR parsers, rather than being deferred. The result is that they allow
231 the programmer to prune possible parses based on the values of run-time
232 expressions.
ca2a6d15 233
d1400569
AD
234** The directive %expect-rr is now an error in non GLR mode
235
236 It used to be an error only if used in non GLR mode, _and_ if there are
237 reduce/reduce conflicts.
238
e436fa67
AD
239* Noteworthy changes in release ?.? (????-??-??) [?]
240
6b4cb804
AD
241** Bug fixes
242
243 Bugs in the test suite have been fixed.
244
245 Some errors in translations have been addressed, and --help now directs
246 users to the appropriate place to report them.
247
248 Stray Info files shipped by accident are removed.
249
250 Incorrect definitions of YY_, issued by yacc.c when no parser header is
251 generated, are removed.
9c26b8fc 252
5a05f42e 253** Changes in the format of error messages
cbaea010
TR
254
255 This used to be the format of many error reports:
256
257 foo.y:5.10-24: result type clash on merge function 'merge': <t3> != <t2>
258 foo.y:4.13-27: previous declaration
259
260 It is now:
261
262 foo.y:5.10-25: result type clash on merge function 'merge': <t3> != <t2>
263 foo.y:4.13-27: previous declaration
264
c9d5bcc9
AD
265** Header guards (yacc.c, glr.c, glr.cc)
266
267 In order to avoid collisions, the header guards are now
268 YY_<PREFIX>_<FILE>_INCLUDED, instead of merely <PREFIX>_<FILE>.
269 For instance the header generated from
270
271 %define api.prefix "calc"
272 %defines "lib/parse.h"
273
274 will use YY_CALC_LIB_PARSE_H_INCLUDED as guard.
275
e1eeecd3 276* Noteworthy changes in release 2.6.2 (2012-08-03) [stable]
9c26b8fc 277
43ca8040
AD
278** Bug fixes
279
280 Buffer overruns, complaints from Flex, and portability issues in the test
281 suite have been fixed.
282
c9d546b2
AD
283** Spaces in %lex- and %parse-param (lalr1.cc, glr.cc)
284
285 Trailing end-of-lines in %parse-param or %lex-param would result in
286 invalid C++. This is fixed.
9c26b8fc 287
dcd5344d
AD
288** Spurious spaces and end-of-lines
289
290 The generated files no longer end (nor start) with empty lines.
291
77b214ef 292* Noteworthy changes in release 2.6.1 (2012-07-30) [stable]
a4107f24 293
8617d87e
AD
294 Bison no longer executes user-specified M4 code when processing a grammar.
295
e20e6a50
AD
296** Future Changes
297
298 In addition to the removal of the features announced in Bison 2.6, the
299 next major release will remove the "Temporary hack for adding a semicolon
300 to the user action", as announced in the release 2.5. Instead of:
301
302 exp: exp "+" exp { $$ = $1 + $3 };
303
304 write:
305
306 exp: exp "+" exp { $$ = $1 + $3; };
307
8617d87e
AD
308** Bug fixes
309
0e164d43
AD
310*** Type names are now properly escaped.
311
312*** glr.cc: set_debug_level and debug_level work as expected.
a4107f24 313
26313726
AD
314*** Stray @ or $ in actions
315
316 While Bison used to warn about stray $ or @ in action rules, it did not
317 for other actions such as printers, destructors, or initial actions. It
318 now does.
319
cd735a8c 320** Type names in actions
4982f078
AD
321
322 For consistency with rule actions, it is now possible to qualify $$ by a
cd735a8c 323 type-name in destructors, printers, and initial actions. For instance:
4982f078
AD
324
325 %printer { fprintf (yyo, "(%d, %f)", $<ival>$, $<fval>$); } <*> <>;
326
327 will display two values for each typed and untyped symbol (provided
cd735a8c 328 that YYSTYPE has both "ival" and "fval" fields).
60aa04a2 329
1505e8bb 330* Noteworthy changes in release 2.6 (2012-07-19) [stable]
0f11eec2 331
9553083c
AD
332** Future changes:
333
55d1006f
AD
334 The next major release of Bison will drop support for the following
335 deprecated features. Please report disagreements to bug-bison@gnu.org.
0f11eec2 336
aaf61036 337*** K&R C parsers
55d1006f
AD
338
339 Support for generating parsers in K&R C will be removed. Parsers
242cc08e 340 generated for C support ISO C90, and are tested with ISO C99 and ISO C11
55d1006f
AD
341 compilers.
342
258cddbc 343*** Features deprecated since Bison 1.875
0f11eec2 344
258cddbc
AD
345 The definitions of yystype and yyltype will be removed; use YYSTYPE and
346 YYLTYPE.
0f11eec2 347
258cddbc
AD
348 YYPARSE_PARAM and YYLEX_PARAM, deprecated in favor of %parse-param and
349 %lex-param, will no longer be supported.
350
351 Support for the preprocessor symbol YYERROR_VERBOSE will be removed, use
352 %error-verbose.
55d1006f
AD
353
354*** The generated header will be included (yacc.c)
0f11eec2
AD
355
356 Instead of duplicating the content of the generated header (definition of
55d1006f
AD
357 YYSTYPE, yyparse declaration etc.), the generated parser will include it,
358 as is already the case for GLR or C++ parsers. This change is deferred
359 because existing versions of ylwrap (e.g., Automake 1.12.1) do not support
360 it.
0f11eec2 361
c2425191 362** Generated Parser Headers
56ca3d8f 363
258cddbc 364*** Guards (yacc.c, glr.c, glr.cc)
c3e9f08f
AD
365
366 The generated headers are now guarded, as is already the case for C++
242cc08e 367 parsers (lalr1.cc). For instance, with --defines=foo.h:
c3e9f08f 368
e29f0771
AD
369 #ifndef YY_FOO_H
370 # define YY_FOO_H
371 ...
372 #endif /* !YY_FOO_H */
c3e9f08f 373
258cddbc 374*** New declarations (yacc.c, glr.c)
56ca3d8f
AD
375
376 The generated header now declares yydebug and yyparse. Both honor
377 --name-prefix=bar_, and yield
378
e29f0771 379 int bar_parse (void);
56ca3d8f
AD
380
381 rather than
382
e29f0771
AD
383 #define yyparse bar_parse
384 int yyparse (void);
56ca3d8f
AD
385
386 in order to facilitate the inclusion of several parser headers inside a
387 single compilation unit.
c3e9f08f 388
258cddbc
AD
389*** Exported symbols in C++
390
391 The symbols YYTOKEN_TABLE and YYERROR_VERBOSE, which were defined in the
392 header, are removed, as they prevent the possibility of including several
393 generated headers from a single compilation unit.
394
694af10c
AD
395*** YYLSP_NEEDED
396
397 For the same reasons, the undocumented and unused macro YYLSP_NEEDED is no
398 longer defined.
399
4b3847c3
AD
400** New %define variable: api.prefix
401
402 Now that the generated headers are more complete and properly protected
403 against multiple inclusions, constant names, such as YYSTYPE are a
404 problem. While yyparse and others are properly renamed by %name-prefix,
405 YYSTYPE, YYDEBUG and others have never been affected by it. Because it
406 would introduce backward compatibility issues in projects not expecting
407 YYSTYPE to be renamed, instead of changing the behavior of %name-prefix,
408 it is deprecated in favor of a new %define variable: api.prefix.
409
410 The following examples compares both:
411
412 %name-prefix "bar_" | %define api.prefix "bar_"
413 %token <ival> FOO %token <ival> FOO
414 %union { int ival; } %union { int ival; }
415 %% %%
416 exp: 'a'; exp: 'a';
417
418 bison generates:
419
420 #ifndef BAR_FOO_H #ifndef BAR_FOO_H
421 # define BAR_FOO_H # define BAR_FOO_H
422
423 /* Enabling traces. */ /* Enabling traces. */
5f108727
AD
424 # ifndef YYDEBUG | # ifndef BAR_DEBUG
425 > # if defined YYDEBUG
426 > # if YYDEBUG
427 > # define BAR_DEBUG 1
428 > # else
429 > # define BAR_DEBUG 0
430 > # endif
431 > # else
432 # define YYDEBUG 0 | # define BAR_DEBUG 0
433 > # endif
434 # endif | # endif
435
436 # if YYDEBUG | # if BAR_DEBUG
4b3847c3
AD
437 extern int bar_debug; extern int bar_debug;
438 # endif # endif
439
440 /* Tokens. */ /* Tokens. */
441 # ifndef YYTOKENTYPE | # ifndef BAR_TOKENTYPE
442 # define YYTOKENTYPE | # define BAR_TOKENTYPE
443 enum yytokentype { | enum bar_tokentype {
444 FOO = 258 FOO = 258
445 }; };
446 # endif # endif
447
448 #if ! defined YYSTYPE \ | #if ! defined BAR_STYPE \
449 && ! defined YYSTYPE_IS_DECLARED | && ! defined BAR_STYPE_IS_DECLARED
450 typedef union YYSTYPE | typedef union BAR_STYPE
451 { {
452 int ival; int ival;
453 } YYSTYPE; | } BAR_STYPE;
454 # define YYSTYPE_IS_DECLARED 1 | # define BAR_STYPE_IS_DECLARED 1
455 #endif #endif
456
457 extern YYSTYPE bar_lval; | extern BAR_STYPE bar_lval;
458
459 int bar_parse (void); int bar_parse (void);
460
461 #endif /* !BAR_FOO_H */ #endif /* !BAR_FOO_H */
462
dfaac272 463* Noteworthy changes in release 2.5.1 (2012-06-05) [stable]
df6e3db0 464
debe2c03 465** Future changes:
765e1bd4 466
e4ab1254 467 The next major release will drop support for generating parsers in K&R C.
041308d0 468
466b4cf2 469** yacc.c: YYBACKUP works as expected.
ef51bfa7 470
d834eca0 471** glr.c improvements:
041308d0 472
d834eca0 473*** Location support is eliminated when not requested:
041308d0 474
e4ab1254
AD
475 GLR parsers used to include location-related code even when locations were
476 not requested, and therefore not even usable.
378e917c 477
d834eca0 478*** __attribute__ is preserved:
d115aad9 479
e4ab1254
AD
480 __attribute__ is no longer disabled when __STRICT_ANSI__ is defined (i.e.,
481 when -std is passed to GCC).
041308d0 482
466b4cf2 483** lalr1.java: several fixes:
041308d0 484
e4ab1254
AD
485 The Java parser no longer throws ArrayIndexOutOfBoundsException if the
486 first token leads to a syntax error. Some minor clean ups.
041308d0 487
22172d47 488** Changes for C++:
ef51bfa7 489
22172d47 490*** C++11 compatibility:
ef51bfa7 491
e4ab1254
AD
492 C and C++ parsers use "nullptr" instead of "0" when __cplusplus is 201103L
493 or higher.
936c88d1 494
22172d47
AD
495*** Header guards
496
497 The header files such as "parser.hh", "location.hh", etc. used a constant
498 name for preprocessor guards, for instance:
499
e29f0771
AD
500 #ifndef BISON_LOCATION_HH
501 # define BISON_LOCATION_HH
502 ...
