3 * Noteworthy changes in release ?.? (????-??-??) [?]
5 ** Incompatible changes
9 Support for YYFAIL is removed (deprecated in Bison 2.4.2).
10 Support for yystype and yyltype (instead of YYSTYPE and YYLTYPE)
11 is removed (deprecated in Bison 1.875).
12 Support for YYPARSE_PARAM is removed (deprecated in Bison 1.875).
16 *** Enhancements of the -Werror option
18 The -Werror=CATEGORY option is now recognized, and will treat specified
19 warnings as errors. The warnings need not have been explicitly activated
20 using the -W option, this is similar to what GCC 4.7 does.
22 For example, given the following command line, Bison will treat both
23 warnings related to POSIX Yacc incompatibilities and S/R conflicts as
24 errors (and only those):
26 $ bison -Werror=yacc,error=conflicts-sr input.y
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:
31 $ bison -Werror -Wnone -Wyacc -Wconflicts-sr input.y
33 (By default -Wconflicts-sr,conflicts-rr,deprecated,other is enabled.)
35 Note that the categories in this -Werror option may not be prefixed with
36 "no-". However, -Wno-error[=CATEGORY] is valid.
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".
42 *** The display of warnings is now richer
44 The option that controls a given warning is now displayed:
46 foo.y:4.6: warning: type clash on default action: <foo> != <bar> [-Wother]
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].
52 For instance, where the previous version of Bison would report (and exit
55 bison: warnings being treated as errors
56 input.y:1.1: warning: stray ',' treated as white space
60 input.y:1.1: error: stray ',' treated as white space [-Werror=other]
62 *** Deprecated constructs
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.
68 *** Useless semantic types
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:
77 %printer {} <type1> <type3>
78 %destructor {} <type2> <type4>
80 nterm: term { $$ = $1; };
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
85 *** Undefined but unused symbols
87 Bison used to raise an error for undefined symbols that are not used in
88 the grammar. This is now only a warning.
91 %destructor {} symbol2
96 *** Useless destructors or printers
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.
103 %token <type1> token1
107 %printer {} token1 <type1> <type3>
108 %destructor {} token2 <type2> <type4>
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:
117 exp: exp '+' exp | '0' | '0';
119 compare the previous version of bison:
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
127 with the new behavior:
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
133 foo.y: error: 1 shift/reduce conflict [-Werror=conflicts-sr]
134 foo.y: error: 2 reduce/reduce conflicts [-Werror=conflicts-rr]
136 When %expect or %expect-rr is used, such as with bar.y:
141 exp: exp '+' exp | '0' | '0';
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
153 bar.y: error: shift/reduce conflicts: 1 found, 0 expected
154 bar.y: error: reduce/reduce conflicts: 2 found, 0 expected
156 ** Additional yylex/yyparse arguments
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
162 %lex-param {arg1_type *arg1}
163 %lex-param {arg2_type *arg2}
164 %parse-param {arg1_type *arg1}
165 %parse-param {arg2_type *arg2}
169 %param {arg1_type *arg1} {arg2_type *arg2}
171 ** Java skeleton improvements
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".
177 ** C++ skeletons improvements
179 *** The parser header is no longer mandatory (lalr1.cc, glr.cc)
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).
185 *** Locations are no longer mandatory (lalr1.cc, glr.cc)
187 Both lalr1.cc and glr.cc no longer require %location.
189 *** syntax_error exception (lalr1.cc)
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).
198 ** Variable api.token.prefix
200 The variable api.token.prefix changes the way tokens are identified in
201 the generated files. This is especially useful to avoid collisions
202 with identifiers in the target language. For instance
204 %token FILE for ERROR
205 %define api.token.prefix "TOK_"
207 start: FILE for ERROR;
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).
214 ** Renamed %define variables
216 The following variables have been renamed for consistency. Backward
217 compatibility is ensured, but upgrading is recommended.
219 lr.default-reductions -> lr.default-reduction
220 lr.keep-unreachable-states -> lr.keep-unreachable-state
221 namespace -> api.namespace
223 ** Variable parse.error
225 This variable controls the verbosity of error messages. The use of the
226 %error-verbose directive is deprecated in favor of "%define parse.error
229 ** Semantic predicates
231 The new, experimental, semantic-predicate feature allows actions of the
232 form "%?{ BOOLEAN-EXPRESSION }", which cause syntax errors (as for
233 YYERROR) if the expression evaluates to 0, and are evaluated immediately
234 in GLR parsers, rather than being deferred. The result is that they allow
235 the programmer to prune possible parses based on the values of run-time
238 ** The directive %expect-rr is now an error in non GLR mode
240 It used to be an error only if used in non GLR mode, _and_ if there are
241 reduce/reduce conflicts.
243 * Noteworthy changes in release ?.? (????-??-??) [?]
245 ** Changes in the format of error messages
247 This used to be the format of many error reports:
249 foo.y:5.10-24: result type clash on merge function 'merge': <t3> != <t2>
250 foo.y:4.13-27: previous declaration
254 foo.y:5.10-25: result type clash on merge function 'merge': <t3> != <t2>
255 foo.y:4.13-27: previous declaration
257 ** Exception safety (lalr1.cc)
259 The parse function now catches exceptions, uses the %destructors to
260 release memory (the lookahead symbol and the symbols pushed on the stack)
261 before re-throwing the exception.
263 This feature is somewhat experimental. User feedback would be
266 ** New %define variable: api.location.type (glr.cc, lalr1.cc, lalr1.java)
268 The %define variable api.location.type defines the name of the type to use
269 for locations. When defined, Bison no longer generates the position.hh
270 and location.hh files, nor does the parser will include them: the user is
271 then responsible to define her type.
273 This can be used in programs with several parsers to factor their location
274 and position files: let one of them generate them, and the others just use
277 This feature was actually introduced, but not documented, in Bison 2.5,
278 under the name "location_type" (which is maintained for backward
281 For consistency, lalr1.java's %define variables location_type and
282 position_type are deprecated in favor of api.location.type and
285 ** Graph improvements in DOT and XSLT
287 The graphical presentation of the states is more readable: their shape is
288 now rectangular, the state number is clearly displayed, and the items are
289 numbered and left-justified.
291 The reductions are now explicitly represented as transitions to other
292 diamond shaped nodes.
294 These changes are present in both --graph output and xml2dot.xsl XSLT
295 processing, with minor (documented) differences.
297 Two nodes were added to the documentation: Xml and Graphviz.
299 * Noteworthy changes in release 2.6.5 (2012-11-07) [stable]
301 We consider compiler warnings about Bison generated parsers to be bugs.
302 Rather than working around them in your own project, please consider
303 reporting them to us.
307 Warnings about uninitialized yylval and/or yylloc for push parsers with a
308 pure interface have been fixed for GCC 4.0 up to 4.8, and Clang 2.9 to
311 Other issues in the test suite have been addressed.
313 Nul characters are correctly displayed in error messages.
315 When possible, yylloc is correctly initialized before calling yylex. It
316 is no longer necessary to initialize it in the %initial-action.
318 * Noteworthy changes in release 2.6.4 (2012-10-23) [stable]
320 Bison 2.6.3's --version was incorrect. This release fixes this issue.
322 * Noteworthy changes in release 2.6.3 (2012-10-22) [stable]
326 Bugs and portability issues in the test suite have been fixed.
328 Some errors in translations have been addressed, and --help now directs
329 users to the appropriate place to report them.
331 Stray Info files shipped by accident are removed.
333 Incorrect definitions of YY_, issued by yacc.c when no parser header is
334 generated, are removed.
336 All the generated headers are self-contained.
338 ** Header guards (yacc.c, glr.c, glr.cc)
340 In order to avoid collisions, the header guards are now
341 YY_<PREFIX>_<FILE>_INCLUDED, instead of merely <PREFIX>_<FILE>.
342 For instance the header generated from
344 %define api.prefix "calc"
345 %defines "lib/parse.h"
347 will use YY_CALC_LIB_PARSE_H_INCLUDED as guard.
349 ** Fix compiler warnings in the generated parser (yacc.c, glr.c)
351 The compilation of pure parsers (%define api.pure) can trigger GCC
354 input.c: In function 'yyparse':
355 input.c:1503:12: warning: 'yylval' may be used uninitialized in this
356 function [-Wmaybe-uninitialized]
360 This is now fixed; pragmas to avoid these warnings are no longer needed.
362 Warnings from clang ("equality comparison with extraneous parentheses" and
363 "function declared 'noreturn' should not return") have also been
366 * Noteworthy changes in release 2.6.2 (2012-08-03) [stable]
370 Buffer overruns, complaints from Flex, and portability issues in the test
371 suite have been fixed.
373 ** Spaces in %lex- and %parse-param (lalr1.cc, glr.cc)
375 Trailing end-of-lines in %parse-param or %lex-param would result in
376 invalid C++. This is fixed.
