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 Changes in version 1.75a, 2002-10-24:
11 * Bison can no longer be built by a K&R C compiler; it requires C89 or
12 later to be built. This change originally took place a few versions
13 ago, but nobody noticed until we recently asked someone to try
14 building Bison with a K&R C compiler.
17 This new directive is preferred over YYERROR_VERBOSE.
19 * %lex-param, %parse-param
20 These new directives are preferred over PARSE_PARAM and LEX_PARAM.
21 In addition, they provide a means for yyerror to remain pure, and
22 to access to the current location.
25 Bison now recognizes #line in its input, and forwards them.
27 Changes in version 1.75, 2002-10-14:
29 * Bison should now work on 64-bit hosts.
31 * Indonesian translation thanks to Tedi Heriyanto.
34 Fix spurious parse errors.
37 Some people redefine yyerror to steal yyparse' private variables.
38 Reenable this trick until an official feature replaces it.
41 In agreement with POSIX and with other Yaccs, leaving a default
42 action is valid when $$ is untyped, and $1 typed:
46 but the converse remains an error:
50 * Values of mid-rule actions
53 foo: { ... } { $$ = $1; } ...
55 was incorrectly rejected: $1 is defined in the second mid-rule
56 action, and is equal to the $$ of the first mid-rule action.
58 Changes in version 1.50, 2002-10-04:
63 causes Bison to produce a Generalized LR (GLR) parser, capable of handling
64 almost any context-free grammar, ambiguous or not. The new declarations
65 %dprec and %merge on grammar rules allow parse-time resolution of
66 ambiguities. Contributed by Paul Hilfinger.
68 Unfortunately Bison 1.50 does not work properly on 64-bit hosts
69 like the Alpha, so please stick to 32-bit hosts for now.
72 When not in Yacc compatibility mode, when the output file was not
73 specified, running `bison foo/bar.y' created `foo/bar.c'. It
77 The undefined token was systematically mapped to 2 which prevented
78 the use of 2 by the user. This is no longer the case.
80 * Unknown token numbers
81 If yylex returned an out of range value, yyparse could die. This is
85 According to POSIX, the error token must be 256.
86 Bison extends this requirement by making it a preference: *if* the
87 user specified that one of her tokens is numbered 256, then error
88 will be mapped onto another number.
90 * Verbose error messages
91 They no longer report `..., expecting error or...' for states where
92 error recovery is possible.
95 Defaults to `$end' instead of `$'.
97 * Error recovery now conforms to documentation and to POSIX
98 When a Bison-generated parser encounters a syntax error, it now pops
99 the stack until it finds a state that allows shifting the error
100 token. Formerly, it popped the stack until it found a state that
101 allowed some non-error action other than a default reduction on the
102 error token. The new behavior has long been the documented behavior,
103 and has long been required by POSIX. For more details, please see
104 <http://mail.gnu.org/pipermail/bug-bison/2002-May/001452.html>.
107 Popped tokens and nonterminals are now reported.
110 Larger grammars are now supported (larger token numbers, larger grammar
111 size (= sum of the LHS and RHS lengths), larger LALR tables).
112 Formerly, many of these numbers ran afoul of 16-bit limits;
113 now these limits are 32 bits on most hosts.
115 * Explicit initial rule
116 Bison used to play hacks with the initial rule, which the user does
117 not write. It is now explicit, and visible in the reports and
121 Before, Bison reported the useless rules, but, although not used,
122 included them in the parsers. They are now actually removed.
124 * Useless rules, useless nonterminals
125 They are now reported, as a warning, with their locations.
127 * Rules never reduced
128 Rules that can never be reduced because of conflicts are now
131 * Incorrect `Token not used'
134 %token useless useful
136 exp: '0' %prec useful;
138 where a token was used to set the precedence of the last rule,
139 bison reported both `useful' and `useless' as useless tokens.
141 * Revert the C++ namespace changes introduced in 1.31
142 as they caused too many portability hassles.
145 By an accident of design, the default computation of @$ was
146 performed after another default computation was performed: @$ = @1.
147 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
148 the computation of @$.
151 The token end of file may be specified by the user, in which case,
152 the user symbol is used in the reports, the graphs, and the verbose
153 error messages instead of `$end', which remains being the default.