503 #endif // !BISON_LOCATION_HH
22172d47
AD
504
505 The inclusion guard is now computed from "PREFIX/FILE-NAME", where lower
506 case characters are converted to upper case, and series of
507 non-alphanumerical characters are converted to an underscore.
508
509 With "bison -o lang++/parser.cc", "location.hh" would now include:
510
e29f0771
AD
511 #ifndef YY_LANG_LOCATION_HH
512 # define YY_LANG_LOCATION_HH
513 ...
514 #endif // !YY_LANG_LOCATION_HH
22172d47
AD
515
516*** C++ locations:
936c88d1 517
e4ab1254
AD
518 The position and location constructors (and their initialize methods)
519 accept new arguments for line and column. Several issues in the
520 documentation were fixed.
936c88d1 521
466b4cf2
AD
522** liby is no longer asking for "rpl_fprintf" on some platforms.
523
7e508a2b
AD
524** Changes in the manual:
525
526*** %printer is documented
527
e4ab1254
AD
528 The "%printer" directive, supported since at least Bison 1.50, is finally
529 documented. The "mfcalc" example is extended to demonstrate it.
7e508a2b 530
e4ab1254
AD
531 For consistency with the C skeletons, the C++ parsers now also support
532 "yyoutput" (as an alias to "debug_stream ()").
7e508a2b
AD
533
534*** Several improvements have been made:
466b4cf2 535
e4ab1254
AD
536 The layout for grammar excerpts was changed to a more compact scheme.
537 Named references are motivated. The description of the automaton
538 description file (*.output) is updated to the current format. Incorrect
539 index entries were fixed. Some other errors were fixed.
466b4cf2 540
86b08b49
AD
541** Building bison:
542
543*** Conflicting prototypes with recent/modified Flex.
466b4cf2 544
e4ab1254
AD
545 Fixed build problems with the current, unreleased, version of Flex, and
546 some modified versions of 2.5.35, which have modified function prototypes.
466b4cf2 547
8ef26c2a
AD
548*** Warnings during the build procedure have been eliminated.
549
550*** Several portability problems in the test suite have been fixed:
466b4cf2 551
e4ab1254
AD
552 This includes warnings with some compilers, unexpected behavior of tools
553 such as diff, warning messages from the test suite itself, etc.
466b4cf2 554
91aadcc7 555*** The install-pdf target works properly:
8ef26c2a 556
e4ab1254
AD
557 Running "make install-pdf" (or -dvi, -html, -info, and -ps) no longer
558 halts in the middle of its course.
8ef26c2a 559
28801043 560* Changes in version 2.5 (2011-05-14):
50cca368 561
82f3355e
JD
562** Grammar symbol names can now contain non-initial dashes:
563
564 Consistently with directives (such as %error-verbose) and with
565 %define variables (e.g. push-pull), grammar symbol names may contain
566 dashes in any position except the beginning. This is a GNU
567 extension over POSIX Yacc. Thus, use of this extension is reported
568 by -Wyacc and rejected in Yacc mode (--yacc).
569
f1b238df 570** Named references:
66381412
AR
571
572 Historically, Yacc and Bison have supported positional references
573 ($n, $$) to allow access to symbol values from inside of semantic
574 actions code.
575
576 Starting from this version, Bison can also accept named references.
577 When no ambiguity is possible, original symbol names may be used
578 as named references:
579
4b568fc0 580 if_stmt : "if" cond_expr "then" then_stmt ';'
66381412
AR
581 { $if_stmt = mk_if_stmt($cond_expr, $then_stmt); }
582
583 In the more common case, explicit names may be declared:
584
4b568fc0 585 stmt[res] : "if" expr[cond] "then" stmt[then] "else" stmt[else] ';'
66381412
AR
586 { $res = mk_if_stmt($cond, $then, $else); }
587
5b1ff423 588 Location information is also accessible using @name syntax. When
66381412
AR
589 accessing symbol names containing dots or dashes, explicit bracketing
590 ($[sym.1]) must be used.
591
5b1ff423 592 These features are experimental in this version. More user feedback
66381412
AR
593 will help to stabilize them.
594
f1b238df 595** IELR(1) and canonical LR(1):
eb45ef3b
JD
596
597 IELR(1) is a minimal LR(1) parser table generation algorithm. That
598 is, given any context-free grammar, IELR(1) generates parser tables
7262f54f 599 with the full language-recognition power of canonical LR(1) but with
f1b238df
JD
600 nearly the same number of parser states as LALR(1). This reduction
601 in parser states is often an order of magnitude. More importantly,
eb45ef3b
JD
602 because canonical LR(1)'s extra parser states may contain duplicate
603 conflicts in the case of non-LR(1) grammars, the number of conflicts
604 for IELR(1) is often an order of magnitude less as well. This can
605 significantly reduce the complexity of developing of a grammar.
606
607 Bison can now generate IELR(1) and canonical LR(1) parser tables in
608 place of its traditional LALR(1) parser tables, which remain the
609 default. You can specify the type of parser tables in the grammar
610 file with these directives:
611
cf499cff
JD
612 %define lr.type lalr
613 %define lr.type ielr
614 %define lr.type canonical-lr
eb45ef3b 615
7fceb615 616 The default-reduction optimization in the parser tables can also be
e4ab1254
AD
617 adjusted using "%define lr.default-reductions". For details on both
618 of these features, see the new section "Tuning LR" in the Bison
7fceb615 619 manual.
eb45ef3b
JD
620
621 These features are experimental. More user feedback will help to
622 stabilize them.
623
7fceb615 624** LAC (Lookahead Correction) for syntax error handling:
fcf834f9
JD
625
626 Canonical LR, IELR, and LALR can suffer from a couple of problems
627 upon encountering a syntax error. First, the parser might perform
628 additional parser stack reductions before discovering the syntax
7fceb615 629 error. Such reductions can perform user semantic actions that are
fcf834f9
JD
630 unexpected because they are based on an invalid token, and they
631 cause error recovery to begin in a different syntactic context than
632 the one in which the invalid token was encountered. Second, when
7fceb615 633 verbose error messages are enabled (with %error-verbose or the
e4ab1254 634 obsolete "#define YYERROR_VERBOSE"), the expected token list in the
7fceb615
JD
635 syntax error message can both contain invalid tokens and omit valid
636 tokens.
fcf834f9
JD
637
638 The culprits for the above problems are %nonassoc, default
639 reductions in inconsistent states, and parser state merging. Thus,
640 IELR and LALR suffer the most. Canonical LR can suffer only if
641 %nonassoc is used or if default reductions are enabled for
642 inconsistent states.
643
7fceb615
JD
644 LAC is a new mechanism within the parsing algorithm that solves
645 these problems for canonical LR, IELR, and LALR without sacrificing
646 %nonassoc, default reductions, or state merging. When LAC is in
647 use, canonical LR and IELR behave almost exactly the same for both
648 syntactically acceptable and syntactically unacceptable input.
fcf834f9
JD
649 While LALR still does not support the full language-recognition
650 power of canonical LR and IELR, LAC at least enables LALR's syntax
651 error handling to correctly reflect LALR's language-recognition
652 power.
653
654 Currently, LAC is only supported for deterministic parsers in C.
655 You can enable LAC with the following directive:
656
657 %define parse.lac full
658
e4ab1254 659 See the new section "LAC" in the Bison manual for additional
7fceb615 660 details including a few caveats.
fcf834f9
JD
661
662 LAC is an experimental feature. More user feedback will help to
663 stabilize it.
664
d397d9f0 665** %define improvements:
cf499cff 666
f1b238df 667*** Can now be invoked via the command line:
50cca368 668
de5ab940 669 Each of these command-line options
50cca368 670
de5ab940
JD
671 -D NAME[=VALUE]
672 --define=NAME[=VALUE]
673
674 -F NAME[=VALUE]
675 --force-define=NAME[=VALUE]
50cca368
JD
676
677 is equivalent to this grammar file declaration
678
de5ab940 679 %define NAME ["VALUE"]
50cca368 680
de5ab940
JD
681 except that the manner in which Bison processes multiple definitions
682 for the same NAME differs. Most importantly, -F and --force-define
683 quietly override %define, but -D and --define do not. For further
e4ab1254 684 details, see the section "Bison Options" in the Bison manual.
50cca368 685
f1b238df 686*** Variables renamed:
67212941
JD
687
688 The following %define variables
689
690 api.push_pull
691 lr.keep_unreachable_states
692
693 have been renamed to
694
695 api.push-pull
696 lr.keep-unreachable-states
697
698 The old names are now deprecated but will be maintained indefinitely
699 for backward compatibility.
700
7262f54f 701*** Values no longer need to be quoted in the grammar file:
cf499cff
JD
702
703 If a %define value is an identifier, it no longer needs to be placed
704 within quotations marks. For example,
705
706 %define api.push-pull "push"
707
708 can be rewritten as
709
710 %define api.push-pull push
711
d397d9f0 712*** Unrecognized variables are now errors not warnings.
cdf3f113 713
d397d9f0
JD
714*** Multiple invocations for any variable is now an error not a warning.
715
716** Unrecognized %code qualifiers are now errors not warnings.
717
718** Character literals not of length one:
719
720 Previously, Bison quietly converted all character literals to length
721 one. For example, without warning, Bison interpreted the operators in
722 the following grammar to be the same token:
723
724 exp: exp '++'
725 | exp '+' exp
726 ;
727
728 Bison now warns when a character literal is not of length one. In
729 some future release, Bison will start reporting an error instead.
730
731** Destructor calls fixed for lookaheads altered in semantic actions:
732
733 Previously for deterministic parsers in C, if a user semantic action
734 altered yychar, the parser in some cases used the old yychar value to
735 determine which destructor to call for the lookahead upon a syntax
736 error or upon parser return. This bug has been fixed.
737
738** C++ parsers use YYRHSLOC:
739
740 Similarly to the C parsers, the C++ parsers now define the YYRHSLOC
741 macro and use it in the default YYLLOC_DEFAULT. You are encouraged
e4ab1254
AD
742 to use it. If, for instance, your location structure has "first"
743 and "last" members, instead of
d397d9f0 744
e29f0771
AD
745 # define YYLLOC_DEFAULT(Current, Rhs, N) \
746 do \
747 if (N) \
748 { \
749 (Current).first = (Rhs)[1].location.first; \
750 (Current).last = (Rhs)[N].location.last; \
751 } \
752 else \
753 { \
754 (Current).first = (Current).last = (Rhs)[0].location.last; \
755 } \
756 while (false)
d397d9f0
JD
757
758 use:
759
e29f0771
AD
760 # define YYLLOC_DEFAULT(Current, Rhs, N) \
761 do \
762 if (N) \
763 { \
764 (Current).first = YYRHSLOC (Rhs, 1).first; \
765 (Current).last = YYRHSLOC (Rhs, N).last; \
766 } \
767 else \
768 { \
769 (Current).first = (Current).last = YYRHSLOC (Rhs, 0).last; \
770 } \
771 while (false)
d397d9f0
JD
772
773** YYLLOC_DEFAULT in C++:
774
775 The default implementation of YYLLOC_DEFAULT used to be issued in
776 the header file. It is now output in the implementation file, after
777 the user %code sections so that its #ifndef guard does not try to
778 override the user's YYLLOC_DEFAULT if provided.
cdf3f113 779
f1b238df 780** YYFAIL now produces warnings and Java parsers no longer implement it:
4395a9ff
JD
781
782 YYFAIL has existed for many years as an undocumented feature of
783 deterministic parsers in C generated by Bison. More recently, it was
784 a documented feature of Bison's experimental Java parsers. As
785 promised in Bison 2.4.2's NEWS entry, any appearance of YYFAIL in a
786 semantic action now produces a deprecation warning, and Java parsers
787 no longer implement YYFAIL at all. For further details, including a
788 discussion of how to suppress C preprocessor warnings about YYFAIL
789 being unused, see the Bison 2.4.2 NEWS entry.