378 ** Spurious spaces and end-of-lines
380 The generated files no longer end (nor start) with empty lines.
382 * Noteworthy changes in release 2.6.1 (2012-07-30) [stable]
384 Bison no longer executes user-specified M4 code when processing a grammar.
388 In addition to the removal of the features announced in Bison 2.6, the
389 next major release will remove the "Temporary hack for adding a semicolon
390 to the user action", as announced in the release 2.5. Instead of:
392 exp: exp "+" exp { $$ = $1 + $3 };
396 exp: exp "+" exp { $$ = $1 + $3; };
400 *** Type names are now properly escaped.
402 *** glr.cc: set_debug_level and debug_level work as expected.
404 *** Stray @ or $ in actions
406 While Bison used to warn about stray $ or @ in action rules, it did not
407 for other actions such as printers, destructors, or initial actions. It
410 ** Type names in actions
412 For consistency with rule actions, it is now possible to qualify $$ by a
413 type-name in destructors, printers, and initial actions. For instance:
415 %printer { fprintf (yyo, "(%d, %f)", $<ival>$, $<fval>$); } <*> <>;
417 will display two values for each typed and untyped symbol (provided
418 that YYSTYPE has both "ival" and "fval" fields).
420 * Noteworthy changes in release 2.6 (2012-07-19) [stable]
424 The next major release of Bison will drop support for the following
425 deprecated features. Please report disagreements to bug-bison@gnu.org.
429 Support for generating parsers in K&R C will be removed. Parsers
430 generated for C support ISO C90, and are tested with ISO C99 and ISO C11
433 *** Features deprecated since Bison 1.875
435 The definitions of yystype and yyltype will be removed; use YYSTYPE and
438 YYPARSE_PARAM and YYLEX_PARAM, deprecated in favor of %parse-param and
439 %lex-param, will no longer be supported.
441 Support for the preprocessor symbol YYERROR_VERBOSE will be removed, use
444 *** The generated header will be included (yacc.c)
446 Instead of duplicating the content of the generated header (definition of
447 YYSTYPE, yyparse declaration etc.), the generated parser will include it,
448 as is already the case for GLR or C++ parsers. This change is deferred
449 because existing versions of ylwrap (e.g., Automake 1.12.1) do not support
452 ** Generated Parser Headers
454 *** Guards (yacc.c, glr.c, glr.cc)
456 The generated headers are now guarded, as is already the case for C++
457 parsers (lalr1.cc). For instance, with --defines=foo.h:
462 #endif /* !YY_FOO_H */
464 *** New declarations (yacc.c, glr.c)
466 The generated header now declares yydebug and yyparse. Both honor
467 --name-prefix=bar_, and yield
469 int bar_parse (void);
473 #define yyparse bar_parse
476 in order to facilitate the inclusion of several parser headers inside a
477 single compilation unit.
479 *** Exported symbols in C++
481 The symbols YYTOKEN_TABLE and YYERROR_VERBOSE, which were defined in the
482 header, are removed, as they prevent the possibility of including several
483 generated headers from a single compilation unit.
487 For the same reasons, the undocumented and unused macro YYLSP_NEEDED is no
490 ** New %define variable: api.prefix
492 Now that the generated headers are more complete and properly protected
493 against multiple inclusions, constant names, such as YYSTYPE are a
494 problem. While yyparse and others are properly renamed by %name-prefix,
495 YYSTYPE, YYDEBUG and others have never been affected by it. Because it
496 would introduce backward compatibility issues in projects not expecting
497 YYSTYPE to be renamed, instead of changing the behavior of %name-prefix,
498 it is deprecated in favor of a new %define variable: api.prefix.
500 The following examples compares both:
502 %name-prefix "bar_" | %define api.prefix "bar_"
503 %token <ival> FOO %token <ival> FOO
504 %union { int ival; } %union { int ival; }
510 #ifndef BAR_FOO_H #ifndef BAR_FOO_H
511 # define BAR_FOO_H # define BAR_FOO_H
513 /* Enabling traces. */ /* Enabling traces. */
514 # ifndef YYDEBUG | # ifndef BAR_DEBUG
515 > # if defined YYDEBUG
517 > # define BAR_DEBUG 1
519 > # define BAR_DEBUG 0
522 # define YYDEBUG 0 | # define BAR_DEBUG 0
526 # if YYDEBUG | # if BAR_DEBUG
527 extern int bar_debug; extern int bar_debug;
530 /* Tokens. */ /* Tokens. */
531 # ifndef YYTOKENTYPE | # ifndef BAR_TOKENTYPE
532 # define YYTOKENTYPE | # define BAR_TOKENTYPE
533 enum yytokentype { | enum bar_tokentype {
538 #if ! defined YYSTYPE \ | #if ! defined BAR_STYPE \
539 && ! defined YYSTYPE_IS_DECLARED | && ! defined BAR_STYPE_IS_DECLARED
540 typedef union YYSTYPE | typedef union BAR_STYPE
543 } YYSTYPE; | } BAR_STYPE;
544 # define YYSTYPE_IS_DECLARED 1 | # define BAR_STYPE_IS_DECLARED 1
547 extern YYSTYPE bar_lval; | extern BAR_STYPE bar_lval;
549 int bar_parse (void); int bar_parse (void);
551 #endif /* !BAR_FOO_H */ #endif /* !BAR_FOO_H */
553 * Noteworthy changes in release 2.5.1 (2012-06-05) [stable]
557 The next major release will drop support for generating parsers in K&R C.
559 ** yacc.c: YYBACKUP works as expected.
561 ** glr.c improvements:
563 *** Location support is eliminated when not requested:
565 GLR parsers used to include location-related code even when locations were
566 not requested, and therefore not even usable.
568 *** __attribute__ is preserved:
570 __attribute__ is no longer disabled when __STRICT_ANSI__ is defined (i.e.,
571 when -std is passed to GCC).
573 ** lalr1.java: several fixes:
575 The Java parser no longer throws ArrayIndexOutOfBoundsException if the
576 first token leads to a syntax error. Some minor clean ups.
580 *** C++11 compatibility:
582 C and C++ parsers use "nullptr" instead of "0" when __cplusplus is 201103L
587 The header files such as "parser.hh", "location.hh", etc. used a constant
588 name for preprocessor guards, for instance:
590 #ifndef BISON_LOCATION_HH
591 # define BISON_LOCATION_HH
593 #endif // !BISON_LOCATION_HH
595 The inclusion guard is now computed from "PREFIX/FILE-NAME", where lower
596 case characters are converted to upper case, and series of
597 non-alphanumerical characters are converted to an underscore.
599 With "bison -o lang++/parser.cc", "location.hh" would now include:
601 #ifndef YY_LANG_LOCATION_HH
602 # define YY_LANG_LOCATION_HH
604 #endif // !YY_LANG_LOCATION_HH
608 The position and location constructors (and their initialize methods)
609 accept new arguments for line and column. Several issues in the
610 documentation were fixed.
612 ** liby is no longer asking for "rpl_fprintf" on some platforms.
614 ** Changes in the manual:
616 *** %printer is documented
618 The "%printer" directive, supported since at least Bison 1.50, is finally
619 documented. The "mfcalc" example is extended to demonstrate it.
621 For consistency with the C skeletons, the C++ parsers now also support
622 "yyoutput" (as an alias to "debug_stream ()").
624 *** Several improvements have been made:
626 The layout for grammar excerpts was changed to a more compact scheme.
627 Named references are motivated. The description of the automaton
628 description file (*.output) is updated to the current format. Incorrect
629 index entries were fixed. Some other errors were fixed.
633 *** Conflicting prototypes with recent/modified Flex.
635 Fixed build problems with the current, unreleased, version of Flex, and
636 some modified versions of 2.5.35, which have modified function prototypes.
638 *** Warnings during the build procedure have been eliminated.
640 *** Several portability problems in the test suite have been fixed:
642 This includes warnings with some compilers, unexpected behavior of tools
643 such as diff, warning messages from the test suite itself, etc.
645 *** The install-pdf target works properly:
647 Running "make install-pdf" (or -dvi, -html, -info, and -ps) no longer
648 halts in the middle of its course.
650 * Changes in version 2.5 (2011-05-14):
652 ** Grammar symbol names can now contain non-initial dashes:
654 Consistently with directives (such as %error-verbose) and with
655 %define variables (e.g. push-pull), grammar symbol names may contain
656 dashes in any position except the beginning. This is a GNU
657 extension over POSIX Yacc. Thus, use of this extension is reported
658 by -Wyacc and rejected in Yacc mode (--yacc).
662 Historically, Yacc and Bison have supported positional references
663 ($n, $$) to allow access to symbol values from inside of semantic
666 Starting from this version, Bison can also accept named references.
667 When no ambiguity is possible, original symbol names may be used
670 if_stmt : "if" cond_expr "then" then_stmt ';'
671 { $if_stmt = mk_if_stmt($cond_expr, $then_stmt); }
673 In the more common case, explicit names may be declared:
675 stmt[res] : "if" expr[cond] "then" stmt[then] "else" stmt[else] ';'
676 { $res = mk_if_stmt($cond, $then, $else); }
678 Location information is also accessible using @name syntax. When
679 accessing symbol names containing dots or dashes, explicit bracketing
680 ($[sym.1]) must be used.
682 These features are experimental in this version. More user feedback
683 will help to stabilize them.
