]>
Commit | Line | Data |
---|---|---|
6780ca7a DM |
1 | Bison News |
2 | ---------- | |
3af4feb2 | 3 | |
74553c98 AD |
4 | * Changes in version ?.? (????-??-??): |
5 | ||
2055a44e AD |
6 | ** Additional yylex/yyparse arguments |
7 | ||
8 | The new directive %param declare additional argument to both yylex | |
bb9191dd | 9 | and yyparse. The %lex-param, %parse-param, and %param directives |
2055a44e AD |
10 | support one or more arguments. Instead of |
11 | ||
12 | %lex-param {arg1_type *arg1} | |
13 | %lex-param {arg2_type *arg2} | |
14 | %parse-param {arg1_type *arg1} | |
15 | %parse-param {arg2_type *arg2} | |
16 | ||
17 | one may now declare | |
18 | ||
19 | %param {arg1_type *arg1} {arg2_type *arg2} | |
20 | ||
21 | ** Java skeleton improvements | |
0ea583d2 AD |
22 | |
23 | The constants for token names were moved to the Lexer interface. | |
24 | Also, it is possible to add code to the parser's constructors using | |
25 | "%code init" and "%define init_throws". | |
26 | ||
4c6622c2 | 27 | ** Variable api.tokens.prefix |
99c08fb6 | 28 | |
4c6622c2 | 29 | The variable api.tokens.prefix changes the way tokens are identified in |
99c08fb6 AD |
30 | the generated files. This is especially useful to avoid collisions |
31 | with identifiers in the target language. For instance | |
32 | ||
33 | %token FILE for ERROR | |
4c6622c2 | 34 | %define api.tokens.prefix "TOK_" |
99c08fb6 AD |
35 | %% |
36 | start: FILE for ERROR; | |
37 | ||
38 | will generate the definition of the symbols TOK_FILE, TOK_for, and | |
39 | TOK_ERROR in the generated sources. In particular, the scanner must | |
40 | use these prefixed token names, although the grammar itself still | |
41 | uses the short names (as in the sample rule given above). | |
42 | ||
67501061 AD |
43 | ** Variable api.namespace |
44 | ||
45 | The "namespace" variable is renamed "api.namespace". Backward | |
46 | compatibility is ensured, but upgrading is recommended. | |
47 | ||
31b850d2 AD |
48 | ** Variable parse.error |
49 | ||
50 | The variable error controls the verbosity of error messages. The | |
51 | use of the %error-verbose directive is deprecated in favor of | |
52 | %define parse.error "verbose". | |
53 | ||
ca2a6d15 PH |
54 | ** Semantic predicates |
55 | ||
56 | The new, experimental, semantic-predicate feature allows actions of | |
57 | the form %?{ BOOLEAN-EXPRESSION }, which cause syntax errors (as for | |
58 | YYERROR) if the expression evaluates to 0, and are evaluated immediately | |
59 | in GLR parsers, rather than being deferred. The result is that they | |
60 | allow the programmer to prune possible parses based on the values of | |
61 | runtime expressions. | |
62 | ||
50cca368 JD |
63 | * Changes in version 2.5 (????-??-??): |
64 | ||
82f3355e JD |
65 | ** Grammar symbol names can now contain non-initial dashes: |
66 | ||
67 | Consistently with directives (such as %error-verbose) and with | |
68 | %define variables (e.g. push-pull), grammar symbol names may contain | |
69 | dashes in any position except the beginning. This is a GNU | |
70 | extension over POSIX Yacc. Thus, use of this extension is reported | |
71 | by -Wyacc and rejected in Yacc mode (--yacc). | |
72 | ||
f1b238df | 73 | ** Named references: |
66381412 AR |
74 | |
75 | Historically, Yacc and Bison have supported positional references | |
76 | ($n, $$) to allow access to symbol values from inside of semantic | |
77 | actions code. | |
78 | ||
79 | Starting from this version, Bison can also accept named references. | |
80 | When no ambiguity is possible, original symbol names may be used | |
81 | as named references: | |
82 | ||
83 | if_stmt : 'if' cond_expr 'then' then_stmt ';' | |
84 | { $if_stmt = mk_if_stmt($cond_expr, $then_stmt); } | |
85 | ||
86 | In the more common case, explicit names may be declared: | |
87 | ||
88 | stmt[res] : 'if' expr[cond] 'then' stmt[then] 'else' stmt[else] ';' | |
89 | { $res = mk_if_stmt($cond, $then, $else); } | |
90 | ||
5b1ff423 | 91 | Location information is also accessible using @name syntax. When |
66381412 AR |
92 | accessing symbol names containing dots or dashes, explicit bracketing |
93 | ($[sym.1]) must be used. | |
94 | ||
5b1ff423 | 95 | These features are experimental in this version. More user feedback |
66381412 AR |
96 | will help to stabilize them. |
97 | ||
f1b238df | 98 | ** IELR(1) and canonical LR(1): |
eb45ef3b JD |
99 | |
100 | IELR(1) is a minimal LR(1) parser table generation algorithm. That | |
101 | is, given any context-free grammar, IELR(1) generates parser tables | |
102 | with the full language recognition power of canonical LR(1) but with | |
f1b238df JD |
103 | nearly the same number of parser states as LALR(1). This reduction |
104 | in parser states is often an order of magnitude. More importantly, | |
eb45ef3b JD |
105 | because canonical LR(1)'s extra parser states may contain duplicate |
106 | conflicts in the case of non-LR(1) grammars, the number of conflicts | |
107 | for IELR(1) is often an order of magnitude less as well. This can | |
108 | significantly reduce the complexity of developing of a grammar. | |
109 | ||
110 | Bison can now generate IELR(1) and canonical LR(1) parser tables in | |
111 | place of its traditional LALR(1) parser tables, which remain the | |
112 | default. You can specify the type of parser tables in the grammar | |
113 | file with these directives: | |
114 | ||
cf499cff JD |
115 | %define lr.type lalr |
116 | %define lr.type ielr | |
117 | %define lr.type canonical-lr | |
eb45ef3b | 118 | |
7fceb615 JD |
119 | The default-reduction optimization in the parser tables can also be |
120 | adjusted using `%define lr.default-reductions'. For details on both | |
121 | of these features, see the new section `Tuning LR' in the Bison | |
122 | manual. | |
eb45ef3b JD |
123 | |
124 | These features are experimental. More user feedback will help to | |
125 | stabilize them. | |
126 | ||
7fceb615 | 127 | ** LAC (Lookahead Correction) for syntax error handling: |
fcf834f9 JD |
128 | |
129 | Canonical LR, IELR, and LALR can suffer from a couple of problems | |
130 | upon encountering a syntax error. First, the parser might perform | |
131 | additional parser stack reductions before discovering the syntax | |
7fceb615 | 132 | error. Such reductions can perform user semantic actions that are |
fcf834f9 JD |
133 | unexpected because they are based on an invalid token, and they |
134 | cause error recovery to begin in a different syntactic context than | |
135 | the one in which the invalid token was encountered. Second, when | |
7fceb615 JD |
136 | verbose error messages are enabled (with %error-verbose or the |
137 | obsolete `#define YYERROR_VERBOSE'), the expected token list in the | |
138 | syntax error message can both contain invalid tokens and omit valid | |
139 | tokens. | |
fcf834f9 JD |
140 | |
141 | The culprits for the above problems are %nonassoc, default | |
142 | reductions in inconsistent states, and parser state merging. Thus, | |
143 | IELR and LALR suffer the most. Canonical LR can suffer only if | |
144 | %nonassoc is used or if default reductions are enabled for | |
145 | inconsistent states. | |
146 | ||
7fceb615 JD |
147 | LAC is a new mechanism within the parsing algorithm that solves |
148 | these problems for canonical LR, IELR, and LALR without sacrificing | |
149 | %nonassoc, default reductions, or state merging. When LAC is in | |
150 | use, canonical LR and IELR behave almost exactly the same for both | |
151 | syntactically acceptable and syntactically unacceptable input. | |
fcf834f9 JD |
152 | While LALR still does not support the full language-recognition |
153 | power of canonical LR and IELR, LAC at least enables LALR's syntax | |
154 | error handling to correctly reflect LALR's language-recognition | |
155 | power. | |
156 | ||
157 | Currently, LAC is only supported for deterministic parsers in C. | |
158 | You can enable LAC with the following directive: | |
159 | ||
160 | %define parse.lac full | |
161 | ||
7fceb615 JD |
162 | See the new section `LAC' in the Bison manual for additional |
163 | details including a few caveats. | |
fcf834f9 JD |
164 | |
165 | LAC is an experimental feature. More user feedback will help to | |
166 | stabilize it. | |
167 | ||
d397d9f0 | 168 | ** %define improvements: |
cf499cff | 169 | |
f1b238df | 170 | *** Can now be invoked via the command line: |
50cca368 | 171 | |
de5ab940 | 172 | Each of these command-line options |
50cca368 | 173 | |
de5ab940 JD |
174 | -D NAME[=VALUE] |
175 | --define=NAME[=VALUE] | |
176 | ||
177 | -F NAME[=VALUE] | |
178 | --force-define=NAME[=VALUE] | |
50cca368 JD |
179 | |
180 | is equivalent to this grammar file declaration | |
181 | ||
de5ab940 | 182 | %define NAME ["VALUE"] |
50cca368 | 183 | |
de5ab940 JD |
184 | except that the manner in which Bison processes multiple definitions |
185 | for the same NAME differs. Most importantly, -F and --force-define | |
186 | quietly override %define, but -D and --define do not. For further | |
187 | details, see the section "Bison Options" in the Bison manual. | |
50cca368 | 188 | |
f1b238df | 189 | *** Variables renamed: |
67212941 JD |
190 | |
191 | The following %define variables | |
192 | ||
193 | api.push_pull | |
194 | lr.keep_unreachable_states | |
195 | ||
196 | have been renamed to | |
197 | ||
198 | api.push-pull | |
199 | lr.keep-unreachable-states | |
200 | ||
201 | The old names are now deprecated but will be maintained indefinitely | |
202 | for backward compatibility. | |
203 | ||
f1b238df | 204 | *** Values no longer need to be quoted in grammar file: |
cf499cff JD |
205 | |
206 | If a %define value is an identifier, it no longer needs to be placed | |
207 | within quotations marks. For example, | |
208 | ||
209 | %define api.push-pull "push" | |
210 | ||
211 | can be rewritten as | |
212 | ||
213 | %define api.push-pull push | |
214 | ||
d397d9f0 | 215 | *** Unrecognized variables are now errors not warnings. |
