]>
Commit | Line | Data |
---|---|---|
ed7658fe | 1 | GNU Bison NEWS |
3af4feb2 | 2 | |
52718713 | 3 | * Noteworthy changes in release ?.? (????-??-??) [?] |
74553c98 | 4 | |
2055a44e AD |
5 | ** Additional yylex/yyparse arguments |
6 | ||
7 | The new directive %param declare additional argument to both yylex | |
bb9191dd | 8 | and yyparse. The %lex-param, %parse-param, and %param directives |
2055a44e AD |
9 | support one or more arguments. Instead of |
10 | ||
e436fa67 AD |
11 | %lex-param {arg1_type *arg1} |
12 | %lex-param {arg2_type *arg2} | |
13 | %parse-param {arg1_type *arg1} | |
14 | %parse-param {arg2_type *arg2} | |
2055a44e AD |
15 | |
16 | one may now declare | |
17 | ||
e436fa67 | 18 | %param {arg1_type *arg1} {arg2_type *arg2} |
2055a44e AD |
19 | |
20 | ** Java skeleton improvements | |
0ea583d2 AD |
21 | |
22 | The constants for token names were moved to the Lexer interface. | |
23 | Also, it is possible to add code to the parser's constructors using | |
24 | "%code init" and "%define init_throws". | |
25 | ||
a6552c5d AD |
26 | ** C++ skeleton improvements |
27 | ||
28 | The C++ parser features a syntax_error exception, which can be | |
29 | thrown from the scanner or from user rules to raise syntax errors. | |
30 | This facilitates reporting errors caught in sub-functions (e.g., | |
31 | rejecting too large integral literals from a conversion function | |
32 | used by the scanner, or rejecting invalid combinations from a | |
33 | factory invoked by the user actions). | |
34 | ||
4c6622c2 | 35 | ** Variable api.tokens.prefix |
99c08fb6 | 36 | |
4c6622c2 | 37 | The variable api.tokens.prefix changes the way tokens are identified in |
99c08fb6 AD |
38 | the generated files. This is especially useful to avoid collisions |
39 | with identifiers in the target language. For instance | |
40 | ||
e436fa67 AD |
41 | %token FILE for ERROR |
42 | %define api.tokens.prefix "TOK_" | |
43 | %% | |
44 | start: FILE for ERROR; | |
99c08fb6 AD |
45 | |
46 | will generate the definition of the symbols TOK_FILE, TOK_for, and | |
47 | TOK_ERROR in the generated sources. In particular, the scanner must | |
48 | use these prefixed token names, although the grammar itself still | |
49 | uses the short names (as in the sample rule given above). | |
50 | ||
67501061 AD |
51 | ** Variable api.namespace |
52 | ||
53 | The "namespace" variable is renamed "api.namespace". Backward | |
54 | compatibility is ensured, but upgrading is recommended. | |
55 | ||
31b850d2 AD |
56 | ** Variable parse.error |
57 | ||
58 | The variable error controls the verbosity of error messages. The | |
59 | use of the %error-verbose directive is deprecated in favor of | |
60 | %define parse.error "verbose". | |
61 | ||
ca2a6d15 PH |
62 | ** Semantic predicates |
63 | ||
64 | The new, experimental, semantic-predicate feature allows actions of | |
65 | the form %?{ BOOLEAN-EXPRESSION }, which cause syntax errors (as for | |
66 | YYERROR) if the expression evaluates to 0, and are evaluated immediately | |
67 | in GLR parsers, rather than being deferred. The result is that they | |
68 | allow the programmer to prune possible parses based on the values of | |
69 | runtime expressions. | |
70 | ||
e436fa67 AD |
71 | * Noteworthy changes in release ?.? (????-??-??) [?] |
72 | ||
0f11eec2 AD |
73 | ** Future changes: |
74 | ||
75 | The next major release will drop support for generating parsers in K&R C, | |
76 | and remove the definition of yystype (removal announced since Bison | |
77 | 1.875). | |
78 | ||
c3e9f08f | 79 | ** The generated header is included (yacc.c) |
0f11eec2 AD |
80 | |
81 | Instead of duplicating the content of the generated header (definition of | |
82 | YYSTYPE, yyltype etc.), the generated parser now includes it, as was | |
83 | already the case for GLR or C++ parsers. | |
84 | ||
56ca3d8f AD |
85 | ** Headers (yacc.c, glr.c, glr.cc) |
86 | ||
87 | *** Guards | |
c3e9f08f AD |
88 | |
89 | The generated headers are now guarded, as is already the case for C++ | |
90 | parsers (lalr1.cc). For intance, with --defines=foo.h: | |
91 | ||
e29f0771 AD |
92 | #ifndef YY_FOO_H |
93 | # define YY_FOO_H | |
94 | ... | |
95 | #endif /* !YY_FOO_H */ | |
c3e9f08f | 96 | |
56ca3d8f AD |
97 | *** New declarations |
98 | ||
99 | The generated header now declares yydebug and yyparse. Both honor | |
100 | --name-prefix=bar_, and yield | |
101 | ||
e29f0771 | 102 | int bar_parse (void); |
56ca3d8f AD |
103 | |
104 | rather than | |
105 | ||
e29f0771 AD |
106 | #define yyparse bar_parse |
107 | int yyparse (void); | |
56ca3d8f AD |
108 | |
109 | in order to facilitate the inclusion of several parser headers inside a | |
110 | single compilation unit. | |
c3e9f08f | 111 | |
dfaac272 | 112 | * Noteworthy changes in release 2.5.1 (2012-06-05) [stable] |
df6e3db0 | 113 | |
debe2c03 | 114 | ** Future changes: |
765e1bd4 | 115 | |
e4ab1254 | 116 | The next major release will drop support for generating parsers in K&R C. |
041308d0 | 117 | |
466b4cf2 | 118 | ** yacc.c: YYBACKUP works as expected. |
ef51bfa7 | 119 | |
d834eca0 | 120 | ** glr.c improvements: |
041308d0 | 121 | |
d834eca0 | 122 | *** Location support is eliminated when not requested: |
041308d0 | 123 | |
e4ab1254 AD |
124 | GLR parsers used to include location-related code even when locations were |
125 | not requested, and therefore not even usable. | |
378e917c | 126 | |
d834eca0 | 127 | *** __attribute__ is preserved: |
d115aad9 | 128 | |
e4ab1254 AD |
129 | __attribute__ is no longer disabled when __STRICT_ANSI__ is defined (i.e., |
130 | when -std is passed to GCC). | |
041308d0 | 131 | |
466b4cf2 | 132 | ** lalr1.java: several fixes: |
041308d0 | 133 | |
e4ab1254 AD |
134 | The Java parser no longer throws ArrayIndexOutOfBoundsException if the |
135 | first token leads to a syntax error. Some minor clean ups. | |
041308d0 | 136 | |
22172d47 | 137 | ** Changes for C++: |
ef51bfa7 | 138 | |
22172d47 | 139 | *** C++11 compatibility: |
ef51bfa7 | 140 | |
e4ab1254 AD |
141 | C and C++ parsers use "nullptr" instead of "0" when __cplusplus is 201103L |
142 | or higher. | |
936c88d1 | 143 | |
22172d47 AD |
144 | *** Header guards |
145 | ||
146 | The header files such as "parser.hh", "location.hh", etc. used a constant | |
147 | name for preprocessor guards, for instance: | |
148 | ||
149 | #ifndef BISON_LOCATION_HH | |
150 | # define BISON_LOCATION_HH | |
151 | ... | |
152 | #endif // !BISON_LOCATION_HH | |
153 | ||
154 | The inclusion guard is now computed from "PREFIX/FILE-NAME", where lower | |
155 | case characters are converted to upper case, and series of | |
156 | non-alphanumerical characters are converted to an underscore. | |
157 | ||
158 | With "bison -o lang++/parser.cc", "location.hh" would now include: | |
159 | ||
160 | #ifndef YY_LANG_LOCATION_HH | |
161 | # define YY_LANG_LOCATION_HH | |
162 | ... | |
163 | #endif // !YY_LANG_LOCATION_HH | |
164 | ||
165 | *** C++ locations: | |
936c88d1 | 166 | |
e4ab1254 AD |
167 | The position and location constructors (and their initialize methods) |
168 | accept new arguments for line and column. Several issues in the | |
169 | documentation were fixed. | |
936c88d1 | 170 | |
466b4cf2 AD |
171 | ** liby is no longer asking for "rpl_fprintf" on some platforms. |
172 | ||
7e508a2b AD |
173 | ** Changes in the manual: |
174 | ||
175 | *** %printer is documented | |
176 | ||
e4ab1254 AD |
177 | The "%printer" directive, supported since at least Bison 1.50, is finally |
178 | documented. The "mfcalc" example is extended to demonstrate it. | |
7e508a2b | 179 | |
e4ab1254 AD |
180 | For consistency with the C skeletons, the C++ parsers now also support |
181 | "yyoutput" (as an alias to "debug_stream ()"). | |
7e508a2b AD |
182 | |
183 | *** Several improvements have been made: | |
466b4cf2 | 184 | |
e4ab1254 AD |
185 | The layout for grammar excerpts was changed to a more compact scheme. |
186 | Named references are motivated. The description of the automaton | |
187 | description file (*.output) is updated to the current format. Incorrect | |
188 | index entries were fixed. Some other errors were fixed. | |
466b4cf2 | 189 | |
86b08b49 AD |
190 | ** Building bison: |
191 | ||
192 | *** Conflicting prototypes with recent/modified Flex. | |
466b4cf2 | 193 | |
e4ab1254 AD |
194 | Fixed build problems with the current, unreleased, version of Flex, and |
195 | some modified versions of 2.5.35, which have modified function prototypes. | |
466b4cf2 | 196 | |
8ef26c2a AD |
197 | *** Warnings during the build procedure have been eliminated. |
198 | ||
199 | *** Several portability problems in the test suite have been fixed: | |
466b4cf2 | 200 | |
e4ab1254 AD |
201 | This includes warnings with some compilers, unexpected behavior of tools |
202 | such as diff, warning messages from the test suite itself, etc. | |
466b4cf2 | 203 | |
91aadcc7 | 204 | *** The install-pdf target works properly: |
8ef26c2a | 205 | |
e4ab1254 AD |
206 | Running "make install-pdf" (or -dvi, -html, -info, and -ps) no longer |
207 | halts in the middle of its course. | |
8ef26c2a | 208 | |
28801043 | 209 | * Changes in version 2.5 (2011-05-14): |
50cca368 | 210 | |
82f3355e JD |
211 | ** Grammar symbol names can now contain non-initial dashes: |
212 | ||
213 | Consistently with directives (such as %error-verbose) and with | |
214 | %define variables (e.g. push-pull), grammar symbol names may contain | |
215 | dashes in any position except the beginning. This is a GNU | |
216 | extension over POSIX Yacc. Thus, use of this extension is reported | |
217 | by -Wyacc and rejected in Yacc mode (--yacc). | |
218 | ||
f1b238df | 219 | ** Named references: |
66381412 AR |
220 | |
221 | Historically, Yacc and Bison have supported positional references | |
222 | ($n, $$) to allow access to symbol values from inside of semantic | |
223 | actions code. | |
224 | ||
225 | Starting from this version, Bison can also accept named references. | |
226 | When no ambiguity is possible, original symbol names may be used | |
227 | as named references: | |
228 | ||
4b568fc0 | 229 | if_stmt : "if" cond_expr "then" then_stmt ';' |
66381412 AR |
230 | { $if_stmt = mk_if_stmt($cond_expr, $then_stmt); } |
231 | ||
232 | In the more common case, explicit names may be declared: | |
233 | ||
4b568fc0 | 234 | stmt[res] : "if" expr[cond] "then" stmt[then] "else" stmt[else] ';' |
66381412 AR |
235 | { $res = mk_if_stmt($cond, $then, $else); } |
236 | ||
5b1ff423 | 237 | Location information is also accessible using @name syntax. When |
66381412 AR |
238 | accessing symbol names containing dots or dashes, explicit bracketing |
239 | ($[sym.1]) must be used. | |
240 | ||
5b1ff423 | 241 | These features are experimental in this version. More user feedback |
66381412 AR |
242 | will help to stabilize them. |
243 | ||
f1b238df | 244 | ** IELR(1) and canonical LR(1): |
eb45ef3b JD |
245 | |
246 | IELR(1) is a minimal LR(1) parser table generation algorithm. That | |
247 | is, given any context-free grammar, IELR(1) generates parser tables | |
7262f54f | 248 | with the full language-recognition power of canonical LR(1) but with |