685 ** IELR(1) and canonical LR(1):
687 IELR(1) is a minimal LR(1) parser table generation algorithm. That
688 is, given any context-free grammar, IELR(1) generates parser tables
689 with the full language-recognition power of canonical LR(1) but with
690 nearly the same number of parser states as LALR(1). This reduction
691 in parser states is often an order of magnitude. More importantly,
692 because canonical LR(1)'s extra parser states may contain duplicate
693 conflicts in the case of non-LR(1) grammars, the number of conflicts
694 for IELR(1) is often an order of magnitude less as well. This can
695 significantly reduce the complexity of developing of a grammar.
697 Bison can now generate IELR(1) and canonical LR(1) parser tables in
698 place of its traditional LALR(1) parser tables, which remain the
699 default. You can specify the type of parser tables in the grammar
700 file with these directives:
704 %define lr.type canonical-lr
706 The default-reduction optimization in the parser tables can also be
707 adjusted using "%define lr.default-reductions". For details on both
708 of these features, see the new section "Tuning LR" in the Bison
711 These features are experimental. More user feedback will help to
714 ** LAC (Lookahead Correction) for syntax error handling:
716 Canonical LR, IELR, and LALR can suffer from a couple of problems
717 upon encountering a syntax error. First, the parser might perform
718 additional parser stack reductions before discovering the syntax
719 error. Such reductions can perform user semantic actions that are
720 unexpected because they are based on an invalid token, and they
721 cause error recovery to begin in a different syntactic context than
722 the one in which the invalid token was encountered. Second, when
723 verbose error messages are enabled (with %error-verbose or the
724 obsolete "#define YYERROR_VERBOSE"), the expected token list in the
725 syntax error message can both contain invalid tokens and omit valid
728 The culprits for the above problems are %nonassoc, default
729 reductions in inconsistent states, and parser state merging. Thus,
730 IELR and LALR suffer the most. Canonical LR can suffer only if
731 %nonassoc is used or if default reductions are enabled for
734 LAC is a new mechanism within the parsing algorithm that solves
735 these problems for canonical LR, IELR, and LALR without sacrificing
736 %nonassoc, default reductions, or state merging. When LAC is in
737 use, canonical LR and IELR behave almost exactly the same for both
738 syntactically acceptable and syntactically unacceptable input.
739 While LALR still does not support the full language-recognition
740 power of canonical LR and IELR, LAC at least enables LALR's syntax
741 error handling to correctly reflect LALR's language-recognition
744 Currently, LAC is only supported for deterministic parsers in C.
745 You can enable LAC with the following directive:
747 %define parse.lac full
749 See the new section "LAC" in the Bison manual for additional
750 details including a few caveats.
752 LAC is an experimental feature. More user feedback will help to
755 ** %define improvements:
757 *** Can now be invoked via the command line:
759 Each of these command-line options
762 --define=NAME[=VALUE]
765 --force-define=NAME[=VALUE]
767 is equivalent to this grammar file declaration
769 %define NAME ["VALUE"]
771 except that the manner in which Bison processes multiple definitions
772 for the same NAME differs. Most importantly, -F and --force-define
773 quietly override %define, but -D and --define do not. For further
774 details, see the section "Bison Options" in the Bison manual.
776 *** Variables renamed:
778 The following %define variables
781 lr.keep_unreachable_states
786 lr.keep-unreachable-states
788 The old names are now deprecated but will be maintained indefinitely
789 for backward compatibility.
791 *** Values no longer need to be quoted in the grammar file:
793 If a %define value is an identifier, it no longer needs to be placed
794 within quotations marks. For example,
796 %define api.push-pull "push"
800 %define api.push-pull push
802 *** Unrecognized variables are now errors not warnings.
804 *** Multiple invocations for any variable is now an error not a warning.
806 ** Unrecognized %code qualifiers are now errors not warnings.
808 ** Character literals not of length one:
810 Previously, Bison quietly converted all character literals to length
811 one. For example, without warning, Bison interpreted the operators in
812 the following grammar to be the same token:
818 Bison now warns when a character literal is not of length one. In
819 some future release, Bison will start reporting an error instead.
821 ** Destructor calls fixed for lookaheads altered in semantic actions:
823 Previously for deterministic parsers in C, if a user semantic action
824 altered yychar, the parser in some cases used the old yychar value to
825 determine which destructor to call for the lookahead upon a syntax
826 error or upon parser return. This bug has been fixed.
828 ** C++ parsers use YYRHSLOC:
830 Similarly to the C parsers, the C++ parsers now define the YYRHSLOC
831 macro and use it in the default YYLLOC_DEFAULT. You are encouraged
832 to use it. If, for instance, your location structure has "first"
833 and "last" members, instead of
835 # define YYLLOC_DEFAULT(Current, Rhs, N) \
839 (Current).first = (Rhs)[1].location.first; \
840 (Current).last = (Rhs)[N].location.last; \
844 (Current).first = (Current).last = (Rhs)[0].location.last; \
850 # define YYLLOC_DEFAULT(Current, Rhs, N) \
854 (Current).first = YYRHSLOC (Rhs, 1).first; \
855 (Current).last = YYRHSLOC (Rhs, N).last; \
859 (Current).first = (Current).last = YYRHSLOC (Rhs, 0).last; \
863 ** YYLLOC_DEFAULT in C++:
865 The default implementation of YYLLOC_DEFAULT used to be issued in
866 the header file. It is now output in the implementation file, after
867 the user %code sections so that its #ifndef guard does not try to
868 override the user's YYLLOC_DEFAULT if provided.
870 ** YYFAIL now produces warnings and Java parsers no longer implement it:
872 YYFAIL has existed for many years as an undocumented feature of
873 deterministic parsers in C generated by Bison. More recently, it was
874 a documented feature of Bison's experimental Java parsers. As
875 promised in Bison 2.4.2's NEWS entry, any appearance of YYFAIL in a
876 semantic action now produces a deprecation warning, and Java parsers
877 no longer implement YYFAIL at all. For further details, including a
878 discussion of how to suppress C preprocessor warnings about YYFAIL
879 being unused, see the Bison 2.4.2 NEWS entry.
881 ** Temporary hack for adding a semicolon to the user action:
883 Previously, Bison appended a semicolon to every user action for
884 reductions when the output language defaulted to C (specifically, when
885 neither %yacc, %language, %skeleton, or equivalent command-line
886 options were specified). This allowed actions such as
888 exp: exp "+" exp { $$ = $1 + $3 };
892 exp: exp "+" exp { $$ = $1 + $3; };
894 As a first step in removing this misfeature, Bison now issues a
895 warning when it appends a semicolon. Moreover, in cases where Bison
896 cannot easily determine whether a semicolon is needed (for example, an
897 action ending with a cpp directive or a braced compound initializer),
898 it no longer appends one. Thus, the C compiler might now complain
899 about a missing semicolon where it did not before. Future releases of
900 Bison will cease to append semicolons entirely.
902 ** Verbose syntax error message fixes:
904 When %error-verbose or the obsolete "#define YYERROR_VERBOSE" is
905 specified, syntax error messages produced by the generated parser
906 include the unexpected token as well as a list of expected tokens.
907 The effect of %nonassoc on these verbose messages has been corrected
908 in two ways, but a more complete fix requires LAC, described above:
910 *** When %nonassoc is used, there can exist parser states that accept no
911 tokens, and so the parser does not always require a lookahead token
912 in order to detect a syntax error. Because no unexpected token or
913 expected tokens can then be reported, the verbose syntax error
914 message described above is suppressed, and the parser instead
915 reports the simpler message, "syntax error". Previously, this
916 suppression was sometimes erroneously triggered by %nonassoc when a
917 lookahead was actually required. Now verbose messages are
918 suppressed only when all previous lookaheads have already been
919 shifted or discarded.
921 *** Previously, the list of expected tokens erroneously included tokens
922 that would actually induce a syntax error because conflicts for them
923 were resolved with %nonassoc in the current parser state. Such
924 tokens are now properly omitted from the list.
926 *** Expected token lists are still often wrong due to state merging
927 (from LALR or IELR) and default reductions, which can both add
928 invalid tokens and subtract valid tokens. Canonical LR almost
929 completely fixes this problem by eliminating state merging and
930 default reductions. However, there is one minor problem left even
931 when using canonical LR and even after the fixes above. That is,
932 if the resolution of a conflict with %nonassoc appears in a later
933 parser state than the one at which some syntax error is
934 discovered, the conflicted token is still erroneously included in
935 the expected token list. Bison's new LAC implementation,
936 described above, eliminates this problem and the need for
937 canonical LR. However, LAC is still experimental and is disabled
940 ** Java skeleton fixes:
942 *** A location handling bug has been fixed.
944 *** The top element of each of the value stack and location stack is now
945 cleared when popped so that it can be garbage collected.
947 *** Parser traces now print the top element of the stack.