cdf3f113 | 216 | |
d397d9f0 JD |
217 | *** Multiple invocations for any variable is now an error not a warning. |
218 | ||
219 | ** Unrecognized %code qualifiers are now errors not warnings. | |
220 | ||
221 | ** Character literals not of length one: | |
222 | ||
223 | Previously, Bison quietly converted all character literals to length | |
224 | one. For example, without warning, Bison interpreted the operators in | |
225 | the following grammar to be the same token: | |
226 | ||
227 | exp: exp '++' | |
228 | | exp '+' exp | |
229 | ; | |
230 | ||
231 | Bison now warns when a character literal is not of length one. In | |
232 | some future release, Bison will start reporting an error instead. | |
233 | ||
234 | ** Destructor calls fixed for lookaheads altered in semantic actions: | |
235 | ||
236 | Previously for deterministic parsers in C, if a user semantic action | |
237 | altered yychar, the parser in some cases used the old yychar value to | |
238 | determine which destructor to call for the lookahead upon a syntax | |
239 | error or upon parser return. This bug has been fixed. | |
240 | ||
241 | ** C++ parsers use YYRHSLOC: | |
242 | ||
243 | Similarly to the C parsers, the C++ parsers now define the YYRHSLOC | |
244 | macro and use it in the default YYLLOC_DEFAULT. You are encouraged | |
245 | to use it. If, for instance, your location structure has "first" | |
246 | and "last" members, instead of | |
247 | ||
248 | # define YYLLOC_DEFAULT(Current, Rhs, N) \ | |
249 | do \ | |
250 | if (N) \ | |
251 | { \ | |
252 | (Current).first = (Rhs)[1].location.first; \ | |
253 | (Current).last = (Rhs)[N].location.last; \ | |
254 | } \ | |
255 | else \ | |
256 | { \ | |
257 | (Current).first = (Current).last = (Rhs)[0].location.last; \ | |
258 | } \ | |
259 | while (false) | |
260 | ||
261 | use: | |
262 | ||
263 | # define YYLLOC_DEFAULT(Current, Rhs, N) \ | |
264 | do \ | |
265 | if (N) \ | |
266 | { \ | |
267 | (Current).first = YYRHSLOC (Rhs, 1).first; \ | |
268 | (Current).last = YYRHSLOC (Rhs, N).last; \ | |
269 | } \ | |
270 | else \ | |
271 | { \ | |
272 | (Current).first = (Current).last = YYRHSLOC (Rhs, 0).last; \ | |
273 | } \ | |
274 | while (false) | |
275 | ||
276 | ** YYLLOC_DEFAULT in C++: | |
277 | ||
278 | The default implementation of YYLLOC_DEFAULT used to be issued in | |
279 | the header file. It is now output in the implementation file, after | |
280 | the user %code sections so that its #ifndef guard does not try to | |
281 | override the user's YYLLOC_DEFAULT if provided. | |
cdf3f113 | 282 | |
f1b238df | 283 | ** YYFAIL now produces warnings and Java parsers no longer implement it: |
4395a9ff JD |
284 | |
285 | YYFAIL has existed for many years as an undocumented feature of | |
286 | deterministic parsers in C generated by Bison. More recently, it was | |
287 | a documented feature of Bison's experimental Java parsers. As | |
288 | promised in Bison 2.4.2's NEWS entry, any appearance of YYFAIL in a | |
289 | semantic action now produces a deprecation warning, and Java parsers | |
290 | no longer implement YYFAIL at all. For further details, including a | |
291 | discussion of how to suppress C preprocessor warnings about YYFAIL | |
292 | being unused, see the Bison 2.4.2 NEWS entry. | |
293 | ||
f1b238df | 294 | ** Temporary hack for adding a semicolon to the user action: |
197b82ba JD |
295 | |
296 | Previously, Bison appended a semicolon to every user action for | |
297 | reductions when the output language defaulted to C (specifically, when | |
298 | neither %yacc, %language, %skeleton, or equivalent command-line | |
299 | options were specified). This allowed actions such as | |
300 | ||
301 | exp: exp "+" exp { $$ = $1 + $3 }; | |
302 | ||
303 | instead of | |
304 | ||
305 | exp: exp "+" exp { $$ = $1 + $3; }; | |
306 | ||
307 | As a first step in removing this misfeature, Bison now issues a | |
308 | warning when it appends a semicolon. Moreover, in cases where Bison | |
309 | cannot easily determine whether a semicolon is needed (for example, an | |
310 | action ending with a cpp directive or a braced compound initializer), | |
311 | it no longer appends one. Thus, the C compiler might now complain | |
312 | about a missing semicolon where it did not before. Future releases of | |
313 | Bison will cease to append semicolons entirely. | |
314 | ||
d2060f06 JD |
315 | ** Verbose syntax error message fixes: |
316 | ||
7fceb615 JD |
317 | When %error-verbose or the obsolete `#define YYERROR_VERBOSE' is |
318 | specified, syntax error messages produced by the generated parser | |
319 | include the unexpected token as well as a list of expected tokens. | |
320 | The effect of %nonassoc on these verbose messages has been corrected | |
321 | in two ways, but a more complete fix requires LAC, described above: | |
d2060f06 JD |
322 | |
323 | *** When %nonassoc is used, there can exist parser states that accept no | |
324 | tokens, and so the parser does not always require a lookahead token | |
325 | in order to detect a syntax error. Because no unexpected token or | |
326 | expected tokens can then be reported, the verbose syntax error | |
327 | message described above is suppressed, and the parser instead | |
328 | reports the simpler message, "syntax error". Previously, this | |
329 | suppression was sometimes erroneously triggered by %nonassoc when a | |
330 | lookahead was actually required. Now verbose messages are | |
331 | suppressed only when all previous lookaheads have already been | |
332 | shifted or discarded. | |
333 | ||
334 | *** Previously, the list of expected tokens erroneously included tokens | |
335 | that would actually induce a syntax error because conflicts for them | |
336 | were resolved with %nonassoc in the current parser state. Such | |
337 | tokens are now properly omitted from the list. | |
338 | ||
339 | *** Expected token lists are still often wrong due to state merging | |
fcf834f9 JD |
340 | (from LALR or IELR) and default reductions, which can both add |
341 | invalid tokens and subtract valid tokens. Canonical LR almost | |
342 | completely fixes this problem by eliminating state merging and | |
343 | default reductions. However, there is one minor problem left even | |
344 | when using canonical LR and even after the fixes above. That is, | |
345 | if the resolution of a conflict with %nonassoc appears in a later | |
346 | parser state than the one at which some syntax error is | |
347 | discovered, the conflicted token is still erroneously included in | |
348 | the expected token list. Bison's new LAC implementation, | |
349 | described above, eliminates this problem and the need for | |
350 | canonical LR. However, LAC is still experimental and is disabled | |
351 | by default. | |
53f036ce | 352 | |
1a33f4f6 JD |
353 | ** Java skeleton fixes: |
354 | ||
355 | *** A location handling bug has been fixed. | |
356 | ||
357 | *** The top element of each of the value stack and location stack is now | |
358 | cleared when popped so that it can be garbage collected. | |
6771a463 | 359 | |
02803d55 JD |
360 | *** Parser traces now print the top element of the stack. |
361 | ||
86408959 JD |
362 | ** -W/--warnings fixes: |
363 | ||
364 | *** Bison now properly recognizes the "no-" versions of categories: | |
365 | ||
366 | For example, given the following command line, Bison now enables all | |
367 | warnings except warnings for incompatibilities with POSIX Yacc: | |
368 | ||
369 | bison -Wall,no-yacc gram.y | |
370 | ||
786743d5 JD |
371 | *** Bison now treats S/R and R/R conflicts like other warnings: |
372 | ||
373 | Previously, conflict reports were independent of Bison's normal | |
374 | warning system. Now, Bison recognizes the warning categories | |
375 | "conflicts-sr" and "conflicts-rr". This change has important | |
376 | consequences for the -W and --warnings command-line options. For | |
377 | example: | |
378 | ||
379 | bison -Wno-conflicts-sr gram.y # S/R conflicts not reported | |
380 | bison -Wno-conflicts-rr gram.y # R/R conflicts not reported | |
381 | bison -Wnone gram.y # no conflicts are reported | |
382 | bison -Werror gram.y # any conflict is an error | |
383 | ||
384 | However, as before, if the %expect or %expect-rr directive is | |
385 | specified, an unexpected number of conflicts is an error, and an | |
386 | expected number of conflicts is not reported, so -W and --warning | |
387 | then have no effect on the conflict report. | |
388 | ||
bf0e44e8 JD |
389 | *** The "none" category no longer disables a preceding "error": |
390 | ||
391 | For example, for the following command line, Bison now reports | |
392 | errors instead of warnings for incompatibilities with POSIX Yacc: | |
393 | ||
394 | bison -Werror,none,yacc gram.y | |
395 | ||
c39014ae JD |
396 | *** The "none" category now disables all Bison warnings. |
397 | ||
398 | Previously, the "none" category disabled only Bison warnings for | |
399 | which there existed a specific -W/--warning category. However, | |
400 | given the following command line, Bison is now guaranteed to | |
401 | suppress all warnings: | |
402 | ||
403 | bison -Wnone gram.y | |
404 | ||
64877e5e | 405 | * Changes in version 2.4.3 (2010-08-05): |
8b9e021f | 406 | |
2bfcac9a JD |
407 | ** Bison now obeys -Werror and --warnings=error for warnings about |
408 | grammar rules that are useless in the parser due to conflicts. | |
409 | ||
8b9e021f JD |
410 | ** Problems with spawning M4 on at least FreeBSD 8 and FreeBSD 9 have |
411 | been fixed. | |
412 | ||
4ad3921d JD |
413 | ** Failures in the test suite for GCC 4.5 have been fixed. |
414 | ||
06cb07d5 JD |
415 | ** Failures in the test suite for some versions of Sun Studio C++ have |
416 | been fixed. | |
417 | ||
9b5049bd JD |
418 | ** Contrary to Bison 2.4.2's NEWS entry, it has been decided that |
419 | warnings about undefined %prec identifiers will not be converted to | |