790
f1b238df 791** Temporary hack for adding a semicolon to the user action:
197b82ba
JD
792
793 Previously, Bison appended a semicolon to every user action for
794 reductions when the output language defaulted to C (specifically, when
795 neither %yacc, %language, %skeleton, or equivalent command-line
796 options were specified). This allowed actions such as
797
798 exp: exp "+" exp { $$ = $1 + $3 };
799
800 instead of
801
802 exp: exp "+" exp { $$ = $1 + $3; };
803
804 As a first step in removing this misfeature, Bison now issues a
805 warning when it appends a semicolon. Moreover, in cases where Bison
806 cannot easily determine whether a semicolon is needed (for example, an
807 action ending with a cpp directive or a braced compound initializer),
808 it no longer appends one. Thus, the C compiler might now complain
809 about a missing semicolon where it did not before. Future releases of
810 Bison will cease to append semicolons entirely.
811
d2060f06
JD
812** Verbose syntax error message fixes:
813
e4ab1254 814 When %error-verbose or the obsolete "#define YYERROR_VERBOSE" is
7fceb615
JD
815 specified, syntax error messages produced by the generated parser
816 include the unexpected token as well as a list of expected tokens.
817 The effect of %nonassoc on these verbose messages has been corrected
818 in two ways, but a more complete fix requires LAC, described above:
d2060f06
JD
819
820*** When %nonassoc is used, there can exist parser states that accept no
821 tokens, and so the parser does not always require a lookahead token
822 in order to detect a syntax error. Because no unexpected token or
823 expected tokens can then be reported, the verbose syntax error
824 message described above is suppressed, and the parser instead
e4ab1254 825 reports the simpler message, "syntax error". Previously, this
d2060f06
JD
826 suppression was sometimes erroneously triggered by %nonassoc when a
827 lookahead was actually required. Now verbose messages are
828 suppressed only when all previous lookaheads have already been
829 shifted or discarded.
830
831*** Previously, the list of expected tokens erroneously included tokens
832 that would actually induce a syntax error because conflicts for them
833 were resolved with %nonassoc in the current parser state. Such
834 tokens are now properly omitted from the list.
835
836*** Expected token lists are still often wrong due to state merging
fcf834f9
JD
837 (from LALR or IELR) and default reductions, which can both add
838 invalid tokens and subtract valid tokens. Canonical LR almost
839 completely fixes this problem by eliminating state merging and
840 default reductions. However, there is one minor problem left even
841 when using canonical LR and even after the fixes above. That is,
842 if the resolution of a conflict with %nonassoc appears in a later
843 parser state than the one at which some syntax error is
844 discovered, the conflicted token is still erroneously included in
845 the expected token list. Bison's new LAC implementation,
846 described above, eliminates this problem and the need for
847 canonical LR. However, LAC is still experimental and is disabled
848 by default.
53f036ce 849
1a33f4f6
JD
850** Java skeleton fixes:
851
852*** A location handling bug has been fixed.
853
854*** The top element of each of the value stack and location stack is now
855 cleared when popped so that it can be garbage collected.
6771a463 856
02803d55
JD
857*** Parser traces now print the top element of the stack.
858
86408959
JD
859** -W/--warnings fixes:
860
e4ab1254 861*** Bison now properly recognizes the "no-" versions of categories:
86408959
JD
862
863 For example, given the following command line, Bison now enables all
864 warnings except warnings for incompatibilities with POSIX Yacc:
865
866 bison -Wall,no-yacc gram.y
867
786743d5
JD
868*** Bison now treats S/R and R/R conflicts like other warnings:
869
870 Previously, conflict reports were independent of Bison's normal
871 warning system. Now, Bison recognizes the warning categories
e4ab1254 872 "conflicts-sr" and "conflicts-rr". This change has important
786743d5
JD
873 consequences for the -W and --warnings command-line options. For
874 example:
875
876 bison -Wno-conflicts-sr gram.y # S/R conflicts not reported
877 bison -Wno-conflicts-rr gram.y # R/R conflicts not reported
878 bison -Wnone gram.y # no conflicts are reported
879 bison -Werror gram.y # any conflict is an error
880
881 However, as before, if the %expect or %expect-rr directive is
882 specified, an unexpected number of conflicts is an error, and an
883 expected number of conflicts is not reported, so -W and --warning
884 then have no effect on the conflict report.
885
e4ab1254 886*** The "none" category no longer disables a preceding "error":
bf0e44e8
JD
887
888 For example, for the following command line, Bison now reports
889 errors instead of warnings for incompatibilities with POSIX Yacc:
890
891 bison -Werror,none,yacc gram.y
892
e4ab1254 893*** The "none" category now disables all Bison warnings:
c39014ae 894
e4ab1254 895 Previously, the "none" category disabled only Bison warnings for
c39014ae
JD
896 which there existed a specific -W/--warning category. However,
897 given the following command line, Bison is now guaranteed to
898 suppress all warnings:
899
900 bison -Wnone gram.y
901
1f36f544
JD
902** Precedence directives can now assign token number 0:
903
904 Since Bison 2.3b, which restored the ability of precedence
905 directives to assign token numbers, doing so for token number 0 has
906 produced an assertion failure. For example:
907
908 %left END 0
909
910 This bug has been fixed.
911
64877e5e 912* Changes in version 2.4.3 (2010-08-05):
8b9e021f 913
2bfcac9a
JD
914** Bison now obeys -Werror and --warnings=error for warnings about
915 grammar rules that are useless in the parser due to conflicts.
916
8b9e021f
JD
917** Problems with spawning M4 on at least FreeBSD 8 and FreeBSD 9 have
918 been fixed.
919
4ad3921d
JD
920** Failures in the test suite for GCC 4.5 have been fixed.
921
06cb07d5
JD
922** Failures in the test suite for some versions of Sun Studio C++ have
923 been fixed.
924
9b5049bd
JD
925** Contrary to Bison 2.4.2's NEWS entry, it has been decided that
926 warnings about undefined %prec identifiers will not be converted to
927 errors in Bison 2.5. They will remain warnings, which should be
928 sufficient for POSIX while avoiding backward compatibility issues.
929
93d7dde9
JD
930** Minor documentation fixes.
931
e19a049c 932* Changes in version 2.4.2 (2010-03-20):
74553c98 933
f39ab286
JD
934** Some portability problems that resulted in failures and livelocks
935 in the test suite on some versions of at least Solaris, AIX, HP-UX,
e19a049c
JD
936 RHEL4, and Tru64 have been addressed. As a result, fatal Bison
937 errors should no longer cause M4 to report a broken pipe on the
f39ab286
JD
938 affected platforms.
939
e4ab1254 940** "%prec IDENTIFIER" requires IDENTIFIER to be defined separately.
8bb3a2e7
JD
941
942 POSIX specifies that an error be reported for any identifier that does
943 not appear on the LHS of a grammar rule and that is not defined by
944 %token, %left, %right, or %nonassoc. Bison 2.3b and later lost this
945 error report for the case when an identifier appears only after a
946 %prec directive. It is now restored. However, for backward
947 compatibility with recent Bison releases, it is only a warning for
948 now. In Bison 2.5 and later, it will return to being an error.
9b5049bd
JD
949 [Between the 2.4.2 and 2.4.3 releases, it was decided that this
950 warning will not be converted to an error in Bison 2.5.]
8bb3a2e7 951
d8911864
EB
952** Detection of GNU M4 1.4.6 or newer during configure is improved.
953
a603c6e0
JD
954** Warnings from gcc's -Wundef option about undefined YYENABLE_NLS,
955 YYLTYPE_IS_TRIVIAL, and __STRICT_ANSI__ in C/C++ parsers are now
956 avoided.
c938d650 957
98a345a2
JD
958** %code is now a permanent feature.
959
960 A traditional Yacc prologue directive is written in the form:
961
962 %{CODE%}
963
964 To provide a more flexible alternative, Bison 2.3b introduced the
965 %code directive with the following forms for C/C++:
966
967 %code {CODE}
968 %code requires {CODE}
969 %code provides {CODE}
970 %code top {CODE}
971
972 These forms are now considered permanent features of Bison. See the
973 %code entries in the section "Bison Declaration Summary" in the Bison
974 manual for a summary of their functionality. See the section
975 "Prologue Alternatives" for a detailed discussion including the
976 advantages of %code over the traditional Yacc prologue directive.
977
978 Bison's Java feature as a whole including its current usage of %code
979 is still considered experimental.
980
1625df5b
JD
981** YYFAIL is deprecated and will eventually be removed.
982
983 YYFAIL has existed for many years as an undocumented feature of
984 deterministic parsers in C generated by Bison. Previously, it was
985 documented for Bison's experimental Java parsers. YYFAIL is no longer
986 documented for Java parsers and is formally deprecated in both cases.
987 Users are strongly encouraged to migrate to YYERROR, which is
988 specified by POSIX.
989
990 Like YYERROR, you can invoke YYFAIL from a semantic action in order to
991 induce a syntax error. The most obvious difference from YYERROR is
992 that YYFAIL will automatically invoke yyerror to report the syntax
993 error so that you don't have to. However, there are several other
994 subtle differences between YYERROR and YYFAIL, and YYFAIL suffers from
e4ab1254 995 inherent flaws when %error-verbose or "#define YYERROR_VERBOSE" is
1625df5b
JD
996 used. For a more detailed discussion, see:
997
998 http://lists.gnu.org/archive/html/bison-patches/2009-12/msg00024.html
999
1000 The upcoming Bison 2.5 will remove YYFAIL from Java parsers, but
1001 deterministic parsers in C will continue to implement it. However,
1002 because YYFAIL is already flawed, it seems futile to try to make new
1003 Bison features compatible with it. Thus, during parser generation,
1004 Bison 2.5 will produce a warning whenever it discovers YYFAIL in a
1005 rule action. In a later release, YYFAIL will be disabled for
e4ab1254 1006 %error-verbose and "#define YYERROR_VERBOSE". Eventually, YYFAIL will
1625df5b
JD
1007 be removed altogether.