949 ** -W/--warnings fixes:
951 *** Bison now properly recognizes the "no-" versions of categories:
953 For example, given the following command line, Bison now enables all
954 warnings except warnings for incompatibilities with POSIX Yacc:
956 bison -Wall,no-yacc gram.y
958 *** Bison now treats S/R and R/R conflicts like other warnings:
960 Previously, conflict reports were independent of Bison's normal
961 warning system. Now, Bison recognizes the warning categories
962 "conflicts-sr" and "conflicts-rr". This change has important
963 consequences for the -W and --warnings command-line options. For
966 bison -Wno-conflicts-sr gram.y # S/R conflicts not reported
967 bison -Wno-conflicts-rr gram.y # R/R conflicts not reported
968 bison -Wnone gram.y # no conflicts are reported
969 bison -Werror gram.y # any conflict is an error
971 However, as before, if the %expect or %expect-rr directive is
972 specified, an unexpected number of conflicts is an error, and an
973 expected number of conflicts is not reported, so -W and --warning
974 then have no effect on the conflict report.
976 *** The "none" category no longer disables a preceding "error":
978 For example, for the following command line, Bison now reports
979 errors instead of warnings for incompatibilities with POSIX Yacc:
981 bison -Werror,none,yacc gram.y
983 *** The "none" category now disables all Bison warnings:
985 Previously, the "none" category disabled only Bison warnings for
986 which there existed a specific -W/--warning category. However,
987 given the following command line, Bison is now guaranteed to
988 suppress all warnings:
992 ** Precedence directives can now assign token number 0:
994 Since Bison 2.3b, which restored the ability of precedence
995 directives to assign token numbers, doing so for token number 0 has
996 produced an assertion failure. For example:
1000 This bug has been fixed.
1002 * Changes in version 2.4.3 (2010-08-05):
1004 ** Bison now obeys -Werror and --warnings=error for warnings about
1005 grammar rules that are useless in the parser due to conflicts.
1007 ** Problems with spawning M4 on at least FreeBSD 8 and FreeBSD 9 have
1010 ** Failures in the test suite for GCC 4.5 have been fixed.
1012 ** Failures in the test suite for some versions of Sun Studio C++ have
1015 ** Contrary to Bison 2.4.2's NEWS entry, it has been decided that
1016 warnings about undefined %prec identifiers will not be converted to
1017 errors in Bison 2.5. They will remain warnings, which should be
1018 sufficient for POSIX while avoiding backward compatibility issues.
1020 ** Minor documentation fixes.
1022 * Changes in version 2.4.2 (2010-03-20):
1024 ** Some portability problems that resulted in failures and livelocks
1025 in the test suite on some versions of at least Solaris, AIX, HP-UX,
1026 RHEL4, and Tru64 have been addressed. As a result, fatal Bison
1027 errors should no longer cause M4 to report a broken pipe on the
1030 ** "%prec IDENTIFIER" requires IDENTIFIER to be defined separately.
1032 POSIX specifies that an error be reported for any identifier that does
1033 not appear on the LHS of a grammar rule and that is not defined by
1034 %token, %left, %right, or %nonassoc. Bison 2.3b and later lost this
1035 error report for the case when an identifier appears only after a
1036 %prec directive. It is now restored. However, for backward
1037 compatibility with recent Bison releases, it is only a warning for
1038 now. In Bison 2.5 and later, it will return to being an error.
1039 [Between the 2.4.2 and 2.4.3 releases, it was decided that this
1040 warning will not be converted to an error in Bison 2.5.]
1042 ** Detection of GNU M4 1.4.6 or newer during configure is improved.
1044 ** Warnings from gcc's -Wundef option about undefined YYENABLE_NLS,
1045 YYLTYPE_IS_TRIVIAL, and __STRICT_ANSI__ in C/C++ parsers are now
1048 ** %code is now a permanent feature.
1050 A traditional Yacc prologue directive is written in the form:
1054 To provide a more flexible alternative, Bison 2.3b introduced the
1055 %code directive with the following forms for C/C++:
1058 %code requires {CODE}
1059 %code provides {CODE}
1062 These forms are now considered permanent features of Bison. See the
1063 %code entries in the section "Bison Declaration Summary" in the Bison
1064 manual for a summary of their functionality. See the section
1065 "Prologue Alternatives" for a detailed discussion including the
1066 advantages of %code over the traditional Yacc prologue directive.
1068 Bison's Java feature as a whole including its current usage of %code
1069 is still considered experimental.
1071 ** YYFAIL is deprecated and will eventually be removed.
1073 YYFAIL has existed for many years as an undocumented feature of
1074 deterministic parsers in C generated by Bison. Previously, it was
1075 documented for Bison's experimental Java parsers. YYFAIL is no longer
1076 documented for Java parsers and is formally deprecated in both cases.
1077 Users are strongly encouraged to migrate to YYERROR, which is
1080 Like YYERROR, you can invoke YYFAIL from a semantic action in order to
1081 induce a syntax error. The most obvious difference from YYERROR is
1082 that YYFAIL will automatically invoke yyerror to report the syntax
1083 error so that you don't have to. However, there are several other
1084 subtle differences between YYERROR and YYFAIL, and YYFAIL suffers from
1085 inherent flaws when %error-verbose or "#define YYERROR_VERBOSE" is
1086 used. For a more detailed discussion, see:
1088 http://lists.gnu.org/archive/html/bison-patches/2009-12/msg00024.html
1090 The upcoming Bison 2.5 will remove YYFAIL from Java parsers, but
1091 deterministic parsers in C will continue to implement it. However,
1092 because YYFAIL is already flawed, it seems futile to try to make new
1093 Bison features compatible with it. Thus, during parser generation,
1094 Bison 2.5 will produce a warning whenever it discovers YYFAIL in a
1095 rule action. In a later release, YYFAIL will be disabled for
1096 %error-verbose and "#define YYERROR_VERBOSE". Eventually, YYFAIL will
1097 be removed altogether.
1099 There exists at least one case where Bison 2.5's YYFAIL warning will
1100 be a false positive. Some projects add phony uses of YYFAIL and other
1101 Bison-defined macros for the sole purpose of suppressing C
1102 preprocessor warnings (from GCC cpp's -Wunused-macros, for example).
1103 To avoid Bison's future warning, such YYFAIL uses can be moved to the
1104 epilogue (that is, after the second "%%") in the Bison input file. In
1105 this release (2.4.2), Bison already generates its own code to suppress
1106 C preprocessor warnings for YYFAIL, so projects can remove their own
1107 phony uses of YYFAIL if compatibility with Bison releases prior to
1108 2.4.2 is not necessary.
1110 ** Internationalization.
1112 Fix a regression introduced in Bison 2.4: Under some circumstances,
1113 message translations were not installed although supported by the
1116 * Changes in version 2.4.1 (2008-12-11):
1118 ** In the GLR defines file, unexpanded M4 macros in the yylval and yylloc
1119 declarations have been fixed.
1121 ** Temporary hack for adding a semicolon to the user action.
1123 Bison used to prepend a trailing semicolon at the end of the user
1124 action for reductions. This allowed actions such as
1126 exp: exp "+" exp { $$ = $1 + $3 };
1130 exp: exp "+" exp { $$ = $1 + $3; };
1132 Some grammars still depend on this "feature". Bison 2.4.1 restores
1133 the previous behavior in the case of C output (specifically, when
1134 neither %language or %skeleton or equivalent command-line options
1135 are used) to leave more time for grammars depending on the old
1136 behavior to be adjusted. Future releases of Bison will disable this
1139 ** A few minor improvements to the Bison manual.
1141 * Changes in version 2.4 (2008-11-02):
1143 ** %language is an experimental feature.
1145 We first introduced this feature in test release 2.3b as a cleaner
1146 alternative to %skeleton. Since then, we have discussed the possibility of
1147 modifying its effect on Bison's output file names. Thus, in this release,
1148 we consider %language to be an experimental feature that will likely evolve
1151 ** Forward compatibility with GNU M4 has been improved.
1153 ** Several bugs in the C++ skeleton and the experimental Java skeleton have been
1156 * Changes in version 2.3b (2008-05-27):
1158 ** The quotes around NAME that used to be required in the following directive
1161 %define NAME "VALUE"
1163 ** The directive "%pure-parser" is now deprecated in favor of:
1167 which has the same effect except that Bison is more careful to warn about
1168 unreasonable usage in the latter case.
1172 Bison can now generate an LALR(1) parser in C with a push interface. That
1173 is, instead of invoking "yyparse", which pulls tokens from "yylex", you can
1174 push one token at a time to the parser using "yypush_parse", which will
1175 return to the caller after processing each token. By default, the push
1176 interface is disabled. Either of the following directives will enable it:
1178 %define api.push_pull "push" // Just push; does not require yylex.
1179 %define api.push_pull "both" // Push and pull; requires yylex.
1181 See the new section "A Push Parser" in the Bison manual for details.
1183 The current push parsing interface is experimental and may evolve. More user
1184 feedback will help to stabilize it.
1186 ** The -g and --graph options now output graphs in Graphviz DOT format,
1187 not VCG format. Like --graph, -g now also takes an optional FILE argument
1188 and thus cannot be bundled with other short options.
1192 Bison can now generate an LALR(1) parser in Java. The skeleton is
1193 "data/lalr1.java". Consider using the new %language directive instead of
1194 %skeleton to select it.