420 | errors in Bison 2.5. They will remain warnings, which should be | |
421 | sufficient for POSIX while avoiding backward compatibility issues. | |
422 | ||
93d7dde9 JD |
423 | ** Minor documentation fixes. |
424 | ||
e19a049c | 425 | * Changes in version 2.4.2 (2010-03-20): |
74553c98 | 426 | |
f39ab286 JD |
427 | ** Some portability problems that resulted in failures and livelocks |
428 | in the test suite on some versions of at least Solaris, AIX, HP-UX, | |
e19a049c JD |
429 | RHEL4, and Tru64 have been addressed. As a result, fatal Bison |
430 | errors should no longer cause M4 to report a broken pipe on the | |
f39ab286 JD |
431 | affected platforms. |
432 | ||
8bb3a2e7 JD |
433 | ** `%prec IDENTIFIER' requires IDENTIFIER to be defined separately. |
434 | ||
435 | POSIX specifies that an error be reported for any identifier that does | |
436 | not appear on the LHS of a grammar rule and that is not defined by | |
437 | %token, %left, %right, or %nonassoc. Bison 2.3b and later lost this | |
438 | error report for the case when an identifier appears only after a | |
439 | %prec directive. It is now restored. However, for backward | |
440 | compatibility with recent Bison releases, it is only a warning for | |
441 | now. In Bison 2.5 and later, it will return to being an error. | |
9b5049bd JD |
442 | [Between the 2.4.2 and 2.4.3 releases, it was decided that this |
443 | warning will not be converted to an error in Bison 2.5.] | |
8bb3a2e7 | 444 | |
d8911864 EB |
445 | ** Detection of GNU M4 1.4.6 or newer during configure is improved. |
446 | ||
a603c6e0 JD |
447 | ** Warnings from gcc's -Wundef option about undefined YYENABLE_NLS, |
448 | YYLTYPE_IS_TRIVIAL, and __STRICT_ANSI__ in C/C++ parsers are now | |
449 | avoided. | |
c938d650 | 450 | |
98a345a2 JD |
451 | ** %code is now a permanent feature. |
452 | ||
453 | A traditional Yacc prologue directive is written in the form: | |
454 | ||
455 | %{CODE%} | |
456 | ||
457 | To provide a more flexible alternative, Bison 2.3b introduced the | |
458 | %code directive with the following forms for C/C++: | |
459 | ||
460 | %code {CODE} | |
461 | %code requires {CODE} | |
462 | %code provides {CODE} | |
463 | %code top {CODE} | |
464 | ||
465 | These forms are now considered permanent features of Bison. See the | |
466 | %code entries in the section "Bison Declaration Summary" in the Bison | |
467 | manual for a summary of their functionality. See the section | |
468 | "Prologue Alternatives" for a detailed discussion including the | |
469 | advantages of %code over the traditional Yacc prologue directive. | |
470 | ||
471 | Bison's Java feature as a whole including its current usage of %code | |
472 | is still considered experimental. | |
473 | ||
1625df5b JD |
474 | ** YYFAIL is deprecated and will eventually be removed. |
475 | ||
476 | YYFAIL has existed for many years as an undocumented feature of | |
477 | deterministic parsers in C generated by Bison. Previously, it was | |
478 | documented for Bison's experimental Java parsers. YYFAIL is no longer | |
479 | documented for Java parsers and is formally deprecated in both cases. | |
480 | Users are strongly encouraged to migrate to YYERROR, which is | |
481 | specified by POSIX. | |
482 | ||
483 | Like YYERROR, you can invoke YYFAIL from a semantic action in order to | |
484 | induce a syntax error. The most obvious difference from YYERROR is | |
485 | that YYFAIL will automatically invoke yyerror to report the syntax | |
486 | error so that you don't have to. However, there are several other | |
487 | subtle differences between YYERROR and YYFAIL, and YYFAIL suffers from | |
488 | inherent flaws when %error-verbose or `#define YYERROR_VERBOSE' is | |
489 | used. For a more detailed discussion, see: | |
490 | ||
491 | http://lists.gnu.org/archive/html/bison-patches/2009-12/msg00024.html | |
492 | ||
493 | The upcoming Bison 2.5 will remove YYFAIL from Java parsers, but | |
494 | deterministic parsers in C will continue to implement it. However, | |
495 | because YYFAIL is already flawed, it seems futile to try to make new | |
496 | Bison features compatible with it. Thus, during parser generation, | |
497 | Bison 2.5 will produce a warning whenever it discovers YYFAIL in a | |
498 | rule action. In a later release, YYFAIL will be disabled for | |
499 | %error-verbose and `#define YYERROR_VERBOSE'. Eventually, YYFAIL will | |
500 | be removed altogether. | |
501 | ||
502 | There exists at least one case where Bison 2.5's YYFAIL warning will | |
503 | be a false positive. Some projects add phony uses of YYFAIL and other | |
504 | Bison-defined macros for the sole purpose of suppressing C | |
505 | preprocessor warnings (from GCC cpp's -Wunused-macros, for example). | |
506 | To avoid Bison's future warning, such YYFAIL uses can be moved to the | |
507 | epilogue (that is, after the second `%%') in the Bison input file. In | |
508 | this release (2.4.2), Bison already generates its own code to suppress | |
509 | C preprocessor warnings for YYFAIL, so projects can remove their own | |
510 | phony uses of YYFAIL if compatibility with Bison releases prior to | |
511 | 2.4.2 is not necessary. | |
512 | ||
2755de8f AD |
513 | ** Internationalization. |
514 | ||
515 | Fix a regression introduced in Bison 2.4: Under some circumstances, | |
516 | message translations were not installed although supported by the | |
517 | host system. | |
518 | ||
74553c98 | 519 | * Changes in version 2.4.1 (2008-12-11): |
c9ba9e59 | 520 | |
0ea583d2 AD |
521 | ** In the GLR defines file, unexpanded M4 macros in the yylval and yylloc |
522 | declarations have been fixed. | |
1979121c | 523 | |
0ea583d2 AD |
524 | ** Temporary hack for adding a semicolon to the user action. |
525 | ||
526 | Bison used to prepend a trailing semicolon at the end of the user | |
527 | action for reductions. This allowed actions such as | |
528 | ||
529 | exp: exp "+" exp { $$ = $1 + $3 }; | |
530 | ||
531 | instead of | |
532 | ||
533 | exp: exp "+" exp { $$ = $1 + $3; }; | |
534 | ||
535 | Some grammars still depend on this `feature'. Bison 2.4.1 restores | |
536 | the previous behavior in the case of C output (specifically, when | |
537 | neither %language or %skeleton or equivalent command-line options | |
538 | are used) to leave more time for grammars depending on the old | |
539 | behavior to be adjusted. Future releases of Bison will disable this | |
540 | feature. | |
541 | ||
542 | ** A few minor improvements to the Bison manual. | |
c9ba9e59 | 543 | |
402b123d | 544 | * Changes in version 2.4 (2008-11-02): |
7bd1665a | 545 | |
402b123d | 546 | ** %language is an experimental feature. |
ed4d67dc JD |
547 | |
548 | We first introduced this feature in test release 2.3b as a cleaner | |
549 | alternative to %skeleton. Since then, we have discussed the possibility of | |
550 | modifying its effect on Bison's output file names. Thus, in this release, | |
551 | we consider %language to be an experimental feature that will likely evolve | |
552 | in future releases. | |
7bd1665a | 553 | |
402b123d | 554 | ** Forward compatibility with GNU M4 has been improved. |
241fda7a | 555 | |
402b123d | 556 | ** Several bugs in the C++ skeleton and the experimental Java skeleton have been |
241fda7a JD |
557 | fixed. |
558 | ||
402b123d | 559 | * Changes in version 2.3b (2008-05-27): |
35fe0834 | 560 | |
402b123d | 561 | ** The quotes around NAME that used to be required in the following directive |
d9df47b6 JD |
562 | are now deprecated: |
563 | ||
564 | %define NAME "VALUE" | |
565 | ||
402b123d | 566 | ** The directive `%pure-parser' is now deprecated in favor of: |
d9df47b6 JD |
567 | |
568 | %define api.pure | |
569 | ||
570 | which has the same effect except that Bison is more careful to warn about | |
571 | unreasonable usage in the latter case. | |
572 | ||
402b123d | 573 | ** Push Parsing |
c373bf8b JD |
574 | |
575 | Bison can now generate an LALR(1) parser in C with a push interface. That | |
ef1b4273 | 576 | is, instead of invoking `yyparse', which pulls tokens from `yylex', you can |
c373bf8b JD |
577 | push one token at a time to the parser using `yypush_parse', which will |
578 | return to the caller after processing each token. By default, the push | |
579 | interface is disabled. Either of the following directives will enable it: | |
580 | ||
581 | %define api.push_pull "push" // Just push; does not require yylex. | |
582 | %define api.push_pull "both" // Push and pull; requires yylex. | |
583 | ||
584 | See the new section `A Push Parser' in the Bison manual for details. | |
585 | ||
59da312b JD |
586 | The current push parsing interface is experimental and may evolve. More user |
587 | feedback will help to stabilize it. | |
588 | ||
402b123d | 589 | ** The -g and --graph options now output graphs in Graphviz DOT format, |
8e55b3aa JD |
590 | not VCG format. Like --graph, -g now also takes an optional FILE argument |
591 | and thus cannot be bundled with other short options. | |
c373bf8b | 592 | |
402b123d | 593 | ** Java |
59da312b JD |
594 | |
595 | Bison can now generate an LALR(1) parser in Java. The skeleton is | |
596 | `data/lalr1.java'. Consider using the new %language directive instead of | |
597 | %skeleton to select it. | |
598 | ||
599 | See the new section `Java Parsers' in the Bison manual for details. | |
600 | ||
601 | The current Java interface is experimental and may evolve. More user | |
602 | feedback will help to stabilize it. | |
603 | ||
402b123d | 604 | ** %language |
59da312b JD |
605 | |
606 | This new directive specifies the programming language of the generated | |
d43f77e7 PB |
607 | parser, which can be C (the default), C++, or Java. Besides the skeleton |
608 | that Bison uses, the directive affects the names of the generated files if | |
609 | the grammar file's name ends in ".y". | |
59da312b | 610 | |
402b123d | 611 | ** XML Automaton Report |
59da312b JD |
612 | |
613 | Bison can now generate an XML report of the LALR(1) automaton using the new | |
