]> git.saurik.com Git - bison.git/blob - NEWS
(strerror_r): Remove decl; not needed.
[bison.git] / NEWS
1 Bison News
2 ----------
3
4 Changes in version 1.49b, 2002-08-12:
5
6 * GLR parsing
7 The declaration
8 %glr-parser
9 causes Bison to produce a Generalized LR (GLR) parser, capable of handling
10 almost any context-free grammar, ambiguous or not. The new declarations
11 %dprec and %merge on grammar rules allow parse-time resolution of
12 ambiguities. Contributed by Paul Hilfinger.
13
14 * Output Directory
15 When not in Yacc compatibility mode, when the output file was not
16 specified, runnning `bison foo/bar.y' created `foo/bar.c'. It
17 now creates `bar.c'.
18
19 * Undefined token
20 The undefined token was systematically mapped to 2 which prevented
21 the use of 2 from the user. This is no longer the case.
22
23 * Unknown token numbers
24 If yylex returned a code out of range, yyparse could die. This is
25 no longer the case.
26
27 * Error token
28 According to POSIX, the error token should be numbered as 256.
29 Bison extends this requirement by making it a preference: *if* the
30 user specified that one of her tokens is numbered 256, then error
31 will be mapped onto another number.
32
33 * Verbose error messages
34 They no longer report `..., expecting error or...' for state where
35 error recovery is possible.
36
37 * End token
38 Defaults to `$end' instead of `$'.
39
40 * Error recovery now conforms to documentation and to POSIX
41 When a Bison-generated parser encounters a syntax error, it now pops
42 the stack until it finds a state that allows shifting the error
43 token. Formerly, it popped the stack until it found a state that
44 allowed some non-error action other than a default reduction on the
45 error token. The new behavior has long been the documented behavior,
46 and has long been required by POSIX. For more details, please see
47 <http://mail.gnu.org/pipermail/bug-bison/2002-May/001452.html>.
48
49 * Traces
50 Popped tokens and nonterminals are now reported.
51
52 * Larger grammars
53 Larger grammars are now supported (larger token numbers, larger grammar
54 size (= sum of the LHS and RHS lengths), larger LALR tables).
55 Formerly, many of these numbers ran afoul of 16-bit limits;
56 now these limits are 32 bits on most hosts.
57
58 * Explicit initial rule
59 Bison used to play hacks with the initial rule, which the user does
60 not write. It is now explicit, and visible in the reports and
61 graphs as rule 0.
62
63 * Useless rules
64 Before, Bison reported the useless rules, but, although not used,
65 included them in the parsers. They are now actually removed.
66
67 * Useless rules, useless nonterminals
68 They are now reported, as a warning, with their locations.
69
70 * Rules never reduced
71 Rules that can never be reduced because of conflicts are now
72 reported.
73
74 * Incorrect `Token not used'
75 On a grammar such as
76
77 %token useless useful
78 %%
79 exp: '0' %prec useful;
80
81 where a token was used to set the precedence of the last rule,
82 bison reported both `useful' and `useless' as useless tokens.
83
84 * Revert the C++ namespace changes introduced in 1.31
85 as they caused too many portability hassles.
86
87 * Default locations
88 By an accident of design, the default computation of @$ was
89 performed after another default computation was performed: @$ = @1.
90 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
91 the computation of @$.
92
93 * Token end-of-file
94 The token end of file may be specified by the user, in which case,
95 the user symbol is used in the reports, the graphs, and the verbose
96 error messages instead of `$end', which remains being the default.
97 For instance
98 %token YYEOF 0
99 or
100 %token YYEOF 0 "end of file"
101
102 * Semantic parser
103 This old option, which has been broken for ages, is removed.
104
105 * New translations
106 Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
107 Croatian, thanks to Denis Lackovic.
108
109 * Incorrect token definitions
110 When fed with `%token 'a' "A"', Bison used to output `#define 'a' 65'.
111
112 * Token definitions as enums
113 Tokens are output both as the traditional #define's, and, provided
114 the compiler supports ANSI C or is a C++ compiler, as enums.
115 This helps debuggers producing symbols instead of values.
116
117 * Reports
118 In addition to --verbose, bison supports --report=THINGS, which
119 produces additional information:
120 - itemset
121 complete the core item sets with their closure
122 - lookahead
123 explicitly associate lookaheads to items
124 - solved
125 describe shift/reduce conflicts solving.
126 Bison used to systematically output this information on top of
127 the report. Solved conflicts are now attached to their states.