1008
1009 There exists at least one case where Bison 2.5's YYFAIL warning will
1010 be a false positive. Some projects add phony uses of YYFAIL and other
1011 Bison-defined macros for the sole purpose of suppressing C
1012 preprocessor warnings (from GCC cpp's -Wunused-macros, for example).
1013 To avoid Bison's future warning, such YYFAIL uses can be moved to the
e4ab1254 1014 epilogue (that is, after the second "%%") in the Bison input file. In
1625df5b
JD
1015 this release (2.4.2), Bison already generates its own code to suppress
1016 C preprocessor warnings for YYFAIL, so projects can remove their own
1017 phony uses of YYFAIL if compatibility with Bison releases prior to
1018 2.4.2 is not necessary.
1019
2755de8f
AD
1020** Internationalization.
1021
1022 Fix a regression introduced in Bison 2.4: Under some circumstances,
1023 message translations were not installed although supported by the
1024 host system.
1025
74553c98 1026* Changes in version 2.4.1 (2008-12-11):
c9ba9e59 1027
0ea583d2
AD
1028** In the GLR defines file, unexpanded M4 macros in the yylval and yylloc
1029 declarations have been fixed.
1979121c 1030
0ea583d2
AD
1031** Temporary hack for adding a semicolon to the user action.
1032
1033 Bison used to prepend a trailing semicolon at the end of the user
1034 action for reductions. This allowed actions such as
1035
1036 exp: exp "+" exp { $$ = $1 + $3 };
1037
1038 instead of
1039
1040 exp: exp "+" exp { $$ = $1 + $3; };
1041
e4ab1254 1042 Some grammars still depend on this "feature". Bison 2.4.1 restores
0ea583d2
AD
1043 the previous behavior in the case of C output (specifically, when
1044 neither %language or %skeleton or equivalent command-line options
1045 are used) to leave more time for grammars depending on the old
1046 behavior to be adjusted. Future releases of Bison will disable this
1047 feature.
1048
1049** A few minor improvements to the Bison manual.
c9ba9e59 1050
402b123d 1051* Changes in version 2.4 (2008-11-02):
7bd1665a 1052
402b123d 1053** %language is an experimental feature.
ed4d67dc
JD
1054
1055 We first introduced this feature in test release 2.3b as a cleaner
1056 alternative to %skeleton. Since then, we have discussed the possibility of
1057 modifying its effect on Bison's output file names. Thus, in this release,
1058 we consider %language to be an experimental feature that will likely evolve
1059 in future releases.
7bd1665a 1060
402b123d 1061** Forward compatibility with GNU M4 has been improved.
241fda7a 1062
402b123d 1063** Several bugs in the C++ skeleton and the experimental Java skeleton have been
241fda7a
JD
1064 fixed.
1065
402b123d 1066* Changes in version 2.3b (2008-05-27):
35fe0834 1067
402b123d 1068** The quotes around NAME that used to be required in the following directive
d9df47b6
JD
1069 are now deprecated:
1070
1071 %define NAME "VALUE"
1072
e4ab1254 1073** The directive "%pure-parser" is now deprecated in favor of:
d9df47b6
JD
1074
1075 %define api.pure
1076
1077 which has the same effect except that Bison is more careful to warn about
1078 unreasonable usage in the latter case.
1079
402b123d 1080** Push Parsing
c373bf8b
JD
1081
1082 Bison can now generate an LALR(1) parser in C with a push interface. That
e4ab1254
AD
1083 is, instead of invoking "yyparse", which pulls tokens from "yylex", you can
1084 push one token at a time to the parser using "yypush_parse", which will
c373bf8b
JD
1085 return to the caller after processing each token. By default, the push
1086 interface is disabled. Either of the following directives will enable it:
1087
1088 %define api.push_pull "push" // Just push; does not require yylex.
1089 %define api.push_pull "both" // Push and pull; requires yylex.
1090
e4ab1254 1091 See the new section "A Push Parser" in the Bison manual for details.
c373bf8b 1092
59da312b
JD
1093 The current push parsing interface is experimental and may evolve. More user
1094 feedback will help to stabilize it.
1095
402b123d 1096** The -g and --graph options now output graphs in Graphviz DOT format,
8e55b3aa
JD
1097 not VCG format. Like --graph, -g now also takes an optional FILE argument
1098 and thus cannot be bundled with other short options.
c373bf8b 1099
402b123d 1100** Java
59da312b
JD
1101
1102 Bison can now generate an LALR(1) parser in Java. The skeleton is
e4ab1254 1103 "data/lalr1.java". Consider using the new %language directive instead of
59da312b
JD
1104 %skeleton to select it.
1105
e4ab1254 1106 See the new section "Java Parsers" in the Bison manual for details.
59da312b
JD
1107
1108 The current Java interface is experimental and may evolve. More user
1109 feedback will help to stabilize it.
1110
402b123d 1111** %language
59da312b
JD
1112
1113 This new directive specifies the programming language of the generated
d43f77e7
PB
1114 parser, which can be C (the default), C++, or Java. Besides the skeleton
1115 that Bison uses, the directive affects the names of the generated files if
1116 the grammar file's name ends in ".y".
59da312b 1117
402b123d 1118** XML Automaton Report
59da312b
JD
1119
1120 Bison can now generate an XML report of the LALR(1) automaton using the new
e4ab1254 1121 "--xml" option. The current XML schema is experimental and may evolve. More
59da312b 1122 user feedback will help to stabilize it.
c373bf8b 1123
402b123d 1124** The grammar file may now specify the name of the parser header file using
c373bf8b
JD
1125 %defines. For example:
1126
1127 %defines "parser.h"
1128
402b123d 1129** When reporting useless rules, useless nonterminals, and unused terminals,
d80fb37a
JD
1130 Bison now employs the terms "useless in grammar" instead of "useless",
1131 "useless in parser" instead of "never reduced", and "unused in grammar"
1132 instead of "unused".
cff03fb2 1133
402b123d 1134** Unreachable State Removal
c373bf8b
JD
1135
1136 Previously, Bison sometimes generated parser tables containing unreachable
31984206
JD
1137 states. A state can become unreachable during conflict resolution if Bison
1138 disables a shift action leading to it from a predecessor state. Bison now:
75ad86ee
JD
1139
1140 1. Removes unreachable states.
1141
1142 2. Does not report any conflicts that appeared in unreachable states.
1143 WARNING: As a result, you may need to update %expect and %expect-rr
1144 directives in existing grammar files.
1145
1146 3. For any rule used only in such states, Bison now reports the rule as
cff03fb2 1147 "useless in parser due to conflicts".
75ad86ee 1148
31984206
JD
1149 This feature can be disabled with the following directive:
1150
1151 %define lr.keep_unreachable_states
1152
e4ab1254 1153 See the %define entry in the "Bison Declaration Summary" in the Bison manual
31984206
JD
1154 for further discussion.
1155
e4ab1254 1156** Lookahead Set Correction in the ".output" Report
b1cc23c4 1157
e4ab1254
AD
1158 When instructed to generate a ".output" file including lookahead sets
1159 (using "--report=lookahead", for example), Bison now prints each reduction's
88c78747
JD
1160 lookahead set only next to the associated state's one item that (1) is
1161 associated with the same rule as the reduction and (2) has its dot at the end
1162 of its RHS. Previously, Bison also erroneously printed the lookahead set
1163 next to all of the state's other items associated with the same rule. This
e4ab1254 1164 bug affected only the ".output" file and not the generated parser source
88c78747
JD
1165 code.
1166
e4ab1254 1167** --report-file=FILE is a new option to override the default ".output" file
59da312b 1168 name.
1bb2bd75 1169
e4ab1254 1170** The "=" that used to be required in the following directives is now
02975b9a
JD
1171 deprecated:
1172
1173 %file-prefix "parser"
1174 %name-prefix "c_"
1175 %output "parser.c"
1176
e4ab1254 1177** An Alternative to "%{...%}" -- "%code QUALIFIER {CODE}"
c373bf8b
JD
1178
1179 Bison 2.3a provided a new set of directives as a more flexible alternative to
8e0a5e9e
JD
1180 the traditional Yacc prologue blocks. Those have now been consolidated into
1181 a single %code directive with an optional qualifier field, which identifies
1182 the purpose of the code and thus the location(s) where Bison should generate
1183 it:
1184
e4ab1254
AD
1185 1. "%code {CODE}" replaces "%after-header {CODE}"
1186 2. "%code requires {CODE}" replaces "%start-header {CODE}"
1187 3. "%code provides {CODE}" replaces "%end-header {CODE}"
1188 4. "%code top {CODE}" replaces "%before-header {CODE}"
8e0a5e9e 1189
e4ab1254
AD
1190 See the %code entries in section "Bison Declaration Summary" in the Bison
1191 manual for a summary of the new functionality. See the new section "Prologue
1192 Alternatives" for a detailed discussion including the advantages of %code
8e0a5e9e
JD
1193 over the traditional Yacc prologues.
1194
1195 The prologue alternatives are experimental. More user feedback will help to
1196 determine whether they should become permanent features.
1197
402b123d 1198** Revised warning: unset or unused mid-rule values
17bd8a73
JD
1199
1200 Since Bison 2.2, Bison has warned about mid-rule values that are set but not
1201 used within any of the actions of the parent rule. For example, Bison warns
1202 about unused $2 in:
1203
1204 exp: '1' { $$ = 1; } '+' exp { $$ = $1 + $4; };
1205
1206 Now, Bison also warns about mid-rule values that are used but not set. For
1207 example, Bison warns about unset $$ in the mid-rule action in:
1208
1209 exp: '1' { $1 = 1; } '+' exp { $$ = $2 + $4; };
1210
1211 However, Bison now disables both of these warnings by default since they
1212 sometimes prove to be false alarms in existing grammars employing the Yacc
1213 constructs $0 or $-N (where N is some positive integer).
1214
e4ab1254
AD
1215 To enable these warnings, specify the option "--warnings=midrule-values" or
1216 "-W", which is a synonym for "--warnings=all".
17bd8a73 1217
e4ab1254 1218** Default %destructor or %printer with "<*>" or "<>"
c373bf8b
JD
1219
1220 Bison now recognizes two separate kinds of default %destructor's and
12e35840
JD
1221 %printer's:
1222
e4ab1254 1223 1. Place "<*>" in a %destructor/%printer symbol list to define a default
12e35840
JD
1224 %destructor/%printer for all grammar symbols for which you have formally
1225 declared semantic type tags.
1226
e4ab1254 1227 2. Place "<>" in a %destructor/%printer symbol list to define a default
12e35840
JD
1228 %destructor/%printer for all grammar symbols without declared semantic
1229 type tags.
1230
e4ab1254
AD
1231 Bison no longer supports the "%symbol-default" notation from Bison 2.3a.
1232 "<*>" and "<>" combined achieve the same effect with one exception: Bison no
12e35840
JD
1233 longer applies any %destructor to a mid-rule value if that mid-rule value is
1234 not actually ever referenced using either $$ or $n in a semantic action.