f1b238df JD |
249 | nearly the same number of parser states as LALR(1). This reduction |
250 | in parser states is often an order of magnitude. More importantly, | |
eb45ef3b JD |
251 | because canonical LR(1)'s extra parser states may contain duplicate |
252 | conflicts in the case of non-LR(1) grammars, the number of conflicts | |
253 | for IELR(1) is often an order of magnitude less as well. This can | |
254 | significantly reduce the complexity of developing of a grammar. | |
255 | ||
256 | Bison can now generate IELR(1) and canonical LR(1) parser tables in | |
257 | place of its traditional LALR(1) parser tables, which remain the | |
258 | default. You can specify the type of parser tables in the grammar | |
259 | file with these directives: | |
260 | ||
cf499cff JD |
261 | %define lr.type lalr |
262 | %define lr.type ielr | |
263 | %define lr.type canonical-lr | |
eb45ef3b | 264 | |
7fceb615 | 265 | The default-reduction optimization in the parser tables can also be |
e4ab1254 AD |
266 | adjusted using "%define lr.default-reductions". For details on both |
267 | of these features, see the new section "Tuning LR" in the Bison | |
7fceb615 | 268 | manual. |
eb45ef3b JD |
269 | |
270 | These features are experimental. More user feedback will help to | |
271 | stabilize them. | |
272 | ||
7fceb615 | 273 | ** LAC (Lookahead Correction) for syntax error handling: |
fcf834f9 JD |
274 | |
275 | Canonical LR, IELR, and LALR can suffer from a couple of problems | |
276 | upon encountering a syntax error. First, the parser might perform | |
277 | additional parser stack reductions before discovering the syntax | |
7fceb615 | 278 | error. Such reductions can perform user semantic actions that are |
fcf834f9 JD |
279 | unexpected because they are based on an invalid token, and they |
280 | cause error recovery to begin in a different syntactic context than | |
281 | the one in which the invalid token was encountered. Second, when | |
7fceb615 | 282 | verbose error messages are enabled (with %error-verbose or the |
e4ab1254 | 283 | obsolete "#define YYERROR_VERBOSE"), the expected token list in the |
7fceb615 JD |
284 | syntax error message can both contain invalid tokens and omit valid |
285 | tokens. | |
fcf834f9 JD |
286 | |
287 | The culprits for the above problems are %nonassoc, default | |
288 | reductions in inconsistent states, and parser state merging. Thus, | |
289 | IELR and LALR suffer the most. Canonical LR can suffer only if | |
290 | %nonassoc is used or if default reductions are enabled for | |
291 | inconsistent states. | |
292 | ||
7fceb615 JD |
293 | LAC is a new mechanism within the parsing algorithm that solves |
294 | these problems for canonical LR, IELR, and LALR without sacrificing | |
295 | %nonassoc, default reductions, or state merging. When LAC is in | |
296 | use, canonical LR and IELR behave almost exactly the same for both | |
297 | syntactically acceptable and syntactically unacceptable input. | |
fcf834f9 JD |
298 | While LALR still does not support the full language-recognition |
299 | power of canonical LR and IELR, LAC at least enables LALR's syntax | |
300 | error handling to correctly reflect LALR's language-recognition | |
301 | power. | |
302 | ||
303 | Currently, LAC is only supported for deterministic parsers in C. | |
304 | You can enable LAC with the following directive: | |
305 | ||
306 | %define parse.lac full | |
307 | ||
e4ab1254 | 308 | See the new section "LAC" in the Bison manual for additional |
7fceb615 | 309 | details including a few caveats. |
fcf834f9 JD |
310 | |
311 | LAC is an experimental feature. More user feedback will help to | |
312 | stabilize it. | |
313 | ||
d397d9f0 | 314 | ** %define improvements: |
cf499cff | 315 | |
f1b238df | 316 | *** Can now be invoked via the command line: |
50cca368 | 317 | |
de5ab940 | 318 | Each of these command-line options |
50cca368 | 319 | |
de5ab940 JD |
320 | -D NAME[=VALUE] |
321 | --define=NAME[=VALUE] | |
322 | ||
323 | -F NAME[=VALUE] | |
324 | --force-define=NAME[=VALUE] | |
50cca368 JD |
325 | |
326 | is equivalent to this grammar file declaration | |
327 | ||
de5ab940 | 328 | %define NAME ["VALUE"] |
50cca368 | 329 | |
de5ab940 JD |
330 | except that the manner in which Bison processes multiple definitions |
331 | for the same NAME differs. Most importantly, -F and --force-define | |
332 | quietly override %define, but -D and --define do not. For further | |
e4ab1254 | 333 | details, see the section "Bison Options" in the Bison manual. |
50cca368 | 334 | |
f1b238df | 335 | *** Variables renamed: |
67212941 JD |
336 | |
337 | The following %define variables | |
338 | ||
339 | api.push_pull | |
340 | lr.keep_unreachable_states | |
341 | ||
342 | have been renamed to | |
343 | ||
344 | api.push-pull | |
345 | lr.keep-unreachable-states | |
346 | ||
347 | The old names are now deprecated but will be maintained indefinitely | |
348 | for backward compatibility. | |
349 | ||
7262f54f | 350 | *** Values no longer need to be quoted in the grammar file: |
cf499cff JD |
351 | |
352 | If a %define value is an identifier, it no longer needs to be placed | |
353 | within quotations marks. For example, | |
354 | ||
355 | %define api.push-pull "push" | |
356 | ||
357 | can be rewritten as | |
358 | ||
359 | %define api.push-pull push | |
360 | ||
d397d9f0 | 361 | *** Unrecognized variables are now errors not warnings. |
cdf3f113 | 362 | |
d397d9f0 JD |
363 | *** Multiple invocations for any variable is now an error not a warning. |
364 | ||
365 | ** Unrecognized %code qualifiers are now errors not warnings. | |
366 | ||
367 | ** Character literals not of length one: | |
368 | ||
369 | Previously, Bison quietly converted all character literals to length | |
370 | one. For example, without warning, Bison interpreted the operators in | |
371 | the following grammar to be the same token: | |
372 | ||
373 | exp: exp '++' | |
374 | | exp '+' exp | |
375 | ; | |
376 | ||
377 | Bison now warns when a character literal is not of length one. In | |
378 | some future release, Bison will start reporting an error instead. | |
379 | ||
380 | ** Destructor calls fixed for lookaheads altered in semantic actions: | |
381 | ||
382 | Previously for deterministic parsers in C, if a user semantic action | |
383 | altered yychar, the parser in some cases used the old yychar value to | |
384 | determine which destructor to call for the lookahead upon a syntax | |
385 | error or upon parser return. This bug has been fixed. | |
386 | ||
387 | ** C++ parsers use YYRHSLOC: | |
388 | ||
389 | Similarly to the C parsers, the C++ parsers now define the YYRHSLOC | |
390 | macro and use it in the default YYLLOC_DEFAULT. You are encouraged | |
e4ab1254 AD |
391 | to use it. If, for instance, your location structure has "first" |
392 | and "last" members, instead of | |
d397d9f0 JD |
393 | |
394 | # define YYLLOC_DEFAULT(Current, Rhs, N) \ | |
395 | do \ | |
396 | if (N) \ | |
397 | { \ | |
398 | (Current).first = (Rhs)[1].location.first; \ | |
399 | (Current).last = (Rhs)[N].location.last; \ | |
400 | } \ | |
401 | else \ | |
402 | { \ | |
403 | (Current).first = (Current).last = (Rhs)[0].location.last; \ | |
404 | } \ | |
405 | while (false) | |
406 | ||
407 | use: | |
408 | ||
409 | # define YYLLOC_DEFAULT(Current, Rhs, N) \ | |
410 | do \ | |
411 | if (N) \ | |
412 | { \ | |
413 | (Current).first = YYRHSLOC (Rhs, 1).first; \ | |
414 | (Current).last = YYRHSLOC (Rhs, N).last; \ | |
415 | } \ | |
416 | else \ | |
417 | { \ | |
418 | (Current).first = (Current).last = YYRHSLOC (Rhs, 0).last; \ | |
419 | } \ | |
420 | while (false) | |
421 | ||
422 | ** YYLLOC_DEFAULT in C++: | |
423 | ||
424 | The default implementation of YYLLOC_DEFAULT used to be issued in | |
425 | the header file. It is now output in the implementation file, after | |
426 | the user %code sections so that its #ifndef guard does not try to | |
427 | override the user's YYLLOC_DEFAULT if provided. | |
cdf3f113 | 428 | |
f1b238df | 429 | ** YYFAIL now produces warnings and Java parsers no longer implement it: |
4395a9ff JD |
430 | |
431 | YYFAIL has existed for many years as an undocumented feature of | |
432 | deterministic parsers in C generated by Bison. More recently, it was | |
433 | a documented feature of Bison's experimental Java parsers. As | |
434 | promised in Bison 2.4.2's NEWS entry, any appearance of YYFAIL in a | |
435 | semantic action now produces a deprecation warning, and Java parsers | |
436 | no longer implement YYFAIL at all. For further details, including a | |
437 | discussion of how to suppress C preprocessor warnings about YYFAIL | |
438 | being unused, see the Bison 2.4.2 NEWS entry. | |
439 | ||
f1b238df | 440 | ** Temporary hack for adding a semicolon to the user action: |
197b82ba JD |
441 | |
442 | Previously, Bison appended a semicolon to every user action for | |
443 | reductions when the output language defaulted to C (specifically, when | |
444 | neither %yacc, %language, %skeleton, or equivalent command-line | |
445 | options were specified). This allowed actions such as | |
446 | ||
447 | exp: exp "+" exp { $$ = $1 + $3 }; | |
448 | ||
449 | instead of | |
450 | ||
451 | exp: exp "+" exp { $$ = $1 + $3; }; | |
452 | ||
453 | As a first step in removing this misfeature, Bison now issues a | |
454 | warning when it appends a semicolon. Moreover, in cases where Bison | |
455 | cannot easily determine whether a semicolon is needed (for example, an | |
456 | action ending with a cpp directive or a braced compound initializer), | |
457 | it no longer appends one. Thus, the C compiler might now complain | |
458 | about a missing semicolon where it did not before. Future releases of | |
459 | Bison will cease to append semicolons entirely. | |
460 | ||
d2060f06 JD |
461 | ** Verbose syntax error message fixes: |
462 | ||
e4ab1254 | 463 | When %error-verbose or the obsolete "#define YYERROR_VERBOSE" is |
7fceb615 JD |
464 | specified, syntax error messages produced by the generated parser |
465 | include the unexpected token as well as a list of expected tokens. | |
466 | The effect of %nonassoc on these verbose messages has been corrected | |
467 | in two ways, but a more complete fix requires LAC, described above: | |
d2060f06 JD |
468 | |
469 | *** When %nonassoc is used, there can exist parser states that accept no | |
470 | tokens, and so the parser does not always require a lookahead token | |
471 | in order to detect a syntax error. Because no unexpected token or | |
472 | expected tokens can then be reported, the verbose syntax error | |
473 | message described above is suppressed, and the parser instead | |
e4ab1254 | 474 | reports the simpler message, "syntax error". Previously, this |
d2060f06 JD |
475 | suppression was sometimes erroneously triggered by %nonassoc when a |
476 | lookahead was actually required. Now verbose messages are | |
477 | suppressed only when all previous lookaheads have already been | |
478 | shifted or discarded. | |
479 | ||
480 | *** Previously, the list of expected tokens erroneously included tokens | |