128
129 * Type clashes
130 Previous versions don't complain when there is a type clash on
131 the default action if the rule has a mid-rule action, such as in:
132
133 %type <foo> bar
134 %%
135 bar: '0' {} '0';
136
137 This is fixed.
138
139 * GNU M4 is now required when using Bison.
140 \f
141 Changes in version 1.35, 2002-03-25:
142
143 * C Skeleton
144 Some projects use Bison's C parser with C++ compilers, and define
145 YYSTYPE as a class. The recent adjustment of C parsers for data
146 alignment and 64 bit architectures made this impossible.
147
148 Because for the time being no real solution for C++ parser
149 generation exists, kludges were implemented in the parser to
150 maintain this use. In the future, when Bison has C++ parsers, this
151 kludge will be disabled.
152
153 This kludge also addresses some C++ problems when the stack was
154 extended.
155
156 \f
157 Changes in version 1.34, 2002-03-12:
158
159 * File name clashes are detected
160 $ bison foo.y -d -o foo.x
161 fatal error: header and parser would both be named `foo.x'
162
163 * A missing `;' at the end of a rule triggers a warning
164 In accordance with POSIX, and in agreement with other
165 Yacc implementations, Bison will mandate this semicolon in the near
166 future. This eases the implementation of a Bison parser of Bison
167 grammars by making this grammar LALR(1) instead of LR(2). To
168 facilitate the transition, this release introduces a warning.
169
170 * Revert the C++ namespace changes introduced in 1.31, as they caused too
171 many portability hassles.
172
173 * DJGPP support added.
174
175 * Fix test suite portability problems.
176 \f
177 Changes in version 1.33, 2002-02-07:
178
179 * Fix C++ issues
180 Groff could not be compiled for the definition of size_t was lacking
181 under some conditions.
182
183 * Catch invalid @n
184 As is done with $n.
185 \f
186 Changes in version 1.32, 2002-01-23:
187
188 * Fix Yacc output file names
189
190 * Portability fixes
191
192 * Italian, Dutch translations
193 \f
194 Changes in version 1.31, 2002-01-14:
195
196 * Many Bug Fixes
197
198 * GNU Gettext and %expect
199 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
200 Bison dies on incorrect %expectations, we fear there will be
201 too many bug reports for Gettext, so _for the time being_, %expect
202 does not trigger an error when the input file is named `plural.y'.
203
204 * Use of alloca in parsers
205 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
206 malloc exclusively. Since 1.29, but was not NEWS'ed.
207
208 alloca is used only when compiled with GCC, to avoid portability
209 problems as on AIX.
210
211 * When the generated parser lacks debugging code, YYDEBUG is now 0
212 (as POSIX requires) instead of being undefined.
213
214 * User Actions
215 Bison has always permitted actions such as { $$ = $1 }: it adds the
216 ending semicolon. Now if in Yacc compatibility mode, the semicolon
217 is no longer output: one has to write { $$ = $1; }.
218
219 * Better C++ compliance
220 The output parsers try to respect C++ namespaces.
221 [This turned out to be a failed experiment, and it was reverted later.]
222
223 * Reduced Grammars
224 Fixed bugs when reporting useless nonterminals.
225
226 * 64 bit hosts
227 The parsers work properly on 64 bit hosts.
228
229 * Error messages
230 Some calls to strerror resulted in scrambled or missing error messages.
231
232 * %expect
233 When the number of shift/reduce conflicts is correct, don't issue
234 any warning.
235
236 * The verbose report includes the rule line numbers.
237
238 * Rule line numbers are fixed in traces.
239
240 * Swedish translation
241
242 * Parse errors
243 Verbose parse error messages from the parsers are better looking.
244 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
245 Now: parse error: unexpected '/', expecting "number" or '-' or '('
246
247 * Fixed parser memory leaks.
248 When the generated parser was using malloc to extend its stacks, the
249 previous allocations were not freed.
250
251 * Fixed verbose output file.
252 Some newlines were missing.
253 Some conflicts in state descriptions were missing.
254
255 * Fixed conflict report.
256 Option -v was needed to get the result.
257
258 * %expect
259 Was not used.
260 Mismatches are errors, not warnings.
261
262 * Fixed incorrect processing of some invalid input.
263
264 * Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
265
266 * Fixed some typos in the documentation.
267
268 * %token MY_EOF 0 is supported.
269 Before, MY_EOF was silently renumbered as 257.
270
271 * doc/refcard.tex is updated.
272
273 * %output, %file-prefix, %name-prefix.