1235
85894313
JD
1236 The default %destructor's and %printer's are experimental. More user
1237 feedback will help to determine whether they should become permanent
1238 features.
1239
e4ab1254 1240 See the section "Freeing Discarded Symbols" in the Bison manual for further
12e35840
JD
1241 details.
1242
402b123d 1243** %left, %right, and %nonassoc can now declare token numbers. This is required
e4ab1254 1244 by POSIX. However, see the end of section "Operator Precedence" in the Bison
ab7f29f8
JD
1245 manual for a caveat concerning the treatment of literal strings.
1246
402b123d 1247** The nonfunctional --no-parser, -n, and %no-parser options have been
b1cc23c4
JD
1248 completely removed from Bison.
1249
402b123d 1250* Changes in version 2.3a, 2006-09-13:
742e4900 1251
402b123d 1252** Instead of %union, you can define and use your own union type
ddc8ede1
PE
1253 YYSTYPE if your grammar contains at least one <type> tag.
1254 Your YYSTYPE need not be a macro; it can be a typedef.
1255 This change is for compatibility with other Yacc implementations,
1256 and is required by POSIX.
1257
402b123d 1258** Locations columns and lines start at 1.
cd48d21d
AD
1259 In accordance with the GNU Coding Standards and Emacs.
1260
402b123d 1261** You may now declare per-type and default %destructor's and %printer's:
ec5479ce
JD
1262
1263 For example:
1264
b2a0b7ca
JD
1265 %union { char *string; }
1266 %token <string> STRING1
1267 %token <string> STRING2
1268 %type <string> string1
1269 %type <string> string2
1270 %union { char character; }
1271 %token <character> CHR
1272 %type <character> chr
1273 %destructor { free ($$); } %symbol-default
1274 %destructor { free ($$); printf ("%d", @$.first_line); } STRING1 string1
1275 %destructor { } <character>
1276
1277 guarantees that, when the parser discards any user-defined symbol that has a
e4ab1254
AD
1278 semantic type tag other than "<character>", it passes its semantic value to
1279 "free". However, when the parser discards a "STRING1" or a "string1", it
1280 also prints its line number to "stdout". It performs only the second
1281 "%destructor" in this case, so it invokes "free" only once.
ec5479ce 1282
85894313
JD
1283 [Although we failed to mention this here in the 2.3a release, the default
1284 %destructor's and %printer's were experimental, and they were rewritten in
1285 future versions.]
1286
e4ab1254
AD
1287** Except for LALR(1) parsers in C with POSIX Yacc emulation enabled (with "-y",
1288 "--yacc", or "%yacc"), Bison no longer generates #define statements for
b931235e
JD
1289 associating token numbers with token names. Removing the #define statements
1290 helps to sanitize the global namespace during preprocessing, but POSIX Yacc
1291 requires them. Bison still generates an enum for token names in all cases.
1292
402b123d 1293** Handling of traditional Yacc prologue blocks is now more consistent but
34f98f46 1294 potentially incompatible with previous releases of Bison.
9bc0dd67
JD
1295
1296 As before, you declare prologue blocks in your grammar file with the
e4ab1254 1297 "%{ ... %}" syntax. To generate the pre-prologue, Bison concatenates all
34f98f46
JD
1298 prologue blocks that you've declared before the first %union. To generate
1299 the post-prologue, Bison concatenates all prologue blocks that you've
ddc8ede1 1300 declared after the first %union.
9bc0dd67 1301
34f98f46 1302 Previous releases of Bison inserted the pre-prologue into both the header
9bc0dd67
JD
1303 file and the code file in all cases except for LALR(1) parsers in C. In the
1304 latter case, Bison inserted it only into the code file. For parsers in C++,
1305 the point of insertion was before any token definitions (which associate
1306 token numbers with names). For parsers in C, the point of insertion was
1307 after the token definitions.
1308
1309 Now, Bison never inserts the pre-prologue into the header file. In the code
1310 file, it always inserts it before the token definitions.
1311
402b123d 1312** Bison now provides a more flexible alternative to the traditional Yacc
34f98f46
JD
1313 prologue blocks: %before-header, %start-header, %end-header, and
1314 %after-header.
1315
1316 For example, the following declaration order in the grammar file reflects the
1317 order in which Bison will output these code blocks. However, you are free to
1318 declare these code blocks in your grammar file in whatever order is most
1319 convenient for you:
1320
1321 %before-header {
1322 /* Bison treats this block like a pre-prologue block: it inserts it into
1323 * the code file before the contents of the header file. It does *not*
1324 * insert it into the header file. This is a good place to put
1325 * #include's that you want at the top of your code file. A common
e4ab1254 1326 * example is '#include "system.h"'. */
34f98f46
JD
1327 }
1328 %start-header {
1329 /* Bison inserts this block into both the header file and the code file.
1330 * In both files, the point of insertion is before any Bison-generated
1331 * token, semantic type, location type, and class definitions. This is a
1332 * good place to define %union dependencies, for example. */
9bc0dd67
JD
1333 }
1334 %union {
34f98f46
JD
1335 /* Unlike the traditional Yacc prologue blocks, the output order for the
1336 * new %*-header blocks is not affected by their declaration position
1337 * relative to any %union in the grammar file. */
9bc0dd67 1338 }
34f98f46
JD
1339 %end-header {
1340 /* Bison inserts this block into both the header file and the code file.
1341 * In both files, the point of insertion is after the Bison-generated
1342 * definitions. This is a good place to declare or define public
1343 * functions or data structures that depend on the Bison-generated
1344 * definitions. */
9bc0dd67 1345 }
34f98f46
JD
1346 %after-header {
1347 /* Bison treats this block like a post-prologue block: it inserts it into
1348 * the code file after the contents of the header file. It does *not*
1349 * insert it into the header file. This is a good place to declare or
1350 * define internal functions or data structures that depend on the
1351 * Bison-generated definitions. */
1352 }
1353
1354 If you have multiple occurrences of any one of the above declarations, Bison
1355 will concatenate the contents in declaration order.
9bc0dd67 1356
85894313
JD
1357 [Although we failed to mention this here in the 2.3a release, the prologue
1358 alternatives were experimental, and they were rewritten in future versions.]
1359
e4ab1254 1360** The option "--report=look-ahead" has been changed to "--report=lookahead".
9e6e7ed2
PE
1361 The old spelling still works, but is not documented and may be removed
1362 in a future release.
742e4900 1363
402b123d 1364* Changes in version 2.3, 2006-06-05:
4ad3ed84 1365
e4ab1254 1366** GLR grammars should now use "YYRECOVERING ()" instead of "YYRECOVERING",
4ad3ed84
PE
1367 for compatibility with LALR(1) grammars.
1368
402b123d 1369** It is now documented that any definition of YYSTYPE or YYLTYPE should
4ad3ed84
PE
1370 be to a type name that does not contain parentheses or brackets.
1371
402b123d 1372* Changes in version 2.2, 2006-05-19:
193d7c70 1373
402b123d 1374** The distribution terms for all Bison-generated parsers now permit
193d7c70
PE
1375 using the parsers in nonfree programs. Previously, this permission
1376 was granted only for Bison-generated LALR(1) parsers in C.
5f4236a0 1377
402b123d 1378** %name-prefix changes the namespace name in C++ outputs.
aa08666d 1379
402b123d 1380** The C++ parsers export their token_type.
5f4236a0 1381
402b123d 1382** Bison now allows multiple %union declarations, and concatenates
d6ca7905
PE
1383 their contents together.
1384
402b123d 1385** New warning: unused values
4d7bc38c
PE
1386 Right-hand side symbols whose values are not used are reported,
1387 if the symbols have destructors. For instance:
affac613 1388
8f3596a6 1389 exp: exp "?" exp ":" exp { $1 ? $1 : $3; }
e9690142
JD
1390 | exp "+" exp
1391 ;
affac613 1392
8f3596a6
AD
1393 will trigger a warning about $$ and $5 in the first rule, and $3 in
1394 the second ($1 is copied to $$ by the default rule). This example
4e26c69e 1395 most likely contains three errors, and could be rewritten as:
affac613 1396
4e26c69e 1397 exp: exp "?" exp ":" exp
e9690142
JD
1398 { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); }
1399 | exp "+" exp
1400 { $$ = $1 ? $1 : $3; if ($1) free ($3); }
1401 ;
affac613 1402
4e26c69e
PE
1403 However, if the original actions were really intended, memory leaks
1404 and all, the warnings can be suppressed by letting Bison believe the
1405 values are used, e.g.:
721be13c 1406
8f3596a6 1407 exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); }
e9690142
JD
1408 | exp "+" exp { $$ = $1; (void) $3; }
1409 ;
721be13c 1410
84866159
AD
1411 If there are mid-rule actions, the warning is issued if no action
1412 uses it. The following triggers no warning: $1 and $3 are used.
1413
1414 exp: exp { push ($1); } '+' exp { push ($3); sum (); };
1415
721be13c
PE
1416 The warning is intended to help catching lost values and memory leaks.
1417 If a value is ignored, its associated memory typically is not reclaimed.
affac613 1418
402b123d 1419** %destructor vs. YYABORT, YYACCEPT, and YYERROR.
9d9b8b70
PE
1420 Destructors are now called when user code invokes YYABORT, YYACCEPT,
1421 and YYERROR, for all objects on the stack, other than objects
1422 corresponding to the right-hand side of the current rule.
a85284cf 1423
402b123d 1424** %expect, %expect-rr
035aa4a0
PE
1425 Incorrect numbers of expected conflicts are now actual errors,
1426 instead of warnings.
1427
402b123d 1428** GLR, YACC parsers.
4e26c69e
PE
1429 The %parse-params are available in the destructors (and the
1430 experimental printers) as per the documentation.
4b367315 1431
e4ab1254 1432** Bison now warns if it finds a stray "$" or "@" in an action.
ad6a9b97 1433
402b123d 1434** %require "VERSION"
4e26c69e
PE
1435 This specifies that the grammar file depends on features implemented
1436 in Bison version VERSION or higher.
b50d2359 1437
402b123d 1438** lalr1.cc: The token and value types are now class members.
e14d0ab6
AD
1439 The tokens were defined as free form enums and cpp macros. YYSTYPE
1440 was defined as a free form union. They are now class members:
e4ab1254
AD
1441 tokens are enumerations of the "yy::parser::token" struct, and the
1442 semantic values have the "yy::parser::semantic_type" type.
fb9712a9
AD
1443
1444 If you do not want or can update to this scheme, the directive
e4ab1254 1445 '%define "global_tokens_and_yystype" "1"' triggers the global
b50d2359
AD
1446 definition of tokens and YYSTYPE. This change is suitable both
1447 for previous releases of Bison, and this one.
fb9712a9 1448
b50d2359 1449 If you wish to update, then make sure older version of Bison will
e4ab1254 1450 fail using '%require "2.2"'.
fb9712a9 1451
402b123d 1452** DJGPP support added.