1196 See the new section "Java Parsers" in the Bison manual for details.
1198 The current Java interface is experimental and may evolve. More user
1199 feedback will help to stabilize it.
1203 This new directive specifies the programming language of the generated
1204 parser, which can be C (the default), C++, or Java. Besides the skeleton
1205 that Bison uses, the directive affects the names of the generated files if
1206 the grammar file's name ends in ".y".
1208 ** XML Automaton Report
1210 Bison can now generate an XML report of the LALR(1) automaton using the new
1211 "--xml" option. The current XML schema is experimental and may evolve. More
1212 user feedback will help to stabilize it.
1214 ** The grammar file may now specify the name of the parser header file using
1215 %defines. For example:
1219 ** When reporting useless rules, useless nonterminals, and unused terminals,
1220 Bison now employs the terms "useless in grammar" instead of "useless",
1221 "useless in parser" instead of "never reduced", and "unused in grammar"
1222 instead of "unused".
1224 ** Unreachable State Removal
1226 Previously, Bison sometimes generated parser tables containing unreachable
1227 states. A state can become unreachable during conflict resolution if Bison
1228 disables a shift action leading to it from a predecessor state. Bison now:
1230 1. Removes unreachable states.
1232 2. Does not report any conflicts that appeared in unreachable states.
1233 WARNING: As a result, you may need to update %expect and %expect-rr
1234 directives in existing grammar files.
1236 3. For any rule used only in such states, Bison now reports the rule as
1237 "useless in parser due to conflicts".
1239 This feature can be disabled with the following directive:
1241 %define lr.keep_unreachable_states
1243 See the %define entry in the "Bison Declaration Summary" in the Bison manual
1244 for further discussion.
1246 ** Lookahead Set Correction in the ".output" Report
1248 When instructed to generate a ".output" file including lookahead sets
1249 (using "--report=lookahead", for example), Bison now prints each reduction's
1250 lookahead set only next to the associated state's one item that (1) is
1251 associated with the same rule as the reduction and (2) has its dot at the end
1252 of its RHS. Previously, Bison also erroneously printed the lookahead set
1253 next to all of the state's other items associated with the same rule. This
1254 bug affected only the ".output" file and not the generated parser source
1257 ** --report-file=FILE is a new option to override the default ".output" file
1260 ** The "=" that used to be required in the following directives is now
1263 %file-prefix "parser"
1267 ** An Alternative to "%{...%}" -- "%code QUALIFIER {CODE}"
1269 Bison 2.3a provided a new set of directives as a more flexible alternative to
1270 the traditional Yacc prologue blocks. Those have now been consolidated into
1271 a single %code directive with an optional qualifier field, which identifies
1272 the purpose of the code and thus the location(s) where Bison should generate
1275 1. "%code {CODE}" replaces "%after-header {CODE}"
1276 2. "%code requires {CODE}" replaces "%start-header {CODE}"
1277 3. "%code provides {CODE}" replaces "%end-header {CODE}"
1278 4. "%code top {CODE}" replaces "%before-header {CODE}"
1280 See the %code entries in section "Bison Declaration Summary" in the Bison
1281 manual for a summary of the new functionality. See the new section "Prologue
1282 Alternatives" for a detailed discussion including the advantages of %code
1283 over the traditional Yacc prologues.
1285 The prologue alternatives are experimental. More user feedback will help to
1286 determine whether they should become permanent features.
1288 ** Revised warning: unset or unused mid-rule values
1290 Since Bison 2.2, Bison has warned about mid-rule values that are set but not
1291 used within any of the actions of the parent rule. For example, Bison warns
1294 exp: '1' { $$ = 1; } '+' exp { $$ = $1 + $4; };
1296 Now, Bison also warns about mid-rule values that are used but not set. For
1297 example, Bison warns about unset $$ in the mid-rule action in:
1299 exp: '1' { $1 = 1; } '+' exp { $$ = $2 + $4; };
1301 However, Bison now disables both of these warnings by default since they
1302 sometimes prove to be false alarms in existing grammars employing the Yacc
1303 constructs $0 or $-N (where N is some positive integer).
1305 To enable these warnings, specify the option "--warnings=midrule-values" or
1306 "-W", which is a synonym for "--warnings=all".
1308 ** Default %destructor or %printer with "<*>" or "<>"
1310 Bison now recognizes two separate kinds of default %destructor's and
1313 1. Place "<*>" in a %destructor/%printer symbol list to define a default
1314 %destructor/%printer for all grammar symbols for which you have formally
1315 declared semantic type tags.
1317 2. Place "<>" in a %destructor/%printer symbol list to define a default
1318 %destructor/%printer for all grammar symbols without declared semantic
1321 Bison no longer supports the "%symbol-default" notation from Bison 2.3a.
1322 "<*>" and "<>" combined achieve the same effect with one exception: Bison no
1323 longer applies any %destructor to a mid-rule value if that mid-rule value is
1324 not actually ever referenced using either $$ or $n in a semantic action.
1326 The default %destructor's and %printer's are experimental. More user
1327 feedback will help to determine whether they should become permanent
1330 See the section "Freeing Discarded Symbols" in the Bison manual for further
1333 ** %left, %right, and %nonassoc can now declare token numbers. This is required
1334 by POSIX. However, see the end of section "Operator Precedence" in the Bison
1335 manual for a caveat concerning the treatment of literal strings.
1337 ** The nonfunctional --no-parser, -n, and %no-parser options have been
1338 completely removed from Bison.
1340 * Changes in version 2.3a, 2006-09-13:
1342 ** Instead of %union, you can define and use your own union type
1343 YYSTYPE if your grammar contains at least one <type> tag.
1344 Your YYSTYPE need not be a macro; it can be a typedef.
1345 This change is for compatibility with other Yacc implementations,
1346 and is required by POSIX.
1348 ** Locations columns and lines start at 1.
1349 In accordance with the GNU Coding Standards and Emacs.
1351 ** You may now declare per-type and default %destructor's and %printer's:
1355 %union { char *string; }
1356 %token <string> STRING1
1357 %token <string> STRING2
1358 %type <string> string1
1359 %type <string> string2
1360 %union { char character; }
1361 %token <character> CHR
1362 %type <character> chr
1363 %destructor { free ($$); } %symbol-default
1364 %destructor { free ($$); printf ("%d", @$.first_line); } STRING1 string1
1365 %destructor { } <character>
1367 guarantees that, when the parser discards any user-defined symbol that has a
1368 semantic type tag other than "<character>", it passes its semantic value to
1369 "free". However, when the parser discards a "STRING1" or a "string1", it
1370 also prints its line number to "stdout". It performs only the second
1371 "%destructor" in this case, so it invokes "free" only once.
1373 [Although we failed to mention this here in the 2.3a release, the default
1374 %destructor's and %printer's were experimental, and they were rewritten in
1377 ** Except for LALR(1) parsers in C with POSIX Yacc emulation enabled (with "-y",
1378 "--yacc", or "%yacc"), Bison no longer generates #define statements for
1379 associating token numbers with token names. Removing the #define statements
1380 helps to sanitize the global namespace during preprocessing, but POSIX Yacc
1381 requires them. Bison still generates an enum for token names in all cases.
1383 ** Handling of traditional Yacc prologue blocks is now more consistent but
1384 potentially incompatible with previous releases of Bison.
1386 As before, you declare prologue blocks in your grammar file with the
1387 "%{ ... %}" syntax. To generate the pre-prologue, Bison concatenates all
1388 prologue blocks that you've declared before the first %union. To generate
1389 the post-prologue, Bison concatenates all prologue blocks that you've
1390 declared after the first %union.
1392 Previous releases of Bison inserted the pre-prologue into both the header
1393 file and the code file in all cases except for LALR(1) parsers in C. In the
1394 latter case, Bison inserted it only into the code file. For parsers in C++,
1395 the point of insertion was before any token definitions (which associate
1396 token numbers with names). For parsers in C, the point of insertion was
1397 after the token definitions.
1399 Now, Bison never inserts the pre-prologue into the header file. In the code
1400 file, it always inserts it before the token definitions.
1402 ** Bison now provides a more flexible alternative to the traditional Yacc
1403 prologue blocks: %before-header, %start-header, %end-header, and
1406 For example, the following declaration order in the grammar file reflects the
1407 order in which Bison will output these code blocks. However, you are free to
1408 declare these code blocks in your grammar file in whatever order is most
1412 /* Bison treats this block like a pre-prologue block: it inserts it into
1413 * the code file before the contents of the header file. It does *not*
1414 * insert it into the header file. This is a good place to put
1415 * #include's that you want at the top of your code file. A common
1416 * example is '#include "system.h"'. */
1419 /* Bison inserts this block into both the header file and the code file.
1420 * In both files, the point of insertion is before any Bison-generated
1421 * token, semantic type, location type, and class definitions. This is a
1422 * good place to define %union dependencies, for example. */
1425 /* Unlike the traditional Yacc prologue blocks, the output order for the
1426 * new %*-header blocks is not affected by their declaration position
1427 * relative to any %union in the grammar file. */
1430 /* Bison inserts this block into both the header file and the code file.