481 | that would actually induce a syntax error because conflicts for them | |
482 | were resolved with %nonassoc in the current parser state. Such | |
483 | tokens are now properly omitted from the list. | |
484 | ||
485 | *** Expected token lists are still often wrong due to state merging | |
fcf834f9 JD |
486 | (from LALR or IELR) and default reductions, which can both add |
487 | invalid tokens and subtract valid tokens. Canonical LR almost | |
488 | completely fixes this problem by eliminating state merging and | |
489 | default reductions. However, there is one minor problem left even | |
490 | when using canonical LR and even after the fixes above. That is, | |
491 | if the resolution of a conflict with %nonassoc appears in a later | |
492 | parser state than the one at which some syntax error is | |
493 | discovered, the conflicted token is still erroneously included in | |
494 | the expected token list. Bison's new LAC implementation, | |
495 | described above, eliminates this problem and the need for | |
496 | canonical LR. However, LAC is still experimental and is disabled | |
497 | by default. | |
53f036ce | 498 | |
1a33f4f6 JD |
499 | ** Java skeleton fixes: |
500 | ||
501 | *** A location handling bug has been fixed. | |
502 | ||
503 | *** The top element of each of the value stack and location stack is now | |
504 | cleared when popped so that it can be garbage collected. | |
6771a463 | 505 | |
02803d55 JD |
506 | *** Parser traces now print the top element of the stack. |
507 | ||
86408959 JD |
508 | ** -W/--warnings fixes: |
509 | ||
e4ab1254 | 510 | *** Bison now properly recognizes the "no-" versions of categories: |
86408959 JD |
511 | |
512 | For example, given the following command line, Bison now enables all | |
513 | warnings except warnings for incompatibilities with POSIX Yacc: | |
514 | ||
515 | bison -Wall,no-yacc gram.y | |
516 | ||
786743d5 JD |
517 | *** Bison now treats S/R and R/R conflicts like other warnings: |
518 | ||
519 | Previously, conflict reports were independent of Bison's normal | |
520 | warning system. Now, Bison recognizes the warning categories | |
e4ab1254 | 521 | "conflicts-sr" and "conflicts-rr". This change has important |
786743d5 JD |
522 | consequences for the -W and --warnings command-line options. For |
523 | example: | |
524 | ||
525 | bison -Wno-conflicts-sr gram.y # S/R conflicts not reported | |
526 | bison -Wno-conflicts-rr gram.y # R/R conflicts not reported | |
527 | bison -Wnone gram.y # no conflicts are reported | |
528 | bison -Werror gram.y # any conflict is an error | |
529 | ||
530 | However, as before, if the %expect or %expect-rr directive is | |
531 | specified, an unexpected number of conflicts is an error, and an | |
532 | expected number of conflicts is not reported, so -W and --warning | |
533 | then have no effect on the conflict report. | |
534 | ||
e4ab1254 | 535 | *** The "none" category no longer disables a preceding "error": |
bf0e44e8 JD |
536 | |
537 | For example, for the following command line, Bison now reports | |
538 | errors instead of warnings for incompatibilities with POSIX Yacc: | |
539 | ||
540 | bison -Werror,none,yacc gram.y | |
541 | ||
e4ab1254 | 542 | *** The "none" category now disables all Bison warnings: |
c39014ae | 543 | |
e4ab1254 | 544 | Previously, the "none" category disabled only Bison warnings for |
c39014ae JD |
545 | which there existed a specific -W/--warning category. However, |
546 | given the following command line, Bison is now guaranteed to | |
547 | suppress all warnings: | |
548 | ||
549 | bison -Wnone gram.y | |
550 | ||
1f36f544 JD |
551 | ** Precedence directives can now assign token number 0: |
552 | ||
553 | Since Bison 2.3b, which restored the ability of precedence | |
554 | directives to assign token numbers, doing so for token number 0 has | |
555 | produced an assertion failure. For example: | |
556 | ||
557 | %left END 0 | |
558 | ||
559 | This bug has been fixed. | |
560 | ||
64877e5e | 561 | * Changes in version 2.4.3 (2010-08-05): |
8b9e021f | 562 | |
2bfcac9a JD |
563 | ** Bison now obeys -Werror and --warnings=error for warnings about |
564 | grammar rules that are useless in the parser due to conflicts. | |
565 | ||
8b9e021f JD |
566 | ** Problems with spawning M4 on at least FreeBSD 8 and FreeBSD 9 have |
567 | been fixed. | |
568 | ||
4ad3921d JD |
569 | ** Failures in the test suite for GCC 4.5 have been fixed. |
570 | ||
06cb07d5 JD |
571 | ** Failures in the test suite for some versions of Sun Studio C++ have |
572 | been fixed. | |
573 | ||
9b5049bd JD |
574 | ** Contrary to Bison 2.4.2's NEWS entry, it has been decided that |
575 | warnings about undefined %prec identifiers will not be converted to | |
576 | errors in Bison 2.5. They will remain warnings, which should be | |
577 | sufficient for POSIX while avoiding backward compatibility issues. | |
578 | ||
93d7dde9 JD |
579 | ** Minor documentation fixes. |
580 | ||
e19a049c | 581 | * Changes in version 2.4.2 (2010-03-20): |
74553c98 | 582 | |
f39ab286 JD |
583 | ** Some portability problems that resulted in failures and livelocks |
584 | in the test suite on some versions of at least Solaris, AIX, HP-UX, | |
e19a049c JD |
585 | RHEL4, and Tru64 have been addressed. As a result, fatal Bison |
586 | errors should no longer cause M4 to report a broken pipe on the | |
f39ab286 JD |
587 | affected platforms. |
588 | ||
e4ab1254 | 589 | ** "%prec IDENTIFIER" requires IDENTIFIER to be defined separately. |
8bb3a2e7 JD |
590 | |
591 | POSIX specifies that an error be reported for any identifier that does | |
592 | not appear on the LHS of a grammar rule and that is not defined by | |
593 | %token, %left, %right, or %nonassoc. Bison 2.3b and later lost this | |
594 | error report for the case when an identifier appears only after a | |
595 | %prec directive. It is now restored. However, for backward | |
596 | compatibility with recent Bison releases, it is only a warning for | |
597 | now. In Bison 2.5 and later, it will return to being an error. | |
9b5049bd JD |
598 | [Between the 2.4.2 and 2.4.3 releases, it was decided that this |
599 | warning will not be converted to an error in Bison 2.5.] | |
8bb3a2e7 | 600 | |
d8911864 EB |
601 | ** Detection of GNU M4 1.4.6 or newer during configure is improved. |
602 | ||
a603c6e0 JD |
603 | ** Warnings from gcc's -Wundef option about undefined YYENABLE_NLS, |
604 | YYLTYPE_IS_TRIVIAL, and __STRICT_ANSI__ in C/C++ parsers are now | |
605 | avoided. | |
c938d650 | 606 | |
98a345a2 JD |
607 | ** %code is now a permanent feature. |
608 | ||
609 | A traditional Yacc prologue directive is written in the form: | |
610 | ||
611 | %{CODE%} | |
612 | ||
613 | To provide a more flexible alternative, Bison 2.3b introduced the | |
614 | %code directive with the following forms for C/C++: | |
615 | ||
616 | %code {CODE} | |
617 | %code requires {CODE} | |
618 | %code provides {CODE} | |
619 | %code top {CODE} | |
620 | ||
621 | These forms are now considered permanent features of Bison. See the | |
622 | %code entries in the section "Bison Declaration Summary" in the Bison | |
623 | manual for a summary of their functionality. See the section | |
624 | "Prologue Alternatives" for a detailed discussion including the | |
625 | advantages of %code over the traditional Yacc prologue directive. | |
626 | ||
627 | Bison's Java feature as a whole including its current usage of %code | |
628 | is still considered experimental. | |
629 | ||
1625df5b JD |
630 | ** YYFAIL is deprecated and will eventually be removed. |
631 | ||
632 | YYFAIL has existed for many years as an undocumented feature of | |
633 | deterministic parsers in C generated by Bison. Previously, it was | |
634 | documented for Bison's experimental Java parsers. YYFAIL is no longer | |
635 | documented for Java parsers and is formally deprecated in both cases. | |
636 | Users are strongly encouraged to migrate to YYERROR, which is | |
637 | specified by POSIX. | |
638 | ||
639 | Like YYERROR, you can invoke YYFAIL from a semantic action in order to | |
640 | induce a syntax error. The most obvious difference from YYERROR is | |
641 | that YYFAIL will automatically invoke yyerror to report the syntax | |
642 | error so that you don't have to. However, there are several other | |
643 | subtle differences between YYERROR and YYFAIL, and YYFAIL suffers from | |
e4ab1254 | 644 | inherent flaws when %error-verbose or "#define YYERROR_VERBOSE" is |
1625df5b JD |
645 | used. For a more detailed discussion, see: |
646 | ||
647 | http://lists.gnu.org/archive/html/bison-patches/2009-12/msg00024.html | |
648 | ||
649 | The upcoming Bison 2.5 will remove YYFAIL from Java parsers, but | |
650 | deterministic parsers in C will continue to implement it. However, | |
651 | because YYFAIL is already flawed, it seems futile to try to make new | |
652 | Bison features compatible with it. Thus, during parser generation, | |
653 | Bison 2.5 will produce a warning whenever it discovers YYFAIL in a | |
654 | rule action. In a later release, YYFAIL will be disabled for | |
e4ab1254 | 655 | %error-verbose and "#define YYERROR_VERBOSE". Eventually, YYFAIL will |
1625df5b JD |
656 | be removed altogether. |
657 | ||
658 | There exists at least one case where Bison 2.5's YYFAIL warning will | |
659 | be a false positive. Some projects add phony uses of YYFAIL and other | |
660 | Bison-defined macros for the sole purpose of suppressing C | |
661 | preprocessor warnings (from GCC cpp's -Wunused-macros, for example). | |
662 | To avoid Bison's future warning, such YYFAIL uses can be moved to the | |
e4ab1254 | 663 | epilogue (that is, after the second "%%") in the Bison input file. In |
1625df5b JD |
664 | this release (2.4.2), Bison already generates its own code to suppress |
665 | C preprocessor warnings for YYFAIL, so projects can remove their own | |
666 | phony uses of YYFAIL if compatibility with Bison releases prior to | |
667 | 2.4.2 is not necessary. | |
668 | ||
2755de8f AD |
669 | ** Internationalization. |
670 | ||
671 | Fix a regression introduced in Bison 2.4: Under some circumstances, | |
672 | message translations were not installed although supported by the | |
673 | host system. | |
674 | ||
74553c98 | 675 | * Changes in version 2.4.1 (2008-12-11): |
c9ba9e59 | 676 | |
0ea583d2 AD |
677 | ** In the GLR defines file, unexpanded M4 macros in the yylval and yylloc |
678 | declarations have been fixed. | |
1979121c | 679 | |
0ea583d2 AD |
680 | ** Temporary hack for adding a semicolon to the user action. |
681 | ||
682 | Bison used to prepend a trailing semicolon at the end of the user | |
683 | action for reductions. This allowed actions such as | |
684 | ||
685 | exp: exp "+" exp { $$ = $1 + $3 }; | |
686 | ||
687 | instead of | |
688 | ||
689 | exp: exp "+" exp { $$ = $1 + $3; }; | |
690 | ||
e4ab1254 | 691 | Some grammars still depend on this "feature". Bison 2.4.1 restores |
0ea583d2 AD |
692 | the previous behavior in the case of C output (specifically, when |
693 | neither %language or %skeleton or equivalent command-line options | |
694 | are used) to leave more time for grammars depending on the old | |
695 | behavior to be adjusted. Future releases of Bison will disable this | |
696 | feature. | |
697 | ||
698 | ** A few minor improvements to the Bison manual. | |
c9ba9e59 | 699 | |
402b123d | 700 | * Changes in version 2.4 (2008-11-02): |
7bd1665a | 701 | |
402b123d | 702 | ** %language is an experimental feature. |
ed4d67dc JD |
703 | |
704 | We first introduced this feature in test release 2.3b as a cleaner | |
705 | alternative to %skeleton. Since then, we have discussed the possibility of | |
706 | modifying its effect on Bison's output file names. Thus, in this release, | |
707 | we consider %language to be an experimental feature that will likely evolve | |
708 | in future releases. | |
7bd1665a | 709 | |
402b123d | 710 | ** Forward compatibility with GNU M4 has been improved. |
241fda7a | 711 | |
402b123d | 712 | ** Several bugs in the C++ skeleton and the experimental Java skeleton have been |
241fda7a JD |
713 | fixed. |
714 | ||
402b123d | 715 | * Changes in version 2.3b (2008-05-27): |
35fe0834 | 716 | |
402b123d | 717 | ** The quotes around NAME that used to be required in the following directive |
d9df47b6 JD |
718 | are now deprecated: |
719 | ||
720 | %define NAME "VALUE" | |
721 | ||
e4ab1254 | 722 | ** The directive "%pure-parser" is now deprecated in favor of: |
d9df47b6 JD |
723 | |
724 | %define api.pure | |
725 | ||
726 | which has the same effect except that Bison is more careful to warn about | |
727 | unreasonable usage in the latter case. | |
728 | ||
402b123d | 729 | ** Push Parsing |
c373bf8b JD |
730 | |
731 | Bison can now generate an LALR(1) parser in C with a push interface. That | |
e4ab1254 AD |
732 | is, instead of invoking "yyparse", which pulls tokens from "yylex", you can |
733 | push one token at a time to the parser using "yypush_parse", which will | |
c373bf8b JD |
734 | return to the caller after processing each token. By default, the push |
735 | interface is disabled. Either of the following directives will enable it: | |
736 | ||
737 | %define api.push_pull "push" // Just push; does not require yylex. | |
738 | %define api.push_pull "both" // Push and pull; requires yylex. | |
739 | ||
e4ab1254 | 740 | See the new section "A Push Parser" in the Bison manual for details. |
c373bf8b | 741 | |
59da312b JD |
742 | The current push parsing interface is experimental and may evolve. More user |
743 | feedback will help to stabilize it. | |
744 | ||
402b123d | 745 | ** The -g and --graph options now output graphs in Graphviz DOT format, |
8e55b3aa JD |
746 | not VCG format. Like --graph, -g now also takes an optional FILE argument |
747 | and thus cannot be bundled with other short options. | |
c373bf8b | 748 | |
402b123d | 749 | ** Java |
59da312b JD |
750 | |
751 | Bison can now generate an LALR(1) parser in Java. The skeleton is | |
e4ab1254 | 752 | "data/lalr1.java". Consider using the new %language directive instead of |
59da312b JD |
753 | %skeleton to select it. |
754 | ||
e4ab1254 | 755 | See the new section "Java Parsers" in the Bison manual for details. |
59da312b JD |
756 | |
757 | The current Java interface is experimental and may evolve. More user | |
758 | feedback will help to stabilize it. | |
759 | ||
402b123d | 760 | ** %language |
59da312b JD |
761 | |
762 | This new directive specifies the programming language of the generated | |
d43f77e7 PB |
763 | parser, which can be C (the default), C++, or Java. Besides the skeleton |
764 | that Bison uses, the directive affects the names of the generated files if | |
765 | the grammar file's name ends in ".y". | |
59da312b | 766 | |
402b123d | 767 | ** XML Automaton Report |
59da312b JD |
768 | |
769 | Bison can now generate an XML report of the LALR(1) automaton using the new | |
e4ab1254 | 770 | "--xml" option. The current XML schema is experimental and may evolve. More |
59da312b | 771 | user feedback will help to stabilize it. |
c373bf8b | 772 | |
402b123d | 773 | ** The grammar file may now specify the name of the parser header file using |
c373bf8b JD |
774 | %defines. For example: |
775 | ||
776 | %defines "parser.h" | |
777 | ||
402b123d | 778 | ** When reporting useless rules, useless nonterminals, and unused terminals, |
d80fb37a JD |
779 | Bison now employs the terms "useless in grammar" instead of "useless", |
780 | "useless in parser" instead of "never reduced", and "unused in grammar" | |
781 | instead of "unused". | |
cff03fb2 | 782 | |
402b123d | 783 | ** Unreachable State Removal |
c373bf8b JD |
784 | |
785 | Previously, Bison sometimes generated parser tables containing unreachable | |
31984206 JD |
786 | states. A state can become unreachable during conflict resolution if Bison |
787 | disables a shift action leading to it from a predecessor state. Bison now: | |
75ad86ee JD |
788 | |
789 | 1. Removes unreachable states. | |
790 | ||
791 | 2. Does not report any conflicts that appeared in unreachable states. | |
792 | WARNING: As a result, you may need to update %expect and %expect-rr | |
793 | directives in existing grammar files. | |
794 | ||
795 | 3. For any rule used only in such states, Bison now reports the rule as | |
cff03fb2 | 796 | "useless in parser due to conflicts". |
75ad86ee | 797 | |
31984206 JD |
798 | This feature can be disabled with the following directive: |
799 | ||
800 | %define lr.keep_unreachable_states | |
801 | ||
e4ab1254 | 802 | See the %define entry in the "Bison Declaration Summary" in the Bison manual |
31984206 JD |
803 | for further discussion. |
804 | ||
e4ab1254 | 805 | ** Lookahead Set Correction in the ".output" Report |
b1cc23c4 | 806 | |
e4ab1254 AD |
807 | When instructed to generate a ".output" file including lookahead sets |
808 | (using "--report=lookahead", for example), Bison now prints each reduction's | |
88c78747 JD |
809 | lookahead set only next to the associated state's one item that (1) is |
810 | associated with the same rule as the reduction and (2) has its dot at the end | |
811 | of its RHS. Previously, Bison also erroneously printed the lookahead set | |
812 | next to all of the state's other items associated with the same rule. This | |
e4ab1254 | 813 | bug affected only the ".output" file and not the generated parser source |
88c78747 JD |
814 | code. |
815 | ||
e4ab1254 | 816 | ** --report-file=FILE is a new option to override the default ".output" file |
59da312b | 817 | name. |
1bb2bd75 | 818 | |
e4ab1254 | 819 | ** The "=" that used to be required in the following directives is now |
02975b9a JD |
820 | deprecated: |
821 | ||
822 | %file-prefix "parser" | |
823 | %name-prefix "c_" | |
824 | %output "parser.c" | |
825 | ||
e4ab1254 | 826 | ** An Alternative to "%{...%}" -- "%code QUALIFIER {CODE}" |
c373bf8b JD |
827 | |
828 | Bison 2.3a provided a new set of directives as a more flexible alternative to | |
8e0a5e9e JD |
829 | the traditional Yacc prologue blocks. Those have now been consolidated into |
830 | a single %code directive with an optional qualifier field, which identifies | |
831 | the purpose of the code and thus the location(s) where Bison should generate | |
832 | it: | |
833 | ||
e4ab1254 AD |
834 | 1. "%code {CODE}" replaces "%after-header {CODE}" |
835 | 2. "%code requires {CODE}" replaces "%start-header {CODE}" | |
836 | 3. "%code provides {CODE}" replaces "%end-header {CODE}" | |
837 | 4. "%code top {CODE}" replaces "%before-header {CODE}" | |
8e0a5e9e | 838 | |
e4ab1254 AD |
839 | See the %code entries in section "Bison Declaration Summary" in the Bison |
840 | manual for a summary of the new functionality. See the new section "Prologue | |
841 | Alternatives" for a detailed discussion including the advantages of %code | |
8e0a5e9e JD |
842 | over the traditional Yacc prologues. |
843 | ||
844 | The prologue alternatives are experimental. More user feedback will help to | |
845 | determine whether they should become permanent features. | |
846 | ||
402b123d | 847 | ** Revised warning: unset or unused mid-rule values |
17bd8a73 JD |
848 | |
849 | Since Bison 2.2, Bison has warned about mid-rule values that are set but not | |
850 | used within any of the actions of the parent rule. For example, Bison warns | |
851 | about unused $2 in: | |
852 | ||
853 | exp: '1' { $$ = 1; } '+' exp { $$ = $1 + $4; }; | |
854 | ||
855 | Now, Bison also warns about mid-rule values that are used but not set. For | |
856 | example, Bison warns about unset $$ in the mid-rule action in: | |
857 | ||
858 | exp: '1' { $1 = 1; } '+' exp { $$ = $2 + $4; }; | |
859 | ||
860 | However, Bison now disables both of these warnings by default since they | |
861 | sometimes prove to be false alarms in existing grammars employing the Yacc | |
862 | constructs $0 or $-N (where N is some positive integer). | |
863 | ||
e4ab1254 AD |
864 | To enable these warnings, specify the option "--warnings=midrule-values" or |
865 | "-W", which is a synonym for "--warnings=all". | |
17bd8a73 | 866 | |
e4ab1254 | 867 | ** Default %destructor or %printer with "<*>" or "<>" |
c373bf8b JD |
868 | |
869 | Bison now recognizes two separate kinds of default %destructor's and | |
12e35840 JD |
870 | %printer's: |
871 | ||
e4ab1254 | 872 | 1. Place "<*>" in a %destructor/%printer symbol list to define a default |
12e35840 JD |
873 | %destructor/%printer for all grammar symbols for which you have formally |
874 | declared semantic type tags. | |
875 | ||
e4ab1254 | 876 | 2. Place "<>" in a %destructor/%printer symbol list to define a default |
12e35840 JD |
877 | %destructor/%printer for all grammar symbols without declared semantic |
878 | type tags. | |
879 | ||
e4ab1254 AD |
880 | Bison no longer supports the "%symbol-default" notation from Bison 2.3a. |
881 | "<*>" and "<>" combined achieve the same effect with one exception: Bison no | |
12e35840 JD |
882 | longer applies any %destructor to a mid-rule value if that mid-rule value is |
883 | not actually ever referenced using either $$ or $n in a semantic action. | |
884 | ||
85894313 JD |
885 | The default %destructor's and %printer's are experimental. More user |
886 | feedback will help to determine whether they should become permanent | |
887 | features. | |
888 | ||
e4ab1254 | 889 | See the section "Freeing Discarded Symbols" in the Bison manual for further |
12e35840 JD |
890 | details. |
891 | ||
402b123d | 892 | ** %left, %right, and %nonassoc can now declare token numbers. This is required |
e4ab1254 | 893 | by POSIX. However, see the end of section "Operator Precedence" in the Bison |
ab7f29f8 JD |
894 | manual for a caveat concerning the treatment of literal strings. |
895 | ||
402b123d | 896 | ** The nonfunctional --no-parser, -n, and %no-parser options have been |
b1cc23c4 JD |
897 | completely removed from Bison. |
898 | ||
402b123d | 899 | * Changes in version 2.3a, 2006-09-13: |
742e4900 | 900 | |
402b123d | 901 | ** Instead of %union, you can define and use your own union type |
ddc8ede1 PE |
902 | YYSTYPE if your grammar contains at least one <type> tag. |
903 | Your YYSTYPE need not be a macro; it can be a typedef. | |
904 | This change is for compatibility with other Yacc implementations, | |
905 | and is required by POSIX. | |
906 | ||
402b123d | 907 | ** Locations columns and lines start at 1. |
cd48d21d AD |
908 | In accordance with the GNU Coding Standards and Emacs. |
909 | ||
402b123d | 910 | ** You may now declare per-type and default %destructor's and %printer's: |
ec5479ce JD |
911 | |
912 | For example: | |
913 | ||
b2a0b7ca JD |
914 | %union { char *string; } |
915 | %token <string> STRING1 | |
916 | %token <string> STRING2 | |
917 | %type <string> string1 | |
918 | %type <string> string2 | |
919 | %union { char character; } | |
920 | %token <character> CHR | |
921 | %type <character> chr | |
922 | %destructor { free ($$); } %symbol-default | |
923 | %destructor { free ($$); printf ("%d", @$.first_line); } STRING1 string1 | |
924 | %destructor { } <character> | |
925 | ||
926 | guarantees that, when the parser discards any user-defined symbol that has a | |
e4ab1254 AD |
927 | semantic type tag other than "<character>", it passes its semantic value to |
928 | "free". However, when the parser discards a "STRING1" or a "string1", it | |
929 | also prints its line number to "stdout". It performs only the second | |
930 | "%destructor" in this case, so it invokes "free" only once. | |
ec5479ce | 931 | |
85894313 JD |
932 | [Although we failed to mention this here in the 2.3a release, the default |
933 | %destructor's and %printer's were experimental, and they were rewritten in | |
934 | future versions.] | |
935 | ||
e4ab1254 AD |
936 | ** Except for LALR(1) parsers in C with POSIX Yacc emulation enabled (with "-y", |
937 | "--yacc", or "%yacc"), Bison no longer generates #define statements for | |
b931235e JD |
938 | associating token numbers with token names. Removing the #define statements |
939 | helps to sanitize the global namespace during preprocessing, but POSIX Yacc | |
940 | requires them. Bison still generates an enum for token names in all cases. | |
941 | ||
402b123d | 942 | ** Handling of traditional Yacc prologue blocks is now more consistent but |
34f98f46 | 943 | potentially incompatible with previous releases of Bison. |
9bc0dd67 JD |
944 | |
945 | As before, you declare prologue blocks in your grammar file with the | |
e4ab1254 | 946 | "%{ ... %}" syntax. To generate the pre-prologue, Bison concatenates all |
34f98f46 JD |
947 | prologue blocks that you've declared before the first %union. To generate |
948 | the post-prologue, Bison concatenates all prologue blocks that you've | |
ddc8ede1 | 949 | declared after the first %union. |
9bc0dd67 | 950 | |
34f98f46 | 951 | Previous releases of Bison inserted the pre-prologue into both the header |
9bc0dd67 JD |
952 | file and the code file in all cases except for LALR(1) parsers in C. In the |
953 | latter case, Bison inserted it only into the code file. For parsers in C++, | |
954 | the point of insertion was before any token definitions (which associate | |
955 | token numbers with names). For parsers in C, the point of insertion was | |
956 | after the token definitions. | |
957 | ||
958 | Now, Bison never inserts the pre-prologue into the header file. In the code | |
959 | file, it always inserts it before the token definitions. | |
960 | ||
402b123d | 961 | ** Bison now provides a more flexible alternative to the traditional Yacc |
34f98f46 JD |
962 | prologue blocks: %before-header, %start-header, %end-header, and |
963 | %after-header. | |
964 | ||
965 | For example, the following declaration order in the grammar file reflects the | |
966 | order in which Bison will output these code blocks. However, you are free to | |
967 | declare these code blocks in your grammar file in whatever order is most | |
968 | convenient for you: | |
969 | ||
970 | %before-header { | |
971 | /* Bison treats this block like a pre-prologue block: it inserts it into | |
972 | * the code file before the contents of the header file. It does *not* | |
973 | * insert it into the header file. This is a good place to put | |
974 | * #include's that you want at the top of your code file. A common | |
e4ab1254 | 975 | * example is '#include "system.h"'. */ |
34f98f46 JD |
976 | } |
977 | %start-header { | |
978 | /* Bison inserts this block into both the header file and the code file. | |
979 | * In both files, the point of insertion is before any Bison-generated | |
980 | * token, semantic type, location type, and class definitions. This is a | |
981 | * good place to define %union dependencies, for example. */ | |
9bc0dd67 JD |
982 | } |
983 | %union { | |
34f98f46 JD |
984 | /* Unlike the traditional Yacc prologue blocks, the output order for the |
985 | * new %*-header blocks is not affected by their declaration position | |
986 | * relative to any %union in the grammar file. */ | |
9bc0dd67 | 987 | } |
34f98f46 JD |
988 | %end-header { |
989 | /* Bison inserts this block into both the header file and the code file. | |
990 | * In both files, the point of insertion is after the Bison-generated | |
991 | * definitions. This is a good place to declare or define public | |
992 | * functions or data structures that depend on the Bison-generated | |
993 | * definitions. */ | |
9bc0dd67 | 994 | } |
34f98f46 JD |
995 | %after-header { |
996 | /* Bison treats this block like a post-prologue block: it inserts it into | |
997 | * the code file after the contents of the header file. It does *not* | |
998 | * insert it into the header file. This is a good place to declare or | |
999 | * define internal functions or data structures that depend on the | |
1000 | * Bison-generated definitions. */ | |
1001 | } | |
1002 | ||
1003 | If you have multiple occurrences of any one of the above declarations, Bison | |
1004 | will concatenate the contents in declaration order. | |
9bc0dd67 | 1005 | |
85894313 JD |
1006 | [Although we failed to mention this here in the 2.3a release, the prologue |
1007 | alternatives were experimental, and they were rewritten in future versions.] | |
1008 | ||
e4ab1254 | 1009 | ** The option "--report=look-ahead" has been changed to "--report=lookahead". |
9e6e7ed2 PE |
1010 | The old spelling still works, but is not documented and may be removed |
1011 | in a future release. | |
742e4900 | 1012 | |
402b123d | 1013 | * Changes in version 2.3, 2006-06-05: |
4ad3ed84 | 1014 | |
e4ab1254 | 1015 | ** GLR grammars should now use "YYRECOVERING ()" instead of "YYRECOVERING", |
4ad3ed84 PE |
1016 | for compatibility with LALR(1) grammars. |
1017 | ||
402b123d | 1018 | ** It is now documented that any definition of YYSTYPE or YYLTYPE should |
4ad3ed84 PE |
1019 | be to a type name that does not contain parentheses or brackets. |
1020 | ||
402b123d | 1021 | * Changes in version 2.2, 2006-05-19: |
193d7c70 | 1022 | |
402b123d | 1023 | ** The distribution terms for all Bison-generated parsers now permit |
193d7c70 PE |
1024 | using the parsers in nonfree programs. Previously, this permission |
1025 | was granted only for Bison-generated LALR(1) parsers in C. | |
5f4236a0 | 1026 | |
402b123d | 1027 | ** %name-prefix changes the namespace name in C++ outputs. |
aa08666d | 1028 | |
402b123d | 1029 | ** The C++ parsers export their token_type. |
5f4236a0 | 1030 | |
402b123d | 1031 | ** Bison now allows multiple %union declarations, and concatenates |
d6ca7905 PE |
1032 | their contents together. |
1033 | ||
402b123d | 1034 | ** New warning: unused values |
4d7bc38c PE |
1035 | Right-hand side symbols whose values are not used are reported, |
1036 | if the symbols have destructors. For instance: | |
affac613 | 1037 | |
8f3596a6 | 1038 | exp: exp "?" exp ":" exp { $1 ? $1 : $3; } |
e9690142 JD |
1039 | | exp "+" exp |
1040 | ; | |
affac613 | 1041 | |
8f3596a6 AD |
1042 | will trigger a warning about $$ and $5 in the first rule, and $3 in |
1043 | the second ($1 is copied to $$ by the default rule). This example | |
4e26c69e | 1044 | most likely contains three errors, and could be rewritten as: |
affac613 | 1045 | |
4e26c69e | 1046 | exp: exp "?" exp ":" exp |
e9690142 JD |
1047 | { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); } |
1048 | | exp "+" exp | |
1049 | { $$ = $1 ? $1 : $3; if ($1) free ($3); } | |
1050 | ; | |
affac613 | 1051 | |
4e26c69e PE |
1052 | However, if the original actions were really intended, memory leaks |
1053 | and all, the warnings can be suppressed by letting Bison believe the | |
1054 | values are used, e.g.: | |
721be13c | 1055 | |
8f3596a6 | 1056 | exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); } |
e9690142 JD |
1057 | | exp "+" exp { $$ = $1; (void) $3; } |
1058 | ; | |
721be13c | 1059 | |
84866159 AD |
1060 | If there are mid-rule actions, the warning is issued if no action |
1061 | uses it. The following triggers no warning: $1 and $3 are used. | |
1062 | ||
1063 | exp: exp { push ($1); } '+' exp { push ($3); sum (); }; | |
1064 | ||
721be13c PE |
1065 | The warning is intended to help catching lost values and memory leaks. |
1066 | If a value is ignored, its associated memory typically is not reclaimed. | |
affac613 | 1067 | |
402b123d | 1068 | ** %destructor vs. YYABORT, YYACCEPT, and YYERROR. |
9d9b8b70 PE |
1069 | Destructors are now called when user code invokes YYABORT, YYACCEPT, |
1070 | and YYERROR, for all objects on the stack, other than objects | |
1071 | corresponding to the right-hand side of the current rule. | |
a85284cf | 1072 | |
402b123d | 1073 | ** %expect, %expect-rr |
035aa4a0 PE |
1074 | Incorrect numbers of expected conflicts are now actual errors, |
1075 | instead of warnings. | |
1076 | ||
402b123d | 1077 | ** GLR, YACC parsers. |
4e26c69e PE |
1078 | The %parse-params are available in the destructors (and the |
1079 | experimental printers) as per the documentation. | |
4b367315 | 1080 | |
e4ab1254 | 1081 | ** Bison now warns if it finds a stray "$" or "@" in an action. |
ad6a9b97 | 1082 | |
402b123d | 1083 | ** %require "VERSION" |
4e26c69e PE |
1084 | This specifies that the grammar file depends on features implemented |
1085 | in Bison version VERSION or higher. | |
b50d2359 | 1086 | |
402b123d | 1087 | ** lalr1.cc: The token and value types are now class members. |
e14d0ab6 AD |
1088 | The tokens were defined as free form enums and cpp macros. YYSTYPE |
1089 | was defined as a free form union. They are now class members: | |
e4ab1254 AD |
1090 | tokens are enumerations of the "yy::parser::token" struct, and the |
1091 | semantic values have the "yy::parser::semantic_type" type. | |
fb9712a9 AD |
1092 | |
1093 | If you do not want or can update to this scheme, the directive | |
e4ab1254 | 1094 | '%define "global_tokens_and_yystype" "1"' triggers the global |
b50d2359 AD |
1095 | definition of tokens and YYSTYPE. This change is suitable both |
1096 | for previous releases of Bison, and this one. | |
fb9712a9 | 1097 | |
b50d2359 | 1098 | If you wish to update, then make sure older version of Bison will |
e4ab1254 | 1099 | fail using '%require "2.2"'. |
fb9712a9 | 1100 | |
402b123d | 1101 | ** DJGPP support added. |
193d7c70 | 1102 | \f |
402b123d | 1103 | * Changes in version 2.1, 2005-09-16: |
1ce59070 | 1104 | |
402b123d | 1105 | ** The C++ lalr1.cc skeleton supports %lex-param. |
e14d0ab6 | 1106 | |
402b123d | 1107 | ** Bison-generated parsers now support the translation of diagnostics like |
baf785db PE |
1108 | "syntax error" into languages other than English. The default |
1109 | language is still English. For details, please see the new | |
0410a6e0 PE |
1110 | Internationalization section of the Bison manual. Software |
1111 | distributors should also see the new PACKAGING file. Thanks to | |
1112 | Bruno Haible for this new feature. | |
1ce59070 | 1113 | |
402b123d | 1114 | ** Wording in the Bison-generated parsers has been changed slightly to |
1a059451 PE |
1115 | simplify translation. In particular, the message "memory exhausted" |
1116 | has replaced "parser stack overflow", as the old message was not | |
1117 | always accurate for modern Bison-generated parsers. | |
1118 | ||
402b123d | 1119 | ** Destructors are now called when the parser aborts, for all symbols left |
258b75ca PE |
1120 | behind on the stack. Also, the start symbol is now destroyed after a |
1121 | successful parse. In both cases, the behavior was formerly inconsistent. | |
1122 | ||
402b123d | 1123 | ** When generating verbose diagnostics, Bison-generated parsers no longer |
72f000b0 PE |
1124 | quote the literal strings associated with tokens. For example, for |
1125 | a syntax error associated with '%token NUM "number"' they might | |
1126 | print 'syntax error, unexpected number' instead of 'syntax error, | |
1127 | unexpected "number"'. | |
193d7c70 | 1128 | \f |
402b123d | 1129 | * Changes in version 2.0, 2004-12-25: |
efeed023 | 1130 | |
402b123d | 1131 | ** Possibly-incompatible changes |
d7e14fc0 | 1132 | |
82de6b0d PE |
1133 | - Bison-generated parsers no longer default to using the alloca function |
1134 | (when available) to extend the parser stack, due to widespread | |
1135 | problems in unchecked stack-overflow detection. You can "#define | |
1136 | YYSTACK_USE_ALLOCA 1" to require the use of alloca, but please read | |
1137 | the manual to determine safe values for YYMAXDEPTH in that case. | |
8dd162d3 | 1138 | |
82de6b0d PE |
1139 | - Error token location. |
1140 | During error recovery, the location of the syntax error is updated | |
1141 | to cover the whole sequence covered by the error token: it includes | |
1142 | the shifted symbols thrown away during the first part of the error | |
1143 | recovery, and the lookahead rejected during the second part. | |
18d192f0 | 1144 | |
82de6b0d PE |
1145 | - Semicolon changes: |
1146 | . Stray semicolons are no longer allowed at the start of a grammar. | |
1147 | . Semicolons are now required after in-grammar declarations. | |
e342c3be | 1148 | |
82de6b0d PE |
1149 | - Unescaped newlines are no longer allowed in character constants or |
1150 | string literals. They were never portable, and GCC 3.4.0 has | |
1151 | dropped support for them. Better diagnostics are now generated if | |
1152 | forget a closing quote. | |
8dd162d3 | 1153 | |
82de6b0d | 1154 | - NUL bytes are no longer allowed in Bison string literals, unfortunately. |
f74b6f91 | 1155 | |
402b123d | 1156 | ** New features |
1452af69 | 1157 | |
82de6b0d | 1158 | - GLR grammars now support locations. |
4febdd96 | 1159 | |
82de6b0d PE |
1160 | - New directive: %initial-action. |
1161 | This directive allows the user to run arbitrary code (including | |
1162 | initializing @$) from yyparse before parsing starts. | |
1452af69 | 1163 | |
82de6b0d PE |
1164 | - A new directive "%expect-rr N" specifies the expected number of |
1165 | reduce/reduce conflicts in GLR parsers. | |
1452af69 | 1166 | |
e4ab1254 | 1167 | - %token numbers can now be hexadecimal integers, e.g., "%token FOO 0x12d". |
82de6b0d | 1168 | This is a GNU extension. |
4febdd96 | 1169 | |
e4ab1254 | 1170 | - The option "--report=lookahead" was changed to "--report=look-ahead". |
9e6e7ed2 | 1171 | [However, this was changed back after 2.3.] |
1452af69 | 1172 | |
82de6b0d | 1173 | - Experimental %destructor support has been added to lalr1.cc. |
1452af69 | 1174 | |
82de6b0d PE |
1175 | - New configure option --disable-yacc, to disable installation of the |
1176 | yacc command and -ly library introduced in 1.875 for POSIX conformance. | |
6040d338 | 1177 | |
402b123d | 1178 | ** Bug fixes |
d5a3fe37 | 1179 | |
82de6b0d PE |
1180 | - For now, %expect-count violations are now just warnings, not errors. |
1181 | This is for compatibility with Bison 1.75 and earlier (when there are | |
1182 | reduce/reduce conflicts) and with Bison 1.30 and earlier (when there | |
1183 | are too many or too few shift/reduce conflicts). However, in future | |
1184 | versions of Bison we plan to improve the %expect machinery so that | |
1185 | these violations will become errors again. | |
3473d0f8 | 1186 | |
82de6b0d PE |
1187 | - Within Bison itself, numbers (e.g., goto numbers) are no longer |
1188 | arbitrarily limited to 16-bit counts. | |
d600ee67 | 1189 | |
82de6b0d | 1190 | - Semicolons are now allowed before "|" in grammar rules, as POSIX requires. |
d600ee67 | 1191 | \f |
402b123d | 1192 | * Changes in version 1.875, 2003-01-01: |
963fcc17 | 1193 | |
402b123d | 1194 | ** The documentation license has been upgraded to version 1.2 |
dc546b0f | 1195 | of the GNU Free Documentation License. |
75eb3bc4 | 1196 | |
402b123d | 1197 | ** syntax error processing |
75eb3bc4 | 1198 | |
dc546b0f PE |
1199 | - In Yacc-style parsers YYLLOC_DEFAULT is now used to compute error |
1200 | locations too. This fixes bugs in error-location computation. | |
75eb3bc4 | 1201 | |
dc546b0f PE |
1202 | - %destructor |
1203 | It is now possible to reclaim the memory associated to symbols | |
1204 | discarded during error recovery. This feature is still experimental. | |
20daca06 | 1205 | |
dc546b0f PE |
1206 | - %error-verbose |
1207 | This new directive is preferred over YYERROR_VERBOSE. | |
74724a70 | 1208 | |
dc546b0f PE |
1209 | - #defining yyerror to steal internal variables is discouraged. |
1210 | It is not guaranteed to work forever. | |
d1de5372 | 1211 | |
402b123d | 1212 | ** POSIX conformance |
d1de5372 | 1213 | |
dc546b0f PE |
1214 | - Semicolons are once again optional at the end of grammar rules. |
1215 | This reverts to the behavior of Bison 1.33 and earlier, and improves | |
1216 | compatibility with Yacc. | |
74724a70 | 1217 | |
e4ab1254 AD |
1218 | - "parse error" -> "syntax error" |
1219 | Bison now uniformly uses the term "syntax error"; formerly, the code | |
1220 | and manual sometimes used the term "parse error" instead. POSIX | |
1221 | requires "syntax error" in diagnostics, and it was thought better to | |
dc546b0f | 1222 | be consistent. |
74724a70 | 1223 | |
dc546b0f PE |
1224 | - The documentation now emphasizes that yylex and yyerror must be |
1225 | declared before use. C99 requires this. | |
d1de5372 | 1226 | |
dc546b0f PE |
1227 | - Bison now parses C99 lexical constructs like UCNs and |
1228 | backslash-newline within C escape sequences, as POSIX 1003.1-2001 requires. | |
d1de5372 | 1229 | |
dc546b0f PE |
1230 | - File names are properly escaped in C output. E.g., foo\bar.y is |
1231 | output as "foo\\bar.y". | |
6780ca7a | 1232 | |
dc546b0f | 1233 | - Yacc command and library now available |
e4ab1254 | 1234 | The Bison distribution now installs a "yacc" command, as POSIX requires. |
dc546b0f PE |
1235 | Also, Bison now installs a small library liby.a containing |
1236 | implementations of Yacc-compatible yyerror and main functions. | |
1237 | This library is normally not useful, but POSIX requires it. | |
6e649e65 | 1238 | |
dc546b0f | 1239 | - Type clashes now generate warnings, not errors. |
6e649e65 | 1240 | |
dc546b0f PE |
1241 | - If the user does not define YYSTYPE as a macro, Bison now declares it |
1242 | using typedef instead of defining it as a macro. | |
1243 | For consistency, YYLTYPE is also declared instead of defined. | |
9501dc6e | 1244 | |
402b123d | 1245 | ** Other compatibility issues |
886a425c | 1246 | |
e4ab1254 AD |
1247 | - %union directives can now have a tag before the "{", e.g., the |
1248 | directive "%union foo {...}" now generates the C code | |
1249 | "typedef union foo { ... } YYSTYPE;"; this is for Yacc compatibility. | |
1250 | The default union tag is "YYSTYPE", for compatibility with Solaris 9 Yacc. | |
1251 | For consistency, YYLTYPE's struct tag is now "YYLTYPE" not "yyltype". | |
dc546b0f | 1252 | This is for compatibility with both Yacc and Bison 1.35. |
72f889cc | 1253 | |
e4ab1254 | 1254 | - ";" is output before the terminating "}" of an action, for |
dc546b0f | 1255 | compatibility with Bison 1.35. |
886a425c | 1256 | |
dc546b0f | 1257 | - Bison now uses a Yacc-style format for conflict reports, e.g., |
e4ab1254 | 1258 | "conflicts: 2 shift/reduce, 1 reduce/reduce". |
437c2d80 | 1259 | |
e4ab1254 | 1260 | - "yystype" and "yyltype" are now obsolescent macros instead of being |
dc546b0f PE |
1261 | typedefs or tags; they are no longer documented and are planned to be |
1262 | withdrawn in a future release. | |
2a8d363a | 1263 | |
402b123d | 1264 | ** GLR parser notes |
2a8d363a | 1265 | |
dc546b0f PE |
1266 | - GLR and inline |
1267 | Users of Bison have to decide how they handle the portability of the | |
e4ab1254 | 1268 | C keyword "inline". |
959e5f51 | 1269 | |
e4ab1254 AD |
1270 | - "parsing stack overflow..." -> "parser stack overflow" |
1271 | GLR parsers now report "parser stack overflow" as per the Bison manual. | |
900c5db5 | 1272 | |
402b123d | 1273 | ** Bison now warns if it detects conflicting outputs to the same file, |
e4ab1254 | 1274 | e.g., it generates a warning for "bison -d -o foo.h foo.y" since |
dc546b0f | 1275 | that command outputs both code and header to foo.h. |
6e40b4eb | 1276 | |
402b123d | 1277 | ** #line in output files |
dc546b0f | 1278 | - --no-line works properly. |
6e40b4eb | 1279 | |
402b123d | 1280 | ** Bison can no longer be built by a K&R C compiler; it requires C89 or |
6e40b4eb AD |
1281 | later to be built. This change originally took place a few versions |
1282 | ago, but nobody noticed until we recently asked someone to try | |
1283 | building Bison with a K&R C compiler. | |
d600ee67 | 1284 | \f |
402b123d | 1285 | * Changes in version 1.75, 2002-10-14: |
7933f2b5 | 1286 | |
402b123d | 1287 | ** Bison should now work on 64-bit hosts. |
7933f2b5 | 1288 | |
402b123d | 1289 | ** Indonesian translation thanks to Tedi Heriyanto. |
7933f2b5 | 1290 | |
402b123d | 1291 | ** GLR parsers |
f50adbbd AD |
1292 | Fix spurious parse errors. |
1293 | ||
402b123d | 1294 | ** Pure parsers |
f50adbbd AD |
1295 | Some people redefine yyerror to steal yyparse' private variables. |
1296 | Reenable this trick until an official feature replaces it. | |
1297 | ||
402b123d | 1298 | ** Type Clashes |
d90c934c AD |
1299 | In agreement with POSIX and with other Yaccs, leaving a default |
1300 | action is valid when $$ is untyped, and $1 typed: | |
1301 | ||
e9690142 | 1302 | untyped: ... typed; |
d90c934c AD |
1303 | |
1304 | but the converse remains an error: | |
1305 | ||
e9690142 | 1306 | typed: ... untyped; |
d90c934c | 1307 | |
402b123d | 1308 | ** Values of mid-rule actions |