193d7c70 1453\f
402b123d 1454* Changes in version 2.1, 2005-09-16:
1ce59070 1455
402b123d 1456** The C++ lalr1.cc skeleton supports %lex-param.
e14d0ab6 1457
402b123d 1458** Bison-generated parsers now support the translation of diagnostics like
baf785db
PE
1459 "syntax error" into languages other than English. The default
1460 language is still English. For details, please see the new
0410a6e0
PE
1461 Internationalization section of the Bison manual. Software
1462 distributors should also see the new PACKAGING file. Thanks to
1463 Bruno Haible for this new feature.
1ce59070 1464
402b123d 1465** Wording in the Bison-generated parsers has been changed slightly to
1a059451
PE
1466 simplify translation. In particular, the message "memory exhausted"
1467 has replaced "parser stack overflow", as the old message was not
1468 always accurate for modern Bison-generated parsers.
1469
402b123d 1470** Destructors are now called when the parser aborts, for all symbols left
258b75ca
PE
1471 behind on the stack. Also, the start symbol is now destroyed after a
1472 successful parse. In both cases, the behavior was formerly inconsistent.
1473
402b123d 1474** When generating verbose diagnostics, Bison-generated parsers no longer
72f000b0
PE
1475 quote the literal strings associated with tokens. For example, for
1476 a syntax error associated with '%token NUM "number"' they might
1477 print 'syntax error, unexpected number' instead of 'syntax error,
1478 unexpected "number"'.
193d7c70 1479\f
402b123d 1480* Changes in version 2.0, 2004-12-25:
efeed023 1481
402b123d 1482** Possibly-incompatible changes
d7e14fc0 1483
82de6b0d
PE
1484 - Bison-generated parsers no longer default to using the alloca function
1485 (when available) to extend the parser stack, due to widespread
1486 problems in unchecked stack-overflow detection. You can "#define
1487 YYSTACK_USE_ALLOCA 1" to require the use of alloca, but please read
1488 the manual to determine safe values for YYMAXDEPTH in that case.
8dd162d3 1489
82de6b0d
PE
1490 - Error token location.
1491 During error recovery, the location of the syntax error is updated
1492 to cover the whole sequence covered by the error token: it includes
1493 the shifted symbols thrown away during the first part of the error
1494 recovery, and the lookahead rejected during the second part.
18d192f0 1495
82de6b0d
PE
1496 - Semicolon changes:
1497 . Stray semicolons are no longer allowed at the start of a grammar.
1498 . Semicolons are now required after in-grammar declarations.
e342c3be 1499
82de6b0d
PE
1500 - Unescaped newlines are no longer allowed in character constants or
1501 string literals. They were never portable, and GCC 3.4.0 has
1502 dropped support for them. Better diagnostics are now generated if
1503 forget a closing quote.
8dd162d3 1504
82de6b0d 1505 - NUL bytes are no longer allowed in Bison string literals, unfortunately.
f74b6f91 1506
402b123d 1507** New features
1452af69 1508
82de6b0d 1509 - GLR grammars now support locations.
4febdd96 1510
82de6b0d
PE
1511 - New directive: %initial-action.
1512 This directive allows the user to run arbitrary code (including
1513 initializing @$) from yyparse before parsing starts.
1452af69 1514
82de6b0d
PE
1515 - A new directive "%expect-rr N" specifies the expected number of
1516 reduce/reduce conflicts in GLR parsers.
1452af69 1517
e4ab1254 1518 - %token numbers can now be hexadecimal integers, e.g., "%token FOO 0x12d".
82de6b0d 1519 This is a GNU extension.
4febdd96 1520
e4ab1254 1521 - The option "--report=lookahead" was changed to "--report=look-ahead".
9e6e7ed2 1522 [However, this was changed back after 2.3.]
1452af69 1523
82de6b0d 1524 - Experimental %destructor support has been added to lalr1.cc.
1452af69 1525
82de6b0d
PE
1526 - New configure option --disable-yacc, to disable installation of the
1527 yacc command and -ly library introduced in 1.875 for POSIX conformance.
6040d338 1528
402b123d 1529** Bug fixes
d5a3fe37 1530
82de6b0d
PE
1531 - For now, %expect-count violations are now just warnings, not errors.
1532 This is for compatibility with Bison 1.75 and earlier (when there are
1533 reduce/reduce conflicts) and with Bison 1.30 and earlier (when there
1534 are too many or too few shift/reduce conflicts). However, in future
1535 versions of Bison we plan to improve the %expect machinery so that
1536 these violations will become errors again.
3473d0f8 1537
82de6b0d
PE
1538 - Within Bison itself, numbers (e.g., goto numbers) are no longer
1539 arbitrarily limited to 16-bit counts.
d600ee67 1540
82de6b0d 1541 - Semicolons are now allowed before "|" in grammar rules, as POSIX requires.
d600ee67 1542\f
402b123d 1543* Changes in version 1.875, 2003-01-01:
963fcc17 1544
402b123d 1545** The documentation license has been upgraded to version 1.2
dc546b0f 1546 of the GNU Free Documentation License.
75eb3bc4 1547
402b123d 1548** syntax error processing
75eb3bc4 1549
dc546b0f
PE
1550 - In Yacc-style parsers YYLLOC_DEFAULT is now used to compute error
1551 locations too. This fixes bugs in error-location computation.
75eb3bc4 1552
dc546b0f
PE
1553 - %destructor
1554 It is now possible to reclaim the memory associated to symbols
1555 discarded during error recovery. This feature is still experimental.
20daca06 1556
dc546b0f
PE
1557 - %error-verbose
1558 This new directive is preferred over YYERROR_VERBOSE.
74724a70 1559
dc546b0f
PE
1560 - #defining yyerror to steal internal variables is discouraged.
1561 It is not guaranteed to work forever.
d1de5372 1562
402b123d 1563** POSIX conformance
d1de5372 1564
dc546b0f
PE
1565 - Semicolons are once again optional at the end of grammar rules.
1566 This reverts to the behavior of Bison 1.33 and earlier, and improves
1567 compatibility with Yacc.
74724a70 1568
e4ab1254
AD
1569 - "parse error" -> "syntax error"
1570 Bison now uniformly uses the term "syntax error"; formerly, the code
1571 and manual sometimes used the term "parse error" instead. POSIX
1572 requires "syntax error" in diagnostics, and it was thought better to
dc546b0f 1573 be consistent.
74724a70 1574
dc546b0f
PE
1575 - The documentation now emphasizes that yylex and yyerror must be
1576 declared before use. C99 requires this.
d1de5372 1577
dc546b0f
PE
1578 - Bison now parses C99 lexical constructs like UCNs and
1579 backslash-newline within C escape sequences, as POSIX 1003.1-2001 requires.
d1de5372 1580
dc546b0f
PE
1581 - File names are properly escaped in C output. E.g., foo\bar.y is
1582 output as "foo\\bar.y".
6780ca7a 1583
dc546b0f 1584 - Yacc command and library now available
e4ab1254 1585 The Bison distribution now installs a "yacc" command, as POSIX requires.
dc546b0f
PE
1586 Also, Bison now installs a small library liby.a containing
1587 implementations of Yacc-compatible yyerror and main functions.
1588 This library is normally not useful, but POSIX requires it.
6e649e65 1589
dc546b0f 1590 - Type clashes now generate warnings, not errors.
6e649e65 1591
dc546b0f
PE
1592 - If the user does not define YYSTYPE as a macro, Bison now declares it
1593 using typedef instead of defining it as a macro.
1594 For consistency, YYLTYPE is also declared instead of defined.
9501dc6e 1595
402b123d 1596** Other compatibility issues
886a425c 1597
e4ab1254
AD
1598 - %union directives can now have a tag before the "{", e.g., the
1599 directive "%union foo {...}" now generates the C code
1600 "typedef union foo { ... } YYSTYPE;"; this is for Yacc compatibility.
1601 The default union tag is "YYSTYPE", for compatibility with Solaris 9 Yacc.
1602 For consistency, YYLTYPE's struct tag is now "YYLTYPE" not "yyltype".
dc546b0f 1603 This is for compatibility with both Yacc and Bison 1.35.
72f889cc 1604
e4ab1254 1605 - ";" is output before the terminating "}" of an action, for
dc546b0f 1606 compatibility with Bison 1.35.
886a425c 1607
dc546b0f 1608 - Bison now uses a Yacc-style format for conflict reports, e.g.,
e4ab1254 1609 "conflicts: 2 shift/reduce, 1 reduce/reduce".
437c2d80 1610
e4ab1254 1611 - "yystype" and "yyltype" are now obsolescent macros instead of being
dc546b0f
PE
1612 typedefs or tags; they are no longer documented and are planned to be
1613 withdrawn in a future release.
2a8d363a 1614
402b123d 1615** GLR parser notes
2a8d363a 1616
dc546b0f
PE
1617 - GLR and inline
1618 Users of Bison have to decide how they handle the portability of the
e4ab1254 1619 C keyword "inline".
959e5f51 1620
e4ab1254
AD
1621 - "parsing stack overflow..." -> "parser stack overflow"
1622 GLR parsers now report "parser stack overflow" as per the Bison manual.
900c5db5 1623
18ad57b3
AD
1624** %parse-param and %lex-param
1625 The macros YYPARSE_PARAM and YYLEX_PARAM provide a means to pass
1626 additional context to yyparse and yylex. They suffer from several
1627 shortcomings:
1628
1629 - a single argument only can be added,
1630 - their types are weak (void *),
242cc08e 1631 - this context is not passed to ancillary functions such as yyerror,
18ad57b3
AD
1632 - only yacc.c parsers support them.
1633
1634 The new %parse-param/%lex-param directives provide a more precise control.
1635 For instance:
1636
1637 %parse-param {int *nastiness}
1638 %lex-param {int *nastiness}
1639 %parse-param {int *randomness}
1640
1641 results in the following signatures:
1642
1643 int yylex (int *nastiness);
1644 int yyparse (int *nastiness, int *randomness);
1645
1646 or, if both %pure-parser and %locations are used:
1647
1648 int yylex (YYSTYPE *lvalp, YYLTYPE *llocp, int *nastiness);
1649 int yyparse (int *nastiness, int *randomness);
1650
402b123d 1651** Bison now warns if it detects conflicting outputs to the same file,
e4ab1254 1652 e.g., it generates a warning for "bison -d -o foo.h foo.y" since
dc546b0f 1653 that command outputs both code and header to foo.h.
6e40b4eb 1654
402b123d 1655** #line in output files
dc546b0f 1656 - --no-line works properly.
6e40b4eb 1657
402b123d 1658** Bison can no longer be built by a K&R C compiler; it requires C89 or
6e40b4eb
AD
1659 later to be built. This change originally took place a few versions
1660 ago, but nobody noticed until we recently asked someone to try
1661 building Bison with a K&R C compiler.
d600ee67 1662\f
402b123d 1663* Changes in version 1.75, 2002-10-14:
7933f2b5 1664
402b123d 1665** Bison should now work on 64-bit hosts.
7933f2b5 1666
402b123d 1667** Indonesian translation thanks to Tedi Heriyanto.