614 | `--xml' option. The current XML schema is experimental and may evolve. More | |
615 | user feedback will help to stabilize it. | |
c373bf8b | 616 | |
402b123d | 617 | ** The grammar file may now specify the name of the parser header file using |
c373bf8b JD |
618 | %defines. For example: |
619 | ||
620 | %defines "parser.h" | |
621 | ||
402b123d | 622 | ** When reporting useless rules, useless nonterminals, and unused terminals, |
d80fb37a JD |
623 | Bison now employs the terms "useless in grammar" instead of "useless", |
624 | "useless in parser" instead of "never reduced", and "unused in grammar" | |
625 | instead of "unused". | |
cff03fb2 | 626 | |
402b123d | 627 | ** Unreachable State Removal |
c373bf8b JD |
628 | |
629 | Previously, Bison sometimes generated parser tables containing unreachable | |
31984206 JD |
630 | states. A state can become unreachable during conflict resolution if Bison |
631 | disables a shift action leading to it from a predecessor state. Bison now: | |
75ad86ee JD |
632 | |
633 | 1. Removes unreachable states. | |
634 | ||
635 | 2. Does not report any conflicts that appeared in unreachable states. | |
636 | WARNING: As a result, you may need to update %expect and %expect-rr | |
637 | directives in existing grammar files. | |
638 | ||
639 | 3. For any rule used only in such states, Bison now reports the rule as | |
cff03fb2 | 640 | "useless in parser due to conflicts". |
75ad86ee | 641 | |
31984206 JD |
642 | This feature can be disabled with the following directive: |
643 | ||
644 | %define lr.keep_unreachable_states | |
645 | ||
646 | See the %define entry in the `Bison Declaration Summary' in the Bison manual | |
647 | for further discussion. | |
648 | ||
402b123d | 649 | ** Lookahead Set Correction in the `.output' Report |
b1cc23c4 | 650 | |
c373bf8b | 651 | When instructed to generate a `.output' file including lookahead sets |
88c78747 JD |
652 | (using `--report=lookahead', for example), Bison now prints each reduction's |
653 | lookahead set only next to the associated state's one item that (1) is | |
654 | associated with the same rule as the reduction and (2) has its dot at the end | |
655 | of its RHS. Previously, Bison also erroneously printed the lookahead set | |
656 | next to all of the state's other items associated with the same rule. This | |
657 | bug affected only the `.output' file and not the generated parser source | |
658 | code. | |
659 | ||
402b123d | 660 | ** --report-file=FILE is a new option to override the default `.output' file |
59da312b | 661 | name. |
1bb2bd75 | 662 | |
402b123d | 663 | ** The `=' that used to be required in the following directives is now |
02975b9a JD |
664 | deprecated: |
665 | ||
666 | %file-prefix "parser" | |
667 | %name-prefix "c_" | |
668 | %output "parser.c" | |
669 | ||
402b123d | 670 | ** An Alternative to `%{...%}' -- `%code QUALIFIER {CODE}' |
c373bf8b JD |
671 | |
672 | Bison 2.3a provided a new set of directives as a more flexible alternative to | |
8e0a5e9e JD |
673 | the traditional Yacc prologue blocks. Those have now been consolidated into |
674 | a single %code directive with an optional qualifier field, which identifies | |
675 | the purpose of the code and thus the location(s) where Bison should generate | |
676 | it: | |
677 | ||
16dc6a9e JD |
678 | 1. `%code {CODE}' replaces `%after-header {CODE}' |
679 | 2. `%code requires {CODE}' replaces `%start-header {CODE}' | |
680 | 3. `%code provides {CODE}' replaces `%end-header {CODE}' | |
681 | 4. `%code top {CODE}' replaces `%before-header {CODE}' | |
8e0a5e9e | 682 | |
61fee93e JD |
683 | See the %code entries in section `Bison Declaration Summary' in the Bison |
684 | manual for a summary of the new functionality. See the new section `Prologue | |
8e0a5e9e JD |
685 | Alternatives' for a detailed discussion including the advantages of %code |
686 | over the traditional Yacc prologues. | |
687 | ||
688 | The prologue alternatives are experimental. More user feedback will help to | |
689 | determine whether they should become permanent features. | |
690 | ||
402b123d | 691 | ** Revised warning: unset or unused mid-rule values |
17bd8a73 JD |
692 | |
693 | Since Bison 2.2, Bison has warned about mid-rule values that are set but not | |
694 | used within any of the actions of the parent rule. For example, Bison warns | |
695 | about unused $2 in: | |
696 | ||
697 | exp: '1' { $$ = 1; } '+' exp { $$ = $1 + $4; }; | |
698 | ||
699 | Now, Bison also warns about mid-rule values that are used but not set. For | |
700 | example, Bison warns about unset $$ in the mid-rule action in: | |
701 | ||
702 | exp: '1' { $1 = 1; } '+' exp { $$ = $2 + $4; }; | |
703 | ||
704 | However, Bison now disables both of these warnings by default since they | |
705 | sometimes prove to be false alarms in existing grammars employing the Yacc | |
706 | constructs $0 or $-N (where N is some positive integer). | |
707 | ||
59da312b | 708 | To enable these warnings, specify the option `--warnings=midrule-values' or |
17bd8a73 JD |
709 | `-W', which is a synonym for `--warnings=all'. |
710 | ||
402b123d | 711 | ** Default %destructor or %printer with `<*>' or `<>' |
c373bf8b JD |
712 | |
713 | Bison now recognizes two separate kinds of default %destructor's and | |
12e35840 JD |
714 | %printer's: |
715 | ||
716 | 1. Place `<*>' in a %destructor/%printer symbol list to define a default | |
717 | %destructor/%printer for all grammar symbols for which you have formally | |
718 | declared semantic type tags. | |
719 | ||
3ebecc24 | 720 | 2. Place `<>' in a %destructor/%printer symbol list to define a default |
12e35840 JD |
721 | %destructor/%printer for all grammar symbols without declared semantic |
722 | type tags. | |
723 | ||
724 | Bison no longer supports the `%symbol-default' notation from Bison 2.3a. | |
3ebecc24 | 725 | `<*>' and `<>' combined achieve the same effect with one exception: Bison no |
12e35840 JD |
726 | longer applies any %destructor to a mid-rule value if that mid-rule value is |
727 | not actually ever referenced using either $$ or $n in a semantic action. | |
728 | ||
85894313 JD |
729 | The default %destructor's and %printer's are experimental. More user |
730 | feedback will help to determine whether they should become permanent | |
731 | features. | |
732 | ||
12e35840 JD |
733 | See the section `Freeing Discarded Symbols' in the Bison manual for further |
734 | details. | |
735 | ||
402b123d | 736 | ** %left, %right, and %nonassoc can now declare token numbers. This is required |
ab7f29f8 JD |
737 | by POSIX. However, see the end of section `Operator Precedence' in the Bison |
738 | manual for a caveat concerning the treatment of literal strings. | |
739 | ||
402b123d | 740 | ** The nonfunctional --no-parser, -n, and %no-parser options have been |
b1cc23c4 JD |
741 | completely removed from Bison. |
742 | ||
402b123d | 743 | * Changes in version 2.3a, 2006-09-13: |
742e4900 | 744 | |
402b123d | 745 | ** Instead of %union, you can define and use your own union type |
ddc8ede1 PE |
746 | YYSTYPE if your grammar contains at least one <type> tag. |
747 | Your YYSTYPE need not be a macro; it can be a typedef. | |
748 | This change is for compatibility with other Yacc implementations, | |
749 | and is required by POSIX. | |
750 | ||
402b123d | 751 | ** Locations columns and lines start at 1. |
cd48d21d AD |
752 | In accordance with the GNU Coding Standards and Emacs. |
753 | ||
402b123d | 754 | ** You may now declare per-type and default %destructor's and %printer's: |
ec5479ce JD |
755 | |
756 | For example: | |
757 | ||
b2a0b7ca JD |
758 | %union { char *string; } |
759 | %token <string> STRING1 | |
760 | %token <string> STRING2 | |
761 | %type <string> string1 | |
762 | %type <string> string2 | |
763 | %union { char character; } | |
764 | %token <character> CHR | |
765 | %type <character> chr | |
766 | %destructor { free ($$); } %symbol-default | |
767 | %destructor { free ($$); printf ("%d", @$.first_line); } STRING1 string1 | |
768 | %destructor { } <character> | |
769 | ||
770 | guarantees that, when the parser discards any user-defined symbol that has a | |
771 | semantic type tag other than `<character>', it passes its semantic value to | |
772 | `free'. However, when the parser discards a `STRING1' or a `string1', it | |
773 | also prints its line number to `stdout'. It performs only the second | |
774 | `%destructor' in this case, so it invokes `free' only once. | |
ec5479ce | 775 | |
85894313 JD |
776 | [Although we failed to mention this here in the 2.3a release, the default |
777 | %destructor's and %printer's were experimental, and they were rewritten in | |
778 | future versions.] | |
779 | ||
402b123d | 780 | ** Except for LALR(1) parsers in C with POSIX Yacc emulation enabled (with `-y', |
b931235e JD |
781 | `--yacc', or `%yacc'), Bison no longer generates #define statements for |
782 | associating token numbers with token names. Removing the #define statements | |
783 | helps to sanitize the global namespace during preprocessing, but POSIX Yacc | |
784 | requires them. Bison still generates an enum for token names in all cases. | |
785 | ||
402b123d | 786 | ** Handling of traditional Yacc prologue blocks is now more consistent but |
34f98f46 | 787 | potentially incompatible with previous releases of Bison. |
9bc0dd67 JD |
788 | |
789 | As before, you declare prologue blocks in your grammar file with the | |
790 | `%{ ... %}' syntax. To generate the pre-prologue, Bison concatenates all | |
34f98f46 JD |
791 | prologue blocks that you've declared before the first %union. To generate |
792 | the post-prologue, Bison concatenates all prologue blocks that you've | |
ddc8ede1 | 793 | declared after the first %union. |
9bc0dd67 | 794 | |
34f98f46 | 795 | Previous releases of Bison inserted the pre-prologue into both the header |
9bc0dd67 JD |
796 | file and the code file in all cases except for LALR(1) parsers in C. In the |
797 | latter case, Bison inserted it only into the code file. For parsers in C++, | |
798 | the point of insertion was before any token definitions (which associate | |
799 | token numbers with names). For parsers in C, the point of insertion was | |
800 | after the token definitions. | |
801 | ||
802 | Now, Bison never inserts the pre-prologue into the header file. In the code | |
803 | file, it always inserts it before the token definitions. | |
804 | ||
402b123d | 805 | ** Bison now provides a more flexible alternative to the traditional Yacc |
34f98f46 JD |
806 | prologue blocks: %before-header, %start-header, %end-header, and |
807 | %after-header. | |
808 | ||
809 | For example, the following declaration order in the grammar file reflects the | |
810 | order in which Bison will output these code blocks. However, you are free to | |
811 | declare these code blocks in your grammar file in whatever order is most | |
812 | convenient for you: | |
813 | ||
814 | %before-header { | |
815 | /* Bison treats this block like a pre-prologue block: it inserts it into | |
816 | * the code file before the contents of the header file. It does *not* | |
817 | * insert it into the header file. This is a good place to put | |
818 | * #include's that you want at the top of your code file. A common | |
819 | * example is `#include "system.h"'. */ | |
820 | } | |
821 | %start-header { | |
822 | /* Bison inserts this block into both the header file and the code file. | |
823 | * In both files, the point of insertion is before any Bison-generated | |
824 | * token, semantic type, location type, and class definitions. This is a | |
825 | * good place to define %union dependencies, for example. */ | |
9bc0dd67 JD |
826 | } |
827 | %union { | |
34f98f46 JD |
828 | /* Unlike the traditional Yacc prologue blocks, the output order for the |
829 | * new %*-header blocks is not affected by their declaration position | |
830 | * relative to any %union in the grammar file. */ | |
9bc0dd67 | 831 | } |
34f98f46 JD |
832 | %end-header { |
833 | /* Bison inserts this block into both the header file and the code file. | |
834 | * In both files, the point of insertion is after the Bison-generated | |
835 | * definitions. This is a good place to declare or define public | |
836 | * functions or data structures that depend on the Bison-generated | |
837 | * definitions. */ | |
9bc0dd67 | 838 | } |
34f98f46 JD |
839 | %after-header { |
840 | /* Bison treats this block like a post-prologue block: it inserts it into | |
841 | * the code file after the contents of the header file. It does *not* | |
842 | * insert it into the header file. This is a good place to declare or | |
843 | * define internal functions or data structures that depend on the | |
844 | * Bison-generated definitions. */ | |
845 | } | |
846 | ||
847 | If you have multiple occurrences of any one of the above declarations, Bison | |
848 | will concatenate the contents in declaration order. | |
9bc0dd67 | 849 | |
85894313 JD |
850 | [Although we failed to mention this here in the 2.3a release, the prologue |
851 | alternatives were experimental, and they were rewritten in future versions.] | |
852 | ||
402b123d | 853 | ** The option `--report=look-ahead' has been changed to `--report=lookahead'. |
9e6e7ed2 PE |
854 | The old spelling still works, but is not documented and may be removed |
855 | in a future release. | |
742e4900 | 856 | |
402b123d | 857 | * Changes in version 2.3, 2006-06-05: |
4ad3ed84 | 858 | |
402b123d | 859 | ** GLR grammars should now use `YYRECOVERING ()' instead of `YYRECOVERING', |
4ad3ed84 PE |
860 | for compatibility with LALR(1) grammars. |
861 | ||
402b123d | 862 | ** It is now documented that any definition of YYSTYPE or YYLTYPE should |
4ad3ed84 PE |
863 | be to a type name that does not contain parentheses or brackets. |
864 | ||
402b123d | 865 | * Changes in version 2.2, 2006-05-19: |
193d7c70 | 866 | |
402b123d | 867 | ** The distribution terms for all Bison-generated parsers now permit |
193d7c70 PE |
868 | using the parsers in nonfree programs. Previously, this permission |
869 | was granted only for Bison-generated LALR(1) parsers in C. | |
5f4236a0 | 870 | |
402b123d | 871 | ** %name-prefix changes the namespace name in C++ outputs. |
aa08666d | 872 | |
402b123d | 873 | ** The C++ parsers export their token_type. |
5f4236a0 | 874 | |
402b123d | 875 | ** Bison now allows multiple %union declarations, and concatenates |
d6ca7905 PE |
876 | their contents together. |
877 | ||
402b123d | 878 | ** New warning: unused values |
4d7bc38c PE |
879 | Right-hand side symbols whose values are not used are reported, |
880 | if the symbols have destructors. For instance: | |
affac613 | 881 | |
8f3596a6 | 882 | exp: exp "?" exp ":" exp { $1 ? $1 : $3; } |
721be13c PE |
883 | | exp "+" exp |
884 | ; | |
affac613 | 885 | |
8f3596a6 AD |
886 | will trigger a warning about $$ and $5 in the first rule, and $3 in |
887 | the second ($1 is copied to $$ by the default rule). This example | |
4e26c69e | 888 | most likely contains three errors, and could be rewritten as: |
affac613 | 889 | |
4e26c69e PE |
890 | exp: exp "?" exp ":" exp |
891 | { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); } | |
892 | | exp "+" exp | |
893 | { $$ = $1 ? $1 : $3; if ($1) free ($3); } | |
721be13c | 894 | ; |
affac613 | 895 | |
4e26c69e PE |
896 | However, if the original actions were really intended, memory leaks |
897 | and all, the warnings can be suppressed by letting Bison believe the | |
898 | values are used, e.g.: | |
721be13c | 899 | |
8f3596a6 | 900 | exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); } |
721be13c PE |
901 | | exp "+" exp { $$ = $1; (void) $3; } |
902 | ; | |
903 | ||
84866159 AD |
904 | If there are mid-rule actions, the warning is issued if no action |
905 | uses it. The following triggers no warning: $1 and $3 are used. | |
906 | ||
907 | exp: exp { push ($1); } '+' exp { push ($3); sum (); }; | |
908 | ||
721be13c PE |
909 | The warning is intended to help catching lost values and memory leaks. |
910 | If a value is ignored, its associated memory typically is not reclaimed. | |
affac613 | 911 | |
402b123d | 912 | ** %destructor vs. YYABORT, YYACCEPT, and YYERROR. |
9d9b8b70 PE |
913 | Destructors are now called when user code invokes YYABORT, YYACCEPT, |
914 | and YYERROR, for all objects on the stack, other than objects | |
915 | corresponding to the right-hand side of the current rule. | |
a85284cf | 916 | |
402b123d | 917 | ** %expect, %expect-rr |
035aa4a0 PE |
918 | Incorrect numbers of expected conflicts are now actual errors, |
919 | instead of warnings. | |
920 | ||
402b123d | 921 | ** GLR, YACC parsers. |
4e26c69e PE |
922 | The %parse-params are available in the destructors (and the |
923 | experimental printers) as per the documentation. | |
4b367315 | 924 | |
402b123d | 925 | ** Bison now warns if it finds a stray `$' or `@' in an action. |
ad6a9b97 | 926 | |
402b123d | 927 | ** %require "VERSION" |
4e26c69e PE |
928 | This specifies that the grammar file depends on features implemented |
929 | in Bison version VERSION or higher. | |
b50d2359 | 930 | |
402b123d | 931 | ** lalr1.cc: The token and value types are now class members. |
e14d0ab6 AD |
932 | The tokens were defined as free form enums and cpp macros. YYSTYPE |
933 | was defined as a free form union. They are now class members: | |
fb9712a9 AD |
934 | tokens are enumerations of the `yy::parser::token' struct, and the |
935 | semantic values have the `yy::parser::semantic_type' type. | |
936 | ||
937 | If you do not want or can update to this scheme, the directive | |
938 | `%define "global_tokens_and_yystype" "1"' triggers the global | |
b50d2359 AD |
939 | definition of tokens and YYSTYPE. This change is suitable both |
940 | for previous releases of Bison, and this one. | |
fb9712a9 | 941 | |
b50d2359 | 942 | If you wish to update, then make sure older version of Bison will |
ab8d9dc5 | 943 | fail using `%require "2.2"'. |
fb9712a9 | 944 | |
402b123d | 945 | ** DJGPP support added. |
193d7c70 | 946 | \f |
402b123d | 947 | * Changes in version 2.1, 2005-09-16: |
1ce59070 | 948 | |
402b123d | 949 | ** The C++ lalr1.cc skeleton supports %lex-param. |
e14d0ab6 | 950 | |
402b123d | 951 | ** Bison-generated parsers now support the translation of diagnostics like |
baf785db PE |
952 | "syntax error" into languages other than English. The default |
953 | language is still English. For details, please see the new | |
0410a6e0 PE |
954 | Internationalization section of the Bison manual. Software |
955 | distributors should also see the new PACKAGING file. Thanks to | |
956 | Bruno Haible for this new feature. | |
1ce59070 | 957 | |
402b123d | 958 | ** Wording in the Bison-generated parsers has been changed slightly to |
1a059451 PE |
959 | simplify translation. In particular, the message "memory exhausted" |
960 | has replaced "parser stack overflow", as the old message was not | |
961 | always accurate for modern Bison-generated parsers. | |
962 | ||
402b123d | 963 | ** Destructors are now called when the parser aborts, for all symbols left |
258b75ca PE |
964 | behind on the stack. Also, the start symbol is now destroyed after a |
965 | successful parse. In both cases, the behavior was formerly inconsistent. | |
966 | ||
402b123d | 967 | ** When generating verbose diagnostics, Bison-generated parsers no longer |
72f000b0 PE |
968 | quote the literal strings associated with tokens. For example, for |
969 | a syntax error associated with '%token NUM "number"' they might | |
970 | print 'syntax error, unexpected number' instead of 'syntax error, | |
971 | unexpected "number"'. | |
193d7c70 | 972 | \f |
402b123d | 973 | * Changes in version 2.0, 2004-12-25: |
efeed023 | 974 | |
402b123d | 975 | ** Possibly-incompatible changes |
d7e14fc0 | 976 | |
82de6b0d PE |
977 | - Bison-generated parsers no longer default to using the alloca function |
978 | (when available) to extend the parser stack, due to widespread | |
979 | problems in unchecked stack-overflow detection. You can "#define | |
980 | YYSTACK_USE_ALLOCA 1" to require the use of alloca, but please read | |
981 | the manual to determine safe values for YYMAXDEPTH in that case. | |
8dd162d3 | 982 | |
82de6b0d PE |
983 | - Error token location. |
984 | During error recovery, the location of the syntax error is updated | |
985 | to cover the whole sequence covered by the error token: it includes | |
986 | the shifted symbols thrown away during the first part of the error | |
987 | recovery, and the lookahead rejected during the second part. | |
18d192f0 | 988 | |
82de6b0d PE |
989 | - Semicolon changes: |
990 | . Stray semicolons are no longer allowed at the start of a grammar. | |
991 | . Semicolons are now required after in-grammar declarations. | |
e342c3be | 992 | |
82de6b0d PE |
993 | - Unescaped newlines are no longer allowed in character constants or |
994 | string literals. They were never portable, and GCC 3.4.0 has | |
995 | dropped support for them. Better diagnostics are now generated if | |
996 | forget a closing quote. | |
8dd162d3 | 997 | |
82de6b0d | 998 | - NUL bytes are no longer allowed in Bison string literals, unfortunately. |
f74b6f91 | 999 | |
402b123d | 1000 | ** New features |
1452af69 | 1001 | |
82de6b0d | 1002 | - GLR grammars now support locations. |
4febdd96 | 1003 | |
82de6b0d PE |
1004 | - New directive: %initial-action. |
1005 | This directive allows the user to run arbitrary code (including | |
1006 | initializing @$) from yyparse before parsing starts. | |
1452af69 | 1007 | |
82de6b0d PE |
1008 | - A new directive "%expect-rr N" specifies the expected number of |
1009 | reduce/reduce conflicts in GLR parsers. | |
1452af69 | 1010 | |
82de6b0d PE |
1011 | - %token numbers can now be hexadecimal integers, e.g., `%token FOO 0x12d'. |
1012 | This is a GNU extension. | |
4febdd96 | 1013 | |
82de6b0d | 1014 | - The option `--report=lookahead' was changed to `--report=look-ahead'. |
9e6e7ed2 | 1015 | [However, this was changed back after 2.3.] |
1452af69 | 1016 | |
82de6b0d | 1017 | - Experimental %destructor support has been added to lalr1.cc. |
1452af69 | 1018 | |
82de6b0d PE |
1019 | - New configure option --disable-yacc, to disable installation of the |
1020 | yacc command and -ly library introduced in 1.875 for POSIX conformance. | |
6040d338 | 1021 | |
402b123d | 1022 | ** Bug fixes |
d5a3fe37 | 1023 | |
82de6b0d PE |
1024 | - For now, %expect-count violations are now just warnings, not errors. |
1025 | This is for compatibility with Bison 1.75 and earlier (when there are | |
1026 | reduce/reduce conflicts) and with Bison 1.30 and earlier (when there | |
1027 | are too many or too few shift/reduce conflicts). However, in future | |
1028 | versions of Bison we plan to improve the %expect machinery so that | |
1029 | these violations will become errors again. | |
3473d0f8 | 1030 | |
82de6b0d PE |
1031 | - Within Bison itself, numbers (e.g., goto numbers) are no longer |
1032 | arbitrarily limited to 16-bit counts. | |
d600ee67 | 1033 | |
82de6b0d | 1034 | - Semicolons are now allowed before "|" in grammar rules, as POSIX requires. |
d600ee67 | 1035 | \f |
402b123d | 1036 | * Changes in version 1.875, 2003-01-01: |
963fcc17 | 1037 | |
402b123d | 1038 | ** The documentation license has been upgraded to version 1.2 |
dc546b0f | 1039 | of the GNU Free Documentation License. |
75eb3bc4 | 1040 | |
402b123d | 1041 | ** syntax error processing |
75eb3bc4 | 1042 | |
dc546b0f PE |
1043 | - In Yacc-style parsers YYLLOC_DEFAULT is now used to compute error |
1044 | locations too. This fixes bugs in error-location computation. | |
75eb3bc4 | 1045 | |
dc546b0f PE |
1046 | - %destructor |
1047 | It is now possible to reclaim the memory associated to symbols | |
1048 | discarded during error recovery. This feature is still experimental. | |
20daca06 | 1049 | |
dc546b0f PE |
1050 | - %error-verbose |
1051 | This new directive is preferred over YYERROR_VERBOSE. | |
74724a70 | 1052 | |
dc546b0f PE |
1053 | - #defining yyerror to steal internal variables is discouraged. |
1054 | It is not guaranteed to work forever. | |
d1de5372 | 1055 | |
402b123d | 1056 | ** POSIX conformance |
d1de5372 | 1057 | |
dc546b0f PE |
1058 | - Semicolons are once again optional at the end of grammar rules. |
1059 | This reverts to the behavior of Bison 1.33 and earlier, and improves | |
1060 | compatibility with Yacc. | |
74724a70 | 1061 | |
dc546b0f PE |
1062 | - `parse error' -> `syntax error' |
1063 | Bison now uniformly uses the term `syntax error'; formerly, the code | |
1064 | and manual sometimes used the term `parse error' instead. POSIX | |
1065 | requires `syntax error' in diagnostics, and it was thought better to | |
1066 | be consistent. | |
74724a70 | 1067 | |
dc546b0f PE |
1068 | - The documentation now emphasizes that yylex and yyerror must be |
1069 | declared before use. C99 requires this. | |
d1de5372 | 1070 | |
dc546b0f PE |
1071 | - Bison now parses C99 lexical constructs like UCNs and |
1072 | backslash-newline within C escape sequences, as POSIX 1003.1-2001 requires. | |
d1de5372 | 1073 | |
dc546b0f PE |
1074 | - File names are properly escaped in C output. E.g., foo\bar.y is |
1075 | output as "foo\\bar.y". | |
6780ca7a | 1076 | |
dc546b0f PE |
1077 | - Yacc command and library now available |
1078 | The Bison distribution now installs a `yacc' command, as POSIX requires. | |
1079 | Also, Bison now installs a small library liby.a containing | |
1080 | implementations of Yacc-compatible yyerror and main functions. | |
1081 | This library is normally not useful, but POSIX requires it. | |
6e649e65 | 1082 | |
dc546b0f | 1083 | - Type clashes now generate warnings, not errors. |
6e649e65 | 1084 | |
dc546b0f PE |
1085 | - If the user does not define YYSTYPE as a macro, Bison now declares it |
1086 | using typedef instead of defining it as a macro. | |
1087 | For consistency, YYLTYPE is also declared instead of defined. | |
9501dc6e | 1088 | |
402b123d | 1089 | ** Other compatibility issues |
886a425c | 1090 | |
dc546b0f PE |
1091 | - %union directives can now have a tag before the `{', e.g., the |
1092 | directive `%union foo {...}' now generates the C code | |
1093 | `typedef union foo { ... } YYSTYPE;'; this is for Yacc compatibility. | |
1094 | The default union tag is `YYSTYPE', for compatibility with Solaris 9 Yacc. | |
1095 | For consistency, YYLTYPE's struct tag is now `YYLTYPE' not `yyltype'. | |
1096 | This is for compatibility with both Yacc and Bison 1.35. | |
72f889cc | 1097 | |
dc546b0f PE |
1098 | - `;' is output before the terminating `}' of an action, for |
1099 | compatibility with Bison 1.35. | |
886a425c | 1100 | |
dc546b0f PE |
1101 | - Bison now uses a Yacc-style format for conflict reports, e.g., |
1102 | `conflicts: 2 shift/reduce, 1 reduce/reduce'. | |
437c2d80 | 1103 | |
dc546b0f PE |
1104 | - `yystype' and `yyltype' are now obsolescent macros instead of being |
1105 | typedefs or tags; they are no longer documented and are planned to be | |
1106 | withdrawn in a future release. | |
2a8d363a | 1107 | |
402b123d | 1108 | ** GLR parser notes |
2a8d363a | 1109 | |
dc546b0f PE |
1110 | - GLR and inline |
1111 | Users of Bison have to decide how they handle the portability of the | |
1112 | C keyword `inline'. | |
959e5f51 | 1113 | |
dc546b0f PE |
1114 | - `parsing stack overflow...' -> `parser stack overflow' |
1115 | GLR parsers now report `parser stack overflow' as per the Bison manual. | |
900c5db5 | 1116 | |
402b123d | 1117 | ** Bison now warns if it detects conflicting outputs to the same file, |
dc546b0f PE |
1118 | e.g., it generates a warning for `bison -d -o foo.h foo.y' since |
1119 | that command outputs both code and header to foo.h. | |
6e40b4eb | 1120 | |
402b123d | 1121 | ** #line in output files |
dc546b0f | 1122 | - --no-line works properly. |
6e40b4eb | 1123 | |
402b123d | 1124 | ** Bison can no longer be built by a K&R C compiler; it requires C89 or |
6e40b4eb AD |
1125 | later to be built. This change originally took place a few versions |
1126 | ago, but nobody noticed until we recently asked someone to try | |
1127 | building Bison with a K&R C compiler. | |
d600ee67 | 1128 | \f |
402b123d | 1129 | * Changes in version 1.75, 2002-10-14: |
7933f2b5 | 1130 | |
402b123d | 1131 | ** Bison should now work on 64-bit hosts. |
7933f2b5 | 1132 | |
402b123d | 1133 | ** Indonesian translation thanks to Tedi Heriyanto. |
7933f2b5 | 1134 | |
402b123d | 1135 | ** GLR parsers |
f50adbbd AD |
1136 | Fix spurious parse errors. |
1137 | ||
402b123d | 1138 | ** Pure parsers |
f50adbbd AD |
1139 | Some people redefine yyerror to steal yyparse' private variables. |
1140 | Reenable this trick until an official feature replaces it. | |
1141 | ||
402b123d | 1142 | ** Type Clashes |
d90c934c AD |
1143 | In agreement with POSIX and with other Yaccs, leaving a default |
1144 | action is valid when $$ is untyped, and $1 typed: | |
1145 | ||
1146 | untyped: ... typed; | |
1147 | ||
1148 | but the converse remains an error: | |
1149 | ||
1150 | typed: ... untyped; | |
1151 | ||
402b123d | 1152 | ** Values of mid-rule actions |
d90c934c AD |
1153 | The following code: |
1154 | ||
1155 | foo: { ... } { $$ = $1; } ... | |
1156 | ||
1157 | was incorrectly rejected: $1 is defined in the second mid-rule | |
1158 | action, and is equal to the $$ of the first mid-rule action. | |
d600ee67 | 1159 | \f |
402b123d | 1160 | * Changes in version 1.50, 2002-10-04: |
adc8c848 | 1161 | |
402b123d | 1162 | ** GLR parsing |
676385e2 PH |
1163 | The declaration |
1164 | %glr-parser | |
1165 | causes Bison to produce a Generalized LR (GLR) parser, capable of handling | |
1166 | almost any context-free grammar, ambiguous or not. The new declarations | |
e8832397 | 1167 | %dprec and %merge on grammar rules allow parse-time resolution of |