d90c934c AD |
1309 | The following code: |
1310 | ||
e9690142 | 1311 | foo: { ... } { $$ = $1; } ... |
d90c934c AD |
1312 | |
1313 | was incorrectly rejected: $1 is defined in the second mid-rule | |
1314 | action, and is equal to the $$ of the first mid-rule action. | |
d600ee67 | 1315 | \f |
402b123d | 1316 | * Changes in version 1.50, 2002-10-04: |
adc8c848 | 1317 | |
402b123d | 1318 | ** GLR parsing |
676385e2 PH |
1319 | The declaration |
1320 | %glr-parser | |
1321 | causes Bison to produce a Generalized LR (GLR) parser, capable of handling | |
1322 | almost any context-free grammar, ambiguous or not. The new declarations | |
e8832397 | 1323 | %dprec and %merge on grammar rules allow parse-time resolution of |
676385e2 PH |
1324 | ambiguities. Contributed by Paul Hilfinger. |
1325 | ||
7933f2b5 | 1326 | Unfortunately Bison 1.50 does not work properly on 64-bit hosts |
420f93c8 PE |
1327 | like the Alpha, so please stick to 32-bit hosts for now. |
1328 | ||
402b123d | 1329 | ** Output Directory |
8c165d89 | 1330 | When not in Yacc compatibility mode, when the output file was not |
e4ab1254 AD |
1331 | specified, running "bison foo/bar.y" created "foo/bar.c". It |
1332 | now creates "bar.c". | |
8c165d89 | 1333 | |
402b123d | 1334 | ** Undefined token |
007a50a4 | 1335 | The undefined token was systematically mapped to 2 which prevented |
e88dbdbf | 1336 | the use of 2 by the user. This is no longer the case. |
007a50a4 | 1337 | |
402b123d | 1338 | ** Unknown token numbers |
e88dbdbf | 1339 | If yylex returned an out of range value, yyparse could die. This is |
007a50a4 AD |
1340 | no longer the case. |
1341 | ||
402b123d | 1342 | ** Error token |
e88dbdbf | 1343 | According to POSIX, the error token must be 256. |
23c5a174 AD |
1344 | Bison extends this requirement by making it a preference: *if* the |
1345 | user specified that one of her tokens is numbered 256, then error | |
1346 | will be mapped onto another number. | |
1347 | ||
402b123d | 1348 | ** Verbose error messages |
e4ab1254 | 1349 | They no longer report "..., expecting error or..." for states where |
217598da AD |
1350 | error recovery is possible. |
1351 | ||
402b123d | 1352 | ** End token |
e4ab1254 | 1353 | Defaults to "$end" instead of "$". |
217598da | 1354 | |
402b123d | 1355 | ** Error recovery now conforms to documentation and to POSIX |
68cd8af3 PE |
1356 | When a Bison-generated parser encounters a syntax error, it now pops |
1357 | the stack until it finds a state that allows shifting the error | |
1358 | token. Formerly, it popped the stack until it found a state that | |
1359 | allowed some non-error action other than a default reduction on the | |
1360 | error token. The new behavior has long been the documented behavior, | |
1361 | and has long been required by POSIX. For more details, please see | |
337116ba PE |
1362 | Paul Eggert, "Reductions during Bison error handling" (2002-05-20) |
1363 | <http://lists.gnu.org/archive/html/bug-bison/2002-05/msg00038.html>. | |
68cd8af3 | 1364 | |
402b123d | 1365 | ** Traces |
5504898e AD |
1366 | Popped tokens and nonterminals are now reported. |
1367 | ||
402b123d | 1368 | ** Larger grammars |
a861a339 PE |
1369 | Larger grammars are now supported (larger token numbers, larger grammar |
1370 | size (= sum of the LHS and RHS lengths), larger LALR tables). | |
1371 | Formerly, many of these numbers ran afoul of 16-bit limits; | |
1372 | now these limits are 32 bits on most hosts. | |
355e7c1c | 1373 | |
402b123d | 1374 | ** Explicit initial rule |
643a5994 AD |
1375 | Bison used to play hacks with the initial rule, which the user does |
1376 | not write. It is now explicit, and visible in the reports and | |
1377 | graphs as rule 0. | |
23c5a174 | 1378 | |
402b123d | 1379 | ** Useless rules |
643a5994 | 1380 | Before, Bison reported the useless rules, but, although not used, |
77714df2 | 1381 | included them in the parsers. They are now actually removed. |
23c5a174 | 1382 | |
402b123d | 1383 | ** Useless rules, useless nonterminals |
6b98e4b5 AD |
1384 | They are now reported, as a warning, with their locations. |
1385 | ||
402b123d | 1386 | ** Rules never reduced |
e8832397 AD |
1387 | Rules that can never be reduced because of conflicts are now |
1388 | reported. | |
1389 | ||
e4ab1254 | 1390 | ** Incorrect "Token not used" |
11652ab3 AD |
1391 | On a grammar such as |
1392 | ||
1393 | %token useless useful | |
1394 | %% | |
1395 | exp: '0' %prec useful; | |
1396 | ||
1397 | where a token was used to set the precedence of the last rule, | |
e4ab1254 | 1398 | bison reported both "useful" and "useless" as useless tokens. |
11652ab3 | 1399 | |
402b123d | 1400 | ** Revert the C++ namespace changes introduced in 1.31 |
77714df2 | 1401 | as they caused too many portability hassles. |
0179dd65 | 1402 | |
402b123d | 1403 | ** Default locations |
b2d52318 AD |
1404 | By an accident of design, the default computation of @$ was |
1405 | performed after another default computation was performed: @$ = @1. | |
1406 | The latter is now removed: YYLLOC_DEFAULT is fully responsible of | |
1407 | the computation of @$. | |
adc8c848 | 1408 | |
402b123d | 1409 | ** Token end-of-file |
b7c49edf AD |
1410 | The token end of file may be specified by the user, in which case, |
1411 | the user symbol is used in the reports, the graphs, and the verbose | |
e4ab1254 | 1412 | error messages instead of "$end", which remains being the default. |
b7c49edf | 1413 | For instance |
7bd6c77e | 1414 | %token MYEOF 0 |
b7c49edf | 1415 | or |
7bd6c77e | 1416 | %token MYEOF 0 "end of file" |
fdbcd8e2 | 1417 | |
402b123d | 1418 | ** Semantic parser |
fdbcd8e2 AD |
1419 | This old option, which has been broken for ages, is removed. |
1420 | ||
402b123d | 1421 | ** New translations |
a861a339 | 1422 | Brazilian Portuguese, thanks to Alexandre Folle de Menezes. |
84614e13 AD |
1423 | Croatian, thanks to Denis Lackovic. |
1424 | ||
402b123d | 1425 | ** Incorrect token definitions |
e4ab1254 AD |
1426 | When given |
1427 | %token 'a' "A" | |
1428 | bison used to output | |
1429 | #define 'a' 65 | |
b87f8b21 | 1430 | |
402b123d | 1431 | ** Token definitions as enums |
77714df2 AD |
1432 | Tokens are output both as the traditional #define's, and, provided |
1433 | the compiler supports ANSI C or is a C++ compiler, as enums. | |
e88dbdbf | 1434 | This lets debuggers display names instead of integers. |
77714df2 | 1435 | |
402b123d | 1436 | ** Reports |
ec3bc396 AD |
1437 | In addition to --verbose, bison supports --report=THINGS, which |
1438 | produces additional information: | |
b408954b AD |
1439 | - itemset |
1440 | complete the core item sets with their closure | |
e4ab1254 | 1441 | - lookahead [changed to "look-ahead" in 1.875e through 2.3, but changed back] |
9e6e7ed2 | 1442 | explicitly associate lookahead tokens to items |
b408954b AD |
1443 | - solved |
1444 | describe shift/reduce conflicts solving. | |
1445 | Bison used to systematically output this information on top of | |
1446 | the report. Solved conflicts are now attached to their states. | |
ec3bc396 | 1447 | |
402b123d | 1448 | ** Type clashes |
9af3fbce AD |
1449 | Previous versions don't complain when there is a type clash on |
1450 | the default action if the rule has a mid-rule action, such as in: | |
1451 | ||
1452 | %type <foo> bar | |
1453 | %% | |
1454 | bar: '0' {} '0'; | |
1455 | ||
1456 | This is fixed. | |
a861a339 | 1457 | |
402b123d | 1458 | ** GNU M4 is now required when using Bison. |
f987e9d2 | 1459 | \f |
402b123d | 1460 | * Changes in version 1.35, 2002-03-25: |
76551463 | 1461 | |
402b123d | 1462 | ** C Skeleton |
76551463 AD |
1463 | Some projects use Bison's C parser with C++ compilers, and define |
1464 | YYSTYPE as a class. The recent adjustment of C parsers for data | |
1465 | alignment and 64 bit architectures made this impossible. | |
1466 | ||
1467 | Because for the time being no real solution for C++ parser | |
1468 | generation exists, kludges were implemented in the parser to | |
1469 | maintain this use. In the future, when Bison has C++ parsers, this | |
1470 | kludge will be disabled. | |
1471 | ||
1472 | This kludge also addresses some C++ problems when the stack was | |
1473 | extended. | |
76551463 | 1474 | \f |
402b123d | 1475 | * Changes in version 1.34, 2002-03-12: |
76551463 | 1476 | |
402b123d | 1477 | ** File name clashes are detected |
76551463 | 1478 | $ bison foo.y -d -o foo.x |
e4ab1254 | 1479 | fatal error: header and parser would both be named "foo.x" |
76551463 | 1480 | |
e4ab1254 | 1481 | ** A missing ";" at the end of a rule triggers a warning |
76551463 AD |
1482 | In accordance with POSIX, and in agreement with other |
1483 | Yacc implementations, Bison will mandate this semicolon in the near | |
1484 | future. This eases the implementation of a Bison parser of Bison | |
1485 | grammars by making this grammar LALR(1) instead of LR(2). To | |
1486 | facilitate the transition, this release introduces a warning. | |
1487 | ||
402b123d | 1488 | ** Revert the C++ namespace changes introduced in 1.31, as they caused too |
76551463 AD |
1489 | many portability hassles. |
1490 | ||
402b123d | 1491 | ** DJGPP support added. |
76551463 | 1492 | |
402b123d | 1493 | ** Fix test suite portability problems. |
76551463 | 1494 | \f |
402b123d | 1495 | * Changes in version 1.33, 2002-02-07: |
76551463 | 1496 | |
402b123d | 1497 | ** Fix C++ issues |
76551463 AD |
1498 | Groff could not be compiled for the definition of size_t was lacking |
1499 | under some conditions. | |
1500 | ||
402b123d | 1501 | ** Catch invalid @n |
76551463 AD |
1502 | As is done with $n. |
1503 | \f | |
402b123d | 1504 | * Changes in version 1.32, 2002-01-23: |
76551463 | 1505 | |
402b123d | 1506 | ** Fix Yacc output file names |
76551463 | 1507 | |
402b123d | 1508 | ** Portability fixes |
76551463 | 1509 | |
402b123d | 1510 | ** Italian, Dutch translations |
76551463 | 1511 | \f |
402b123d | 1512 | * Changes in version 1.31, 2002-01-14: |
52d1aeee | 1513 | |
402b123d | 1514 | ** Many Bug Fixes |
52d1aeee | 1515 | |
402b123d | 1516 | ** GNU Gettext and %expect |
52d1aeee MA |
1517 | GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that |
1518 | Bison dies on incorrect %expectations, we fear there will be | |
1519 | too many bug reports for Gettext, so _for the time being_, %expect | |
e4ab1254 | 1520 | does not trigger an error when the input file is named "plural.y". |
52d1aeee | 1521 | |
402b123d | 1522 | ** Use of alloca in parsers |
52d1aeee MA |
1523 | If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use |
1524 | malloc exclusively. Since 1.29, but was not NEWS'ed. | |
1525 | ||
1526 | alloca is used only when compiled with GCC, to avoid portability | |
1527 | problems as on AIX. | |
1528 | ||
402b123d | 1529 | ** yyparse now returns 2 if memory is exhausted; formerly it dumped core. |
b47dbebe | 1530 | |
402b123d | 1531 | ** When the generated parser lacks debugging code, YYDEBUG is now 0 |
52d1aeee MA |
1532 | (as POSIX requires) instead of being undefined. |
1533 | ||
402b123d | 1534 | ** User Actions |
52d1aeee MA |