7933f2b5 1668
402b123d 1669** GLR parsers
f50adbbd
AD
1670 Fix spurious parse errors.
1671
402b123d 1672** Pure parsers
f50adbbd
AD
1673 Some people redefine yyerror to steal yyparse' private variables.
1674 Reenable this trick until an official feature replaces it.
1675
402b123d 1676** Type Clashes
d90c934c
AD
1677 In agreement with POSIX and with other Yaccs, leaving a default
1678 action is valid when $$ is untyped, and $1 typed:
1679
e9690142 1680 untyped: ... typed;
d90c934c
AD
1681
1682 but the converse remains an error:
1683
e9690142 1684 typed: ... untyped;
d90c934c 1685
402b123d 1686** Values of mid-rule actions
d90c934c
AD
1687 The following code:
1688
e9690142 1689 foo: { ... } { $$ = $1; } ...
d90c934c
AD
1690
1691 was incorrectly rejected: $1 is defined in the second mid-rule
1692 action, and is equal to the $$ of the first mid-rule action.
d600ee67 1693\f
402b123d 1694* Changes in version 1.50, 2002-10-04:
adc8c848 1695
402b123d 1696** GLR parsing
676385e2
PH
1697 The declaration
1698 %glr-parser
1699 causes Bison to produce a Generalized LR (GLR) parser, capable of handling
1700 almost any context-free grammar, ambiguous or not. The new declarations
e8832397 1701 %dprec and %merge on grammar rules allow parse-time resolution of
676385e2
PH
1702 ambiguities. Contributed by Paul Hilfinger.
1703
7933f2b5 1704 Unfortunately Bison 1.50 does not work properly on 64-bit hosts
420f93c8
PE
1705 like the Alpha, so please stick to 32-bit hosts for now.
1706
402b123d 1707** Output Directory
8c165d89 1708 When not in Yacc compatibility mode, when the output file was not
e4ab1254
AD
1709 specified, running "bison foo/bar.y" created "foo/bar.c". It
1710 now creates "bar.c".
8c165d89 1711
402b123d 1712** Undefined token
007a50a4 1713 The undefined token was systematically mapped to 2 which prevented
e88dbdbf 1714 the use of 2 by the user. This is no longer the case.
007a50a4 1715
402b123d 1716** Unknown token numbers
e88dbdbf 1717 If yylex returned an out of range value, yyparse could die. This is
007a50a4
AD
1718 no longer the case.
1719
402b123d 1720** Error token
e88dbdbf 1721 According to POSIX, the error token must be 256.
23c5a174
AD
1722 Bison extends this requirement by making it a preference: *if* the
1723 user specified that one of her tokens is numbered 256, then error
1724 will be mapped onto another number.
1725
402b123d 1726** Verbose error messages
e4ab1254 1727 They no longer report "..., expecting error or..." for states where
217598da
AD
1728 error recovery is possible.
1729
402b123d 1730** End token
e4ab1254 1731 Defaults to "$end" instead of "$".
217598da 1732
402b123d 1733** Error recovery now conforms to documentation and to POSIX
68cd8af3
PE
1734 When a Bison-generated parser encounters a syntax error, it now pops
1735 the stack until it finds a state that allows shifting the error
1736 token. Formerly, it popped the stack until it found a state that
1737 allowed some non-error action other than a default reduction on the
1738 error token. The new behavior has long been the documented behavior,
1739 and has long been required by POSIX. For more details, please see
337116ba
PE
1740 Paul Eggert, "Reductions during Bison error handling" (2002-05-20)
1741 <http://lists.gnu.org/archive/html/bug-bison/2002-05/msg00038.html>.
68cd8af3 1742
402b123d 1743** Traces
5504898e
AD
1744 Popped tokens and nonterminals are now reported.
1745
402b123d 1746** Larger grammars
a861a339
PE
1747 Larger grammars are now supported (larger token numbers, larger grammar
1748 size (= sum of the LHS and RHS lengths), larger LALR tables).
1749 Formerly, many of these numbers ran afoul of 16-bit limits;
1750 now these limits are 32 bits on most hosts.
355e7c1c 1751
402b123d 1752** Explicit initial rule
643a5994
AD
1753 Bison used to play hacks with the initial rule, which the user does
1754 not write. It is now explicit, and visible in the reports and
1755 graphs as rule 0.
23c5a174 1756
402b123d 1757** Useless rules
643a5994 1758 Before, Bison reported the useless rules, but, although not used,
77714df2 1759 included them in the parsers. They are now actually removed.
23c5a174 1760
402b123d 1761** Useless rules, useless nonterminals
6b98e4b5
AD
1762 They are now reported, as a warning, with their locations.
1763
402b123d 1764** Rules never reduced
e8832397
AD
1765 Rules that can never be reduced because of conflicts are now
1766 reported.
1767
e4ab1254 1768** Incorrect "Token not used"
11652ab3
AD
1769 On a grammar such as
1770
e29f0771
AD
1771 %token useless useful
1772 %%
1773 exp: '0' %prec useful;
11652ab3
AD
1774
1775 where a token was used to set the precedence of the last rule,
e4ab1254 1776 bison reported both "useful" and "useless" as useless tokens.
11652ab3 1777
402b123d 1778** Revert the C++ namespace changes introduced in 1.31
77714df2 1779 as they caused too many portability hassles.
0179dd65 1780
402b123d 1781** Default locations
b2d52318
AD
1782 By an accident of design, the default computation of @$ was
1783 performed after another default computation was performed: @$ = @1.
1784 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
1785 the computation of @$.
adc8c848 1786
402b123d 1787** Token end-of-file
b7c49edf
AD
1788 The token end of file may be specified by the user, in which case,
1789 the user symbol is used in the reports, the graphs, and the verbose
e4ab1254 1790 error messages instead of "$end", which remains being the default.
b7c49edf 1791 For instance
e29f0771 1792 %token MYEOF 0
b7c49edf 1793 or
e29f0771 1794 %token MYEOF 0 "end of file"
fdbcd8e2 1795
402b123d 1796** Semantic parser
fdbcd8e2
AD
1797 This old option, which has been broken for ages, is removed.
1798
402b123d 1799** New translations
a861a339 1800 Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
84614e13
AD
1801 Croatian, thanks to Denis Lackovic.
1802
402b123d 1803** Incorrect token definitions
e4ab1254
AD
1804 When given
1805 %token 'a' "A"
1806 bison used to output
1807 #define 'a' 65
b87f8b21 1808
402b123d 1809** Token definitions as enums
77714df2
AD
1810 Tokens are output both as the traditional #define's, and, provided
1811 the compiler supports ANSI C or is a C++ compiler, as enums.
e88dbdbf 1812 This lets debuggers display names instead of integers.
77714df2 1813
402b123d 1814** Reports
ec3bc396
AD
1815 In addition to --verbose, bison supports --report=THINGS, which
1816 produces additional information:
b408954b
AD
1817 - itemset
1818 complete the core item sets with their closure
e4ab1254 1819 - lookahead [changed to "look-ahead" in 1.875e through 2.3, but changed back]
9e6e7ed2 1820 explicitly associate lookahead tokens to items
b408954b
AD
1821 - solved
1822 describe shift/reduce conflicts solving.
1823 Bison used to systematically output this information on top of
1824 the report. Solved conflicts are now attached to their states.
ec3bc396 1825
402b123d 1826** Type clashes
9af3fbce
AD
1827 Previous versions don't complain when there is a type clash on
1828 the default action if the rule has a mid-rule action, such as in:
1829
e29f0771
AD
1830 %type <foo> bar
1831 %%
1832 bar: '0' {} '0';
9af3fbce
AD
1833
1834 This is fixed.
a861a339 1835
402b123d 1836** GNU M4 is now required when using Bison.
f987e9d2 1837\f
402b123d 1838* Changes in version 1.35, 2002-03-25:
76551463 1839
402b123d 1840** C Skeleton
76551463
AD
1841 Some projects use Bison's C parser with C++ compilers, and define
1842 YYSTYPE as a class. The recent adjustment of C parsers for data
1843 alignment and 64 bit architectures made this impossible.
1844
1845 Because for the time being no real solution for C++ parser
1846 generation exists, kludges were implemented in the parser to
1847 maintain this use. In the future, when Bison has C++ parsers, this
1848 kludge will be disabled.
1849
1850 This kludge also addresses some C++ problems when the stack was
1851 extended.
76551463 1852\f
402b123d 1853* Changes in version 1.34, 2002-03-12:
76551463 1854
402b123d 1855** File name clashes are detected
76551463 1856 $ bison foo.y -d -o foo.x
e4ab1254 1857 fatal error: header and parser would both be named "foo.x"
76551463 1858
e4ab1254 1859** A missing ";" at the end of a rule triggers a warning
76551463
AD
1860 In accordance with POSIX, and in agreement with other
1861 Yacc implementations, Bison will mandate this semicolon in the near
1862 future. This eases the implementation of a Bison parser of Bison
1863 grammars by making this grammar LALR(1) instead of LR(2). To
1864 facilitate the transition, this release introduces a warning.
1865
402b123d 1866** Revert the C++ namespace changes introduced in 1.31, as they caused too
76551463
AD
1867 many portability hassles.
1868
402b123d 1869** DJGPP support added.
76551463 1870
402b123d 1871** Fix test suite portability problems.
76551463 1872\f
402b123d 1873* Changes in version 1.33, 2002-02-07:
76551463 1874
402b123d 1875** Fix C++ issues
76551463
AD
1876 Groff could not be compiled for the definition of size_t was lacking
1877 under some conditions.
1878
402b123d 1879** Catch invalid @n
76551463
AD
1880 As is done with $n.
1881\f
402b123d 1882* Changes in version 1.32, 2002-01-23:
76551463 1883
402b123d 1884** Fix Yacc output file names
76551463 1885
402b123d 1886** Portability fixes
76551463 1887
402b123d 1888** Italian, Dutch translations
76551463 1889\f
402b123d 1890* Changes in version 1.31, 2002-01-14:
52d1aeee 1891
402b123d 1892** Many Bug Fixes
52d1aeee 1893
402b123d 1894** GNU Gettext and %expect
52d1aeee
MA
1895 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
1896 Bison dies on incorrect %expectations, we fear there will be
1897 too many bug reports for Gettext, so _for the time being_, %expect
e4ab1254 1898 does not trigger an error when the input file is named "plural.y".
52d1aeee 1899
402b123d 1900** Use of alloca in parsers
52d1aeee
MA
1901 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
1902 malloc exclusively. Since 1.29, but was not NEWS'ed.
1903
1904 alloca is used only when compiled with GCC, to avoid portability
1905 problems as on AIX.
1906
402b123d 1907** yyparse now returns 2 if memory is exhausted; formerly it dumped core.
b47dbebe 1908
402b123d 1909** When the generated parser lacks debugging code, YYDEBUG is now 0
52d1aeee
MA
1910 (as POSIX requires) instead of being undefined.
1911
402b123d 1912** User Actions
52d1aeee
MA
1913 Bison has always permitted actions such as { $$ = $1 }: it adds the
1914 ending semicolon. Now if in Yacc compatibility mode, the semicolon
1915 is no longer output: one has to write { $$ = $1; }.
