4 Changes in version 1.75b:
6 * Bison now parses C99 lexical constructs like digraphs, UCNs, and
7 backslash-newline within C escape sequences, as POSIX now requires.
9 * --no-line works properly.
12 This new directive is preferred over YYERROR_VERBOSE.
14 * %lex-param, %parse-param
15 These new directives are preferred over PARSE_PARAM and LEX_PARAM.
16 In addition, they provide a means for yyerror to remain pure, and
17 to access to the current location.
19 #defining yyerror to steal internal variables is strongly
20 discouraged. It is not guaranteed to work for ever.
23 Bison now recognizes #line in its input, and forwards them.
26 File names are properly escaped. E.g. foo\bar.y give #line 123 "foo\\bar.y".
28 Changes in version 1.75a, 2002-10-24:
30 * Bison can no longer be built by a K&R C compiler; it requires C89 or
31 later to be built. This change originally took place a few versions
32 ago, but nobody noticed until we recently asked someone to try
33 building Bison with a K&R C compiler.
35 Changes in version 1.75, 2002-10-14:
37 * Bison should now work on 64-bit hosts.
39 * Indonesian translation thanks to Tedi Heriyanto.
42 Fix spurious parse errors.
45 Some people redefine yyerror to steal yyparse' private variables.
46 Reenable this trick until an official feature replaces it.
49 In agreement with POSIX and with other Yaccs, leaving a default
50 action is valid when $$ is untyped, and $1 typed:
54 but the converse remains an error:
58 * Values of mid-rule actions
61 foo: { ... } { $$ = $1; } ...
63 was incorrectly rejected: $1 is defined in the second mid-rule
64 action, and is equal to the $$ of the first mid-rule action.
66 Changes in version 1.50, 2002-10-04:
71 causes Bison to produce a Generalized LR (GLR) parser, capable of handling
72 almost any context-free grammar, ambiguous or not. The new declarations
73 %dprec and %merge on grammar rules allow parse-time resolution of
74 ambiguities. Contributed by Paul Hilfinger.
76 Unfortunately Bison 1.50 does not work properly on 64-bit hosts
77 like the Alpha, so please stick to 32-bit hosts for now.
80 When not in Yacc compatibility mode, when the output file was not
81 specified, running `bison foo/bar.y' created `foo/bar.c'. It
85 The undefined token was systematically mapped to 2 which prevented
86 the use of 2 by the user. This is no longer the case.
88 * Unknown token numbers
89 If yylex returned an out of range value, yyparse could die. This is
93 According to POSIX, the error token must be 256.
94 Bison extends this requirement by making it a preference: *if* the
95 user specified that one of her tokens is numbered 256, then error
96 will be mapped onto another number.
98 * Verbose error messages
99 They no longer report `..., expecting error or...' for states where
100 error recovery is possible.
103 Defaults to `$end' instead of `$'.
105 * Error recovery now conforms to documentation and to POSIX
106 When a Bison-generated parser encounters a syntax error, it now pops
107 the stack until it finds a state that allows shifting the error
108 token. Formerly, it popped the stack until it found a state that
109 allowed some non-error action other than a default reduction on the
110 error token. The new behavior has long been the documented behavior,
111 and has long been required by POSIX. For more details, please see
112 <http://mail.gnu.org/pipermail/bug-bison/2002-May/001452.html>.
115 Popped tokens and nonterminals are now reported.
118 Larger grammars are now supported (larger token numbers, larger grammar
119 size (= sum of the LHS and RHS lengths), larger LALR tables).
120 Formerly, many of these numbers ran afoul of 16-bit limits;
121 now these limits are 32 bits on most hosts.
123 * Explicit initial rule
124 Bison used to play hacks with the initial rule, which the user does
125 not write. It is now explicit, and visible in the reports and
129 Before, Bison reported the useless rules, but, although not used,
130 included them in the parsers. They are now actually removed.
132 * Useless rules, useless nonterminals
133 They are now reported, as a warning, with their locations.
135 * Rules never reduced
136 Rules that can never be reduced because of conflicts are now
139 * Incorrect `Token not used'
142 %token useless useful
144 exp: '0' %prec useful;
146 where a token was used to set the precedence of the last rule,
147 bison reported both `useful' and `useless' as useless tokens.
149 * Revert the C++ namespace changes introduced in 1.31
150 as they caused too many portability hassles.