274 New.
275
276 * --output
277 New, aliasing `--output-file'.
278 \f
279 Changes in version 1.30, 2001-10-26:
280
281 * `--defines' and `--graph' have now an optionnal argument which is the
282 output file name. `-d' and `-g' do not change, they do not take any
283 argument.
284
285 * `%source_extension' and `%header_extension' are removed, failed
286 experiment.
287
288 * Portability fixes.
289 \f
290 Changes in version 1.29, 2001-09-07:
291
292 * The output file does not define const, as this caused problems when used
293 with common autoconfiguration schemes. If you still use ancient compilers
294 that lack const, compile with the equivalent of the C compiler option
295 `-Dconst='. autoconf's AC_C_CONST macro provides one way to do this.
296
297 * Added `-g' and `--graph'.
298
299 * The Bison manual is now distributed under the terms of the GNU FDL.
300
301 * The input and the output files has automatically a similar extension.
302
303 * Russian translation added.
304
305 * NLS support updated; should hopefully be less troublesome.
306
307 * Added the old Bison reference card.
308
309 * Added `--locations' and `%locations'.
310
311 * Added `-S' and `--skeleton'.
312
313 * `%raw', `-r', `--raw' is disabled.
314
315 * Special characters are escaped when output. This solves the problems
316 of the #line lines with path names including backslashes.
317
318 * New directives.
319 `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose',
320 `%debug', `%source_extension' and `%header_extension'.
321
322 * @$
323 Automatic location tracking.
324 \f
325 Changes in version 1.28, 1999-07-06:
326
327 * Should compile better now with K&R compilers.
328
329 * Added NLS.
330
331 * Fixed a problem with escaping the double quote character.
332
333 * There is now a FAQ.
334 \f
335 Changes in version 1.27:
336
337 * The make rule which prevented bison.simple from being created on
338 some systems has been fixed.
339 \f
340 Changes in version 1.26:
341
342 * Bison now uses automake.
343
344 * New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
345
346 * Token numbers now start at 257 as previously documented, not 258.
347
348 * Bison honors the TMPDIR environment variable.
349
350 * A couple of buffer overruns have been fixed.
351
352 * Problems when closing files should now be reported.
353
354 * Generated parsers should now work even on operating systems which do
355 not provide alloca().
356 \f
357 Changes in version 1.25, 1995-10-16:
358
359 * Errors in the input grammar are not fatal; Bison keeps reading
360 the grammar file, and reports all the errors found in it.
361
362 * Tokens can now be specified as multiple-character strings: for
363 example, you could use "<=" for a token which looks like <=, instead
364 of chosing a name like LESSEQ.
365
366 * The %token_table declaration says to write a table of tokens (names
367 and numbers) into the parser file. The yylex function can use this
368 table to recognize multiple-character string tokens, or for other
369 purposes.
370
371 * The %no_lines declaration says not to generate any #line preprocessor
372 directives in the parser file.
373
374 * The %raw declaration says to use internal Bison token numbers, not
375 Yacc-compatible token numbers, when token names are defined as macros.
376
377 * The --no-parser option produces the parser tables without including
378 the parser engine; a project can now use its own parser engine.
379 The actions go into a separate file called NAME.act, in the form of
380 a switch statement body.
381 \f
382 Changes in version 1.23:
383
384 The user can define YYPARSE_PARAM as the name of an argument to be
385 passed into yyparse. The argument should have type void *. It should
386 actually point to an object. Grammar actions can access the variable
387 by casting it to the proper pointer type.
388
389 Line numbers in output file corrected.
390 \f
391 Changes in version 1.22:
392
393 --help option added.
394 \f
395 Changes in version 1.20:
396
397 Output file does not redefine const for C++.
398
399 Local Variables:
400 mode: outline
401 End:
402
403 -----
404
405 Copyright (C) 2001, 2002 Free Software Foundation, Inc.
406
407 This file is part of GNU Autoconf.
408
409 GNU Autoconf is free software; you can redistribute it and/or modify
410 it under the terms of the GNU General Public License as published by
411 the Free Software Foundation; either version 2, or (at your option)
412 any later version.
413
414 GNU Autoconf is distributed in the hope that it will be useful,
415 but WITHOUT ANY WARRANTY; without even the implied warranty of
416 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
417 GNU General Public License for more details.
418
419 You should have received a copy of the GNU General Public License
420 along with autoconf; see the file COPYING. If not, write to
421 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
422 Boston, MA 02111-1307, USA.