157 %token YYEOF 0 "end of file"
160 This old option, which has been broken for ages, is removed.
163 Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
164 Croatian, thanks to Denis Lackovic.
166 * Incorrect token definitions
167 When given `%token 'a' "A"', Bison used to output `#define 'a' 65'.
169 * Token definitions as enums
170 Tokens are output both as the traditional #define's, and, provided
171 the compiler supports ANSI C or is a C++ compiler, as enums.
172 This lets debuggers display names instead of integers.
175 In addition to --verbose, bison supports --report=THINGS, which
176 produces additional information:
178 complete the core item sets with their closure
180 explicitly associate lookaheads to items
182 describe shift/reduce conflicts solving.
183 Bison used to systematically output this information on top of
184 the report. Solved conflicts are now attached to their states.
187 Previous versions don't complain when there is a type clash on
188 the default action if the rule has a mid-rule action, such as in:
196 * GNU M4 is now required when using Bison.
198 Changes in version 1.35, 2002-03-25:
201 Some projects use Bison's C parser with C++ compilers, and define
202 YYSTYPE as a class. The recent adjustment of C parsers for data
203 alignment and 64 bit architectures made this impossible.
205 Because for the time being no real solution for C++ parser
206 generation exists, kludges were implemented in the parser to
207 maintain this use. In the future, when Bison has C++ parsers, this
208 kludge will be disabled.
210 This kludge also addresses some C++ problems when the stack was
213 Changes in version 1.34, 2002-03-12:
215 * File name clashes are detected
216 $ bison foo.y -d -o foo.x
217 fatal error: header and parser would both be named `foo.x'
219 * A missing `;' at the end of a rule triggers a warning
220 In accordance with POSIX, and in agreement with other
221 Yacc implementations, Bison will mandate this semicolon in the near
222 future. This eases the implementation of a Bison parser of Bison
223 grammars by making this grammar LALR(1) instead of LR(2). To
224 facilitate the transition, this release introduces a warning.
226 * Revert the C++ namespace changes introduced in 1.31, as they caused too
227 many portability hassles.
229 * DJGPP support added.
231 * Fix test suite portability problems.
233 Changes in version 1.33, 2002-02-07:
236 Groff could not be compiled for the definition of size_t was lacking
237 under some conditions.
242 Changes in version 1.32, 2002-01-23:
244 * Fix Yacc output file names
248 * Italian, Dutch translations
250 Changes in version 1.31, 2002-01-14:
254 * GNU Gettext and %expect
255 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
256 Bison dies on incorrect %expectations, we fear there will be
257 too many bug reports for Gettext, so _for the time being_, %expect
258 does not trigger an error when the input file is named `plural.y'.
260 * Use of alloca in parsers
261 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
262 malloc exclusively. Since 1.29, but was not NEWS'ed.
264 alloca is used only when compiled with GCC, to avoid portability
267 * When the generated parser lacks debugging code, YYDEBUG is now 0
268 (as POSIX requires) instead of being undefined.
271 Bison has always permitted actions such as { $$ = $1 }: it adds the
272 ending semicolon. Now if in Yacc compatibility mode, the semicolon
273 is no longer output: one has to write { $$ = $1; }.
275 * Better C++ compliance
276 The output parsers try to respect C++ namespaces.
277 [This turned out to be a failed experiment, and it was reverted later.]
280 Fixed bugs when reporting useless nonterminals.
283 The parsers work properly on 64 bit hosts.
286 Some calls to strerror resulted in scrambled or missing error messages.
289 When the number of shift/reduce conflicts is correct, don't issue
292 * The verbose report includes the rule line numbers.
294 * Rule line numbers are fixed in traces.
296 * Swedish translation
299 Verbose parse error messages from the parsers are better looking.
300 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
301 Now: parse error: unexpected '/', expecting "number" or '-' or '('
303 * Fixed parser memory leaks.
304 When the generated parser was using malloc to extend its stacks, the
305 previous allocations were not freed.
307 * Fixed verbose output file.
308 Some newlines were missing.
309 Some conflicts in state descriptions were missing.
311 * Fixed conflict report.
312 Option -v was needed to get the result.
316 Mismatches are errors, not warnings.
318 * Fixed incorrect processing of some invalid input.
320 * Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
322 * Fixed some typos in the documentation.
324 * %token MY_EOF 0 is supported.
325 Before, MY_EOF was silently renumbered as 257.
327 * doc/refcard.tex is updated.
329 * %output, %file-prefix, %name-prefix.
333 New, aliasing `--output-file'.
335 Changes in version 1.30, 2001-10-26:
337 * `--defines' and `--graph' have now an optionnal argument which is the
338 output file name. `-d' and `-g' do not change, they do not take any
341 * `%source_extension' and `%header_extension' are removed, failed
346 Changes in version 1.29, 2001-09-07:
348 * The output file does not define const, as this caused problems when used
349 with common autoconfiguration schemes. If you still use ancient compilers
350 that lack const, compile with the equivalent of the C compiler option
351 `-Dconst='. autoconf's AC_C_CONST macro provides one way to do this.
353 * Added `-g' and `--graph'.
355 * The Bison manual is now distributed under the terms of the GNU FDL.
357 * The input and the output files has automatically a similar extension.
359 * Russian translation added.
361 * NLS support updated; should hopefully be less troublesome.
363 * Added the old Bison reference card.
365 * Added `--locations' and `%locations'.
367 * Added `-S' and `--skeleton'.
369 * `%raw', `-r', `--raw' is disabled.
371 * Special characters are escaped when output. This solves the problems
372 of the #line lines with path names including backslashes.
375 `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose',
376 `%debug', `%source_extension' and `%header_extension'.
379 Automatic location tracking.
381 Changes in version 1.28, 1999-07-06:
383 * Should compile better now with K&R compilers.
387 * Fixed a problem with escaping the double quote character.
389 * There is now a FAQ.
391 Changes in version 1.27:
393 * The make rule which prevented bison.simple from being created on
394 some systems has been fixed.
396 Changes in version 1.26:
398 * Bison now uses automake.
400 * New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
402 * Token numbers now start at 257 as previously documented, not 258.
404 * Bison honors the TMPDIR environment variable.
406 * A couple of buffer overruns have been fixed.
408 * Problems when closing files should now be reported.
410 * Generated parsers should now work even on operating systems which do
411 not provide alloca().
413 Changes in version 1.25, 1995-10-16:
415 * Errors in the input grammar are not fatal; Bison keeps reading
416 the grammar file, and reports all the errors found in it.
418 * Tokens can now be specified as multiple-character strings: for
419 example, you could use "<=" for a token which looks like <=, instead
420 of chosing a name like LESSEQ.
422 * The %token_table declaration says to write a table of tokens (names
423 and numbers) into the parser file. The yylex function can use this
424 table to recognize multiple-character string tokens, or for other
427 * The %no_lines declaration says not to generate any #line preprocessor
428 directives in the parser file.
430 * The %raw declaration says to use internal Bison token numbers, not
431 Yacc-compatible token numbers, when token names are defined as macros.
433 * The --no-parser option produces the parser tables without including
434 the parser engine; a project can now use its own parser engine.
435 The actions go into a separate file called NAME.act, in the form of
436 a switch statement body.
438 Changes in version 1.23:
440 The user can define YYPARSE_PARAM as the name of an argument to be
441 passed into yyparse. The argument should have type void *. It should
442 actually point to an object. Grammar actions can access the variable
443 by casting it to the proper pointer type.
445 Line numbers in output file corrected.
447 Changes in version 1.22:
451 Changes in version 1.20:
453 Output file does not redefine const for C++.
461 Copyright (C) 2001, 2002 Free Software Foundation, Inc.
463 This file is part of GNU Autoconf.
465 GNU Autoconf is free software; you can redistribute it and/or modify
466 it under the terms of the GNU General Public License as published by
467 the Free Software Foundation; either version 2, or (at your option)
470 GNU Autoconf is distributed in the hope that it will be useful,
471 but WITHOUT ANY WARRANTY; without even the implied warranty of
472 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
473 GNU General Public License for more details.
475 You should have received a copy of the GNU General Public License
476 along with autoconf; see the file COPYING. If not, write to
477 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
478 Boston, MA 02111-1307, USA.