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.
11 Changes in version 1.75a, 2002-10-24:
13 * Bison can no longer be built by a K&R C compiler; it requires C89 or
14 later to be built. This change originally took place a few versions
15 ago, but nobody noticed until we recently asked someone to try
16 building Bison with a K&R C compiler.
19 This new directive is preferred over YYERROR_VERBOSE.
21 * %lex-param, %parse-param
22 These new directives are preferred over PARSE_PARAM and LEX_PARAM.
23 In addition, they provide a means for yyerror to remain pure, and
24 to access to the current location.
27 Bison now recognizes #line in its input, and forwards them.
29 Changes in version 1.75, 2002-10-14:
31 * Bison should now work on 64-bit hosts.
33 * Indonesian translation thanks to Tedi Heriyanto.
36 Fix spurious parse errors.
39 Some people redefine yyerror to steal yyparse' private variables.
40 Reenable this trick until an official feature replaces it.
43 In agreement with POSIX and with other Yaccs, leaving a default
44 action is valid when $$ is untyped, and $1 typed:
48 but the converse remains an error:
52 * Values of mid-rule actions
55 foo: { ... } { $$ = $1; } ...
57 was incorrectly rejected: $1 is defined in the second mid-rule
58 action, and is equal to the $$ of the first mid-rule action.
60 Changes in version 1.50, 2002-10-04:
65 causes Bison to produce a Generalized LR (GLR) parser, capable of handling
66 almost any context-free grammar, ambiguous or not. The new declarations
67 %dprec and %merge on grammar rules allow parse-time resolution of
68 ambiguities. Contributed by Paul Hilfinger.
70 Unfortunately Bison 1.50 does not work properly on 64-bit hosts
71 like the Alpha, so please stick to 32-bit hosts for now.
74 When not in Yacc compatibility mode, when the output file was not
75 specified, running `bison foo/bar.y' created `foo/bar.c'. It
79 The undefined token was systematically mapped to 2 which prevented
80 the use of 2 by the user. This is no longer the case.
82 * Unknown token numbers
83 If yylex returned an out of range value, yyparse could die. This is
87 According to POSIX, the error token must be 256.
88 Bison extends this requirement by making it a preference: *if* the
89 user specified that one of her tokens is numbered 256, then error
90 will be mapped onto another number.
92 * Verbose error messages
93 They no longer report `..., expecting error or...' for states where
94 error recovery is possible.
97 Defaults to `$end' instead of `$'.
99 * Error recovery now conforms to documentation and to POSIX
100 When a Bison-generated parser encounters a syntax error, it now pops
101 the stack until it finds a state that allows shifting the error
102 token. Formerly, it popped the stack until it found a state that
103 allowed some non-error action other than a default reduction on the
104 error token. The new behavior has long been the documented behavior,
105 and has long been required by POSIX. For more details, please see
106 <http://mail.gnu.org/pipermail/bug-bison/2002-May/001452.html>.
109 Popped tokens and nonterminals are now reported.
112 Larger grammars are now supported (larger token numbers, larger grammar
113 size (= sum of the LHS and RHS lengths), larger LALR tables).
114 Formerly, many of these numbers ran afoul of 16-bit limits;
115 now these limits are 32 bits on most hosts.
117 * Explicit initial rule
118 Bison used to play hacks with the initial rule, which the user does
119 not write. It is now explicit, and visible in the reports and
123 Before, Bison reported the useless rules, but, although not used,
124 included them in the parsers. They are now actually removed.
126 * Useless rules, useless nonterminals
127 They are now reported, as a warning, with their locations.
129 * Rules never reduced
130 Rules that can never be reduced because of conflicts are now
133 * Incorrect `Token not used'
136 %token useless useful
138 exp: '0' %prec useful;
140 where a token was used to set the precedence of the last rule,
141 bison reported both `useful' and `useless' as useless tokens.
143 * Revert the C++ namespace changes introduced in 1.31
144 as they caused too many portability hassles.
147 By an accident of design, the default computation of @$ was
148 performed after another default computation was performed: @$ = @1.
149 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
150 the computation of @$.
153 The token end of file may be specified by the user, in which case,
154 the user symbol is used in the reports, the graphs, and the verbose
155 error messages instead of `$end', which remains being the default.