1916
402b123d 1917** Better C++ compliance
52d1aeee 1918 The output parsers try to respect C++ namespaces.
76551463 1919 [This turned out to be a failed experiment, and it was reverted later.]
52d1aeee 1920
402b123d 1921** Reduced Grammars
52d1aeee
MA
1922 Fixed bugs when reporting useless nonterminals.
1923
402b123d 1924** 64 bit hosts
52d1aeee
MA
1925 The parsers work properly on 64 bit hosts.
1926
402b123d 1927** Error messages
52d1aeee
MA
1928 Some calls to strerror resulted in scrambled or missing error messages.
1929
402b123d 1930** %expect
52d1aeee
MA
1931 When the number of shift/reduce conflicts is correct, don't issue
1932 any warning.
1933
402b123d 1934** The verbose report includes the rule line numbers.
52d1aeee 1935
402b123d 1936** Rule line numbers are fixed in traces.
52d1aeee 1937
402b123d 1938** Swedish translation
52d1aeee 1939
402b123d 1940** Parse errors
52d1aeee
MA
1941 Verbose parse error messages from the parsers are better looking.
1942 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
1943 Now: parse error: unexpected '/', expecting "number" or '-' or '('
1944
402b123d 1945** Fixed parser memory leaks.
52d1aeee
MA
1946 When the generated parser was using malloc to extend its stacks, the
1947 previous allocations were not freed.
1948
402b123d 1949** Fixed verbose output file.
52d1aeee
MA
1950 Some newlines were missing.
1951 Some conflicts in state descriptions were missing.
1952
402b123d 1953** Fixed conflict report.
52d1aeee
MA
1954 Option -v was needed to get the result.
1955
402b123d 1956** %expect
52d1aeee
MA
1957 Was not used.
1958 Mismatches are errors, not warnings.
1959
402b123d 1960** Fixed incorrect processing of some invalid input.
52d1aeee 1961
402b123d 1962** Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
52d1aeee 1963
402b123d 1964** Fixed some typos in the documentation.
52d1aeee 1965
402b123d 1966** %token MY_EOF 0 is supported.
52d1aeee
MA
1967 Before, MY_EOF was silently renumbered as 257.
1968
402b123d 1969** doc/refcard.tex is updated.
52d1aeee 1970
402b123d 1971** %output, %file-prefix, %name-prefix.
52d1aeee
MA
1972 New.
1973
402b123d 1974** --output
e4ab1254 1975 New, aliasing "--output-file".
52d1aeee 1976\f
402b123d 1977* Changes in version 1.30, 2001-10-26:
342b8b6e 1978
e4ab1254
AD
1979** "--defines" and "--graph" have now an optional argument which is the
1980 output file name. "-d" and "-g" do not change; they do not take any
342b8b6e
AD
1981 argument.
1982
e4ab1254 1983** "%source_extension" and "%header_extension" are removed, failed
342b8b6e
AD
1984 experiment.
1985
402b123d 1986** Portability fixes.
f987e9d2 1987\f
402b123d 1988* Changes in version 1.29, 2001-09-07:
342b8b6e 1989
402b123d 1990** The output file does not define const, as this caused problems when used
342b8b6e
AD
1991 with common autoconfiguration schemes. If you still use ancient compilers
1992 that lack const, compile with the equivalent of the C compiler option
e4ab1254 1993 "-Dconst=". Autoconf's AC_C_CONST macro provides one way to do this.
342b8b6e 1994
e4ab1254 1995** Added "-g" and "--graph".
f87a2205 1996
402b123d 1997** The Bison manual is now distributed under the terms of the GNU FDL.
f2b5126e 1998
402b123d 1999** The input and the output files has automatically a similar extension.
234a3be3 2000
402b123d 2001** Russian translation added.
f87a2205 2002
402b123d 2003** NLS support updated; should hopefully be less troublesome.
f87a2205 2004
402b123d 2005** Added the old Bison reference card.
c33638bb 2006
e4ab1254 2007** Added "--locations" and "%locations".
6deb4447 2008
e4ab1254 2009** Added "-S" and "--skeleton".
cd5bd6ac 2010
e4ab1254 2011** "%raw", "-r", "--raw" is disabled.
62ab6972 2012
402b123d 2013** Special characters are escaped when output. This solves the problems
cd5bd6ac
AD
2014 of the #line lines with path names including backslashes.
2015
402b123d 2016** New directives.
e4ab1254
AD
2017 "%yacc", "%fixed_output_files", "%defines", "%no_parser", "%verbose",
2018 "%debug", "%source_extension" and "%header_extension".
f987e9d2 2019
402b123d 2020** @$
f987e9d2 2021 Automatic location tracking.
f87a2205 2022\f
402b123d 2023* Changes in version 1.28, 1999-07-06:
d2e00347 2024
402b123d 2025** Should compile better now with K&R compilers.
d2e00347 2026
402b123d 2027** Added NLS.
d2e00347 2028
402b123d 2029** Fixed a problem with escaping the double quote character.
d2e00347 2030
402b123d 2031** There is now a FAQ.
d2e00347 2032\f
402b123d 2033* Changes in version 1.27:
5c31c3c2 2034
402b123d 2035** The make rule which prevented bison.simple from being created on
5c31c3c2
JT
2036 some systems has been fixed.
2037\f
402b123d 2038* Changes in version 1.26:
4be07551 2039
7e508a2b 2040** Bison now uses Automake.
4be07551 2041
402b123d 2042** New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
4be07551 2043
402b123d 2044** Token numbers now start at 257 as previously documented, not 258.
4be07551 2045
402b123d 2046** Bison honors the TMPDIR environment variable.
4be07551 2047
402b123d 2048** A couple of buffer overruns have been fixed.
f51dbca1 2049
402b123d 2050** Problems when closing files should now be reported.
f51dbca1 2051
402b123d 2052** Generated parsers should now work even on operating systems which do
f51dbca1 2053 not provide alloca().
4be07551 2054\f
402b123d 2055* Changes in version 1.25, 1995-10-16:
df8878c5 2056
402b123d 2057** Errors in the input grammar are not fatal; Bison keeps reading
df8878c5 2058the grammar file, and reports all the errors found in it.
8c44d3ec 2059
402b123d 2060** Tokens can now be specified as multiple-character strings: for
df8878c5 2061example, you could use "<=" for a token which looks like <=, instead
7e508a2b 2062of choosing a name like LESSEQ.
df8878c5 2063
402b123d 2064** The %token_table declaration says to write a table of tokens (names
df8878c5
RS
2065and numbers) into the parser file. The yylex function can use this
2066table to recognize multiple-character string tokens, or for other
2067purposes.
2068
402b123d 2069** The %no_lines declaration says not to generate any #line preprocessor
df8878c5
RS
2070directives in the parser file.
2071
402b123d 2072** The %raw declaration says to use internal Bison token numbers, not
df8878c5
RS
2073Yacc-compatible token numbers, when token names are defined as macros.
2074
402b123d 2075** The --no-parser option produces the parser tables without including
df8878c5
RS
2076the parser engine; a project can now use its own parser engine.
2077The actions go into a separate file called NAME.act, in the form of
2078a switch statement body.
2079\f
402b123d 2080* Changes in version 1.23:
6780ca7a 2081
4d019228
DM
2082The user can define YYPARSE_PARAM as the name of an argument to be
2083passed into yyparse. The argument should have type void *. It should
2084actually point to an object. Grammar actions can access the variable
2085by casting it to the proper pointer type.
6780ca7a 2086
6780ca7a 2087Line numbers in output file corrected.
6780ca7a 2088\f
402b123d 2089* Changes in version 1.22:
6780ca7a
DM
2090
2091--help option added.
6780ca7a 2092\f
402b123d 2093* Changes in version 1.20:
6780ca7a
DM
2094
2095Output file does not redefine const for C++.
9f4503d6 2096
76551463
AD
2097-----
2098
34136e65 2099Copyright (C) 1995-2012 Free Software Foundation, Inc.
76551463 2100
74553c98 2101This file is part of Bison, the GNU Parser Generator.
76551463 2102
f16b0819 2103This program is free software: you can redistribute it and/or modify
76551463 2104it under the terms of the GNU General Public License as published by
f16b0819
PE
2105the Free Software Foundation, either version 3 of the License, or
2106(at your option) any later version.
76551463 2107
f16b0819 2108This program is distributed in the hope that it will be useful,
76551463
AD
2109but WITHOUT ANY WARRANTY; without even the implied warranty of
2110MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
2111GNU General Public License for more details.
2112
2113You should have received a copy of the GNU General Public License
f16b0819 2114along with this program. If not, see <http://www.gnu.org/licenses/>.
7e508a2b
AD
2115
2116 LocalWords: yacc YYBACKUP glr GCC lalr ArrayIndexOutOfBoundsException nullptr
2117 LocalWords: cplusplus liby rpl fprintf mfcalc Wyacc stmt cond expr mk sym lr
2118 LocalWords: IELR ielr Lookahead YYERROR nonassoc LALR's api lookaheads yychar
2119 LocalWords: destructor lookahead YYRHSLOC YYLLOC Rhs ifndef YYFAIL cpp sr rr
2120 LocalWords: preprocessor initializer Wno Wnone Werror FreeBSD prec livelocks
2121 LocalWords: Solaris AIX UX RHEL Tru LHS gcc's Wundef YYENABLE NLS YYLTYPE VCG
2122 LocalWords: yyerror cpp's Wunused yylval yylloc prepend yyparse yylex yypush
2123 LocalWords: Graphviz xml nonterminals midrule destructor's YYSTYPE typedef ly
2124 LocalWords: CHR chr printf stdout namespace preprocessing enum pre include's
2125 LocalWords: YYRECOVERING nonfree destructors YYABORT YYACCEPT params enums de
2126 LocalWords: struct yystype DJGPP lex param Haible NUM alloca YYSTACK NUL goto
2127 LocalWords: YYMAXDEPTH Unescaped UCNs YYLTYPE's yyltype typedefs inline Yaccs
2128 LocalWords: Heriyanto Reenable dprec Hilfinger Eggert MYEOF Folle Menezes EOF
242cc08e 2129 LocalWords: Lackovic define's itemset Groff Gettext malloc NEWS'ed YYDEBUG YY
7e508a2b 2130 LocalWords: namespaces strerror const autoconfiguration Dconst Autoconf's FDL
242cc08e
AD
2131 LocalWords: Automake TMPDIR LESSEQ ylwrap endif yydebug YYTOKEN YYLSP ival hh
2132 LocalWords: extern YYTOKENTYPE TOKENTYPE yytokentype tokentype STYPE lval pdf
dcb366b1
AD
2133 LocalWords: lang yyoutput dvi html ps POSIX lvalp llocp Wother nterm arg init
2134 LocalWords: TOK
7e508a2b
AD
2135
2136Local Variables:
2137mode: outline
e4ab1254 2138fill-column: 76
7e508a2b 2139End: