4 Changes in version 1.49b:
7 The undefined token was systematically mapped to 2 which prevented
8 the use of 2 from the user. This is no longer the case.
10 * Unknown token numbers
11 If yylex returned a code out of range, yyparse could die. This is
15 According to POSIX, the error token should be numbered as 256.
16 Bison extends this requirement by making it a preference: *if* the
17 user specified that one of her tokens is numbered 256, then error
18 will be mapped onto another number.
20 * Error recovery now conforms to documentation and to POSIX
21 When a Bison-generated parser encounters a syntax error, it now pops
22 the stack until it finds a state that allows shifting the error
23 token. Formerly, it popped the stack until it found a state that
24 allowed some non-error action other than a default reduction on the
25 error token. The new behavior has long been the documented behavior,
26 and has long been required by POSIX. For more details, please see
27 <http://mail.gnu.org/pipermail/bug-bison/2002-May/001452.html>.
30 Popped tokens and nonterminals are now reported.
33 Large grammars are now supported (large token numbers, large grammar
34 size (= sum of the LHS and RHS lengths), large LALR tables).
36 * Explicit initial rule
37 Bison used to play hacks with the initial rule, which the user does
38 not write. It is now explicit, and visible in the reports and
42 Before, Bison reported the useless rules, but, although not used,
43 included them in the parsers. They are now actually removed.
45 * Incorrect `Token not used'
50 exp: '0' %prec useful;
52 where a token was used to set the precedence of the last rule,
53 bison reported both `useful' and `useless' as useless tokens.
55 * Revert the C++ namespace changes introduced in 1.31
56 as they caused too many portability hassles.
59 By an accident of design, the default computation of @$ was
60 performed after another default computation was performed: @$ = @1.
61 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
62 the computation of @$.
65 The token end of file may be specified by the user, in which case,
66 the user symbol is used in the reports, the graphs, and the verbose
67 error messages instead of `$', which remains being the defaults.
71 %token YYEOF 0 "end of file"
74 This old option, which has been broken for ages, is removed.
77 Croatian, thanks to Denis Lackovic.
79 * Incorrect token definitions
80 When fed with `%token 'a' "A"', Bison used to output `#define 'a' 65'.
82 * Token definitions as enums
83 Tokens are output both as the traditional #define's, and, provided
84 the compiler supports ANSI C or is a C++ compiler, as enums.
85 This helps debuggers producing symbols instead of values.
88 In addition to --verbose, bison supports --report=THINGS, which
89 produces additional information:
91 complete the core item sets with their closure
93 explicitly associate lookaheads to items
95 describe shift/reduce conflicts solving.
96 Bison used to systematically output this information on top of
97 the report. Solved conflicts are now attached to their states.
100 Previous versions don't complain when there is a type clash on
101 the default action if the rule has a mid-rule action, such as in:
109 Changes in version 1.35, 2002-03-25:
112 Some projects use Bison's C parser with C++ compilers, and define
113 YYSTYPE as a class. The recent adjustment of C parsers for data
114 alignment and 64 bit architectures made this impossible.
116 Because for the time being no real solution for C++ parser
117 generation exists, kludges were implemented in the parser to
118 maintain this use. In the future, when Bison has C++ parsers, this
119 kludge will be disabled.
121 This kludge also addresses some C++ problems when the stack was
125 Changes in version 1.34, 2002-03-12:
127 * File name clashes are detected
128 $ bison foo.y -d -o foo.x
129 fatal error: header and parser would both be named `foo.x'
131 * A missing `;' at the end of a rule triggers a warning
132 In accordance with POSIX, and in agreement with other
133 Yacc implementations, Bison will mandate this semicolon in the near
134 future. This eases the implementation of a Bison parser of Bison
135 grammars by making this grammar LALR(1) instead of LR(2). To
136 facilitate the transition, this release introduces a warning.
138 * Revert the C++ namespace changes introduced in 1.31, as they caused too
139 many portability hassles.
141 * DJGPP support added.
143 * Fix test suite portability problems.
145 Changes in version 1.33, 2002-02-07:
148 Groff could not be compiled for the definition of size_t was lacking
149 under some conditions.
154 Changes in version 1.32, 2002-01-23:
156 * Fix Yacc output file names
160 * Italian, Dutch translations
162 Changes in version 1.31, 2002-01-14:
166 * GNU Gettext and %expect
167 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
168 Bison dies on incorrect %expectations, we fear there will be
169 too many bug reports for Gettext, so _for the time being_, %expect
170 does not trigger an error when the input file is named `plural.y'.
172 * Use of alloca in parsers
173 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
174 malloc exclusively. Since 1.29, but was not NEWS'ed.
176 alloca is used only when compiled with GCC, to avoid portability
179 * When the generated parser lacks debugging code, YYDEBUG is now 0
180 (as POSIX requires) instead of being undefined.
183 Bison has always permitted actions such as { $$ = $1 }: it adds the
184 ending semicolon. Now if in Yacc compatibility mode, the semicolon
185 is no longer output: one has to write { $$ = $1; }.
187 * Better C++ compliance
188 The output parsers try to respect C++ namespaces.
189 [This turned out to be a failed experiment, and it was reverted later.]
192 Fixed bugs when reporting useless nonterminals.
195 The parsers work properly on 64 bit hosts.
198 Some calls to strerror resulted in scrambled or missing error messages.
201 When the number of shift/reduce conflicts is correct, don't issue
204 * The verbose report includes the rule line numbers.
206 * Rule line numbers are fixed in traces.
208 * Swedish translation
211 Verbose parse error messages from the parsers are better looking.
212 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
213 Now: parse error: unexpected '/', expecting "number" or '-' or '('
215 * Fixed parser memory leaks.
216 When the generated parser was using malloc to extend its stacks, the
217 previous allocations were not freed.
219 * Fixed verbose output file.
220 Some newlines were missing.
221 Some conflicts in state descriptions were missing.
223 * Fixed conflict report.
224 Option -v was needed to get the result.
228 Mismatches are errors, not warnings.
230 * Fixed incorrect processing of some invalid input.
232 * Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
234 * Fixed some typos in the documentation.
236 * %token MY_EOF 0 is supported.
237 Before, MY_EOF was silently renumbered as 257.
239 * doc/refcard.tex is updated.
241 * %output, %file-prefix, %name-prefix.
245 New, aliasing `--output-file'.
247 Changes in version 1.30, 2001-10-26:
249 * `--defines' and `--graph' have now an optionnal argument which is the
250 output file name. `-d' and `-g' do not change, they do not take any
253 * `%source_extension' and `%header_extension' are removed, failed
258 Changes in version 1.29, 2001-09-07:
260 * The output file does not define const, as this caused problems when used
261 with common autoconfiguration schemes. If you still use ancient compilers
262 that lack const, compile with the equivalent of the C compiler option
263 `-Dconst='. autoconf's AC_C_CONST macro provides one way to do this.
265 * Added `-g' and `--graph'.
267 * The Bison manual is now distributed under the terms of the GNU FDL.
269 * The input and the output files has automatically a similar extension.
271 * Russian translation added.
273 * NLS support updated; should hopefully be less troublesome.
275 * Added the old Bison reference card.
277 * Added `--locations' and `%locations'.
279 * Added `-S' and `--skeleton'.
281 * `%raw', `-r', `--raw' is disabled.
283 * Special characters are escaped when output. This solves the problems
284 of the #line lines with path names including backslashes.
287 `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose',
288 `%debug', `%source_extension' and `%header_extension'.
291 Automatic location tracking.
293 Changes in version 1.28, 1999-07-06:
295 * Should compile better now with K&R compilers.
299 * Fixed a problem with escaping the double quote character.
301 * There is now a FAQ.
303 Changes in version 1.27:
305 * The make rule which prevented bison.simple from being created on
306 some systems has been fixed.
308 Changes in version 1.26:
310 * Bison now uses automake.
312 * New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
314 * Token numbers now start at 257 as previously documented, not 258.
316 * Bison honors the TMPDIR environment variable.
318 * A couple of buffer overruns have been fixed.
320 * Problems when closing files should now be reported.
322 * Generated parsers should now work even on operating systems which do
323 not provide alloca().
325 Changes in version 1.25, 1995-10-16:
327 * Errors in the input grammar are not fatal; Bison keeps reading
328 the grammar file, and reports all the errors found in it.
330 * Tokens can now be specified as multiple-character strings: for
331 example, you could use "<=" for a token which looks like <=, instead
332 of chosing a name like LESSEQ.
334 * The %token_table declaration says to write a table of tokens (names
335 and numbers) into the parser file. The yylex function can use this
336 table to recognize multiple-character string tokens, or for other
339 * The %no_lines declaration says not to generate any #line preprocessor
340 directives in the parser file.
342 * The %raw declaration says to use internal Bison token numbers, not
343 Yacc-compatible token numbers, when token names are defined as macros.
345 * The --no-parser option produces the parser tables without including
346 the parser engine; a project can now use its own parser engine.
347 The actions go into a separate file called NAME.act, in the form of
348 a switch statement body.
350 Changes in version 1.23:
352 The user can define YYPARSE_PARAM as the name of an argument to be
353 passed into yyparse. The argument should have type void *. It should
354 actually point to an object. Grammar actions can access the variable
355 by casting it to the proper pointer type.
357 Line numbers in output file corrected.
359 Changes in version 1.22:
363 Changes in version 1.20:
365 Output file does not redefine const for C++.
373 Copyright (C) 2001, 2002 Free Software Foundation, Inc.
375 This file is part of GNU Autoconf.
377 GNU Autoconf is free software; you can redistribute it and/or modify
378 it under the terms of the GNU General Public License as published by
379 the Free Software Foundation; either version 2, or (at your option)
382 GNU Autoconf is distributed in the hope that it will be useful,
383 but WITHOUT ANY WARRANTY; without even the implied warranty of
384 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
385 GNU General Public License for more details.
387 You should have received a copy of the GNU General Public License
388 along with autoconf; see the file COPYING. If not, write to
389 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
390 Boston, MA 02111-1307, USA.