159 %token YYEOF 0 "end of file"
162 This old option, which has been broken for ages, is removed.
165 Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
166 Croatian, thanks to Denis Lackovic.
168 * Incorrect token definitions
169 When given `%token 'a' "A"', Bison used to output `#define 'a' 65'.
171 * Token definitions as enums
172 Tokens are output both as the traditional #define's, and, provided
173 the compiler supports ANSI C or is a C++ compiler, as enums.
174 This lets debuggers display names instead of integers.
177 In addition to --verbose, bison supports --report=THINGS, which
178 produces additional information:
180 complete the core item sets with their closure
182 explicitly associate lookaheads to items
184 describe shift/reduce conflicts solving.
185 Bison used to systematically output this information on top of
186 the report. Solved conflicts are now attached to their states.
189 Previous versions don't complain when there is a type clash on
190 the default action if the rule has a mid-rule action, such as in:
198 * GNU M4 is now required when using Bison.
200 Changes in version 1.35, 2002-03-25:
203 Some projects use Bison's C parser with C++ compilers, and define
204 YYSTYPE as a class. The recent adjustment of C parsers for data
205 alignment and 64 bit architectures made this impossible.
207 Because for the time being no real solution for C++ parser
208 generation exists, kludges were implemented in the parser to
209 maintain this use. In the future, when Bison has C++ parsers, this
210 kludge will be disabled.
212 This kludge also addresses some C++ problems when the stack was
215 Changes in version 1.34, 2002-03-12:
217 * File name clashes are detected
218 $ bison foo.y -d -o foo.x
219 fatal error: header and parser would both be named `foo.x'
221 * A missing `;' at the end of a rule triggers a warning
222 In accordance with POSIX, and in agreement with other
223 Yacc implementations, Bison will mandate this semicolon in the near
224 future. This eases the implementation of a Bison parser of Bison
225 grammars by making this grammar LALR(1) instead of LR(2). To
226 facilitate the transition, this release introduces a warning.
228 * Revert the C++ namespace changes introduced in 1.31, as they caused too
229 many portability hassles.
231 * DJGPP support added.
233 * Fix test suite portability problems.
235 Changes in version 1.33, 2002-02-07:
238 Groff could not be compiled for the definition of size_t was lacking
239 under some conditions.
244 Changes in version 1.32, 2002-01-23:
246 * Fix Yacc output file names
250 * Italian, Dutch translations
252 Changes in version 1.31, 2002-01-14:
256 * GNU Gettext and %expect
257 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
258 Bison dies on incorrect %expectations, we fear there will be
259 too many bug reports for Gettext, so _for the time being_, %expect
260 does not trigger an error when the input file is named `plural.y'.
262 * Use of alloca in parsers
263 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
264 malloc exclusively. Since 1.29, but was not NEWS'ed.
266 alloca is used only when compiled with GCC, to avoid portability
269 * When the generated parser lacks debugging code, YYDEBUG is now 0
270 (as POSIX requires) instead of being undefined.
273 Bison has always permitted actions such as { $$ = $1 }: it adds the
274 ending semicolon. Now if in Yacc compatibility mode, the semicolon
275 is no longer output: one has to write { $$ = $1; }.
277 * Better C++ compliance
278 The output parsers try to respect C++ namespaces.
279 [This turned out to be a failed experiment, and it was reverted later.]
282 Fixed bugs when reporting useless nonterminals.
285 The parsers work properly on 64 bit hosts.
288 Some calls to strerror resulted in scrambled or missing error messages.
291 When the number of shift/reduce conflicts is correct, don't issue
294 * The verbose report includes the rule line numbers.
296 * Rule line numbers are fixed in traces.
298 * Swedish translation
301 Verbose parse error messages from the parsers are better looking.
302 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
303 Now: parse error: unexpected '/', expecting "number" or '-' or '('
305 * Fixed parser memory leaks.
306 When the generated parser was using malloc to extend its stacks, the
307 previous allocations were not freed.
309 * Fixed verbose output file.
310 Some newlines were missing.
311 Some conflicts in state descriptions were missing.
313 * Fixed conflict report.
