]>
Commit | Line | Data |
---|---|---|
ed7658fe | 1 | GNU Bison NEWS |
3af4feb2 | 2 | |
52718713 | 3 | * Noteworthy changes in release ?.? (????-??-??) [?] |
74553c98 | 4 | |
8458a411 | 5 | ** WARNING: Future backward-incompatibilities! |
597afd73 | 6 | |
597afd73 AD |
7 | Like other GNU packages, Bison will start using some of the C99 features |
8 | for its own code, especially the definition of variables after statements. | |
9 | The generated C parsers still aim at C90. | |
10 | ||
8458a411 | 11 | ** Backward incompatible changes |
47db7ed1 AD |
12 | |
13 | *** Obsolete features | |
14 | ||
40bb6f78 AD |
15 | Support for YYFAIL is removed (deprecated in Bison 2.4.2): use YYERROR. |
16 | ||
17 | Support for yystype and yyltype is removed (deprecated in Bison 1.875): | |
18 | use YYSTYPE and YYLTYPE. | |
19 | ||
20 | Support for YYLEX_PARAM and YYPARSE_PARAM is removed (deprecated in Bison | |
21 | 1.875): use %lex-param, %parse-param, or %param. | |
47db7ed1 | 22 | |
05e25f23 AD |
23 | Missing semicolons at the end of actions are no longer added (as announced |
24 | in the release 2.5). | |
25 | ||
597afd73 AD |
26 | ** Bug fixes |
27 | ||
c21e515e | 28 | *** The epilogue is no longer affected by internal #defines (glr.c) |
597afd73 AD |
29 | |
30 | The glr.c skeleton uses defines such as #define yylval (yystackp->yyval) in | |
31 | generated code. These weren't properly undefined before the inclusion of | |
32 | the user epilogue, so functions such as the following were butchered by the | |
33 | preprocessor expansion: | |
34 | ||
35 | int yylex (YYSTYPE *yylval); | |
36 | ||
6c7022f7 | 37 | This is fixed: yylval, yynerrs, yychar, and yylloc are now valid |
597afd73 AD |
38 | identifiers for user-provided variables. |
39 | ||
f0f95a50 AD |
40 | *** stdio.h is no longer needed when locations are enabled (yacc.c) |
41 | ||
42 | Changes in Bison 2.7 introduced a dependency on FILE and fprintf when | |
43 | locations are enabled. This is fixed. | |
44 | ||
597afd73 AD |
45 | ** Diagnostics reported by Bison |
46 | ||
47 | Most of these features were contributed by Théophile Ranquet and Victor | |
48 | Santet. | |
73370a9d | 49 | |
016426c1 TR |
50 | *** Carets |
51 | ||
52 | Version 2.7 introduced caret errors, for a prettier output. These are now | |
53 | activated by default. The old format can still be used by invoking Bison | |
54 | with -fno-caret (or -fnone). | |
55 | ||
fec5f3c0 AD |
56 | Some error messages that reproduced excerpts of the grammar are now using |
57 | the caret information only. For instance on: | |
58 | ||
59 | %% | |
60 | exp: 'a' | 'a'; | |
61 | ||
62 | Bison 2.7 reports: | |
63 | ||
64 | in.y: warning: 1 reduce/reduce conflict [-Wconflicts-rr] | |
65 | in.y:2.12-14: warning: rule useless in parser due to conflicts: exp: 'a' [-Wother] | |
66 | ||
67 | Now bison reports: | |
68 | ||
69 | in.y: warning: 1 reduce/reduce conflict [-Wconflicts-rr] | |
70 | in.y:2.12-14: warning: rule useless in parser due to conflicts [-Wother] | |
71 | exp: 'a' | 'a'; | |
72 | ^^^ | |
73 | ||
74 | and "bison -fno-caret" reports: | |
75 | ||
76 | in.y: warning: 1 reduce/reduce conflict [-Wconflicts-rr] | |
77 | in.y:2.12-14: warning: rule useless in parser due to conflicts [-Wother] | |
78 | ||
1048a1c9 | 79 | *** Enhancements of the -Werror option |
518e8830 | 80 | |
1048a1c9 | 81 | The -Werror=CATEGORY option is now recognized, and will treat specified |
d949eefd AD |
82 | warnings as errors. The warnings need not have been explicitly activated |
83 | using the -W option, this is similar to what GCC 4.7 does. | |
1048a1c9 AD |
84 | |
85 | For example, given the following command line, Bison will treat both | |
d949eefd | 86 | warnings related to POSIX Yacc incompatibilities and S/R conflicts as |
1048a1c9 AD |
87 | errors (and only those): |
88 | ||
89 | $ bison -Werror=yacc,error=conflicts-sr input.y | |
90 | ||
91 | If no categories are specified, -Werror will make all active warnings into | |
92 | errors. For example, the following line does the same the previous example: | |
93 | ||
94 | $ bison -Werror -Wnone -Wyacc -Wconflicts-sr input.y | |
95 | ||
96 | (By default -Wconflicts-sr,conflicts-rr,deprecated,other is enabled.) | |
97 | ||
98 | Note that the categories in this -Werror option may not be prefixed with | |
99 | "no-". However, -Wno-error[=CATEGORY] is valid. | |
100 | ||
101 | Note that -y enables -Werror=yacc. Therefore it is now possible to require | |
102 | Yacc-like behavior (e.g., always generate y.tab.c), but to report | |
103 | incompatibilities as warnings: "-y -Wno-error=yacc". | |
518e8830 | 104 | |
46bdb8ec | 105 | *** The display of warnings is now richer |
73370a9d | 106 | |
46bdb8ec | 107 | The option that controls a given warning is now displayed: |
73370a9d | 108 | |
46bdb8ec | 109 | foo.y:4.6: warning: type clash on default action: <foo> != <bar> [-Wother] |
73370a9d | 110 | |
46bdb8ec TR |
111 | In the case of warnings treated as errors, the prefix is changed from |
112 | "warning: " to "error: ", and the suffix is displayed, in a manner similar | |
d949eefd | 113 | to GCC, as [-Werror=CATEGORY]. |
1048a1c9 | 114 | |
46bdb8ec TR |
115 | For instance, where the previous version of Bison would report (and exit |
116 | with failure): | |
1048a1c9 AD |
117 | |
118 | bison: warnings being treated as errors | |
46bdb8ec | 119 | input.y:1.1: warning: stray ',' treated as white space |
1048a1c9 | 120 | |
46bdb8ec | 121 | it now reports: |
1048a1c9 | 122 | |
1048a1c9 AD |
123 | input.y:1.1: error: stray ',' treated as white space [-Werror=other] |
124 | ||
125 | *** Deprecated constructs | |
126 | ||
127 | The new 'deprecated' warning category flags obsolete constructs whose | |
128 | support will be discontinued. It is enabled by default. These warnings | |
129 | used to be reported as 'other' warnings. | |
130 | ||
73370a9d | 131 | *** Useless semantic types |
9641b918 VS |
132 | |
133 | Bison now warns about useless (uninhabited) semantic types. Since | |
134 | semantic types are not declared to Bison (they are defined in the opaque | |
135 | %union structure), it is %printer/%destructor directives about useless | |
136 | types that trigger the warning: | |
137 | ||
138 | %token <type1> term | |
139 | %type <type2> nterm | |
140 | %printer {} <type1> <type3> | |
141 | %destructor {} <type2> <type4> | |
142 | %% | |
143 | nterm: term { $$ = $1; }; | |
144 | ||
145 | 3.28-34: warning: type <type3> is used, but is not associated to any symbol | |
146 | 4.28-34: warning: type <type4> is used, but is not associated to any symbol | |
147 | ||
31557b9e | 148 | *** Undefined but unused symbols |
b921d92f | 149 | |
31557b9e AD |
150 | Bison used to raise an error for undefined symbols that are not used in |
151 | the grammar. This is now only a warning. | |
b921d92f VS |
152 | |
153 | %printer {} symbol1 | |
154 | %destructor {} symbol2 | |
31557b9e | 155 | %type <type> symbol3 |
b921d92f VS |
156 | %% |
157 | exp: "a"; | |
158 | ||
73370a9d | 159 | *** Useless destructors or printers |
ea9a35c6 VS |
160 | |
161 | Bison now warns about useless destructors or printers. In the following | |
162 | example, the printer for <type1>, and the destructor for <type2> are | |
163 | useless: all symbols of <type1> (token1) already have a printer, and all | |
164 | symbols of type <type2> (token2) already have a destructor. | |
165 | ||
166 | %token <type1> token1 | |
167 | <type2> token2 | |
168 | <type3> token3 | |
169 | <type4> token4 | |
170 | %printer {} token1 <type1> <type3> | |
171 | %destructor {} token2 <type2> <type4> | |
172 | ||
d87ea54c AD |
173 | *** Conflicts |
174 | ||
175 | The warnings and error messages about shift/reduce and reduce/reduce | |
176 | conflicts have been normalized. For instance on the following foo.y file: | |
177 | ||
178 | %glr-parser | |
179 | %% | |
180 | exp: exp '+' exp | '0' | '0'; | |
181 | ||
182 | compare the previous version of bison: | |
183 | ||
184 | $ bison foo.y | |
185 | foo.y: conflicts: 1 shift/reduce, 2 reduce/reduce | |
186 | $ bison -Werror foo.y | |
187 | bison: warnings being treated as errors | |
188 | foo.y: conflicts: 1 shift/reduce, 2 reduce/reduce | |
189 | ||
190 | with the new behavior: | |
191 | ||
192 | $ bison foo.y | |
193 | foo.y: warning: 1 shift/reduce conflict [-Wconflicts-sr] | |
194 | foo.y: warning: 2 reduce/reduce conflicts [-Wconflicts-rr] | |
195 | $ bison -Werror foo.y | |
9503b0a4 TR |
196 | foo.y: error: 1 shift/reduce conflict [-Werror=conflicts-sr] |
197 | foo.y: error: 2 reduce/reduce conflicts [-Werror=conflicts-rr] | |
d87ea54c AD |
198 | |
199 | When %expect or %expect-rr is used, such as with bar.y: | |
200 | ||
201 | %expect 0 | |
202 | %glr-parser | |
203 | %% | |
204 | exp: exp '+' exp | '0' | '0'; | |
205 | ||
206 | Former behavior: | |
207 | ||
208 | $ bison bar.y | |
209 | bar.y: conflicts: 1 shift/reduce, 2 reduce/reduce | |
210 | bar.y: expected 0 shift/reduce conflicts | |
211 | bar.y: expected 0 reduce/reduce conflicts | |
212 | ||
213 | New one: | |
214 | ||
215 | $ bison bar.y | |
9503b0a4 TR |
216 | bar.y: error: shift/reduce conflicts: 1 found, 0 expected |
217 | bar.y: error: reduce/reduce conflicts: 2 found, 0 expected | |
d87ea54c | 218 | |
f24695ef AD |
219 | ** Incompatibilities with POSIX Yacc |
220 | ||
221 | The 'yacc' category is no longer part of '-Wall', enable it explicitly | |
222 | with '-Wyacc'. | |
223 | ||
2055a44e AD |
224 | ** Additional yylex/yyparse arguments |
225 | ||
6dc4663d AD |
226 | The new directive %param declares additional arguments to both yylex and |
227 | yyparse. The %lex-param, %parse-param, and %param directives support one | |
228 | or more arguments. Instead of | |
2055a44e | 229 | |
e436fa67 AD |
230 | %lex-param {arg1_type *arg1} |
231 | %lex-param {arg2_type *arg2} | |
232 | %parse-param {arg1_type *arg1} | |
233 | %parse-param {arg2_type *arg2} | |
2055a44e AD |
234 | |
235 | one may now declare | |
236 | ||
e436fa67 | 237 | %param {arg1_type *arg1} {arg2_type *arg2} |
2055a44e | 238 | |
2a6b66c5 | 239 | ** Variable api.token.prefix |
99c08fb6 | 240 | |
2a6b66c5 | 241 | The variable api.token.prefix changes the way tokens are identified in |
99c08fb6 AD |
242 | the generated files. This is especially useful to avoid collisions |
243 | with identifiers in the target language. For instance | |
244 | ||
e436fa67 | 245 | %token FILE for ERROR |
2a6b66c5 | 246 | %define api.token.prefix "TOK_" |
e436fa67 AD |
247 | %% |
248 | start: FILE for ERROR; | |
99c08fb6 AD |
249 | |
250 | will generate the definition of the symbols TOK_FILE, TOK_for, and | |
251 | TOK_ERROR in the generated sources. In particular, the scanner must | |
252 | use these prefixed token names, although the grammar itself still | |
253 | uses the short names (as in the sample rule given above). | |
254 | ||
31b850d2 AD |
255 | ** Variable parse.error |
256 | ||
1f77b2e0 AD |
257 | This variable controls the verbosity of error messages. The use of the |
258 | %error-verbose directive is deprecated in favor of "%define parse.error | |
259 | verbose". | |
31b850d2 | 260 | |
c21e515e AD |
261 | ** Renamed %define variables |
262 | ||
263 | The following variables have been renamed for consistency. Backward | |
264 | compatibility is ensured, but upgrading is recommended. | |
265 | ||
266 | lr.default-reductions -> lr.default-reduction | |
267 | lr.keep-unreachable-states -> lr.keep-unreachable-state | |
268 | namespace -> api.namespace | |
269 | stype -> api.value.type | |
270 | ||
ca2a6d15 PH |
271 | ** Semantic predicates |
272 | ||
597afd73 AD |
273 | Contributed by Paul Hilfinger. |
274 | ||
1f77b2e0 AD |
275 | The new, experimental, semantic-predicate feature allows actions of the |
276 | form "%?{ BOOLEAN-EXPRESSION }", which cause syntax errors (as for | |
ca2a6d15 | 277 | YYERROR) if the expression evaluates to 0, and are evaluated immediately |
1f77b2e0 AD |
278 | in GLR parsers, rather than being deferred. The result is that they allow |
279 | the programmer to prune possible parses based on the values of run-time | |
280 | expressions. | |
ca2a6d15 | 281 | |
d1400569 AD |
282 | ** The directive %expect-rr is now an error in non GLR mode |
283 | ||
284 | It used to be an error only if used in non GLR mode, _and_ if there are | |
285 | reduce/reduce conflicts. | |
286 | ||
5202b6ac | 287 | ** Tokens are numbered in their order of appearance |
93561c21 | 288 | |
5202b6ac VT |
289 | Contributed by Valentin Tolmer. |
290 | ||
291 | With '%token A B', A had a number less than the one of B. However, | |
292 | precedence declarations used to generate a reversed order. This is now | |
293 | fixed, and introducing tokens with any of %token, %left, %right, | |
294 | %precedence, or %nonassoc yields the same result. | |
295 | ||
296 | When mixing declarations of tokens with a litteral character (e.g., 'a') | |
297 | or with an identifier (e.g., B) in a precedence declaration, Bison | |
298 | numbered the litteral characters first. For example | |
299 | ||
300 | %right A B 'c' 'd' | |
301 | ||
302 | would lead to the tokens declared in this order: 'c' 'd' A B. Again, the | |
303 | input order is now preserved. | |
304 | ||
305 | These changes were made so that one can remove useless precedence and | |
306 | associativity declarations (i.e., map %nonassoc, %left or %right to | |
307 | %precedence, or to %token) and get exactly the same output. | |
93561c21 | 308 | |
cc2235ac VT |
309 | ** Useless precedence and associativity |
310 | ||
d2f9ae18 AD |
311 | Contributed by Valentin Tolmer. |
312 | ||
1282c124 | 313 | When developing and maintaining a grammar, useless associativity and |
cc2235ac VT |
314 | precedence directives are common. They can be a nuisance: new ambiguities |
315 | arising are sometimes masked because their conflicts are resolved due to | |
316 | the extra precedence or associativity information. Furthermore, it can | |
317 | hinder the comprehension of a new grammar: one will wonder about the role | |
318 | of a precedence, where in fact it is useless. The following changes aim | |
319 | at detecting and reporting these extra directives. | |
320 | ||
321 | *** Precedence warning category | |
322 | ||
323 | A new category of warning, -Wprecedence, was introduced. It flags the | |
324 | useless precedence and associativity directives. | |
325 | ||
326 | *** Useless associativity | |
327 | ||
328 | Bison now warns about symbols with a declared associativity that is never | |
329 | used to resolve conflicts. In that case, using %precedence is sufficient; | |
330 | the parsing tables will remain unchanged. Solving these warnings may raise | |
331 | useless precedence warnings, as the symbols no longer have associativity. | |
332 | For example: | |
333 | ||
334 | %left '+' | |
335 | %left '*' | |
336 | %% | |
337 | exp: | |
1282c124 AD |
338 | "number" |
339 | | exp '+' "number" | |
cc2235ac VT |
340 | | exp '*' exp |
341 | ; | |
342 | ||
343 | will produce a | |
344 | ||
345 | warning: useless associativity for '+', use %precedence [-Wprecedence] | |
346 | %left '+' | |
347 | ^^^ | |
348 | ||
349 | *** Useless precedence | |
350 | ||
351 | Bison now warns about symbols with a declared precedence and no declared | |
352 | associativity (i.e., declared with %precedence), and whose precedence is | |
353 | never used. In that case, the symbol can be safely declared with %token | |
354 | instead, without modifying the parsing tables. For example: | |
355 | ||
356 | %precedence '=' | |
357 | %% | |
1282c124 | 358 | exp: "var" '=' "number"; |
cc2235ac VT |
359 | |
360 | will produce a | |
361 | ||
362 | warning: useless precedence for '=' [-Wprecedence] | |
363 | %precedence '=' | |
364 | ^^^ | |
365 | ||
366 | *** Useless precedence and associativity | |
367 | ||
368 | In case of both useless precedence and associativity, the issue is flagged | |
369 | as follows: | |
370 | ||
371 | %nonassoc '=' | |
372 | %% | |
1282c124 | 373 | exp: "var" '=' "number"; |
cc2235ac VT |
374 | |
375 | The warning is: | |
376 | ||
377 | warning: useless precedence and associativity for '=' [-Wprecedence] | |
378 | %nonassoc '=' | |
379 | ^^^ | |
09add9c2 AD |
380 | |
381 | ** Empty rules | |
382 | ||
6240346a AD |
383 | With help from Joel E. Denny and Gabriel Rassoul. |
384 | ||
09add9c2 AD |
385 | Empty rules (i.e., with an empty right-hand side) can now be explicitly |
386 | marked by the new %empty directive. Using %empty on a non-empty rule is | |
387 | an error. The new -Wempty-rule warning reports empty rules without | |
388 | %empty. On the following grammar: | |
389 | ||
390 | %% | |
391 | s: a b c; | |
392 | a: ; | |
393 | b: %empty; | |
394 | c: 'a' %empty; | |
395 | ||
396 | bison reports: | |
397 | ||
398 | 3.4-5: warning: empty rule without %empty [-Wempty-rule] | |
399 | a: {} | |
400 | ^^ | |
401 | 5.8-13: error: %empty on non-empty rule | |
402 | c: 'a' %empty {}; | |
403 | ^^^^^^ | |
cc2235ac | 404 | |
c21e515e AD |
405 | ** Java skeleton improvements |
406 | ||
407 | Contributed by Paolo Bonzini. | |
408 | ||
409 | The constants for token names were moved to the Lexer interface. Also, it | |
410 | is possible to add code to the parser's constructors using "%code init" | |
411 | and "%define init_throws". | |
412 | ||
413 | ** C++ skeletons improvements | |
414 | ||
415 | *** The parser header is no longer mandatory (lalr1.cc, glr.cc) | |
416 | ||
417 | Using %defines is now optional. Without it, the needed support classes | |
418 | are defined in the generated parser, instead of additional files (such as | |
419 | location.hh, position.hh and stack.hh). | |
420 | ||
421 | *** Locations are no longer mandatory (lalr1.cc, glr.cc) | |
422 | ||
423 | Both lalr1.cc and glr.cc no longer require %location. | |
424 | ||
425 | *** syntax_error exception (lalr1.cc) | |
426 | ||
427 | The C++ parser features a syntax_error exception, which can be | |
428 | thrown from the scanner or from user rules to raise syntax errors. | |
429 | This facilitates reporting errors caught in sub-functions (e.g., | |
430 | rejecting too large integral literals from a conversion function | |
431 | used by the scanner, or rejecting invalid combinations from a | |
432 | factory invoked by the user actions). | |
433 | ||
434 | *** %define api.value.type variant | |
435 | ||
436 | This is based on a submission from Michiel De Wilde. With help | |
437 | from Théophile Ranquet. | |
438 | ||
439 | In this mode, complex C++ objects can be used as semantic values. For | |
440 | instance: | |
441 | ||
442 | %token <::std::string> TEXT; | |
443 | %token <int> NUMBER; | |
444 | %token SEMICOLON ";" | |
445 | %type <::std::string> item; | |
446 | %type <::std::list<std::string>> list; | |
447 | %% | |
448 | result: | |
449 | list { std::cout << $1 << std::endl; } | |
450 | ; | |
451 | ||
452 | list: | |
6240346a | 453 | %empty { /* Generates an empty string list. */ } |
c21e515e AD |
454 | | list item ";" { std::swap ($$, $1); $$.push_back ($2); } |
455 | ; | |
456 | ||
457 | item: | |
458 | TEXT { std::swap ($$, $1); } | |
459 | | NUMBER { $$ = string_cast ($1); } | |
460 | ; | |
461 | ||
462 | *** %define api.token.constructor | |
463 | ||
464 | When variants are enabled, Bison can generate functions to build the | |
465 | tokens. This guarantees that the token type (e.g., NUMBER) is consistent | |
466 | with the semantic value (e.g., int): | |
467 | ||
468 | parser::symbol_type yylex () | |
469 | { | |
470 | parser::location_type loc = ...; | |
471 | ... | |
472 | return parser::make_TEXT ("Hello, world!", loc); | |
473 | ... | |
474 | return parser::make_NUMBER (42, loc); | |
475 | ... | |
476 | return parser::make_SEMICOLON (loc); | |
477 | ... | |
478 | } | |
479 | ||
c13bb348 | 480 | * Noteworthy changes in release 2.7 (2012-12-12) [stable] |
effd30c0 | 481 | |
edf9a06f | 482 | ** Bug fixes |
7bada535 | 483 | |
edf9a06f | 484 | Warnings about uninitialized yylloc in yyparse have been fixed. |
7bada535 | 485 | |
1127a75a AD |
486 | Restored C90 compliance (yet no report was ever made). |
487 | ||
d4fe9e88 | 488 | ** Diagnostics are improved |
7bada535 | 489 | |
8458a411 AD |
490 | Contributed by Théophile Ranquet. |
491 | ||
d4fe9e88 | 492 | *** Changes in the format of error messages |
7bada535 | 493 | |
d4fe9e88 | 494 | This used to be the format of many error reports: |
1f1bd572 | 495 | |
d4fe9e88 AD |
496 | input.y:2.7-12: %type redeclaration for exp |
497 | input.y:1.7-12: previous declaration | |
1f1bd572 | 498 | |
d4fe9e88 | 499 | It is now: |
1f1bd572 | 500 | |
d4fe9e88 AD |
501 | input.y:2.7-12: error: %type redeclaration for exp |
502 | input.y:1.7-12: previous declaration | |
cbaea010 | 503 | |
d4fe9e88 | 504 | *** New format for error reports: carets |
cbaea010 | 505 | |
d4fe9e88 | 506 | Caret errors have been added to Bison: |
cbaea010 | 507 | |
d4fe9e88 AD |
508 | input.y:2.7-12: error: %type redeclaration for exp |
509 | %type <sval> exp | |
510 | ^^^^^^ | |
511 | input.y:1.7-12: previous declaration | |
512 | %type <ival> exp | |
513 | ^^^^^^ | |
cbaea010 | 514 | |
d4fe9e88 | 515 | or |
cbaea010 | 516 | |
7bada535 | 517 | input.y:3.20-23: error: ambiguous reference: '$exp' |
fb6040f0 | 518 | exp: exp '+' exp { $exp = $1 + $3; }; |
7bada535 | 519 | ^^^^ |
fb6040f0 TR |
520 | input.y:3.1-3: refers to: $exp at $$ |
521 | exp: exp '+' exp { $exp = $1 + $3; }; | |
522 | ^^^ | |
523 | input.y:3.6-8: refers to: $exp at $1 | |
524 | exp: exp '+' exp { $exp = $1 + $3; }; | |
525 | ^^^ | |
526 | input.y:3.14-16: refers to: $exp at $3 | |
527 | exp: exp '+' exp { $exp = $1 + $3; }; | |
528 | ^^^ | |
7bada535 | 529 | |
1282c124 AD |
530 | The default behavior for now is still not to display these unless |
531 | explicitly asked with -fcaret (or -fall). However, in a later release, it | |
d4fe9e88 AD |
532 | will be made the default behavior (but may still be deactivated with |
533 | -fno-caret). | |
d3e4409a | 534 | |
1f1bd572 | 535 | ** New value for %define variable: api.pure full |
d3e4409a | 536 | |
1f1bd572 | 537 | The %define variable api.pure requests a pure (reentrant) parser. However, |
d4fe9e88 AD |
538 | for historical reasons, using it in a location-tracking Yacc parser |
539 | resulted in a yyerror function that did not take a location as a | |
540 | parameter. With this new value, the user may request a better pure parser, | |
541 | where yyerror does take a location as a parameter (in location-tracking | |
542 | parsers). | |
1f1bd572 TR |
543 | |
544 | The use of "%define api.pure true" is deprecated in favor of this new | |
545 | "%define api.pure full". | |
d3e4409a | 546 | |
7287be84 | 547 | ** New %define variable: api.location.type (glr.cc, lalr1.cc, lalr1.java) |
db8ab2be AD |
548 | |
549 | The %define variable api.location.type defines the name of the type to use | |
550 | for locations. When defined, Bison no longer generates the position.hh | |
551 | and location.hh files, nor does the parser will include them: the user is | |
552 | then responsible to define her type. | |
553 | ||
554 | This can be used in programs with several parsers to factor their location | |
7287be84 AD |
555 | and position files: let one of them generate them, and the others just use |
556 | them. | |
db8ab2be AD |
557 | |
558 | This feature was actually introduced, but not documented, in Bison 2.5, | |
559 | under the name "location_type" (which is maintained for backward | |
560 | compatibility). | |
561 | ||
7287be84 AD |
562 | For consistency, lalr1.java's %define variables location_type and |
563 | position_type are deprecated in favor of api.location.type and | |
564 | api.position.type. | |
565 | ||
d4fe9e88 AD |
566 | ** Exception safety (lalr1.cc) |
567 | ||
568 | The parse function now catches exceptions, uses the %destructors to | |
569 | release memory (the lookahead symbol and the symbols pushed on the stack) | |
570 | before re-throwing the exception. | |
571 | ||
572 | This feature is somewhat experimental. User feedback would be | |
573 | appreciated. | |
574 | ||
9c16d399 | 575 | ** Graph improvements in DOT and XSLT |
fc4fdd62 | 576 | |
8458a411 AD |
577 | Contributed by Théophile Ranquet. |
578 | ||
fc4fdd62 TR |
579 | The graphical presentation of the states is more readable: their shape is |
580 | now rectangular, the state number is clearly displayed, and the items are | |
581 | numbered and left-justified. | |
582 | ||
583 | The reductions are now explicitly represented as transitions to other | |
584 | diamond shaped nodes. | |
585 | ||
9c16d399 TR |
586 | These changes are present in both --graph output and xml2dot.xsl XSLT |
587 | processing, with minor (documented) differences. | |
588 | ||
d4fe9e88 | 589 | ** %language is no longer an experimental feature. |
fb4c8a7c | 590 | |
d4fe9e88 AD |
591 | The introduction of this feature, in 2.4, was four years ago. The |
592 | --language option and the %language directive are no longer experimental. | |
fb4c8a7c | 593 | |
53e2cd1e AD |
594 | ** Documentation |
595 | ||
596 | The sections about shift/reduce and reduce/reduce conflicts resolution | |
597 | have been fixed and extended. | |
9d2423f5 | 598 | |
d4fe9e88 AD |
599 | Although introduced more than four years ago, XML and Graphviz reports |
600 | were not properly documented. | |
601 | ||
be22823e AD |
602 | The translation of mid-rule actions is now described. |
603 | ||
9d3f7eaf | 604 | * Noteworthy changes in release 2.6.5 (2012-11-07) [stable] |
6f1360bd | 605 | |
a68b1f23 AD |
606 | We consider compiler warnings about Bison generated parsers to be bugs. |
607 | Rather than working around them in your own project, please consider | |
608 | reporting them to us. | |
609 | ||
610 | ** Bug fixes | |
611 | ||
612 | Warnings about uninitialized yylval and/or yylloc for push parsers with a | |
613 | pure interface have been fixed for GCC 4.0 up to 4.8, and Clang 2.9 to | |
614 | 3.2. | |
615 | ||
616 | Other issues in the test suite have been addressed. | |
6f1360bd | 617 | |
1282c124 | 618 | Null characters are correctly displayed in error messages. |
95066e92 | 619 | |
a1d1ab50 AD |
620 | When possible, yylloc is correctly initialized before calling yylex. It |
621 | is no longer necessary to initialize it in the %initial-action. | |
622 | ||
0ac15849 | 623 | * Noteworthy changes in release 2.6.4 (2012-10-23) [stable] |
a4eb820f | 624 | |
468455e1 | 625 | Bison 2.6.3's --version was incorrect. This release fixes this issue. |
a4eb820f | 626 | |
6eb8f74f | 627 | * Noteworthy changes in release 2.6.3 (2012-10-22) [stable] |
933ec544 | 628 | |
6b4cb804 AD |
629 | ** Bug fixes |
630 | ||
a1a77e1f | 631 | Bugs and portability issues in the test suite have been fixed. |
6b4cb804 AD |
632 | |
633 | Some errors in translations have been addressed, and --help now directs | |
634 | users to the appropriate place to report them. | |
635 | ||
636 | Stray Info files shipped by accident are removed. | |
637 | ||
638 | Incorrect definitions of YY_, issued by yacc.c when no parser header is | |
639 | generated, are removed. | |
9c26b8fc | 640 | |
a2b3f101 TR |
641 | All the generated headers are self-contained. |
642 | ||
c9d5bcc9 AD |
643 | ** Header guards (yacc.c, glr.c, glr.cc) |
644 | ||
645 | In order to avoid collisions, the header guards are now | |
646 | YY_<PREFIX>_<FILE>_INCLUDED, instead of merely <PREFIX>_<FILE>. | |
647 | For instance the header generated from | |
648 | ||
649 | %define api.prefix "calc" | |
650 | %defines "lib/parse.h" | |
651 | ||
652 | will use YY_CALC_LIB_PARSE_H_INCLUDED as guard. | |
653 | ||
c12c4c50 | 654 | ** Fix compiler warnings in the generated parser (yacc.c, glr.c) |
321d3e35 AD |
655 | |
656 | The compilation of pure parsers (%define api.pure) can trigger GCC | |
657 | warnings such as: | |
658 | ||
659 | input.c: In function 'yyparse': | |
660 | input.c:1503:12: warning: 'yylval' may be used uninitialized in this | |
661 | function [-Wmaybe-uninitialized] | |
662 | *++yyvsp = yylval; | |
663 | ^ | |
664 | ||
665 | This is now fixed; pragmas to avoid these warnings are no longer needed. | |
666 | ||
c12c4c50 AD |
667 | Warnings from clang ("equality comparison with extraneous parentheses" and |
668 | "function declared 'noreturn' should not return") have also been | |
669 | addressed. | |
670 | ||
e1eeecd3 | 671 | * Noteworthy changes in release 2.6.2 (2012-08-03) [stable] |
9c26b8fc | 672 | |
43ca8040 AD |
673 | ** Bug fixes |
674 | ||
675 | Buffer overruns, complaints from Flex, and portability issues in the test | |
676 | suite have been fixed. | |
677 | ||
c9d546b2 AD |
678 | ** Spaces in %lex- and %parse-param (lalr1.cc, glr.cc) |
679 | ||
680 | Trailing end-of-lines in %parse-param or %lex-param would result in | |
681 | invalid C++. This is fixed. | |
9c26b8fc | 682 | |
dcd5344d AD |
683 | ** Spurious spaces and end-of-lines |
684 | ||
685 | The generated files no longer end (nor start) with empty lines. | |
686 | ||
77b214ef | 687 | * Noteworthy changes in release 2.6.1 (2012-07-30) [stable] |
a4107f24 | 688 | |
8617d87e AD |
689 | Bison no longer executes user-specified M4 code when processing a grammar. |
690 | ||
e20e6a50 AD |
691 | ** Future Changes |
692 | ||
693 | In addition to the removal of the features announced in Bison 2.6, the | |
694 | next major release will remove the "Temporary hack for adding a semicolon | |
695 | to the user action", as announced in the release 2.5. Instead of: | |
696 | ||
697 | exp: exp "+" exp { $$ = $1 + $3 }; | |
698 | ||
699 | write: | |
700 | ||
701 | exp: exp "+" exp { $$ = $1 + $3; }; | |
702 | ||
8617d87e AD |
703 | ** Bug fixes |
704 | ||
0e164d43 AD |
705 | *** Type names are now properly escaped. |
706 | ||
707 | *** glr.cc: set_debug_level and debug_level work as expected. | |
a4107f24 | 708 | |
26313726 AD |
709 | *** Stray @ or $ in actions |
710 | ||
711 | While Bison used to warn about stray $ or @ in action rules, it did not | |
712 | for other actions such as printers, destructors, or initial actions. It | |
713 | now does. | |
714 | ||
cd735a8c | 715 | ** Type names in actions |
4982f078 AD |
716 | |
717 | For consistency with rule actions, it is now possible to qualify $$ by a | |
cd735a8c | 718 | type-name in destructors, printers, and initial actions. For instance: |
4982f078 AD |
719 | |
720 | %printer { fprintf (yyo, "(%d, %f)", $<ival>$, $<fval>$); } <*> <>; | |
721 | ||
722 | will display two values for each typed and untyped symbol (provided | |
cd735a8c | 723 | that YYSTYPE has both "ival" and "fval" fields). |
60aa04a2 | 724 | |
1505e8bb | 725 | * Noteworthy changes in release 2.6 (2012-07-19) [stable] |
0f11eec2 | 726 | |
d0a30438 | 727 | ** Future changes |
9553083c | 728 | |
55d1006f AD |
729 | The next major release of Bison will drop support for the following |
730 | deprecated features. Please report disagreements to bug-bison@gnu.org. | |
0f11eec2 | 731 | |
aaf61036 | 732 | *** K&R C parsers |
55d1006f AD |
733 | |
734 | Support for generating parsers in K&R C will be removed. Parsers | |
242cc08e | 735 | generated for C support ISO C90, and are tested with ISO C99 and ISO C11 |
55d1006f AD |
736 | compilers. |
737 | ||
258cddbc | 738 | *** Features deprecated since Bison 1.875 |
0f11eec2 | 739 | |
258cddbc AD |
740 | The definitions of yystype and yyltype will be removed; use YYSTYPE and |
741 | YYLTYPE. | |
0f11eec2 | 742 | |
258cddbc AD |
743 | YYPARSE_PARAM and YYLEX_PARAM, deprecated in favor of %parse-param and |
744 | %lex-param, will no longer be supported. | |
745 | ||
746 | Support for the preprocessor symbol YYERROR_VERBOSE will be removed, use | |
747 | %error-verbose. | |
55d1006f AD |
748 | |
749 | *** The generated header will be included (yacc.c) | |
0f11eec2 AD |
750 | |
751 | Instead of duplicating the content of the generated header (definition of | |
55d1006f AD |
752 | YYSTYPE, yyparse declaration etc.), the generated parser will include it, |
753 | as is already the case for GLR or C++ parsers. This change is deferred | |
754 | because existing versions of ylwrap (e.g., Automake 1.12.1) do not support | |
755 | it. | |
0f11eec2 | 756 | |
c2425191 | 757 | ** Generated Parser Headers |
56ca3d8f | 758 | |
258cddbc | 759 | *** Guards (yacc.c, glr.c, glr.cc) |
c3e9f08f AD |
760 | |
761 | The generated headers are now guarded, as is already the case for C++ | |
242cc08e | 762 | parsers (lalr1.cc). For instance, with --defines=foo.h: |
c3e9f08f | 763 | |
e29f0771 AD |
764 | #ifndef YY_FOO_H |
765 | # define YY_FOO_H | |
766 | ... | |
767 | #endif /* !YY_FOO_H */ | |
c3e9f08f | 768 | |
258cddbc | 769 | *** New declarations (yacc.c, glr.c) |
56ca3d8f AD |
770 | |
771 | The generated header now declares yydebug and yyparse. Both honor | |
772 | --name-prefix=bar_, and yield | |
773 | ||
e29f0771 | 774 | int bar_parse (void); |
56ca3d8f AD |
775 | |
776 | rather than | |
777 | ||
e29f0771 AD |
778 | #define yyparse bar_parse |
779 | int yyparse (void); | |
56ca3d8f AD |
780 | |
781 | in order to facilitate the inclusion of several parser headers inside a | |
782 | single compilation unit. | |
c3e9f08f | 783 | |
258cddbc AD |
784 | *** Exported symbols in C++ |
785 | ||
786 | The symbols YYTOKEN_TABLE and YYERROR_VERBOSE, which were defined in the | |
787 | header, are removed, as they prevent the possibility of including several | |
788 | generated headers from a single compilation unit. | |
789 | ||
694af10c AD |
790 | *** YYLSP_NEEDED |
791 | ||
792 | For the same reasons, the undocumented and unused macro YYLSP_NEEDED is no | |
793 | longer defined. | |
794 | ||
4b3847c3 AD |
795 | ** New %define variable: api.prefix |
796 | ||
797 | Now that the generated headers are more complete and properly protected | |
798 | against multiple inclusions, constant names, such as YYSTYPE are a | |
799 | problem. While yyparse and others are properly renamed by %name-prefix, | |
800 | YYSTYPE, YYDEBUG and others have never been affected by it. Because it | |
801 | would introduce backward compatibility issues in projects not expecting | |
802 | YYSTYPE to be renamed, instead of changing the behavior of %name-prefix, | |
803 | it is deprecated in favor of a new %define variable: api.prefix. | |
804 | ||
805 | The following examples compares both: | |
806 | ||
807 | %name-prefix "bar_" | %define api.prefix "bar_" | |
808 | %token <ival> FOO %token <ival> FOO | |
809 | %union { int ival; } %union { int ival; } | |
810 | %% %% | |
811 | exp: 'a'; exp: 'a'; | |
812 | ||
813 | bison generates: | |
814 | ||
815 | #ifndef BAR_FOO_H #ifndef BAR_FOO_H | |
816 | # define BAR_FOO_H # define BAR_FOO_H | |
817 | ||
818 | /* Enabling traces. */ /* Enabling traces. */ | |
5f108727 AD |
819 | # ifndef YYDEBUG | # ifndef BAR_DEBUG |
820 | > # if defined YYDEBUG | |
821 | > # if YYDEBUG | |
822 | > # define BAR_DEBUG 1 | |
823 | > # else | |
824 | > # define BAR_DEBUG 0 | |
825 | > # endif | |
826 | > # else | |
827 | # define YYDEBUG 0 | # define BAR_DEBUG 0 | |
828 | > # endif | |
829 | # endif | # endif | |
830 | ||
831 | # if YYDEBUG | # if BAR_DEBUG | |
4b3847c3 AD |
832 | extern int bar_debug; extern int bar_debug; |
833 | # endif # endif | |
834 | ||
835 | /* Tokens. */ /* Tokens. */ | |
836 | # ifndef YYTOKENTYPE | # ifndef BAR_TOKENTYPE | |
837 | # define YYTOKENTYPE | # define BAR_TOKENTYPE | |
838 | enum yytokentype { | enum bar_tokentype { | |
839 | FOO = 258 FOO = 258 | |
840 | }; }; | |
841 | # endif # endif | |
842 | ||
843 | #if ! defined YYSTYPE \ | #if ! defined BAR_STYPE \ | |
844 | && ! defined YYSTYPE_IS_DECLARED | && ! defined BAR_STYPE_IS_DECLARED | |
845 | typedef union YYSTYPE | typedef union BAR_STYPE | |
846 | { { | |
847 | int ival; int ival; | |
848 | } YYSTYPE; | } BAR_STYPE; | |
849 | # define YYSTYPE_IS_DECLARED 1 | # define BAR_STYPE_IS_DECLARED 1 | |
850 | #endif #endif | |
851 | ||
852 | extern YYSTYPE bar_lval; | extern BAR_STYPE bar_lval; | |
853 | ||
854 | int bar_parse (void); int bar_parse (void); | |
855 | ||
856 | #endif /* !BAR_FOO_H */ #endif /* !BAR_FOO_H */ | |
857 | ||
dfaac272 | 858 | * Noteworthy changes in release 2.5.1 (2012-06-05) [stable] |
df6e3db0 | 859 | |
debe2c03 | 860 | ** Future changes: |
765e1bd4 | 861 | |
e4ab1254 | 862 | The next major release will drop support for generating parsers in K&R C. |
041308d0 | 863 | |
466b4cf2 | 864 | ** yacc.c: YYBACKUP works as expected. |
ef51bfa7 | 865 | |
d834eca0 | 866 | ** glr.c improvements: |
041308d0 | 867 | |
d834eca0 | 868 | *** Location support is eliminated when not requested: |
041308d0 | 869 | |
e4ab1254 AD |
870 | GLR parsers used to include location-related code even when locations were |
871 | not requested, and therefore not even usable. | |
378e917c | 872 | |
d834eca0 | 873 | *** __attribute__ is preserved: |
d115aad9 | 874 | |
e4ab1254 AD |
875 | __attribute__ is no longer disabled when __STRICT_ANSI__ is defined (i.e., |
876 | when -std is passed to GCC). | |
041308d0 | 877 | |
466b4cf2 | 878 | ** lalr1.java: several fixes: |
041308d0 | 879 | |
e4ab1254 AD |
880 | The Java parser no longer throws ArrayIndexOutOfBoundsException if the |
881 | first token leads to a syntax error. Some minor clean ups. | |
041308d0 | 882 | |
22172d47 | 883 | ** Changes for C++: |
ef51bfa7 | 884 | |
22172d47 | 885 | *** C++11 compatibility: |
ef51bfa7 | 886 | |
e4ab1254 AD |
887 | C and C++ parsers use "nullptr" instead of "0" when __cplusplus is 201103L |
888 | or higher. | |
936c88d1 | 889 | |
22172d47 AD |
890 | *** Header guards |
891 | ||
892 | The header files such as "parser.hh", "location.hh", etc. used a constant | |
893 | name for preprocessor guards, for instance: | |
894 | ||
e29f0771 AD |
895 | #ifndef BISON_LOCATION_HH |
896 | # define BISON_LOCATION_HH | |
897 | ... | |
898 | #endif // !BISON_LOCATION_HH | |
22172d47 AD |
899 | |
900 | The inclusion guard is now computed from "PREFIX/FILE-NAME", where lower | |
901 | case characters are converted to upper case, and series of | |
902 | non-alphanumerical characters are converted to an underscore. | |
903 | ||
904 | With "bison -o lang++/parser.cc", "location.hh" would now include: | |
905 | ||
e29f0771 AD |
906 | #ifndef YY_LANG_LOCATION_HH |
907 | # define YY_LANG_LOCATION_HH | |
908 | ... | |
909 | #endif // !YY_LANG_LOCATION_HH | |
22172d47 AD |
910 | |
911 | *** C++ locations: | |
936c88d1 | 912 | |
e4ab1254 AD |
913 | The position and location constructors (and their initialize methods) |
914 | accept new arguments for line and column. Several issues in the | |
915 | documentation were fixed. | |
936c88d1 | 916 | |
466b4cf2 AD |
917 | ** liby is no longer asking for "rpl_fprintf" on some platforms. |
918 | ||
7e508a2b AD |
919 | ** Changes in the manual: |
920 | ||
921 | *** %printer is documented | |
922 | ||
e4ab1254 AD |
923 | The "%printer" directive, supported since at least Bison 1.50, is finally |
924 | documented. The "mfcalc" example is extended to demonstrate it. | |
7e508a2b | 925 | |
e4ab1254 AD |
926 | For consistency with the C skeletons, the C++ parsers now also support |
927 | "yyoutput" (as an alias to "debug_stream ()"). | |
7e508a2b AD |
928 | |
929 | *** Several improvements have been made: | |
466b4cf2 | 930 | |
e4ab1254 AD |
931 | The layout for grammar excerpts was changed to a more compact scheme. |
932 | Named references are motivated. The description of the automaton | |
933 | description file (*.output) is updated to the current format. Incorrect | |
934 | index entries were fixed. Some other errors were fixed. | |
466b4cf2 | 935 | |
86b08b49 AD |
936 | ** Building bison: |
937 | ||
938 | *** Conflicting prototypes with recent/modified Flex. | |
466b4cf2 | 939 | |
e4ab1254 AD |
940 | Fixed build problems with the current, unreleased, version of Flex, and |
941 | some modified versions of 2.5.35, which have modified function prototypes. | |
466b4cf2 | 942 | |
8ef26c2a AD |
943 | *** Warnings during the build procedure have been eliminated. |
944 | ||
945 | *** Several portability problems in the test suite have been fixed: | |
466b4cf2 | 946 | |
e4ab1254 AD |
947 | This includes warnings with some compilers, unexpected behavior of tools |
948 | such as diff, warning messages from the test suite itself, etc. | |
466b4cf2 | 949 | |
91aadcc7 | 950 | *** The install-pdf target works properly: |
8ef26c2a | 951 | |
e4ab1254 AD |
952 | Running "make install-pdf" (or -dvi, -html, -info, and -ps) no longer |
953 | halts in the middle of its course. | |
8ef26c2a | 954 | |
28801043 | 955 | * Changes in version 2.5 (2011-05-14): |
50cca368 | 956 | |
82f3355e JD |
957 | ** Grammar symbol names can now contain non-initial dashes: |
958 | ||
959 | Consistently with directives (such as %error-verbose) and with | |
960 | %define variables (e.g. push-pull), grammar symbol names may contain | |
961 | dashes in any position except the beginning. This is a GNU | |
962 | extension over POSIX Yacc. Thus, use of this extension is reported | |
963 | by -Wyacc and rejected in Yacc mode (--yacc). | |
964 | ||
f1b238df | 965 | ** Named references: |
66381412 AR |
966 | |
967 | Historically, Yacc and Bison have supported positional references | |
968 | ($n, $$) to allow access to symbol values from inside of semantic | |
969 | actions code. | |
970 | ||
971 | Starting from this version, Bison can also accept named references. | |
972 | When no ambiguity is possible, original symbol names may be used | |
973 | as named references: | |
974 | ||
4b568fc0 | 975 | if_stmt : "if" cond_expr "then" then_stmt ';' |
66381412 AR |
976 | { $if_stmt = mk_if_stmt($cond_expr, $then_stmt); } |
977 | ||
978 | In the more common case, explicit names may be declared: | |
979 | ||
4b568fc0 | 980 | stmt[res] : "if" expr[cond] "then" stmt[then] "else" stmt[else] ';' |
66381412 AR |
981 | { $res = mk_if_stmt($cond, $then, $else); } |
982 | ||
5b1ff423 | 983 | Location information is also accessible using @name syntax. When |
66381412 AR |
984 | accessing symbol names containing dots or dashes, explicit bracketing |
985 | ($[sym.1]) must be used. | |
986 | ||
5b1ff423 | 987 | These features are experimental in this version. More user feedback |
66381412 | 988 | will help to stabilize them. |
2bd435c3 | 989 | Contributed by Alex Rozenman. |
66381412 | 990 | |
f1b238df | 991 | ** IELR(1) and canonical LR(1): |
eb45ef3b JD |
992 | |
993 | IELR(1) is a minimal LR(1) parser table generation algorithm. That | |
994 | is, given any context-free grammar, IELR(1) generates parser tables | |
7262f54f | 995 | with the full language-recognition power of canonical LR(1) but with |
f1b238df JD |
996 | nearly the same number of parser states as LALR(1). This reduction |
997 | in parser states is often an order of magnitude. More importantly, | |
eb45ef3b JD |
998 | because canonical LR(1)'s extra parser states may contain duplicate |
999 | conflicts in the case of non-LR(1) grammars, the number of conflicts | |
1000 | for IELR(1) is often an order of magnitude less as well. This can | |
1001 | significantly reduce the complexity of developing of a grammar. | |
1002 | ||
1003 | Bison can now generate IELR(1) and canonical LR(1) parser tables in | |
1004 | place of its traditional LALR(1) parser tables, which remain the | |
1005 | default. You can specify the type of parser tables in the grammar | |
1006 | file with these directives: | |
1007 | ||
cf499cff JD |
1008 | %define lr.type lalr |
1009 | %define lr.type ielr | |
1010 | %define lr.type canonical-lr | |
eb45ef3b | 1011 | |
7fceb615 | 1012 | The default-reduction optimization in the parser tables can also be |
e4ab1254 AD |
1013 | adjusted using "%define lr.default-reductions". For details on both |
1014 | of these features, see the new section "Tuning LR" in the Bison | |
7fceb615 | 1015 | manual. |
eb45ef3b JD |
1016 | |
1017 | These features are experimental. More user feedback will help to | |
1018 | stabilize them. | |
1019 | ||
8458a411 AD |
1020 | ** LAC (Lookahead Correction) for syntax error handling |
1021 | ||
1022 | Contributed by Joel E. Denny. | |
fcf834f9 JD |
1023 | |
1024 | Canonical LR, IELR, and LALR can suffer from a couple of problems | |
1025 | upon encountering a syntax error. First, the parser might perform | |
1026 | additional parser stack reductions before discovering the syntax | |
7fceb615 | 1027 | error. Such reductions can perform user semantic actions that are |
fcf834f9 JD |
1028 | unexpected because they are based on an invalid token, and they |
1029 | cause error recovery to begin in a different syntactic context than | |
1030 | the one in which the invalid token was encountered. Second, when | |
7fceb615 | 1031 | verbose error messages are enabled (with %error-verbose or the |
e4ab1254 | 1032 | obsolete "#define YYERROR_VERBOSE"), the expected token list in the |
7fceb615 JD |
1033 | syntax error message can both contain invalid tokens and omit valid |
1034 | tokens. | |
fcf834f9 JD |
1035 | |
1036 | The culprits for the above problems are %nonassoc, default | |
1037 | reductions in inconsistent states, and parser state merging. Thus, | |
1038 | IELR and LALR suffer the most. Canonical LR can suffer only if | |
1039 | %nonassoc is used or if default reductions are enabled for | |
1040 | inconsistent states. | |
1041 | ||
7fceb615 JD |
1042 | LAC is a new mechanism within the parsing algorithm that solves |
1043 | these problems for canonical LR, IELR, and LALR without sacrificing | |
1044 | %nonassoc, default reductions, or state merging. When LAC is in | |
1045 | use, canonical LR and IELR behave almost exactly the same for both | |
1046 | syntactically acceptable and syntactically unacceptable input. | |
fcf834f9 JD |
1047 | While LALR still does not support the full language-recognition |
1048 | power of canonical LR and IELR, LAC at least enables LALR's syntax | |
1049 | error handling to correctly reflect LALR's language-recognition | |
1050 | power. | |
1051 | ||
1052 | Currently, LAC is only supported for deterministic parsers in C. | |
1053 | You can enable LAC with the following directive: | |
1054 | ||
1055 | %define parse.lac full | |
1056 | ||
e4ab1254 | 1057 | See the new section "LAC" in the Bison manual for additional |
7fceb615 | 1058 | details including a few caveats. |
fcf834f9 JD |
1059 | |
1060 | LAC is an experimental feature. More user feedback will help to | |
1061 | stabilize it. | |
1062 | ||
d397d9f0 | 1063 | ** %define improvements: |
cf499cff | 1064 | |
f1b238df | 1065 | *** Can now be invoked via the command line: |
50cca368 | 1066 | |
de5ab940 | 1067 | Each of these command-line options |
50cca368 | 1068 | |
de5ab940 JD |
1069 | -D NAME[=VALUE] |
1070 | --define=NAME[=VALUE] | |
1071 | ||
1072 | -F NAME[=VALUE] | |
1073 | --force-define=NAME[=VALUE] | |
50cca368 JD |
1074 | |
1075 | is equivalent to this grammar file declaration | |
1076 | ||
de5ab940 | 1077 | %define NAME ["VALUE"] |
50cca368 | 1078 | |
de5ab940 JD |
1079 | except that the manner in which Bison processes multiple definitions |
1080 | for the same NAME differs. Most importantly, -F and --force-define | |
1081 | quietly override %define, but -D and --define do not. For further | |
e4ab1254 | 1082 | details, see the section "Bison Options" in the Bison manual. |
50cca368 | 1083 | |
f1b238df | 1084 | *** Variables renamed: |
67212941 JD |
1085 | |
1086 | The following %define variables | |
1087 | ||
1088 | api.push_pull | |
1089 | lr.keep_unreachable_states | |
1090 | ||
1091 | have been renamed to | |
1092 | ||
1093 | api.push-pull | |
1094 | lr.keep-unreachable-states | |
1095 | ||
1096 | The old names are now deprecated but will be maintained indefinitely | |
1097 | for backward compatibility. | |
1098 | ||
7262f54f | 1099 | *** Values no longer need to be quoted in the grammar file: |
cf499cff JD |
1100 | |
1101 | If a %define value is an identifier, it no longer needs to be placed | |
1102 | within quotations marks. For example, | |
1103 | ||
1104 | %define api.push-pull "push" | |
1105 | ||
1106 | can be rewritten as | |
1107 | ||
1108 | %define api.push-pull push | |
1109 | ||
d397d9f0 | 1110 | *** Unrecognized variables are now errors not warnings. |
cdf3f113 | 1111 | |
d397d9f0 JD |
1112 | *** Multiple invocations for any variable is now an error not a warning. |
1113 | ||
1114 | ** Unrecognized %code qualifiers are now errors not warnings. | |
1115 | ||
1116 | ** Character literals not of length one: | |
1117 | ||
1118 | Previously, Bison quietly converted all character literals to length | |
1119 | one. For example, without warning, Bison interpreted the operators in | |
1120 | the following grammar to be the same token: | |
1121 | ||
1122 | exp: exp '++' | |
1123 | | exp '+' exp | |
1124 | ; | |
1125 | ||
1126 | Bison now warns when a character literal is not of length one. In | |
1127 | some future release, Bison will start reporting an error instead. | |
1128 | ||
1129 | ** Destructor calls fixed for lookaheads altered in semantic actions: | |
1130 | ||
1131 | Previously for deterministic parsers in C, if a user semantic action | |
1132 | altered yychar, the parser in some cases used the old yychar value to | |
1133 | determine which destructor to call for the lookahead upon a syntax | |
1134 | error or upon parser return. This bug has been fixed. | |
1135 | ||
1136 | ** C++ parsers use YYRHSLOC: | |
1137 | ||
1138 | Similarly to the C parsers, the C++ parsers now define the YYRHSLOC | |
1139 | macro and use it in the default YYLLOC_DEFAULT. You are encouraged | |
e4ab1254 AD |
1140 | to use it. If, for instance, your location structure has "first" |
1141 | and "last" members, instead of | |
d397d9f0 | 1142 | |
e29f0771 AD |
1143 | # define YYLLOC_DEFAULT(Current, Rhs, N) \ |
1144 | do \ | |
1145 | if (N) \ | |
1146 | { \ | |
1147 | (Current).first = (Rhs)[1].location.first; \ | |
1148 | (Current).last = (Rhs)[N].location.last; \ | |
1149 | } \ | |
1150 | else \ | |
1151 | { \ | |
1152 | (Current).first = (Current).last = (Rhs)[0].location.last; \ | |
1153 | } \ | |
1154 | while (false) | |
d397d9f0 JD |
1155 | |
1156 | use: | |
1157 | ||
e29f0771 AD |
1158 | # define YYLLOC_DEFAULT(Current, Rhs, N) \ |
1159 | do \ | |
1160 | if (N) \ | |
1161 | { \ | |
1162 | (Current).first = YYRHSLOC (Rhs, 1).first; \ | |
1163 | (Current).last = YYRHSLOC (Rhs, N).last; \ | |
1164 | } \ | |
1165 | else \ | |
1166 | { \ | |
1167 | (Current).first = (Current).last = YYRHSLOC (Rhs, 0).last; \ | |
1168 | } \ | |
1169 | while (false) | |
d397d9f0 JD |
1170 | |
1171 | ** YYLLOC_DEFAULT in C++: | |
1172 | ||
1173 | The default implementation of YYLLOC_DEFAULT used to be issued in | |
1174 | the header file. It is now output in the implementation file, after | |
1175 | the user %code sections so that its #ifndef guard does not try to | |
1176 | override the user's YYLLOC_DEFAULT if provided. | |
cdf3f113 | 1177 | |
f1b238df | 1178 | ** YYFAIL now produces warnings and Java parsers no longer implement it: |
4395a9ff JD |
1179 | |
1180 | YYFAIL has existed for many years as an undocumented feature of | |
1181 | deterministic parsers in C generated by Bison. More recently, it was | |
1182 | a documented feature of Bison's experimental Java parsers. As | |
1183 | promised in Bison 2.4.2's NEWS entry, any appearance of YYFAIL in a | |
1184 | semantic action now produces a deprecation warning, and Java parsers | |
1185 | no longer implement YYFAIL at all. For further details, including a | |
1186 | discussion of how to suppress C preprocessor warnings about YYFAIL | |
1187 | being unused, see the Bison 2.4.2 NEWS entry. | |
1188 | ||
f1b238df | 1189 | ** Temporary hack for adding a semicolon to the user action: |
197b82ba JD |
1190 | |
1191 | Previously, Bison appended a semicolon to every user action for | |
1192 | reductions when the output language defaulted to C (specifically, when | |
1193 | neither %yacc, %language, %skeleton, or equivalent command-line | |
1194 | options were specified). This allowed actions such as | |
1195 | ||
1196 | exp: exp "+" exp { $$ = $1 + $3 }; | |
1197 | ||
1198 | instead of | |
1199 | ||
1200 | exp: exp "+" exp { $$ = $1 + $3; }; | |
1201 | ||
1202 | As a first step in removing this misfeature, Bison now issues a | |
1203 | warning when it appends a semicolon. Moreover, in cases where Bison | |
1204 | cannot easily determine whether a semicolon is needed (for example, an | |
1205 | action ending with a cpp directive or a braced compound initializer), | |
1206 | it no longer appends one. Thus, the C compiler might now complain | |
1207 | about a missing semicolon where it did not before. Future releases of | |
1208 | Bison will cease to append semicolons entirely. | |
1209 | ||
d2060f06 JD |
1210 | ** Verbose syntax error message fixes: |
1211 | ||
e4ab1254 | 1212 | When %error-verbose or the obsolete "#define YYERROR_VERBOSE" is |
7fceb615 JD |
1213 | specified, syntax error messages produced by the generated parser |
1214 | include the unexpected token as well as a list of expected tokens. | |
1215 | The effect of %nonassoc on these verbose messages has been corrected | |
1216 | in two ways, but a more complete fix requires LAC, described above: | |
d2060f06 JD |
1217 | |
1218 | *** When %nonassoc is used, there can exist parser states that accept no | |
1219 | tokens, and so the parser does not always require a lookahead token | |
1220 | in order to detect a syntax error. Because no unexpected token or | |
1221 | expected tokens can then be reported, the verbose syntax error | |
1222 | message described above is suppressed, and the parser instead | |
e4ab1254 | 1223 | reports the simpler message, "syntax error". Previously, this |
d2060f06 JD |
1224 | suppression was sometimes erroneously triggered by %nonassoc when a |
1225 | lookahead was actually required. Now verbose messages are | |
1226 | suppressed only when all previous lookaheads have already been | |
1227 | shifted or discarded. | |
1228 | ||
1229 | *** Previously, the list of expected tokens erroneously included tokens | |
1230 | that would actually induce a syntax error because conflicts for them | |
1231 | were resolved with %nonassoc in the current parser state. Such | |
1232 | tokens are now properly omitted from the list. | |
1233 | ||
1234 | *** Expected token lists are still often wrong due to state merging | |
fcf834f9 JD |
1235 | (from LALR or IELR) and default reductions, which can both add |
1236 | invalid tokens and subtract valid tokens. Canonical LR almost | |
1237 | completely fixes this problem by eliminating state merging and | |
1238 | default reductions. However, there is one minor problem left even | |
1239 | when using canonical LR and even after the fixes above. That is, | |
1240 | if the resolution of a conflict with %nonassoc appears in a later | |
1241 | parser state than the one at which some syntax error is | |
1242 | discovered, the conflicted token is still erroneously included in | |
1243 | the expected token list. Bison's new LAC implementation, | |
1244 | described above, eliminates this problem and the need for | |
1245 | canonical LR. However, LAC is still experimental and is disabled | |
1246 | by default. | |
53f036ce | 1247 | |
1a33f4f6 JD |
1248 | ** Java skeleton fixes: |
1249 | ||
1250 | *** A location handling bug has been fixed. | |
1251 | ||
1252 | *** The top element of each of the value stack and location stack is now | |
1253 | cleared when popped so that it can be garbage collected. | |
6771a463 | 1254 | |
02803d55 JD |
1255 | *** Parser traces now print the top element of the stack. |
1256 | ||
86408959 JD |
1257 | ** -W/--warnings fixes: |
1258 | ||
e4ab1254 | 1259 | *** Bison now properly recognizes the "no-" versions of categories: |
86408959 JD |
1260 | |
1261 | For example, given the following command line, Bison now enables all | |
1262 | warnings except warnings for incompatibilities with POSIX Yacc: | |
1263 | ||
1264 | bison -Wall,no-yacc gram.y | |
1265 | ||
786743d5 JD |
1266 | *** Bison now treats S/R and R/R conflicts like other warnings: |
1267 | ||
1268 | Previously, conflict reports were independent of Bison's normal | |
1269 | warning system. Now, Bison recognizes the warning categories | |
e4ab1254 | 1270 | "conflicts-sr" and "conflicts-rr". This change has important |
786743d5 JD |
1271 | consequences for the -W and --warnings command-line options. For |
1272 | example: | |
1273 | ||
1274 | bison -Wno-conflicts-sr gram.y # S/R conflicts not reported | |
1275 | bison -Wno-conflicts-rr gram.y # R/R conflicts not reported | |
1276 | bison -Wnone gram.y # no conflicts are reported | |