1535 | Bison has always permitted actions such as { $$ = $1 }: it adds the |
1536 | ending semicolon. Now if in Yacc compatibility mode, the semicolon | |
1537 | is no longer output: one has to write { $$ = $1; }. | |
1538 | ||
402b123d | 1539 | ** Better C++ compliance |
52d1aeee | 1540 | The output parsers try to respect C++ namespaces. |
76551463 | 1541 | [This turned out to be a failed experiment, and it was reverted later.] |
52d1aeee | 1542 | |
402b123d | 1543 | ** Reduced Grammars |
52d1aeee MA |
1544 | Fixed bugs when reporting useless nonterminals. |
1545 | ||
402b123d | 1546 | ** 64 bit hosts |
52d1aeee MA |
1547 | The parsers work properly on 64 bit hosts. |
1548 | ||
402b123d | 1549 | ** Error messages |
52d1aeee MA |
1550 | Some calls to strerror resulted in scrambled or missing error messages. |
1551 | ||
402b123d | 1552 | ** %expect |
52d1aeee MA |
1553 | When the number of shift/reduce conflicts is correct, don't issue |
1554 | any warning. | |
1555 | ||
402b123d | 1556 | ** The verbose report includes the rule line numbers. |
52d1aeee | 1557 | |
402b123d | 1558 | ** Rule line numbers are fixed in traces. |
52d1aeee | 1559 | |
402b123d | 1560 | ** Swedish translation |
52d1aeee | 1561 | |
402b123d | 1562 | ** Parse errors |
52d1aeee MA |
1563 | Verbose parse error messages from the parsers are better looking. |
1564 | Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'('' | |
1565 | Now: parse error: unexpected '/', expecting "number" or '-' or '(' | |
1566 | ||
402b123d | 1567 | ** Fixed parser memory leaks. |
52d1aeee MA |
1568 | When the generated parser was using malloc to extend its stacks, the |
1569 | previous allocations were not freed. | |
1570 | ||
402b123d | 1571 | ** Fixed verbose output file. |
52d1aeee MA |
1572 | Some newlines were missing. |
1573 | Some conflicts in state descriptions were missing. | |
1574 | ||
402b123d | 1575 | ** Fixed conflict report. |
52d1aeee MA |
1576 | Option -v was needed to get the result. |
1577 | ||
402b123d | 1578 | ** %expect |
52d1aeee MA |
1579 | Was not used. |
1580 | Mismatches are errors, not warnings. | |
1581 | ||
402b123d | 1582 | ** Fixed incorrect processing of some invalid input. |
52d1aeee | 1583 | |
402b123d | 1584 | ** Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H. |
52d1aeee | 1585 | |
402b123d | 1586 | ** Fixed some typos in the documentation. |
52d1aeee | 1587 | |
402b123d | 1588 | ** %token MY_EOF 0 is supported. |
52d1aeee MA |
1589 | Before, MY_EOF was silently renumbered as 257. |
1590 | ||
402b123d | 1591 | ** doc/refcard.tex is updated. |
52d1aeee | 1592 | |
402b123d | 1593 | ** %output, %file-prefix, %name-prefix. |
52d1aeee MA |
1594 | New. |
1595 | ||
402b123d | 1596 | ** --output |
e4ab1254 | 1597 | New, aliasing "--output-file". |
52d1aeee | 1598 | \f |
402b123d | 1599 | * Changes in version 1.30, 2001-10-26: |
342b8b6e | 1600 | |
e4ab1254 AD |
1601 | ** "--defines" and "--graph" have now an optional argument which is the |
1602 | output file name. "-d" and "-g" do not change; they do not take any | |
342b8b6e AD |
1603 | argument. |
1604 | ||
e4ab1254 | 1605 | ** "%source_extension" and "%header_extension" are removed, failed |
342b8b6e AD |
1606 | experiment. |
1607 | ||
402b123d | 1608 | ** Portability fixes. |
f987e9d2 | 1609 | \f |
402b123d | 1610 | * Changes in version 1.29, 2001-09-07: |
342b8b6e | 1611 | |
402b123d | 1612 | ** The output file does not define const, as this caused problems when used |
342b8b6e AD |
1613 | with common autoconfiguration schemes. If you still use ancient compilers |
1614 | that lack const, compile with the equivalent of the C compiler option | |
e4ab1254 | 1615 | "-Dconst=". Autoconf's AC_C_CONST macro provides one way to do this. |
342b8b6e | 1616 | |
e4ab1254 | 1617 | ** Added "-g" and "--graph". |
f87a2205 | 1618 | |
402b123d | 1619 | ** The Bison manual is now distributed under the terms of the GNU FDL. |
f2b5126e | 1620 | |
402b123d | 1621 | ** The input and the output files has automatically a similar extension. |
234a3be3 | 1622 | |
402b123d | 1623 | ** Russian translation added. |
f87a2205 | 1624 | |
402b123d | 1625 | ** NLS support updated; should hopefully be less troublesome. |
f87a2205 | 1626 | |
402b123d | 1627 | ** Added the old Bison reference card. |
c33638bb | 1628 | |
e4ab1254 | 1629 | ** Added "--locations" and "%locations". |
6deb4447 | 1630 | |
e4ab1254 | 1631 | ** Added "-S" and "--skeleton". |
cd5bd6ac | 1632 | |
e4ab1254 | 1633 | ** "%raw", "-r", "--raw" is disabled. |
62ab6972 | 1634 | |
402b123d | 1635 | ** Special characters are escaped when output. This solves the problems |
cd5bd6ac AD |
1636 | of the #line lines with path names including backslashes. |
1637 | ||
402b123d | 1638 | ** New directives. |
e4ab1254 AD |
1639 | "%yacc", "%fixed_output_files", "%defines", "%no_parser", "%verbose", |
1640 | "%debug", "%source_extension" and "%header_extension". | |
f987e9d2 | 1641 | |
402b123d | 1642 | ** @$ |
f987e9d2 | 1643 | Automatic location tracking. |
f87a2205 | 1644 | \f |
402b123d | 1645 | * Changes in version 1.28, 1999-07-06: |
d2e00347 | 1646 | |
402b123d | 1647 | ** Should compile better now with K&R compilers. |
d2e00347 | 1648 | |
402b123d | 1649 | ** Added NLS. |
d2e00347 | 1650 | |
402b123d | 1651 | ** Fixed a problem with escaping the double quote character. |
d2e00347 | 1652 | |
402b123d | 1653 | ** There is now a FAQ. |
d2e00347 | 1654 | \f |
402b123d | 1655 | * Changes in version 1.27: |
5c31c3c2 | 1656 | |
402b123d | 1657 | ** The make rule which prevented bison.simple from being created on |
5c31c3c2 JT |
1658 | some systems has been fixed. |
1659 | \f | |
402b123d | 1660 | * Changes in version 1.26: |
4be07551 | 1661 | |
7e508a2b | 1662 | ** Bison now uses Automake. |
4be07551 | 1663 | |
402b123d | 1664 | ** New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>. |
4be07551 | 1665 | |
402b123d | 1666 | ** Token numbers now start at 257 as previously documented, not 258. |
4be07551 | 1667 | |
402b123d | 1668 | ** Bison honors the TMPDIR environment variable. |
4be07551 | 1669 | |
402b123d | 1670 | ** A couple of buffer overruns have been fixed. |
f51dbca1 | 1671 | |
402b123d | 1672 | ** Problems when closing files should now be reported. |
f51dbca1 | 1673 | |
402b123d | 1674 | ** Generated parsers should now work even on operating systems which do |
f51dbca1 | 1675 | not provide alloca(). |
4be07551 | 1676 | \f |
402b123d | 1677 | * Changes in version 1.25, 1995-10-16: |
df8878c5 | 1678 | |
402b123d | 1679 | ** Errors in the input grammar are not fatal; Bison keeps reading |
df8878c5 | 1680 | the grammar file, and reports all the errors found in it. |
8c44d3ec | 1681 | |
402b123d | 1682 | ** Tokens can now be specified as multiple-character strings: for |
df8878c5 | 1683 | example, you could use "<=" for a token which looks like <=, instead |
7e508a2b | 1684 | of choosing a name like LESSEQ. |
df8878c5 | 1685 | |
402b123d | 1686 | ** The %token_table declaration says to write a table of tokens (names |
df8878c5 RS |
1687 | and numbers) into the parser file. The yylex function can use this |
1688 | table to recognize multiple-character string tokens, or for other | |
1689 | purposes. | |
1690 | ||
402b123d | 1691 | ** The %no_lines declaration says not to generate any #line preprocessor |
df8878c5 RS |
1692 | directives in the parser file. |
1693 | ||
402b123d | 1694 | ** The %raw declaration says to use internal Bison token numbers, not |
df8878c5 RS |
1695 | Yacc-compatible token numbers, when token names are defined as macros. |
1696 | ||
402b123d | 1697 | ** The --no-parser option produces the parser tables without including |
df8878c5 RS |
1698 | the parser engine; a project can now use its own parser engine. |
1699 | The actions go into a separate file called NAME.act, in the form of | |
1700 | a switch statement body. | |
1701 | \f | |
402b123d | 1702 | * Changes in version 1.23: |
6780ca7a | 1703 | |
4d019228 DM |
1704 | The user can define YYPARSE_PARAM as the name of an argument to be |
1705 | passed into yyparse. The argument should have type void *. It should | |
1706 | actually point to an object. Grammar actions can access the variable | |
1707 | by casting it to the proper pointer type. | |
6780ca7a | 1708 | |
6780ca7a | 1709 | Line numbers in output file corrected. |
6780ca7a | 1710 | \f |
402b123d | 1711 | * Changes in version 1.22: |
6780ca7a DM |
1712 | |
1713 | --help option added. | |
6780ca7a | 1714 | \f |
402b123d | 1715 | * Changes in version 1.20: |
6780ca7a DM |
1716 | |
1717 | Output file does not redefine const for C++. | |
9f4503d6 | 1718 | |
76551463 AD |
1719 | ----- |
1720 | ||
34136e65 | 1721 | Copyright (C) 1995-2012 Free Software Foundation, Inc. |
76551463 | 1722 | |
74553c98 | 1723 | This file is part of Bison, the GNU Parser Generator. |
76551463 | 1724 | |
f16b0819 | 1725 | This program is free software: you can redistribute it and/or modify |
76551463 | 1726 | it under the terms of the GNU General Public License as published by |
f16b0819 PE |
1727 | the Free Software Foundation, either version 3 of the License, or |
1728 | (at your option) any later version. | |
76551463 | 1729 | |
f16b0819 | 1730 | This program is distributed in the hope that it will be useful, |
76551463 AD |
1731 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
1732 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
1733 | GNU General Public License for more details. | |
1734 | ||
1735 | You should have received a copy of the GNU General Public License | |
f16b0819 | 1736 | along with this program. If not, see <http://www.gnu.org/licenses/>. |
7e508a2b AD |
1737 | |
1738 | LocalWords: yacc YYBACKUP glr GCC lalr ArrayIndexOutOfBoundsException nullptr | |
1739 | LocalWords: cplusplus liby rpl fprintf mfcalc Wyacc stmt cond expr mk sym lr | |
1740 | LocalWords: IELR ielr Lookahead YYERROR nonassoc LALR's api lookaheads yychar | |
1741 | LocalWords: destructor lookahead YYRHSLOC YYLLOC Rhs ifndef YYFAIL cpp sr rr | |
1742 | LocalWords: preprocessor initializer Wno Wnone Werror FreeBSD prec livelocks | |
1743 | LocalWords: Solaris AIX UX RHEL Tru LHS gcc's Wundef YYENABLE NLS YYLTYPE VCG | |
1744 | LocalWords: yyerror cpp's Wunused yylval yylloc prepend yyparse yylex yypush | |
1745 | LocalWords: Graphviz xml nonterminals midrule destructor's YYSTYPE typedef ly | |
1746 | LocalWords: CHR chr printf stdout namespace preprocessing enum pre include's | |
1747 | LocalWords: YYRECOVERING nonfree destructors YYABORT YYACCEPT params enums de | |
1748 | LocalWords: struct yystype DJGPP lex param Haible NUM alloca YYSTACK NUL goto | |
1749 | LocalWords: YYMAXDEPTH Unescaped UCNs YYLTYPE's yyltype typedefs inline Yaccs | |
1750 | LocalWords: Heriyanto Reenable dprec Hilfinger Eggert MYEOF Folle Menezes EOF | |
1751 | LocalWords: Lackovic define's itemset Groff Gettext malloc NEWS'ed YYDEBUG | |
1752 | LocalWords: namespaces strerror const autoconfiguration Dconst Autoconf's FDL | |
1753 | LocalWords: Automake TMPDIR LESSEQ | |
1754 | ||
1755 | Local Variables: | |
1756 | mode: outline | |
e4ab1254 | 1757 | fill-column: 76 |
7e508a2b | 1758 | End: |