1431 * In both files, the point of insertion is after the Bison-generated
1432 * definitions. This is a good place to declare or define public
1433 * functions or data structures that depend on the Bison-generated
1437 /* Bison treats this block like a post-prologue block: it inserts it into
1438 * the code file after the contents of the header file. It does *not*
1439 * insert it into the header file. This is a good place to declare or
1440 * define internal functions or data structures that depend on the
1441 * Bison-generated definitions. */
1444 If you have multiple occurrences of any one of the above declarations, Bison
1445 will concatenate the contents in declaration order.
1447 [Although we failed to mention this here in the 2.3a release, the prologue
1448 alternatives were experimental, and they were rewritten in future versions.]
1450 ** The option "--report=look-ahead" has been changed to "--report=lookahead".
1451 The old spelling still works, but is not documented and may be removed
1452 in a future release.
1454 * Changes in version 2.3, 2006-06-05:
1456 ** GLR grammars should now use "YYRECOVERING ()" instead of "YYRECOVERING",
1457 for compatibility with LALR(1) grammars.
1459 ** It is now documented that any definition of YYSTYPE or YYLTYPE should
1460 be to a type name that does not contain parentheses or brackets.
1462 * Changes in version 2.2, 2006-05-19:
1464 ** The distribution terms for all Bison-generated parsers now permit
1465 using the parsers in nonfree programs. Previously, this permission
1466 was granted only for Bison-generated LALR(1) parsers in C.
1468 ** %name-prefix changes the namespace name in C++ outputs.
1470 ** The C++ parsers export their token_type.
1472 ** Bison now allows multiple %union declarations, and concatenates
1473 their contents together.
1475 ** New warning: unused values
1476 Right-hand side symbols whose values are not used are reported,
1477 if the symbols have destructors. For instance:
1479 exp: exp "?" exp ":" exp { $1 ? $1 : $3; }
1483 will trigger a warning about $$ and $5 in the first rule, and $3 in
1484 the second ($1 is copied to $$ by the default rule). This example
1485 most likely contains three errors, and could be rewritten as:
1487 exp: exp "?" exp ":" exp
1488 { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); }
1490 { $$ = $1 ? $1 : $3; if ($1) free ($3); }
1493 However, if the original actions were really intended, memory leaks
1494 and all, the warnings can be suppressed by letting Bison believe the
1495 values are used, e.g.:
1497 exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); }
1498 | exp "+" exp { $$ = $1; (void) $3; }
1501 If there are mid-rule actions, the warning is issued if no action
1502 uses it. The following triggers no warning: $1 and $3 are used.
1504 exp: exp { push ($1); } '+' exp { push ($3); sum (); };
1506 The warning is intended to help catching lost values and memory leaks.
1507 If a value is ignored, its associated memory typically is not reclaimed.
1509 ** %destructor vs. YYABORT, YYACCEPT, and YYERROR.
1510 Destructors are now called when user code invokes YYABORT, YYACCEPT,
1511 and YYERROR, for all objects on the stack, other than objects
1512 corresponding to the right-hand side of the current rule.
1514 ** %expect, %expect-rr
1515 Incorrect numbers of expected conflicts are now actual errors,
1516 instead of warnings.
1518 ** GLR, YACC parsers.
1519 The %parse-params are available in the destructors (and the
1520 experimental printers) as per the documentation.
1522 ** Bison now warns if it finds a stray "$" or "@" in an action.
1524 ** %require "VERSION"
1525 This specifies that the grammar file depends on features implemented
1526 in Bison version VERSION or higher.
1528 ** lalr1.cc: The token and value types are now class members.
1529 The tokens were defined as free form enums and cpp macros. YYSTYPE
1530 was defined as a free form union. They are now class members:
1531 tokens are enumerations of the "yy::parser::token" struct, and the
1532 semantic values have the "yy::parser::semantic_type" type.
1534 If you do not want or can update to this scheme, the directive
1535 '%define "global_tokens_and_yystype" "1"' triggers the global
1536 definition of tokens and YYSTYPE. This change is suitable both
1537 for previous releases of Bison, and this one.
1539 If you wish to update, then make sure older version of Bison will
1540 fail using '%require "2.2"'.
1542 ** DJGPP support added.
1544 * Changes in version 2.1, 2005-09-16:
1546 ** The C++ lalr1.cc skeleton supports %lex-param.
1548 ** Bison-generated parsers now support the translation of diagnostics like
1549 "syntax error" into languages other than English. The default
1550 language is still English. For details, please see the new
1551 Internationalization section of the Bison manual. Software
1552 distributors should also see the new PACKAGING file. Thanks to
1553 Bruno Haible for this new feature.
1555 ** Wording in the Bison-generated parsers has been changed slightly to
1556 simplify translation. In particular, the message "memory exhausted"
1557 has replaced "parser stack overflow", as the old message was not
1558 always accurate for modern Bison-generated parsers.
1560 ** Destructors are now called when the parser aborts, for all symbols left
1561 behind on the stack. Also, the start symbol is now destroyed after a
1562 successful parse. In both cases, the behavior was formerly inconsistent.
1564 ** When generating verbose diagnostics, Bison-generated parsers no longer
1565 quote the literal strings associated with tokens. For example, for
1566 a syntax error associated with '%token NUM "number"' they might
1567 print 'syntax error, unexpected number' instead of 'syntax error,
1568 unexpected "number"'.
1570 * Changes in version 2.0, 2004-12-25:
1572 ** Possibly-incompatible changes
1574 - Bison-generated parsers no longer default to using the alloca function
1575 (when available) to extend the parser stack, due to widespread
1576 problems in unchecked stack-overflow detection. You can "#define
1577 YYSTACK_USE_ALLOCA 1" to require the use of alloca, but please read
1578 the manual to determine safe values for YYMAXDEPTH in that case.
1580 - Error token location.
1581 During error recovery, the location of the syntax error is updated
1582 to cover the whole sequence covered by the error token: it includes
1583 the shifted symbols thrown away during the first part of the error
1584 recovery, and the lookahead rejected during the second part.
1586 - Semicolon changes:
1587 . Stray semicolons are no longer allowed at the start of a grammar.
1588 . Semicolons are now required after in-grammar declarations.
1590 - Unescaped newlines are no longer allowed in character constants or
1591 string literals. They were never portable, and GCC 3.4.0 has
1592 dropped support for them. Better diagnostics are now generated if
1593 forget a closing quote.
1595 - NUL bytes are no longer allowed in Bison string literals, unfortunately.
1599 - GLR grammars now support locations.
1601 - New directive: %initial-action.
1602 This directive allows the user to run arbitrary code (including
1603 initializing @$) from yyparse before parsing starts.
1605 - A new directive "%expect-rr N" specifies the expected number of
1606 reduce/reduce conflicts in GLR parsers.
1608 - %token numbers can now be hexadecimal integers, e.g., "%token FOO 0x12d".
1609 This is a GNU extension.
1611 - The option "--report=lookahead" was changed to "--report=look-ahead".
1612 [However, this was changed back after 2.3.]
1614 - Experimental %destructor support has been added to lalr1.cc.
1616 - New configure option --disable-yacc, to disable installation of the
1617 yacc command and -ly library introduced in 1.875 for POSIX conformance.
1621 - For now, %expect-count violations are now just warnings, not errors.
1622 This is for compatibility with Bison 1.75 and earlier (when there are
1623 reduce/reduce conflicts) and with Bison 1.30 and earlier (when there
1624 are too many or too few shift/reduce conflicts). However, in future
1625 versions of Bison we plan to improve the %expect machinery so that
1626 these violations will become errors again.
1628 - Within Bison itself, numbers (e.g., goto numbers) are no longer
1629 arbitrarily limited to 16-bit counts.
1631 - Semicolons are now allowed before "|" in grammar rules, as POSIX requires.
1633 * Changes in version 1.875, 2003-01-01:
1635 ** The documentation license has been upgraded to version 1.2
1636 of the GNU Free Documentation License.
1638 ** syntax error processing
1640 - In Yacc-style parsers YYLLOC_DEFAULT is now used to compute error
1641 locations too. This fixes bugs in error-location computation.
1644 It is now possible to reclaim the memory associated to symbols
1645 discarded during error recovery. This feature is still experimental.
1648 This new directive is preferred over YYERROR_VERBOSE.
1650 - #defining yyerror to steal internal variables is discouraged.
1651 It is not guaranteed to work forever.
1653 ** POSIX conformance
1655 - Semicolons are once again optional at the end of grammar rules.
1656 This reverts to the behavior of Bison 1.33 and earlier, and improves
1657 compatibility with Yacc.
1659 - "parse error" -> "syntax error"
1660 Bison now uniformly uses the term "syntax error"; formerly, the code
1661 and manual sometimes used the term "parse error" instead. POSIX
1662 requires "syntax error" in diagnostics, and it was thought better to
1665 - The documentation now emphasizes that yylex and yyerror must be
1666 declared before use. C99 requires this.
1668 - Bison now parses C99 lexical constructs like UCNs and
1669 backslash-newline within C escape sequences, as POSIX 1003.1-2001 requires.