153 By an accident of design, the default computation of @$ was
154 performed after another default computation was performed: @$ = @1.
155 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
156 the computation of @$.
159 The token end of file may be specified by the user, in which case,
160 the user symbol is used in the reports, the graphs, and the verbose
161 error messages instead of `$end', which remains being the default.
165 %token YYEOF 0 "end of file"
168 This old option, which has been broken for ages, is removed.
171 Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
172 Croatian, thanks to Denis Lackovic.
174 * Incorrect token definitions
175 When given `%token 'a' "A"', Bison used to output `#define 'a' 65'.
177 * Token definitions as enums
178 Tokens are output both as the traditional #define's, and, provided
179 the compiler supports ANSI C or is a C++ compiler, as enums.
180 This lets debuggers display names instead of integers.
183 In addition to --verbose, bison supports --report=THINGS, which
184 produces additional information:
186 complete the core item sets with their closure
188 explicitly associate lookaheads to items
190 describe shift/reduce conflicts solving.
191 Bison used to systematically output this information on top of
192 the report. Solved conflicts are now attached to their states.
195 Previous versions don't complain when there is a type clash on
196 the default action if the rule has a mid-rule action, such as in:
204 * GNU M4 is now required when using Bison.
206 Changes in version 1.35, 2002-03-25:
209 Some projects use Bison's C parser with C++ compilers, and define
210 YYSTYPE as a class. The recent adjustment of C parsers for data
211 alignment and 64 bit architectures made this impossible.
213 Because for the time being no real solution for C++ parser
214 generation exists, kludges were implemented in the parser to
215 maintain this use. In the future, when Bison has C++ parsers, this
216 kludge will be disabled.
218 This kludge also addresses some C++ problems when the stack was
221 Changes in version 1.34, 2002-03-12:
223 * File name clashes are detected
224 $ bison foo.y -d -o foo.x
225 fatal error: header and parser would both be named `foo.x'
227 * A missing `;' at the end of a rule triggers a warning
228 In accordance with POSIX, and in agreement with other
229 Yacc implementations, Bison will mandate this semicolon in the near
230 future. This eases the implementation of a Bison parser of Bison
231 grammars by making this grammar LALR(1) instead of LR(2). To
232 facilitate the transition, this release introduces a warning.
234 * Revert the C++ namespace changes introduced in 1.31, as they caused too
235 many portability hassles.
237 * DJGPP support added.
239 * Fix test suite portability problems.
241 Changes in version 1.33, 2002-02-07:
244 Groff could not be compiled for the definition of size_t was lacking
245 under some conditions.
250 Changes in version 1.32, 2002-01-23:
252 * Fix Yacc output file names
256 * Italian, Dutch translations
258 Changes in version 1.31, 2002-01-14:
262 * GNU Gettext and %expect
263 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
264 Bison dies on incorrect %expectations, we fear there will be
265 too many bug reports for Gettext, so _for the time being_, %expect
266 does not trigger an error when the input file is named `plural.y'.
268 * Use of alloca in parsers
269 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
270 malloc exclusively. Since 1.29, but was not NEWS'ed.
272 alloca is used only when compiled with GCC, to avoid portability
275 * When the generated parser lacks debugging code, YYDEBUG is now 0
276 (as POSIX requires) instead of being undefined.
279 Bison has always permitted actions such as { $$ = $1 }: it adds the
280 ending semicolon. Now if in Yacc compatibility mode, the semicolon
281 is no longer output: one has to write { $$ = $1; }.
283 * Better C++ compliance
284 The output parsers try to respect C++ namespaces.
285 [This turned out to be a failed experiment, and it was reverted later.]
288 Fixed bugs when reporting useless nonterminals.
291 The parsers work properly on 64 bit hosts.
294 Some calls to strerror resulted in scrambled or missing error messages.
297 When the number of shift/reduce conflicts is correct, don't issue
300 * The verbose report includes the rule line numbers.
302 * Rule line numbers are fixed in traces.
304 * Swedish translation
307 Verbose parse error messages from the parsers are better looking.
308 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
309 Now: parse error: unexpected '/', expecting "number" or '-' or '('
311 * Fixed parser memory leaks.
312 When the generated parser was using malloc to extend its stacks, the
313 previous allocations were not freed.
315 * Fixed verbose output file.
