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