1671 - File names are properly escaped in C output. E.g., foo\bar.y is
1672 output as "foo\\bar.y".
1674 - Yacc command and library now available
1675 The Bison distribution now installs a "yacc" command, as POSIX requires.
1676 Also, Bison now installs a small library liby.a containing
1677 implementations of Yacc-compatible yyerror and main functions.
1678 This library is normally not useful, but POSIX requires it.
1680 - Type clashes now generate warnings, not errors.
1682 - If the user does not define YYSTYPE as a macro, Bison now declares it
1683 using typedef instead of defining it as a macro.
1684 For consistency, YYLTYPE is also declared instead of defined.
1686 ** Other compatibility issues
1688 - %union directives can now have a tag before the "{", e.g., the
1689 directive "%union foo {...}" now generates the C code
1690 "typedef union foo { ... } YYSTYPE;"; this is for Yacc compatibility.
1691 The default union tag is "YYSTYPE", for compatibility with Solaris 9 Yacc.
1692 For consistency, YYLTYPE's struct tag is now "YYLTYPE" not "yyltype".
1693 This is for compatibility with both Yacc and Bison 1.35.
1695 - ";" is output before the terminating "}" of an action, for
1696 compatibility with Bison 1.35.
1698 - Bison now uses a Yacc-style format for conflict reports, e.g.,
1699 "conflicts: 2 shift/reduce, 1 reduce/reduce".
1701 - "yystype" and "yyltype" are now obsolescent macros instead of being
1702 typedefs or tags; they are no longer documented and are planned to be
1703 withdrawn in a future release.
1708 Users of Bison have to decide how they handle the portability of the
1711 - "parsing stack overflow..." -> "parser stack overflow"
1712 GLR parsers now report "parser stack overflow" as per the Bison manual.
1714 ** %parse-param and %lex-param
1715 The macros YYPARSE_PARAM and YYLEX_PARAM provide a means to pass
1716 additional context to yyparse and yylex. They suffer from several
1719 - a single argument only can be added,
1720 - their types are weak (void *),
1721 - this context is not passed to ancillary functions such as yyerror,
1722 - only yacc.c parsers support them.
1724 The new %parse-param/%lex-param directives provide a more precise control.
1727 %parse-param {int *nastiness}
1728 %lex-param {int *nastiness}
1729 %parse-param {int *randomness}
1731 results in the following signatures:
1733 int yylex (int *nastiness);
1734 int yyparse (int *nastiness, int *randomness);
1736 or, if both %pure-parser and %locations are used:
1738 int yylex (YYSTYPE *lvalp, YYLTYPE *llocp, int *nastiness);
1739 int yyparse (int *nastiness, int *randomness);
1741 ** Bison now warns if it detects conflicting outputs to the same file,
1742 e.g., it generates a warning for "bison -d -o foo.h foo.y" since
1743 that command outputs both code and header to foo.h.
1745 ** #line in output files
1746 - --no-line works properly.
1748 ** Bison can no longer be built by a K&R C compiler; it requires C89 or
1749 later to be built. This change originally took place a few versions
1750 ago, but nobody noticed until we recently asked someone to try
1751 building Bison with a K&R C compiler.
1753 * Changes in version 1.75, 2002-10-14:
1755 ** Bison should now work on 64-bit hosts.
1757 ** Indonesian translation thanks to Tedi Heriyanto.
1760 Fix spurious parse errors.
1763 Some people redefine yyerror to steal yyparse' private variables.
1764 Reenable this trick until an official feature replaces it.
1767 In agreement with POSIX and with other Yaccs, leaving a default
1768 action is valid when $$ is untyped, and $1 typed:
1772 but the converse remains an error:
1776 ** Values of mid-rule actions
1779 foo: { ... } { $$ = $1; } ...
1781 was incorrectly rejected: $1 is defined in the second mid-rule
1782 action, and is equal to the $$ of the first mid-rule action.
1784 * Changes in version 1.50, 2002-10-04:
1789 causes Bison to produce a Generalized LR (GLR) parser, capable of handling
1790 almost any context-free grammar, ambiguous or not. The new declarations
1791 %dprec and %merge on grammar rules allow parse-time resolution of
1792 ambiguities. Contributed by Paul Hilfinger.
1794 Unfortunately Bison 1.50 does not work properly on 64-bit hosts
1795 like the Alpha, so please stick to 32-bit hosts for now.
1798 When not in Yacc compatibility mode, when the output file was not
1799 specified, running "bison foo/bar.y" created "foo/bar.c". It
1800 now creates "bar.c".
1803 The undefined token was systematically mapped to 2 which prevented
1804 the use of 2 by the user. This is no longer the case.
1806 ** Unknown token numbers
1807 If yylex returned an out of range value, yyparse could die. This is
1811 According to POSIX, the error token must be 256.
1812 Bison extends this requirement by making it a preference: *if* the
1813 user specified that one of her tokens is numbered 256, then error
1814 will be mapped onto another number.
1816 ** Verbose error messages
1817 They no longer report "..., expecting error or..." for states where
1818 error recovery is possible.
1821 Defaults to "$end" instead of "$".
1823 ** Error recovery now conforms to documentation and to POSIX
1824 When a Bison-generated parser encounters a syntax error, it now pops
1825 the stack until it finds a state that allows shifting the error
1826 token. Formerly, it popped the stack until it found a state that
1827 allowed some non-error action other than a default reduction on the
1828 error token. The new behavior has long been the documented behavior,
1829 and has long been required by POSIX. For more details, please see
1830 Paul Eggert, "Reductions during Bison error handling" (2002-05-20)
1831 <http://lists.gnu.org/archive/html/bug-bison/2002-05/msg00038.html>.
1834 Popped tokens and nonterminals are now reported.
1837 Larger grammars are now supported (larger token numbers, larger grammar
1838 size (= sum of the LHS and RHS lengths), larger LALR tables).
1839 Formerly, many of these numbers ran afoul of 16-bit limits;
1840 now these limits are 32 bits on most hosts.
1842 ** Explicit initial rule
1843 Bison used to play hacks with the initial rule, which the user does
1844 not write. It is now explicit, and visible in the reports and
1848 Before, Bison reported the useless rules, but, although not used,
1849 included them in the parsers. They are now actually removed.
1851 ** Useless rules, useless nonterminals
1852 They are now reported, as a warning, with their locations.
1854 ** Rules never reduced
1855 Rules that can never be reduced because of conflicts are now
1858 ** Incorrect "Token not used"
1859 On a grammar such as
1861 %token useless useful
1863 exp: '0' %prec useful;
1865 where a token was used to set the precedence of the last rule,
1866 bison reported both "useful" and "useless" as useless tokens.
1868 ** Revert the C++ namespace changes introduced in 1.31
1869 as they caused too many portability hassles.
1871 ** Default locations
1872 By an accident of design, the default computation of @$ was
1873 performed after another default computation was performed: @$ = @1.
1874 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
1875 the computation of @$.
1877 ** Token end-of-file
1878 The token end of file may be specified by the user, in which case,
1879 the user symbol is used in the reports, the graphs, and the verbose
1880 error messages instead of "$end", which remains being the default.
1884 %token MYEOF 0 "end of file"
1887 This old option, which has been broken for ages, is removed.
1890 Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
1891 Croatian, thanks to Denis Lackovic.
1893 ** Incorrect token definitions
1896 bison used to output
1899 ** Token definitions as enums
1900 Tokens are output both as the traditional #define's, and, provided
1901 the compiler supports ANSI C or is a C++ compiler, as enums.
1902 This lets debuggers display names instead of integers.
1905 In addition to --verbose, bison supports --report=THINGS, which
1906 produces additional information:
1908 complete the core item sets with their closure
1909 - lookahead [changed to "look-ahead" in 1.875e through 2.3, but changed back]
1910 explicitly associate lookahead tokens to items
1912 describe shift/reduce conflicts solving.
1913 Bison used to systematically output this information on top of
1914 the report. Solved conflicts are now attached to their states.
1917 Previous versions don't complain when there is a type clash on
1918 the default action if the rule has a mid-rule action, such as in:
1926 ** GNU M4 is now required when using Bison.
1928 * Changes in version 1.35, 2002-03-25:
1931 Some projects use Bison's C parser with C++ compilers, and define
1932 YYSTYPE as a class. The recent adjustment of C parsers for data
1933 alignment and 64 bit architectures made this impossible.
1935 Because for the time being no real solution for C++ parser
1936 generation exists, kludges were implemented in the parser to
1937 maintain this use. In the future, when Bison has C++ parsers, this
1938 kludge will be disabled.
1940 This kludge also addresses some C++ problems when the stack was
1943 * Changes in version 1.34, 2002-03-12:
1945 ** File name clashes are detected
1946 $ bison foo.y -d -o foo.x
1947 fatal error: header and parser would both be named "foo.x"
1949 ** A missing ";" at the end of a rule triggers a warning
1950 In accordance with POSIX, and in agreement with other
1951 Yacc implementations, Bison will mandate this semicolon in the near
1952 future. This eases the implementation of a Bison parser of Bison
1953 grammars by making this grammar LALR(1) instead of LR(2). To
1954 facilitate the transition, this release introduces a warning.