316 Some newlines were missing.
317 Some conflicts in state descriptions were missing.
319 * Fixed conflict report.
320 Option -v was needed to get the result.
324 Mismatches are errors, not warnings.
326 * Fixed incorrect processing of some invalid input.
328 * Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
330 * Fixed some typos in the documentation.
332 * %token MY_EOF 0 is supported.
333 Before, MY_EOF was silently renumbered as 257.
335 * doc/refcard.tex is updated.
337 * %output, %file-prefix, %name-prefix.
341 New, aliasing `--output-file'.
343 Changes in version 1.30, 2001-10-26:
345 * `--defines' and `--graph' have now an optional argument which is the
346 output file name. `-d' and `-g' do not change; they do not take any
349 * `%source_extension' and `%header_extension' are removed, failed
354 Changes in version 1.29, 2001-09-07:
356 * The output file does not define const, as this caused problems when used
357 with common autoconfiguration schemes. If you still use ancient compilers
358 that lack const, compile with the equivalent of the C compiler option
359 `-Dconst='. autoconf's AC_C_CONST macro provides one way to do this.
361 * Added `-g' and `--graph'.
363 * The Bison manual is now distributed under the terms of the GNU FDL.
365 * The input and the output files has automatically a similar extension.
367 * Russian translation added.
369 * NLS support updated; should hopefully be less troublesome.
371 * Added the old Bison reference card.
373 * Added `--locations' and `%locations'.
375 * Added `-S' and `--skeleton'.
377 * `%raw', `-r', `--raw' is disabled.
379 * Special characters are escaped when output. This solves the problems
380 of the #line lines with path names including backslashes.
383 `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose',
384 `%debug', `%source_extension' and `%header_extension'.
387 Automatic location tracking.
389 Changes in version 1.28, 1999-07-06:
391 * Should compile better now with K&R compilers.
395 * Fixed a problem with escaping the double quote character.
397 * There is now a FAQ.
399 Changes in version 1.27:
401 * The make rule which prevented bison.simple from being created on
402 some systems has been fixed.
404 Changes in version 1.26:
406 * Bison now uses automake.
408 * New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
410 * Token numbers now start at 257 as previously documented, not 258.
412 * Bison honors the TMPDIR environment variable.
414 * A couple of buffer overruns have been fixed.
416 * Problems when closing files should now be reported.
418 * Generated parsers should now work even on operating systems which do
419 not provide alloca().
421 Changes in version 1.25, 1995-10-16:
423 * Errors in the input grammar are not fatal; Bison keeps reading
424 the grammar file, and reports all the errors found in it.
426 * Tokens can now be specified as multiple-character strings: for
427 example, you could use "<=" for a token which looks like <=, instead
428 of chosing a name like LESSEQ.
430 * The %token_table declaration says to write a table of tokens (names
431 and numbers) into the parser file. The yylex function can use this
432 table to recognize multiple-character string tokens, or for other
435 * The %no_lines declaration says not to generate any #line preprocessor
436 directives in the parser file.
438 * The %raw declaration says to use internal Bison token numbers, not
439 Yacc-compatible token numbers, when token names are defined as macros.
441 * The --no-parser option produces the parser tables without including
442 the parser engine; a project can now use its own parser engine.
443 The actions go into a separate file called NAME.act, in the form of
444 a switch statement body.
446 Changes in version 1.23:
448 The user can define YYPARSE_PARAM as the name of an argument to be
449 passed into yyparse. The argument should have type void *. It should
450 actually point to an object. Grammar actions can access the variable
451 by casting it to the proper pointer type.
453 Line numbers in output file corrected.
455 Changes in version 1.22:
459 Changes in version 1.20:
461 Output file does not redefine const for C++.
469 Copyright (C) 2001, 2002 Free Software Foundation, Inc.
471 This file is part of GNU Autoconf.
473 GNU Autoconf is free software; you can redistribute it and/or modify
474 it under the terms of the GNU General Public License as published by
475 the Free Software Foundation; either version 2, or (at your option)
478 GNU Autoconf is distributed in the hope that it will be useful,
479 but WITHOUT ANY WARRANTY; without even the implied warranty of
480 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
481 GNU General Public License for more details.
483 You should have received a copy of the GNU General Public License
484 along with autoconf; see the file COPYING. If not, write to
485 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
486 Boston, MA 02111-1307, USA.