314 Option -v was needed to get the result.
318 Mismatches are errors, not warnings.
320 * Fixed incorrect processing of some invalid input.
322 * Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
324 * Fixed some typos in the documentation.
326 * %token MY_EOF 0 is supported.
327 Before, MY_EOF was silently renumbered as 257.
329 * doc/refcard.tex is updated.
331 * %output, %file-prefix, %name-prefix.
335 New, aliasing `--output-file'.
337 Changes in version 1.30, 2001-10-26:
339 * `--defines' and `--graph' have now an optionnal argument which is the
340 output file name. `-d' and `-g' do not change, they do not take any
343 * `%source_extension' and `%header_extension' are removed, failed
348 Changes in version 1.29, 2001-09-07:
350 * The output file does not define const, as this caused problems when used
351 with common autoconfiguration schemes. If you still use ancient compilers
352 that lack const, compile with the equivalent of the C compiler option
353 `-Dconst='. autoconf's AC_C_CONST macro provides one way to do this.
355 * Added `-g' and `--graph'.
357 * The Bison manual is now distributed under the terms of the GNU FDL.
359 * The input and the output files has automatically a similar extension.
361 * Russian translation added.
363 * NLS support updated; should hopefully be less troublesome.
365 * Added the old Bison reference card.
367 * Added `--locations' and `%locations'.
369 * Added `-S' and `--skeleton'.
371 * `%raw', `-r', `--raw' is disabled.
373 * Special characters are escaped when output. This solves the problems
374 of the #line lines with path names including backslashes.
377 `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose',
378 `%debug', `%source_extension' and `%header_extension'.
381 Automatic location tracking.
383 Changes in version 1.28, 1999-07-06:
385 * Should compile better now with K&R compilers.
389 * Fixed a problem with escaping the double quote character.
391 * There is now a FAQ.
393 Changes in version 1.27:
395 * The make rule which prevented bison.simple from being created on
396 some systems has been fixed.
398 Changes in version 1.26:
400 * Bison now uses automake.
402 * New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
404 * Token numbers now start at 257 as previously documented, not 258.
406 * Bison honors the TMPDIR environment variable.
408 * A couple of buffer overruns have been fixed.
410 * Problems when closing files should now be reported.
412 * Generated parsers should now work even on operating systems which do
413 not provide alloca().
415 Changes in version 1.25, 1995-10-16:
417 * Errors in the input grammar are not fatal; Bison keeps reading
418 the grammar file, and reports all the errors found in it.
420 * Tokens can now be specified as multiple-character strings: for
421 example, you could use "<=" for a token which looks like <=, instead
422 of chosing a name like LESSEQ.
424 * The %token_table declaration says to write a table of tokens (names
425 and numbers) into the parser file. The yylex function can use this
426 table to recognize multiple-character string tokens, or for other
429 * The %no_lines declaration says not to generate any #line preprocessor
430 directives in the parser file.
432 * The %raw declaration says to use internal Bison token numbers, not
433 Yacc-compatible token numbers, when token names are defined as macros.
435 * The --no-parser option produces the parser tables without including
436 the parser engine; a project can now use its own parser engine.
437 The actions go into a separate file called NAME.act, in the form of
438 a switch statement body.
440 Changes in version 1.23:
442 The user can define YYPARSE_PARAM as the name of an argument to be
443 passed into yyparse. The argument should have type void *. It should
444 actually point to an object. Grammar actions can access the variable
445 by casting it to the proper pointer type.
447 Line numbers in output file corrected.
449 Changes in version 1.22:
453 Changes in version 1.20:
455 Output file does not redefine const for C++.
463 Copyright (C) 2001, 2002 Free Software Foundation, Inc.
465 This file is part of GNU Autoconf.
467 GNU Autoconf is free software; you can redistribute it and/or modify
468 it under the terms of the GNU General Public License as published by
469 the Free Software Foundation; either version 2, or (at your option)
472 GNU Autoconf is distributed in the hope that it will be useful,
473 but WITHOUT ANY WARRANTY; without even the implied warranty of
474 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
475 GNU General Public License for more details.
477 You should have received a copy of the GNU General Public License
478 along with autoconf; see the file COPYING. If not, write to
479 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
480 Boston, MA 02111-1307, USA.