1956 ** Revert the C++ namespace changes introduced in 1.31, as they caused too
1957 many portability hassles.
1959 ** DJGPP support added.
1961 ** Fix test suite portability problems.
1963 * Changes in version 1.33, 2002-02-07:
1966 Groff could not be compiled for the definition of size_t was lacking
1967 under some conditions.
1972 * Changes in version 1.32, 2002-01-23:
1974 ** Fix Yacc output file names
1976 ** Portability fixes
1978 ** Italian, Dutch translations
1980 * Changes in version 1.31, 2002-01-14:
1984 ** GNU Gettext and %expect
1985 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
1986 Bison dies on incorrect %expectations, we fear there will be
1987 too many bug reports for Gettext, so _for the time being_, %expect
1988 does not trigger an error when the input file is named "plural.y".
1990 ** Use of alloca in parsers
1991 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
1992 malloc exclusively. Since 1.29, but was not NEWS'ed.
1994 alloca is used only when compiled with GCC, to avoid portability
1997 ** yyparse now returns 2 if memory is exhausted; formerly it dumped core.
1999 ** When the generated parser lacks debugging code, YYDEBUG is now 0
2000 (as POSIX requires) instead of being undefined.
2003 Bison has always permitted actions such as { $$ = $1 }: it adds the
2004 ending semicolon. Now if in Yacc compatibility mode, the semicolon
2005 is no longer output: one has to write { $$ = $1; }.
2007 ** Better C++ compliance
2008 The output parsers try to respect C++ namespaces.
2009 [This turned out to be a failed experiment, and it was reverted later.]
2012 Fixed bugs when reporting useless nonterminals.
2015 The parsers work properly on 64 bit hosts.
2018 Some calls to strerror resulted in scrambled or missing error messages.
2021 When the number of shift/reduce conflicts is correct, don't issue
2024 ** The verbose report includes the rule line numbers.
2026 ** Rule line numbers are fixed in traces.
2028 ** Swedish translation
2031 Verbose parse error messages from the parsers are better looking.
2032 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
2033 Now: parse error: unexpected '/', expecting "number" or '-' or '('
2035 ** Fixed parser memory leaks.
2036 When the generated parser was using malloc to extend its stacks, the
2037 previous allocations were not freed.
2039 ** Fixed verbose output file.
2040 Some newlines were missing.
2041 Some conflicts in state descriptions were missing.
2043 ** Fixed conflict report.
2044 Option -v was needed to get the result.
2048 Mismatches are errors, not warnings.
2050 ** Fixed incorrect processing of some invalid input.
2052 ** Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
2054 ** Fixed some typos in the documentation.
2056 ** %token MY_EOF 0 is supported.
2057 Before, MY_EOF was silently renumbered as 257.
2059 ** doc/refcard.tex is updated.
2061 ** %output, %file-prefix, %name-prefix.
2065 New, aliasing "--output-file".
2067 * Changes in version 1.30, 2001-10-26:
2069 ** "--defines" and "--graph" have now an optional argument which is the
2070 output file name. "-d" and "-g" do not change; they do not take any
2073 ** "%source_extension" and "%header_extension" are removed, failed
2076 ** Portability fixes.
2078 * Changes in version 1.29, 2001-09-07:
2080 ** The output file does not define const, as this caused problems when used
2081 with common autoconfiguration schemes. If you still use ancient compilers
2082 that lack const, compile with the equivalent of the C compiler option
2083 "-Dconst=". Autoconf's AC_C_CONST macro provides one way to do this.
2085 ** Added "-g" and "--graph".
2087 ** The Bison manual is now distributed under the terms of the GNU FDL.
2089 ** The input and the output files has automatically a similar extension.
2091 ** Russian translation added.
2093 ** NLS support updated; should hopefully be less troublesome.
2095 ** Added the old Bison reference card.
2097 ** Added "--locations" and "%locations".
2099 ** Added "-S" and "--skeleton".
2101 ** "%raw", "-r", "--raw" is disabled.
2103 ** Special characters are escaped when output. This solves the problems
2104 of the #line lines with path names including backslashes.
2107 "%yacc", "%fixed_output_files", "%defines", "%no_parser", "%verbose",
2108 "%debug", "%source_extension" and "%header_extension".
2111 Automatic location tracking.
2113 * Changes in version 1.28, 1999-07-06:
2115 ** Should compile better now with K&R compilers.
2119 ** Fixed a problem with escaping the double quote character.
2121 ** There is now a FAQ.
2123 * Changes in version 1.27:
2125 ** The make rule which prevented bison.simple from being created on
2126 some systems has been fixed.
2128 * Changes in version 1.26:
2130 ** Bison now uses Automake.
2132 ** New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
2134 ** Token numbers now start at 257 as previously documented, not 258.
2136 ** Bison honors the TMPDIR environment variable.
2138 ** A couple of buffer overruns have been fixed.
2140 ** Problems when closing files should now be reported.
2142 ** Generated parsers should now work even on operating systems which do
2143 not provide alloca().
2145 * Changes in version 1.25, 1995-10-16:
2147 ** Errors in the input grammar are not fatal; Bison keeps reading
2148 the grammar file, and reports all the errors found in it.
2150 ** Tokens can now be specified as multiple-character strings: for
2151 example, you could use "<=" for a token which looks like <=, instead
2152 of choosing a name like LESSEQ.
2154 ** The %token_table declaration says to write a table of tokens (names
2155 and numbers) into the parser file. The yylex function can use this
2156 table to recognize multiple-character string tokens, or for other
2159 ** The %no_lines declaration says not to generate any #line preprocessor
2160 directives in the parser file.
2162 ** The %raw declaration says to use internal Bison token numbers, not
2163 Yacc-compatible token numbers, when token names are defined as macros.
2165 ** The --no-parser option produces the parser tables without including
2166 the parser engine; a project can now use its own parser engine.
2167 The actions go into a separate file called NAME.act, in the form of
2168 a switch statement body.
2170 * Changes in version 1.23:
2172 The user can define YYPARSE_PARAM as the name of an argument to be
2173 passed into yyparse. The argument should have type void *. It should
2174 actually point to an object. Grammar actions can access the variable
2175 by casting it to the proper pointer type.
2177 Line numbers in output file corrected.
2179 * Changes in version 1.22:
2181 --help option added.
2183 * Changes in version 1.20:
2185 Output file does not redefine const for C++.
2189 Copyright (C) 1995-2012 Free Software Foundation, Inc.
2191 This file is part of Bison, the GNU Parser Generator.
2193 This program is free software: you can redistribute it and/or modify
2194 it under the terms of the GNU General Public License as published by
2195 the Free Software Foundation, either version 3 of the License, or
2196 (at your option) any later version.
2198 This program is distributed in the hope that it will be useful,
2199 but WITHOUT ANY WARRANTY; without even the implied warranty of
2200 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
2201 GNU General Public License for more details.
2203 You should have received a copy of the GNU General Public License
2204 along with this program. If not, see <http://www.gnu.org/licenses/>.
2206 LocalWords: yacc YYBACKUP glr GCC lalr ArrayIndexOutOfBoundsException nullptr
2207 LocalWords: cplusplus liby rpl fprintf mfcalc Wyacc stmt cond expr mk sym lr
2208 LocalWords: IELR ielr Lookahead YYERROR nonassoc LALR's api lookaheads yychar
2209 LocalWords: destructor lookahead YYRHSLOC YYLLOC Rhs ifndef YYFAIL cpp sr rr
2210 LocalWords: preprocessor initializer Wno Wnone Werror FreeBSD prec livelocks
2211 LocalWords: Solaris AIX UX RHEL Tru LHS gcc's Wundef YYENABLE NLS YYLTYPE VCG
2212 LocalWords: yyerror cpp's Wunused yylval yylloc prepend yyparse yylex yypush
2213 LocalWords: Graphviz xml nonterminals midrule destructor's YYSTYPE typedef ly
2214 LocalWords: CHR chr printf stdout namespace preprocessing enum pre include's
2215 LocalWords: YYRECOVERING nonfree destructors YYABORT YYACCEPT params enums de
2216 LocalWords: struct yystype DJGPP lex param Haible NUM alloca YYSTACK NUL goto
2217 LocalWords: YYMAXDEPTH Unescaped UCNs YYLTYPE's yyltype typedefs inline Yaccs
2218 LocalWords: Heriyanto Reenable dprec Hilfinger Eggert MYEOF Folle Menezes EOF
2219 LocalWords: Lackovic define's itemset Groff Gettext malloc NEWS'ed YYDEBUG YY
2220 LocalWords: namespaces strerror const autoconfiguration Dconst Autoconf's FDL
2221 LocalWords: Automake TMPDIR LESSEQ ylwrap endif yydebug YYTOKEN YYLSP ival hh
2222 LocalWords: extern YYTOKENTYPE TOKENTYPE yytokentype tokentype STYPE lval pdf
2223 LocalWords: lang yyoutput dvi html ps POSIX lvalp llocp Wother nterm arg init
2224 LocalWords: TOK calc yyo fval Wconflicts