1277 | bison -Werror gram.y # any conflict is an error | |
1278 | ||
1279 | However, as before, if the %expect or %expect-rr directive is | |
1280 | specified, an unexpected number of conflicts is an error, and an | |
1281 | expected number of conflicts is not reported, so -W and --warning | |
1282 | then have no effect on the conflict report. | |
1283 | ||
e4ab1254 | 1284 | *** The "none" category no longer disables a preceding "error": |
bf0e44e8 JD |
1285 | |
1286 | For example, for the following command line, Bison now reports | |
1287 | errors instead of warnings for incompatibilities with POSIX Yacc: | |
1288 | ||
1289 | bison -Werror,none,yacc gram.y | |
1290 | ||
e4ab1254 | 1291 | *** The "none" category now disables all Bison warnings: |
c39014ae | 1292 | |
e4ab1254 | 1293 | Previously, the "none" category disabled only Bison warnings for |
c39014ae JD |
1294 | which there existed a specific -W/--warning category. However, |
1295 | given the following command line, Bison is now guaranteed to | |
1296 | suppress all warnings: | |
1297 | ||
1298 | bison -Wnone gram.y | |
1299 | ||
1f36f544 JD |
1300 | ** Precedence directives can now assign token number 0: |
1301 | ||
1302 | Since Bison 2.3b, which restored the ability of precedence | |
1303 | directives to assign token numbers, doing so for token number 0 has | |
1304 | produced an assertion failure. For example: | |
1305 | ||
1306 | %left END 0 | |
1307 | ||
1308 | This bug has been fixed. | |
1309 | ||
64877e5e | 1310 | * Changes in version 2.4.3 (2010-08-05): |
8b9e021f | 1311 | |
2bfcac9a JD |
1312 | ** Bison now obeys -Werror and --warnings=error for warnings about |
1313 | grammar rules that are useless in the parser due to conflicts. | |
1314 | ||
8b9e021f JD |
1315 | ** Problems with spawning M4 on at least FreeBSD 8 and FreeBSD 9 have |
1316 | been fixed. | |
1317 | ||
4ad3921d JD |
1318 | ** Failures in the test suite for GCC 4.5 have been fixed. |
1319 | ||
06cb07d5 JD |
1320 | ** Failures in the test suite for some versions of Sun Studio C++ have |
1321 | been fixed. | |
1322 | ||
9b5049bd JD |
1323 | ** Contrary to Bison 2.4.2's NEWS entry, it has been decided that |
1324 | warnings about undefined %prec identifiers will not be converted to | |
1325 | errors in Bison 2.5. They will remain warnings, which should be | |
1326 | sufficient for POSIX while avoiding backward compatibility issues. | |
1327 | ||
93d7dde9 JD |
1328 | ** Minor documentation fixes. |
1329 | ||
e19a049c | 1330 | * Changes in version 2.4.2 (2010-03-20): |
74553c98 | 1331 | |
f39ab286 JD |
1332 | ** Some portability problems that resulted in failures and livelocks |
1333 | in the test suite on some versions of at least Solaris, AIX, HP-UX, | |
e19a049c JD |
1334 | RHEL4, and Tru64 have been addressed. As a result, fatal Bison |
1335 | errors should no longer cause M4 to report a broken pipe on the | |
f39ab286 JD |
1336 | affected platforms. |
1337 | ||
e4ab1254 | 1338 | ** "%prec IDENTIFIER" requires IDENTIFIER to be defined separately. |
8bb3a2e7 JD |
1339 | |
1340 | POSIX specifies that an error be reported for any identifier that does | |
1341 | not appear on the LHS of a grammar rule and that is not defined by | |
1342 | %token, %left, %right, or %nonassoc. Bison 2.3b and later lost this | |
1343 | error report for the case when an identifier appears only after a | |
1344 | %prec directive. It is now restored. However, for backward | |
1345 | compatibility with recent Bison releases, it is only a warning for | |
1346 | now. In Bison 2.5 and later, it will return to being an error. | |
9b5049bd JD |
1347 | [Between the 2.4.2 and 2.4.3 releases, it was decided that this |
1348 | warning will not be converted to an error in Bison 2.5.] | |
8bb3a2e7 | 1349 | |
d8911864 EB |
1350 | ** Detection of GNU M4 1.4.6 or newer during configure is improved. |
1351 | ||
a603c6e0 JD |
1352 | ** Warnings from gcc's -Wundef option about undefined YYENABLE_NLS, |
1353 | YYLTYPE_IS_TRIVIAL, and __STRICT_ANSI__ in C/C++ parsers are now | |
1354 | avoided. | |
c938d650 | 1355 | |
98a345a2 JD |
1356 | ** %code is now a permanent feature. |
1357 | ||
1358 | A traditional Yacc prologue directive is written in the form: | |
1359 | ||
1360 | %{CODE%} | |
1361 | ||
1362 | To provide a more flexible alternative, Bison 2.3b introduced the | |
1363 | %code directive with the following forms for C/C++: | |
1364 | ||
1365 | %code {CODE} | |
1366 | %code requires {CODE} | |
1367 | %code provides {CODE} | |
1368 | %code top {CODE} | |
1369 | ||
1370 | These forms are now considered permanent features of Bison. See the | |
1371 | %code entries in the section "Bison Declaration Summary" in the Bison | |
1372 | manual for a summary of their functionality. See the section | |
1373 | "Prologue Alternatives" for a detailed discussion including the | |
1374 | advantages of %code over the traditional Yacc prologue directive. | |
1375 | ||
1376 | Bison's Java feature as a whole including its current usage of %code | |
1377 | is still considered experimental. | |
1378 | ||
1625df5b JD |
1379 | ** YYFAIL is deprecated and will eventually be removed. |
1380 | ||
1381 | YYFAIL has existed for many years as an undocumented feature of | |
1382 | deterministic parsers in C generated by Bison. Previously, it was | |
1383 | documented for Bison's experimental Java parsers. YYFAIL is no longer | |
1384 | documented for Java parsers and is formally deprecated in both cases. | |
1385 | Users are strongly encouraged to migrate to YYERROR, which is | |
1386 | specified by POSIX. | |
1387 | ||
1388 | Like YYERROR, you can invoke YYFAIL from a semantic action in order to | |
1389 | induce a syntax error. The most obvious difference from YYERROR is | |
1390 | that YYFAIL will automatically invoke yyerror to report the syntax | |
1391 | error so that you don't have to. However, there are several other | |
1392 | subtle differences between YYERROR and YYFAIL, and YYFAIL suffers from | |
e4ab1254 | 1393 | inherent flaws when %error-verbose or "#define YYERROR_VERBOSE" is |
1625df5b JD |
1394 | used. For a more detailed discussion, see: |
1395 | ||
1396 | http://lists.gnu.org/archive/html/bison-patches/2009-12/msg00024.html | |
1397 | ||
1398 | The upcoming Bison 2.5 will remove YYFAIL from Java parsers, but | |
1399 | deterministic parsers in C will continue to implement it. However, | |
1400 | because YYFAIL is already flawed, it seems futile to try to make new | |
1401 | Bison features compatible with it. Thus, during parser generation, | |
1402 | Bison 2.5 will produce a warning whenever it discovers YYFAIL in a | |
1403 | rule action. In a later release, YYFAIL will be disabled for | |
e4ab1254 | 1404 | %error-verbose and "#define YYERROR_VERBOSE". Eventually, YYFAIL will |
1625df5b JD |
1405 | be removed altogether. |
1406 | ||
1407 | There exists at least one case where Bison 2.5's YYFAIL warning will | |
1408 | be a false positive. Some projects add phony uses of YYFAIL and other | |
1409 | Bison-defined macros for the sole purpose of suppressing C | |
1410 | preprocessor warnings (from GCC cpp's -Wunused-macros, for example). | |
1411 | To avoid Bison's future warning, such YYFAIL uses can be moved to the | |
e4ab1254 | 1412 | epilogue (that is, after the second "%%") in the Bison input file. In |
1625df5b JD |
1413 | this release (2.4.2), Bison already generates its own code to suppress |
1414 | C preprocessor warnings for YYFAIL, so projects can remove their own | |
1415 | phony uses of YYFAIL if compatibility with Bison releases prior to | |
1416 | 2.4.2 is not necessary. | |
1417 | ||
2755de8f AD |
1418 | ** Internationalization. |
1419 | ||
1420 | Fix a regression introduced in Bison 2.4: Under some circumstances, | |
1421 | message translations were not installed although supported by the | |
1422 | host system. | |
1423 | ||
74553c98 | 1424 | * Changes in version 2.4.1 (2008-12-11): |
c9ba9e59 | 1425 | |
0ea583d2 AD |
1426 | ** In the GLR defines file, unexpanded M4 macros in the yylval and yylloc |
1427 | declarations have been fixed. | |
1979121c | 1428 | |
0ea583d2 AD |
1429 | ** Temporary hack for adding a semicolon to the user action. |
1430 | ||
1431 | Bison used to prepend a trailing semicolon at the end of the user | |
1432 | action for reductions. This allowed actions such as | |
1433 | ||
1434 | exp: exp "+" exp { $$ = $1 + $3 }; | |
1435 | ||
1436 | instead of | |
1437 | ||
1438 | exp: exp "+" exp { $$ = $1 + $3; }; | |
1439 | ||
e4ab1254 | 1440 | Some grammars still depend on this "feature". Bison 2.4.1 restores |
0ea583d2 AD |
1441 | the previous behavior in the case of C output (specifically, when |
1442 | neither %language or %skeleton or equivalent command-line options | |
1443 | are used) to leave more time for grammars depending on the old | |
1444 | behavior to be adjusted. Future releases of Bison will disable this | |
1445 | feature. | |
1446 | ||
1447 | ** A few minor improvements to the Bison manual. | |
c9ba9e59 | 1448 | |
402b123d | 1449 | * Changes in version 2.4 (2008-11-02): |
7bd1665a | 1450 | |
402b123d | 1451 | ** %language is an experimental feature. |
ed4d67dc JD |
1452 | |
1453 | We first introduced this feature in test release 2.3b as a cleaner | |
1454 | alternative to %skeleton. Since then, we have discussed the possibility of | |
1455 | modifying its effect on Bison's output file names. Thus, in this release, | |
1456 | we consider %language to be an experimental feature that will likely evolve | |
1457 | in future releases. | |
7bd1665a | 1458 | |
402b123d | 1459 | ** Forward compatibility with GNU M4 has been improved. |
241fda7a | 1460 | |
402b123d | 1461 | ** Several bugs in the C++ skeleton and the experimental Java skeleton have been |
241fda7a JD |
1462 | fixed. |
1463 | ||
402b123d | 1464 | * Changes in version 2.3b (2008-05-27): |
35fe0834 | 1465 | |
402b123d | 1466 | ** The quotes around NAME that used to be required in the following directive |
d9df47b6 JD |
1467 | are now deprecated: |
1468 | ||
1469 | %define NAME "VALUE" | |
1470 | ||
e4ab1254 | 1471 | ** The directive "%pure-parser" is now deprecated in favor of: |
d9df47b6 JD |
1472 | |
1473 | %define api.pure | |
1474 | ||
1475 | which has the same effect except that Bison is more careful to warn about | |
1476 | unreasonable usage in the latter case. | |
1477 | ||
402b123d | 1478 | ** Push Parsing |
c373bf8b JD |
1479 | |
1480 | Bison can now generate an LALR(1) parser in C with a push interface. That | |
e4ab1254 AD |
1481 | is, instead of invoking "yyparse", which pulls tokens from "yylex", you can |
1482 | push one token at a time to the parser using "yypush_parse", which will | |
c373bf8b JD |
1483 | return to the caller after processing each token. By default, the push |
1484 | interface is disabled. Either of the following directives will enable it: | |
1485 | ||
1486 | %define api.push_pull "push" // Just push; does not require yylex. | |
1487 | %define api.push_pull "both" // Push and pull; requires yylex. | |
1488 | ||
e4ab1254 | 1489 | See the new section "A Push Parser" in the Bison manual for details. |
c373bf8b | 1490 | |
59da312b JD |
1491 | The current push parsing interface is experimental and may evolve. More user |
1492 | feedback will help to stabilize it. | |
1493 | ||
402b123d | 1494 | ** The -g and --graph options now output graphs in Graphviz DOT format, |
8e55b3aa JD |
1495 | not VCG format. Like --graph, -g now also takes an optional FILE argument |
1496 | and thus cannot be bundled with other short options. | |
c373bf8b | 1497 | |
402b123d | 1498 | ** Java |
59da312b JD |
1499 | |
1500 | Bison can now generate an LALR(1) parser in Java. The skeleton is | |
e4ab1254 | 1501 | "data/lalr1.java". Consider using the new %language directive instead of |
59da312b JD |
1502 | %skeleton to select it. |
1503 | ||
e4ab1254 | 1504 | See the new section "Java Parsers" in the Bison manual for details. |
59da312b JD |
1505 | |
1506 | The current Java interface is experimental and may evolve. More user | |
1507 | feedback will help to stabilize it. | |
2bd435c3 | 1508 | Contributed by Paolo Bonzini. |
59da312b | 1509 | |
402b123d | 1510 | ** %language |
59da312b JD |
1511 | |
1512 | This new directive specifies the programming language of the generated | |
d43f77e7 PB |
1513 | parser, which can be C (the default), C++, or Java. Besides the skeleton |
1514 | that Bison uses, the directive affects the names of the generated files if | |
1515 | the grammar file's name ends in ".y". | |
59da312b | 1516 | |
402b123d | 1517 | ** XML Automaton Report |
59da312b JD |
1518 | |
1519 | Bison can now generate an XML report of the LALR(1) automaton using the new | |
e4ab1254 | 1520 | "--xml" option. The current XML schema is experimental and may evolve. More |
59da312b | 1521 | user feedback will help to stabilize it. |
2bd435c3 | 1522 | Contributed by Wojciech Polak. |
c373bf8b | 1523 | |
402b123d | 1524 | ** The grammar file may now specify the name of the parser header file using |
c373bf8b JD |
1525 | %defines. For example: |
1526 | ||
1527 | %defines "parser.h" | |
1528 | ||
402b123d | 1529 | ** When reporting useless rules, useless nonterminals, and unused terminals, |
d80fb37a JD |
1530 | Bison now employs the terms "useless in grammar" instead of "useless", |
1531 | "useless in parser" instead of "never reduced", and "unused in grammar" | |
1532 | instead of "unused". | |
cff03fb2 | 1533 | |
402b123d | 1534 | ** Unreachable State Removal |
c373bf8b JD |
1535 | |
1536 | Previously, Bison sometimes generated parser tables containing unreachable | |
31984206 JD |
1537 | states. A state can become unreachable during conflict resolution if Bison |
1538 | disables a shift action leading to it from a predecessor state. Bison now: | |
75ad86ee JD |
1539 | |
1540 | 1. Removes unreachable states. | |
1541 | ||
1542 | 2. Does not report any conflicts that appeared in unreachable states. | |
1543 | WARNING: As a result, you may need to update %expect and %expect-rr | |
1544 | directives in existing grammar files. | |
1545 | ||
1546 | 3. For any rule used only in such states, Bison now reports the rule as | |
cff03fb2 | 1547 | "useless in parser due to conflicts". |
75ad86ee | 1548 | |
31984206 JD |
1549 | This feature can be disabled with the following directive: |
1550 | ||
1551 | %define lr.keep_unreachable_states | |
1552 | ||
e4ab1254 | 1553 | See the %define entry in the "Bison Declaration Summary" in the Bison manual |
31984206 JD |
1554 | for further discussion. |
1555 | ||
e4ab1254 | 1556 | ** Lookahead Set Correction in the ".output" Report |
b1cc23c4 | 1557 | |
e4ab1254 AD |
1558 | When instructed to generate a ".output" file including lookahead sets |
1559 | (using "--report=lookahead", for example), Bison now prints each reduction's | |
88c78747 JD |
1560 | lookahead set only next to the associated state's one item that (1) is |
1561 | associated with the same rule as the reduction and (2) has its dot at the end | |
1562 | of its RHS. Previously, Bison also erroneously printed the lookahead set | |
1563 | next to all of the state's other items associated with the same rule. This | |
e4ab1254 | 1564 | bug affected only the ".output" file and not the generated parser source |
88c78747 JD |
1565 | code. |
1566 | ||
e4ab1254 | 1567 | ** --report-file=FILE is a new option to override the default ".output" file |
59da312b | 1568 | name. |
1bb2bd75 | 1569 | |
e4ab1254 | 1570 | ** The "=" that used to be required in the following directives is now |
02975b9a JD |
1571 | deprecated: |
1572 | ||
1573 | %file-prefix "parser" | |
1574 | %name-prefix "c_" | |
1575 | %output "parser.c" | |
1576 | ||
e4ab1254 | 1577 | ** An Alternative to "%{...%}" -- "%code QUALIFIER {CODE}" |
c373bf8b JD |
1578 | |
1579 | Bison 2.3a provided a new set of directives as a more flexible alternative to | |
8e0a5e9e JD |
1580 | the traditional Yacc prologue blocks. Those have now been consolidated into |
1581 | a single %code directive with an optional qualifier field, which identifies | |
1582 | the purpose of the code and thus the location(s) where Bison should generate | |
1583 | it: | |
1584 | ||
e4ab1254 AD |
1585 | 1. "%code {CODE}" replaces "%after-header {CODE}" |
1586 | 2. "%code requires {CODE}" replaces "%start-header {CODE}" | |
1587 | 3. "%code provides {CODE}" replaces "%end-header {CODE}" | |
1588 | 4. "%code top {CODE}" replaces "%before-header {CODE}" | |
8e0a5e9e | 1589 | |
e4ab1254 AD |
1590 | See the %code entries in section "Bison Declaration Summary" in the Bison |
1591 | manual for a summary of the new functionality. See the new section "Prologue | |
1592 | Alternatives" for a detailed discussion including the advantages of %code | |
8e0a5e9e JD |
1593 | over the traditional Yacc prologues. |
1594 | ||
1595 | The prologue alternatives are experimental. More user feedback will help to | |
1596 | determine whether they should become permanent features. | |
1597 | ||
402b123d | 1598 | ** Revised warning: unset or unused mid-rule values |
17bd8a73 JD |
1599 | |
1600 | Since Bison 2.2, Bison has warned about mid-rule values that are set but not | |
1601 | used within any of the actions of the parent rule. For example, Bison warns | |
1602 | about unused $2 in: | |
1603 | ||
1604 | exp: '1' { $$ = 1; } '+' exp { $$ = $1 + $4; }; | |
1605 | ||
1606 | Now, Bison also warns about mid-rule values that are used but not set. For | |
1607 | example, Bison warns about unset $$ in the mid-rule action in: | |
1608 | ||
1609 | exp: '1' { $1 = 1; } '+' exp { $$ = $2 + $4; }; | |
1610 | ||
1611 | However, Bison now disables both of these warnings by default since they | |
1612 | sometimes prove to be false alarms in existing grammars employing the Yacc | |
1613 | constructs $0 or $-N (where N is some positive integer). | |
1614 | ||
e4ab1254 AD |
1615 | To enable these warnings, specify the option "--warnings=midrule-values" or |
1616 | "-W", which is a synonym for "--warnings=all". | |
17bd8a73 | 1617 | |
e4ab1254 | 1618 | ** Default %destructor or %printer with "<*>" or "<>" |
c373bf8b JD |
1619 | |
1620 | Bison now recognizes two separate kinds of default %destructor's and | |
12e35840 JD |
1621 | %printer's: |
1622 | ||
e4ab1254 | 1623 | 1. Place "<*>" in a %destructor/%printer symbol list to define a default |
12e35840 JD |
1624 | %destructor/%printer for all grammar symbols for which you have formally |
1625 | declared semantic type tags. | |
1626 | ||
e4ab1254 | 1627 | 2. Place "<>" in a %destructor/%printer symbol list to define a default |
12e35840 JD |
1628 | %destructor/%printer for all grammar symbols without declared semantic |
1629 | type tags. | |
1630 | ||
e4ab1254 AD |
1631 | Bison no longer supports the "%symbol-default" notation from Bison 2.3a. |
1632 | "<*>" and "<>" combined achieve the same effect with one exception: Bison no | |
12e35840 JD |
1633 | longer applies any %destructor to a mid-rule value if that mid-rule value is |
1634 | not actually ever referenced using either $$ or $n in a semantic action. | |
1635 | ||
85894313 JD |
1636 | The default %destructor's and %printer's are experimental. More user |
1637 | feedback will help to determine whether they should become permanent | |
1638 | features. | |
1639 | ||
e4ab1254 | 1640 | See the section "Freeing Discarded Symbols" in the Bison manual for further |
12e35840 JD |
1641 | details. |
1642 | ||
402b123d | 1643 | ** %left, %right, and %nonassoc can now declare token numbers. This is required |
e4ab1254 | 1644 | by POSIX. However, see the end of section "Operator Precedence" in the Bison |
ab7f29f8 JD |
1645 | manual for a caveat concerning the treatment of literal strings. |
1646 | ||
402b123d | 1647 | ** The nonfunctional --no-parser, -n, and %no-parser options have been |
b1cc23c4 JD |
1648 | completely removed from Bison. |
1649 | ||
402b123d | 1650 | * Changes in version 2.3a, 2006-09-13: |
742e4900 | 1651 | |
402b123d | 1652 | ** Instead of %union, you can define and use your own union type |
ddc8ede1 PE |
1653 | YYSTYPE if your grammar contains at least one <type> tag. |
1654 | Your YYSTYPE need not be a macro; it can be a typedef. | |
1655 | This change is for compatibility with other Yacc implementations, | |
1656 | and is required by POSIX. | |
1657 | ||
402b123d | 1658 | ** Locations columns and lines start at 1. |
cd48d21d AD |
1659 | In accordance with the GNU Coding Standards and Emacs. |
1660 | ||
402b123d | 1661 | ** You may now declare per-type and default %destructor's and %printer's: |
ec5479ce JD |
1662 | |
1663 | For example: | |
1664 | ||
b2a0b7ca JD |
1665 | %union { char *string; } |
1666 | %token <string> STRING1 | |
1667 | %token <string> STRING2 | |
1668 | %type <string> string1 | |
1669 | %type <string> string2 | |
1670 | %union { char character; } | |
1671 | %token <character> CHR | |
1672 | %type <character> chr | |
1673 | %destructor { free ($$); } %symbol-default | |
1674 | %destructor { free ($$); printf ("%d", @$.first_line); } STRING1 string1 | |
1675 | %destructor { } <character> | |
1676 | ||
1677 | guarantees that, when the parser discards any user-defined symbol that has a | |
e4ab1254 AD |
1678 | semantic type tag other than "<character>", it passes its semantic value to |
1679 | "free". However, when the parser discards a "STRING1" or a "string1", it | |
1680 | also prints its line number to "stdout". It performs only the second | |
1681 | "%destructor" in this case, so it invokes "free" only once. | |
ec5479ce | 1682 | |
85894313 JD |
1683 | [Although we failed to mention this here in the 2.3a release, the default |
1684 | %destructor's and %printer's were experimental, and they were rewritten in | |
1685 | future versions.] | |
1686 | ||
e4ab1254 AD |
1687 | ** Except for LALR(1) parsers in C with POSIX Yacc emulation enabled (with "-y", |
1688 | "--yacc", or "%yacc"), Bison no longer generates #define statements for | |
b931235e JD |
1689 | associating token numbers with token names. Removing the #define statements |
1690 | helps to sanitize the global namespace during preprocessing, but POSIX Yacc | |
1691 | requires them. Bison still generates an enum for token names in all cases. | |
1692 | ||
402b123d | 1693 | ** Handling of traditional Yacc prologue blocks is now more consistent but |
34f98f46 | 1694 | potentially incompatible with previous releases of Bison. |
9bc0dd67 JD |
1695 | |
1696 | As before, you declare prologue blocks in your grammar file with the | |
e4ab1254 | 1697 | "%{ ... %}" syntax. To generate the pre-prologue, Bison concatenates all |