676385e2 PH |
1168 | ambiguities. Contributed by Paul Hilfinger. |
1169 | ||
7933f2b5 | 1170 | Unfortunately Bison 1.50 does not work properly on 64-bit hosts |
420f93c8 PE |
1171 | like the Alpha, so please stick to 32-bit hosts for now. |
1172 | ||
402b123d | 1173 | ** Output Directory |
8c165d89 | 1174 | When not in Yacc compatibility mode, when the output file was not |
e88dbdbf | 1175 | specified, running `bison foo/bar.y' created `foo/bar.c'. It |
8c165d89 AD |
1176 | now creates `bar.c'. |
1177 | ||
402b123d | 1178 | ** Undefined token |
007a50a4 | 1179 | The undefined token was systematically mapped to 2 which prevented |
e88dbdbf | 1180 | the use of 2 by the user. This is no longer the case. |
007a50a4 | 1181 | |
402b123d | 1182 | ** Unknown token numbers |
e88dbdbf | 1183 | If yylex returned an out of range value, yyparse could die. This is |
007a50a4 AD |
1184 | no longer the case. |
1185 | ||
402b123d | 1186 | ** Error token |
e88dbdbf | 1187 | According to POSIX, the error token must be 256. |
23c5a174 AD |
1188 | Bison extends this requirement by making it a preference: *if* the |
1189 | user specified that one of her tokens is numbered 256, then error | |
1190 | will be mapped onto another number. | |
1191 | ||
402b123d | 1192 | ** Verbose error messages |
e88dbdbf | 1193 | They no longer report `..., expecting error or...' for states where |
217598da AD |
1194 | error recovery is possible. |
1195 | ||
402b123d | 1196 | ** End token |
217598da AD |
1197 | Defaults to `$end' instead of `$'. |
1198 | ||
402b123d | 1199 | ** Error recovery now conforms to documentation and to POSIX |
68cd8af3 PE |
1200 | When a Bison-generated parser encounters a syntax error, it now pops |
1201 | the stack until it finds a state that allows shifting the error | |
1202 | token. Formerly, it popped the stack until it found a state that | |
1203 | allowed some non-error action other than a default reduction on the | |
1204 | error token. The new behavior has long been the documented behavior, | |
1205 | and has long been required by POSIX. For more details, please see | |
337116ba PE |
1206 | Paul Eggert, "Reductions during Bison error handling" (2002-05-20) |
1207 | <http://lists.gnu.org/archive/html/bug-bison/2002-05/msg00038.html>. | |
68cd8af3 | 1208 | |
402b123d | 1209 | ** Traces |
5504898e AD |
1210 | Popped tokens and nonterminals are now reported. |
1211 | ||
402b123d | 1212 | ** Larger grammars |
a861a339 PE |
1213 | Larger grammars are now supported (larger token numbers, larger grammar |
1214 | size (= sum of the LHS and RHS lengths), larger LALR tables). | |
1215 | Formerly, many of these numbers ran afoul of 16-bit limits; | |
1216 | now these limits are 32 bits on most hosts. | |
355e7c1c | 1217 | |
402b123d | 1218 | ** Explicit initial rule |
643a5994 AD |
1219 | Bison used to play hacks with the initial rule, which the user does |
1220 | not write. It is now explicit, and visible in the reports and | |
1221 | graphs as rule 0. | |
23c5a174 | 1222 | |
402b123d | 1223 | ** Useless rules |
643a5994 | 1224 | Before, Bison reported the useless rules, but, although not used, |
77714df2 | 1225 | included them in the parsers. They are now actually removed. |
23c5a174 | 1226 | |
402b123d | 1227 | ** Useless rules, useless nonterminals |
6b98e4b5 AD |
1228 | They are now reported, as a warning, with their locations. |
1229 | ||
402b123d | 1230 | ** Rules never reduced |
e8832397 AD |
1231 | Rules that can never be reduced because of conflicts are now |
1232 | reported. | |
1233 | ||
402b123d | 1234 | ** Incorrect `Token not used' |
11652ab3 AD |
1235 | On a grammar such as |
1236 | ||
1237 | %token useless useful | |
1238 | %% | |
1239 | exp: '0' %prec useful; | |
1240 | ||
1241 | where a token was used to set the precedence of the last rule, | |
1242 | bison reported both `useful' and `useless' as useless tokens. | |
1243 | ||
402b123d | 1244 | ** Revert the C++ namespace changes introduced in 1.31 |
77714df2 | 1245 | as they caused too many portability hassles. |
0179dd65 | 1246 | |
402b123d | 1247 | ** Default locations |
b2d52318 AD |
1248 | By an accident of design, the default computation of @$ was |
1249 | performed after another default computation was performed: @$ = @1. | |
1250 | The latter is now removed: YYLLOC_DEFAULT is fully responsible of | |
1251 | the computation of @$. | |
adc8c848 | 1252 | |
402b123d | 1253 | ** Token end-of-file |
b7c49edf AD |
1254 | The token end of file may be specified by the user, in which case, |
1255 | the user symbol is used in the reports, the graphs, and the verbose | |
a861a339 | 1256 | error messages instead of `$end', which remains being the default. |
b7c49edf | 1257 | For instance |
7bd6c77e | 1258 | %token MYEOF 0 |
b7c49edf | 1259 | or |
7bd6c77e | 1260 | %token MYEOF 0 "end of file" |
fdbcd8e2 | 1261 | |
402b123d | 1262 | ** Semantic parser |
fdbcd8e2 AD |
1263 | This old option, which has been broken for ages, is removed. |
1264 | ||
402b123d | 1265 | ** New translations |
a861a339 | 1266 | Brazilian Portuguese, thanks to Alexandre Folle de Menezes. |
84614e13 AD |
1267 | Croatian, thanks to Denis Lackovic. |
1268 | ||
402b123d | 1269 | ** Incorrect token definitions |
e88dbdbf | 1270 | When given `%token 'a' "A"', Bison used to output `#define 'a' 65'. |
b87f8b21 | 1271 | |
402b123d | 1272 | ** Token definitions as enums |
77714df2 AD |
1273 | Tokens are output both as the traditional #define's, and, provided |
1274 | the compiler supports ANSI C or is a C++ compiler, as enums. | |
e88dbdbf | 1275 | This lets debuggers display names instead of integers. |
77714df2 | 1276 | |
402b123d | 1277 | ** Reports |
ec3bc396 AD |
1278 | In addition to --verbose, bison supports --report=THINGS, which |
1279 | produces additional information: | |
b408954b AD |
1280 | - itemset |
1281 | complete the core item sets with their closure | |
9e6e7ed2 PE |
1282 | - lookahead [changed to `look-ahead' in 1.875e through 2.3, but changed back] |
1283 | explicitly associate lookahead tokens to items | |
b408954b AD |
1284 | - solved |
1285 | describe shift/reduce conflicts solving. | |
1286 | Bison used to systematically output this information on top of | |
1287 | the report. Solved conflicts are now attached to their states. | |
ec3bc396 | 1288 | |
402b123d | 1289 | ** Type clashes |
9af3fbce AD |
1290 | Previous versions don't complain when there is a type clash on |
1291 | the default action if the rule has a mid-rule action, such as in: | |
1292 | ||
1293 | %type <foo> bar | |
1294 | %% | |
1295 | bar: '0' {} '0'; | |
1296 | ||
1297 | This is fixed. | |
a861a339 | 1298 | |
402b123d | 1299 | ** GNU M4 is now required when using Bison. |
f987e9d2 | 1300 | \f |
402b123d | 1301 | * Changes in version 1.35, 2002-03-25: |
76551463 | 1302 | |
402b123d | 1303 | ** C Skeleton |
76551463 AD |
1304 | Some projects use Bison's C parser with C++ compilers, and define |
1305 | YYSTYPE as a class. The recent adjustment of C parsers for data | |
1306 | alignment and 64 bit architectures made this impossible. | |
1307 | ||
1308 | Because for the time being no real solution for C++ parser | |
1309 | generation exists, kludges were implemented in the parser to | |
1310 | maintain this use. In the future, when Bison has C++ parsers, this | |
1311 | kludge will be disabled. | |
1312 | ||
1313 | This kludge also addresses some C++ problems when the stack was | |
1314 | extended. | |
76551463 | 1315 | \f |
402b123d | 1316 | * Changes in version 1.34, 2002-03-12: |
76551463 | 1317 | |
402b123d | 1318 | ** File name clashes are detected |
76551463 AD |
1319 | $ bison foo.y -d -o foo.x |
1320 | fatal error: header and parser would both be named `foo.x' | |
1321 | ||
402b123d | 1322 | ** A missing `;' at the end of a rule triggers a warning |
76551463 AD |
1323 | In accordance with POSIX, and in agreement with other |
1324 | Yacc implementations, Bison will mandate this semicolon in the near | |
1325 | future. This eases the implementation of a Bison parser of Bison | |
1326 | grammars by making this grammar LALR(1) instead of LR(2). To | |
1327 | facilitate the transition, this release introduces a warning. | |
1328 | ||
402b123d | 1329 | ** Revert the C++ namespace changes introduced in 1.31, as they caused too |
76551463 AD |
1330 | many portability hassles. |
1331 | ||
402b123d | 1332 | ** DJGPP support added. |
76551463 | 1333 | |
402b123d | 1334 | ** Fix test suite portability problems. |
76551463 | 1335 | \f |
402b123d | 1336 | * Changes in version 1.33, 2002-02-07: |
76551463 | 1337 | |
402b123d | 1338 | ** Fix C++ issues |
76551463 AD |
1339 | Groff could not be compiled for the definition of size_t was lacking |
1340 | under some conditions. | |
1341 | ||
402b123d | 1342 | ** Catch invalid @n |
76551463 AD |
1343 | As is done with $n. |
1344 | \f | |
402b123d | 1345 | * Changes in version 1.32, 2002-01-23: |
76551463 | 1346 | |
402b123d | 1347 | ** Fix Yacc output file names |
76551463 | 1348 | |
402b123d | 1349 | ** Portability fixes |
76551463 | 1350 | |
402b123d | 1351 | ** Italian, Dutch translations |
76551463 | 1352 | \f |
402b123d | 1353 | * Changes in version 1.31, 2002-01-14: |
52d1aeee | 1354 | |
402b123d | 1355 | ** Many Bug Fixes |
52d1aeee | 1356 | |
402b123d | 1357 | ** GNU Gettext and %expect |
52d1aeee MA |
1358 | GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that |
1359 | Bison dies on incorrect %expectations, we fear there will be | |
1360 | too many bug reports for Gettext, so _for the time being_, %expect | |
1361 | does not trigger an error when the input file is named `plural.y'. | |
1362 | ||
402b123d | 1363 | ** Use of alloca in parsers |
52d1aeee MA |
1364 | If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use |
1365 | malloc exclusively. Since 1.29, but was not NEWS'ed. | |
1366 | ||
1367 | alloca is used only when compiled with GCC, to avoid portability | |
1368 | problems as on AIX. | |
1369 | ||
402b123d | 1370 | ** yyparse now returns 2 if memory is exhausted; formerly it dumped core. |
b47dbebe | 1371 | |
402b123d | 1372 | ** When the generated parser lacks debugging code, YYDEBUG is now 0 |
52d1aeee MA |
1373 | (as POSIX requires) instead of being undefined. |
1374 | ||
402b123d | 1375 | ** User Actions |
52d1aeee MA |
1376 | Bison has always permitted actions such as { $$ = $1 }: it adds the |
1377 | ending semicolon. Now if in Yacc compatibility mode, the semicolon | |
1378 | is no longer output: one has to write { $$ = $1; }. | |
1379 | ||
402b123d | 1380 | ** Better C++ compliance |
52d1aeee | 1381 | The output parsers try to respect C++ namespaces. |
76551463 | 1382 | [This turned out to be a failed experiment, and it was reverted later.] |
52d1aeee | 1383 | |
402b123d | 1384 | ** Reduced Grammars |
52d1aeee MA |
1385 | Fixed bugs when reporting useless nonterminals. |
1386 | ||
402b123d | 1387 | ** 64 bit hosts |
52d1aeee MA |
1388 | The parsers work properly on 64 bit hosts. |
1389 | ||
402b123d | 1390 | ** Error messages |
52d1aeee MA |
1391 | Some calls to strerror resulted in scrambled or missing error messages. |
1392 | ||
402b123d | 1393 | ** %expect |
52d1aeee MA |
1394 | When the number of shift/reduce conflicts is correct, don't issue |
1395 | any warning. | |
1396 | ||
402b123d | 1397 | ** The verbose report includes the rule line numbers. |
52d1aeee | 1398 | |
402b123d | 1399 | ** Rule line numbers are fixed in traces. |
52d1aeee | 1400 | |
402b123d | 1401 | ** Swedish translation |
52d1aeee | 1402 | |
402b123d | 1403 | ** Parse errors |
52d1aeee MA |
1404 | Verbose parse error messages from the parsers are better looking. |
1405 | Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'('' | |
1406 | Now: parse error: unexpected '/', expecting "number" or '-' or '(' | |
1407 | ||
402b123d | 1408 | ** Fixed parser memory leaks. |
52d1aeee MA |
1409 | When the generated parser was using malloc to extend its stacks, the |
1410 | previous allocations were not freed. | |
1411 | ||
402b123d | 1412 | ** Fixed verbose output file. |
52d1aeee MA |
1413 | Some newlines were missing. |
1414 | Some conflicts in state descriptions were missing. | |
1415 | ||
402b123d | 1416 | ** Fixed conflict report. |
52d1aeee MA |
1417 | Option -v was needed to get the result. |
1418 | ||
402b123d | 1419 | ** %expect |
52d1aeee MA |
1420 | Was not used. |
1421 | Mismatches are errors, not warnings. | |
1422 | ||
402b123d | 1423 | ** Fixed incorrect processing of some invalid input. |
52d1aeee | 1424 | |
402b123d | 1425 | ** Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H. |
52d1aeee | 1426 | |
402b123d | 1427 | ** Fixed some typos in the documentation. |
52d1aeee | 1428 | |
402b123d | 1429 | ** %token MY_EOF 0 is supported. |
52d1aeee MA |
1430 | Before, MY_EOF was silently renumbered as 257. |
1431 | ||
402b123d | 1432 | ** doc/refcard.tex is updated. |
52d1aeee | 1433 | |
402b123d | 1434 | ** %output, %file-prefix, %name-prefix. |
52d1aeee MA |
1435 | New. |
1436 | ||
402b123d | 1437 | ** --output |
52d1aeee MA |
1438 | New, aliasing `--output-file'. |
1439 | \f | |
402b123d | 1440 | * Changes in version 1.30, 2001-10-26: |
342b8b6e | 1441 | |
402b123d | 1442 | ** `--defines' and `--graph' have now an optional argument which is the |
fdac0091 | 1443 | output file name. `-d' and `-g' do not change; they do not take any |
342b8b6e AD |
1444 | argument. |
1445 | ||
402b123d | 1446 | ** `%source_extension' and `%header_extension' are removed, failed |
342b8b6e AD |
1447 | experiment. |
1448 | ||
402b123d | 1449 | ** Portability fixes. |
f987e9d2 | 1450 | \f |
402b123d | 1451 | * Changes in version 1.29, 2001-09-07: |
342b8b6e | 1452 | |
402b123d | 1453 | ** The output file does not define const, as this caused problems when used |
342b8b6e AD |
1454 | with common autoconfiguration schemes. If you still use ancient compilers |
1455 | that lack const, compile with the equivalent of the C compiler option | |
1456 | `-Dconst='. autoconf's AC_C_CONST macro provides one way to do this. | |
1457 | ||
402b123d | 1458 | ** Added `-g' and `--graph'. |
f87a2205 | 1459 | |
402b123d | 1460 | ** The Bison manual is now distributed under the terms of the GNU FDL. |
f2b5126e | 1461 | |
402b123d | 1462 | ** The input and the output files has automatically a similar extension. |
234a3be3 | 1463 | |
402b123d | 1464 | ** Russian translation added. |
f87a2205 | 1465 | |
402b123d | 1466 | ** NLS support updated; should hopefully be less troublesome. |
f87a2205 | 1467 | |
402b123d | 1468 | ** Added the old Bison reference card. |
c33638bb | 1469 | |
402b123d | 1470 | ** Added `--locations' and `%locations'. |
6deb4447 | 1471 | |
402b123d | 1472 | ** Added `-S' and `--skeleton'. |
cd5bd6ac | 1473 | |
402b123d | 1474 | ** `%raw', `-r', `--raw' is disabled. |
62ab6972 | 1475 | |
402b123d | 1476 | ** Special characters are escaped when output. This solves the problems |
cd5bd6ac AD |
1477 | of the #line lines with path names including backslashes. |
1478 | ||
402b123d | 1479 | ** New directives. |
4ecbf796 MA |
1480 | `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose', |
1481 | `%debug', `%source_extension' and `%header_extension'. | |
f987e9d2 | 1482 | |
402b123d | 1483 | ** @$ |
f987e9d2 | 1484 | Automatic location tracking. |
f87a2205 | 1485 | \f |
402b123d | 1486 | * Changes in version 1.28, 1999-07-06: |
d2e00347 | 1487 | |
402b123d | 1488 | ** Should compile better now with K&R compilers. |
d2e00347 | 1489 | |
402b123d | 1490 | ** Added NLS. |
d2e00347 | 1491 | |
402b123d | 1492 | ** Fixed a problem with escaping the double quote character. |
d2e00347 | 1493 | |
402b123d | 1494 | ** There is now a FAQ. |
d2e00347 | 1495 | \f |
402b123d | 1496 | * Changes in version 1.27: |
5c31c3c2 | 1497 | |
402b123d | 1498 | ** The make rule which prevented bison.simple from being created on |
5c31c3c2 JT |
1499 | some systems has been fixed. |
1500 | \f | |
402b123d | 1501 | * Changes in version 1.26: |
4be07551 | 1502 | |
402b123d | 1503 | ** Bison now uses automake. |
4be07551 | 1504 | |
402b123d | 1505 | ** New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>. |
4be07551 | 1506 | |
402b123d | 1507 | ** Token numbers now start at 257 as previously documented, not 258. |
4be07551 | 1508 | |
402b123d | 1509 | ** Bison honors the TMPDIR environment variable. |
4be07551 | 1510 | |
402b123d | 1511 | ** A couple of buffer overruns have been fixed. |
f51dbca1 | 1512 | |
402b123d | 1513 | ** Problems when closing files should now be reported. |
f51dbca1 | 1514 | |
402b123d | 1515 | ** Generated parsers should now work even on operating systems which do |
f51dbca1 | 1516 | not provide alloca(). |
4be07551 | 1517 | \f |
402b123d | 1518 | * Changes in version 1.25, 1995-10-16: |
df8878c5 | 1519 | |
402b123d | 1520 | ** Errors in the input grammar are not fatal; Bison keeps reading |
df8878c5 | 1521 | the grammar file, and reports all the errors found in it. |
8c44d3ec | 1522 | |
402b123d | 1523 | ** Tokens can now be specified as multiple-character strings: for |
df8878c5 RS |
1524 | example, you could use "<=" for a token which looks like <=, instead |
1525 | of chosing a name like LESSEQ. | |
1526 | ||
402b123d | 1527 | ** The %token_table declaration says to write a table of tokens (names |
df8878c5 RS |
1528 | and numbers) into the parser file. The yylex function can use this |
1529 | table to recognize multiple-character string tokens, or for other | |
1530 | purposes. | |
1531 | ||
402b123d | 1532 | ** The %no_lines declaration says not to generate any #line preprocessor |
df8878c5 RS |
1533 | directives in the parser file. |
1534 | ||
402b123d | 1535 | ** The %raw declaration says to use internal Bison token numbers, not |
df8878c5 RS |
1536 | Yacc-compatible token numbers, when token names are defined as macros. |
1537 | ||
402b123d | 1538 | ** The --no-parser option produces the parser tables without including |
df8878c5 RS |
1539 | the parser engine; a project can now use its own parser engine. |
1540 | The actions go into a separate file called NAME.act, in the form of | |
1541 | a switch statement body. | |
1542 | \f | |
402b123d | 1543 | * Changes in version 1.23: |
6780ca7a | 1544 | |
4d019228 DM |
1545 | The user can define YYPARSE_PARAM as the name of an argument to be |
1546 | passed into yyparse. The argument should have type void *. It should | |
1547 | actually point to an object. Grammar actions can access the variable | |
1548 | by casting it to the proper pointer type. | |
6780ca7a | 1549 | |
6780ca7a | 1550 | Line numbers in output file corrected. |
6780ca7a | 1551 | \f |
402b123d | 1552 | * Changes in version 1.22: |
6780ca7a DM |
1553 | |
1554 | --help option added. | |
6780ca7a | 1555 | \f |
402b123d | 1556 | * Changes in version 1.20: |
6780ca7a DM |
1557 | |
1558 | Output file does not redefine const for C++. | |
9f4503d6 AD |
1559 | |
1560 | Local Variables: | |
1561 | mode: outline | |
1562 | End: | |
76551463 AD |
1563 | |
1564 | ----- | |
1565 | ||
575619af | 1566 | Copyright (C) 1995-2011 Free Software Foundation, Inc. |
76551463 | 1567 | |
74553c98 | 1568 | This file is part of Bison, the GNU Parser Generator. |
76551463 | 1569 | |
f16b0819 | 1570 | This program is free software: you can redistribute it and/or modify |
76551463 | 1571 | it under the terms of the GNU General Public License as published by |
f16b0819 PE |
1572 | the Free Software Foundation, either version 3 of the License, or |
1573 | (at your option) any later version. | |
76551463 | 1574 | |
f16b0819 | 1575 | This program is distributed in the hope that it will be useful, |
76551463 AD |
1576 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
1577 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
1578 | GNU General Public License for more details. | |
1579 | ||
1580 | You should have received a copy of the GNU General Public License | |
f16b0819 | 1581 | along with this program. If not, see <http://www.gnu.org/licenses/>. |