34f98f46 JD |
1698 | prologue blocks that you've declared before the first %union. To generate |
1699 | the post-prologue, Bison concatenates all prologue blocks that you've | |
ddc8ede1 | 1700 | declared after the first %union. |
9bc0dd67 | 1701 | |
34f98f46 | 1702 | Previous releases of Bison inserted the pre-prologue into both the header |
9bc0dd67 JD |
1703 | file and the code file in all cases except for LALR(1) parsers in C. In the |
1704 | latter case, Bison inserted it only into the code file. For parsers in C++, | |
1705 | the point of insertion was before any token definitions (which associate | |
1706 | token numbers with names). For parsers in C, the point of insertion was | |
1707 | after the token definitions. | |
1708 | ||
1709 | Now, Bison never inserts the pre-prologue into the header file. In the code | |
1710 | file, it always inserts it before the token definitions. | |
1711 | ||
402b123d | 1712 | ** Bison now provides a more flexible alternative to the traditional Yacc |
34f98f46 JD |
1713 | prologue blocks: %before-header, %start-header, %end-header, and |
1714 | %after-header. | |
1715 | ||
1716 | For example, the following declaration order in the grammar file reflects the | |
1717 | order in which Bison will output these code blocks. However, you are free to | |
1718 | declare these code blocks in your grammar file in whatever order is most | |
1719 | convenient for you: | |
1720 | ||
1721 | %before-header { | |
1722 | /* Bison treats this block like a pre-prologue block: it inserts it into | |
1723 | * the code file before the contents of the header file. It does *not* | |
1724 | * insert it into the header file. This is a good place to put | |
1725 | * #include's that you want at the top of your code file. A common | |
e4ab1254 | 1726 | * example is '#include "system.h"'. */ |
34f98f46 JD |
1727 | } |
1728 | %start-header { | |
1729 | /* Bison inserts this block into both the header file and the code file. | |
1730 | * In both files, the point of insertion is before any Bison-generated | |
1731 | * token, semantic type, location type, and class definitions. This is a | |
1732 | * good place to define %union dependencies, for example. */ | |
9bc0dd67 JD |
1733 | } |
1734 | %union { | |
34f98f46 JD |
1735 | /* Unlike the traditional Yacc prologue blocks, the output order for the |
1736 | * new %*-header blocks is not affected by their declaration position | |
1737 | * relative to any %union in the grammar file. */ | |
9bc0dd67 | 1738 | } |
34f98f46 JD |
1739 | %end-header { |
1740 | /* Bison inserts this block into both the header file and the code file. | |
1741 | * In both files, the point of insertion is after the Bison-generated | |
1742 | * definitions. This is a good place to declare or define public | |
1743 | * functions or data structures that depend on the Bison-generated | |
1744 | * definitions. */ | |
9bc0dd67 | 1745 | } |
34f98f46 JD |
1746 | %after-header { |
1747 | /* Bison treats this block like a post-prologue block: it inserts it into | |
1748 | * the code file after the contents of the header file. It does *not* | |
1749 | * insert it into the header file. This is a good place to declare or | |
1750 | * define internal functions or data structures that depend on the | |
1751 | * Bison-generated definitions. */ | |
1752 | } | |
1753 | ||
1754 | If you have multiple occurrences of any one of the above declarations, Bison | |
1755 | will concatenate the contents in declaration order. | |
9bc0dd67 | 1756 | |
85894313 JD |
1757 | [Although we failed to mention this here in the 2.3a release, the prologue |
1758 | alternatives were experimental, and they were rewritten in future versions.] | |
1759 | ||
e4ab1254 | 1760 | ** The option "--report=look-ahead" has been changed to "--report=lookahead". |
9e6e7ed2 PE |
1761 | The old spelling still works, but is not documented and may be removed |
1762 | in a future release. | |
742e4900 | 1763 | |
402b123d | 1764 | * Changes in version 2.3, 2006-06-05: |
4ad3ed84 | 1765 | |
e4ab1254 | 1766 | ** GLR grammars should now use "YYRECOVERING ()" instead of "YYRECOVERING", |
4ad3ed84 PE |
1767 | for compatibility with LALR(1) grammars. |
1768 | ||
402b123d | 1769 | ** It is now documented that any definition of YYSTYPE or YYLTYPE should |
4ad3ed84 PE |
1770 | be to a type name that does not contain parentheses or brackets. |
1771 | ||
402b123d | 1772 | * Changes in version 2.2, 2006-05-19: |
193d7c70 | 1773 | |
402b123d | 1774 | ** The distribution terms for all Bison-generated parsers now permit |
193d7c70 PE |
1775 | using the parsers in nonfree programs. Previously, this permission |
1776 | was granted only for Bison-generated LALR(1) parsers in C. | |
5f4236a0 | 1777 | |
402b123d | 1778 | ** %name-prefix changes the namespace name in C++ outputs. |
aa08666d | 1779 | |
402b123d | 1780 | ** The C++ parsers export their token_type. |
5f4236a0 | 1781 | |
402b123d | 1782 | ** Bison now allows multiple %union declarations, and concatenates |
d6ca7905 PE |
1783 | their contents together. |
1784 | ||
402b123d | 1785 | ** New warning: unused values |
4d7bc38c PE |
1786 | Right-hand side symbols whose values are not used are reported, |
1787 | if the symbols have destructors. For instance: | |
affac613 | 1788 | |
8f3596a6 | 1789 | exp: exp "?" exp ":" exp { $1 ? $1 : $3; } |
e9690142 JD |
1790 | | exp "+" exp |
1791 | ; | |
affac613 | 1792 | |
8f3596a6 AD |
1793 | will trigger a warning about $$ and $5 in the first rule, and $3 in |
1794 | the second ($1 is copied to $$ by the default rule). This example | |
4e26c69e | 1795 | most likely contains three errors, and could be rewritten as: |
affac613 | 1796 | |
4e26c69e | 1797 | exp: exp "?" exp ":" exp |
e9690142 JD |
1798 | { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); } |
1799 | | exp "+" exp | |
1800 | { $$ = $1 ? $1 : $3; if ($1) free ($3); } | |
1801 | ; | |
affac613 | 1802 | |
4e26c69e PE |
1803 | However, if the original actions were really intended, memory leaks |
1804 | and all, the warnings can be suppressed by letting Bison believe the | |
1805 | values are used, e.g.: | |
721be13c | 1806 | |
8f3596a6 | 1807 | exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); } |
e9690142 JD |
1808 | | exp "+" exp { $$ = $1; (void) $3; } |
1809 | ; | |
721be13c | 1810 | |
84866159 AD |
1811 | If there are mid-rule actions, the warning is issued if no action |
1812 | uses it. The following triggers no warning: $1 and $3 are used. | |
1813 | ||
1814 | exp: exp { push ($1); } '+' exp { push ($3); sum (); }; | |
1815 | ||
721be13c PE |
1816 | The warning is intended to help catching lost values and memory leaks. |
1817 | If a value is ignored, its associated memory typically is not reclaimed. | |
affac613 | 1818 | |
402b123d | 1819 | ** %destructor vs. YYABORT, YYACCEPT, and YYERROR. |
9d9b8b70 PE |
1820 | Destructors are now called when user code invokes YYABORT, YYACCEPT, |
1821 | and YYERROR, for all objects on the stack, other than objects | |
1822 | corresponding to the right-hand side of the current rule. | |
a85284cf | 1823 | |
402b123d | 1824 | ** %expect, %expect-rr |
035aa4a0 PE |
1825 | Incorrect numbers of expected conflicts are now actual errors, |
1826 | instead of warnings. | |
1827 | ||
402b123d | 1828 | ** GLR, YACC parsers. |
4e26c69e PE |
1829 | The %parse-params are available in the destructors (and the |
1830 | experimental printers) as per the documentation. | |
4b367315 | 1831 | |
e4ab1254 | 1832 | ** Bison now warns if it finds a stray "$" or "@" in an action. |
ad6a9b97 | 1833 | |
402b123d | 1834 | ** %require "VERSION" |
4e26c69e PE |
1835 | This specifies that the grammar file depends on features implemented |
1836 | in Bison version VERSION or higher. | |
b50d2359 | 1837 | |
402b123d | 1838 | ** lalr1.cc: The token and value types are now class members. |
e14d0ab6 AD |
1839 | The tokens were defined as free form enums and cpp macros. YYSTYPE |
1840 | was defined as a free form union. They are now class members: | |
e4ab1254 AD |
1841 | tokens are enumerations of the "yy::parser::token" struct, and the |
1842 | semantic values have the "yy::parser::semantic_type" type. | |
fb9712a9 AD |
1843 | |
1844 | If you do not want or can update to this scheme, the directive | |
e4ab1254 | 1845 | '%define "global_tokens_and_yystype" "1"' triggers the global |
b50d2359 AD |
1846 | definition of tokens and YYSTYPE. This change is suitable both |
1847 | for previous releases of Bison, and this one. | |
fb9712a9 | 1848 | |
b50d2359 | 1849 | If you wish to update, then make sure older version of Bison will |
e4ab1254 | 1850 | fail using '%require "2.2"'. |
fb9712a9 | 1851 | |
402b123d | 1852 | ** DJGPP support added. |
193d7c70 | 1853 | \f |
402b123d | 1854 | * Changes in version 2.1, 2005-09-16: |
1ce59070 | 1855 | |
402b123d | 1856 | ** The C++ lalr1.cc skeleton supports %lex-param. |
e14d0ab6 | 1857 | |
402b123d | 1858 | ** Bison-generated parsers now support the translation of diagnostics like |
baf785db PE |
1859 | "syntax error" into languages other than English. The default |
1860 | language is still English. For details, please see the new | |
0410a6e0 PE |
1861 | Internationalization section of the Bison manual. Software |
1862 | distributors should also see the new PACKAGING file. Thanks to | |
1863 | Bruno Haible for this new feature. | |
1ce59070 | 1864 | |
402b123d | 1865 | ** Wording in the Bison-generated parsers has been changed slightly to |
1a059451 PE |
1866 | simplify translation. In particular, the message "memory exhausted" |
1867 | has replaced "parser stack overflow", as the old message was not | |
1868 | always accurate for modern Bison-generated parsers. | |
1869 | ||
402b123d | 1870 | ** Destructors are now called when the parser aborts, for all symbols left |
258b75ca PE |
1871 | behind on the stack. Also, the start symbol is now destroyed after a |
1872 | successful parse. In both cases, the behavior was formerly inconsistent. | |
1873 | ||
402b123d | 1874 | ** When generating verbose diagnostics, Bison-generated parsers no longer |
72f000b0 PE |
1875 | quote the literal strings associated with tokens. For example, for |
1876 | a syntax error associated with '%token NUM "number"' they might | |
1877 | print 'syntax error, unexpected number' instead of 'syntax error, | |
1878 | unexpected "number"'. | |
193d7c70 | 1879 | \f |
402b123d | 1880 | * Changes in version 2.0, 2004-12-25: |
efeed023 | 1881 | |
402b123d | 1882 | ** Possibly-incompatible changes |
d7e14fc0 | 1883 | |
82de6b0d PE |
1884 | - Bison-generated parsers no longer default to using the alloca function |
1885 | (when available) to extend the parser stack, due to widespread | |
1886 | problems in unchecked stack-overflow detection. You can "#define | |
1887 | YYSTACK_USE_ALLOCA 1" to require the use of alloca, but please read | |
1888 | the manual to determine safe values for YYMAXDEPTH in that case. | |
8dd162d3 | 1889 | |
82de6b0d PE |
1890 | - Error token location. |
1891 | During error recovery, the location of the syntax error is updated | |
1892 | to cover the whole sequence covered by the error token: it includes | |
1893 | the shifted symbols thrown away during the first part of the error | |
1894 | recovery, and the lookahead rejected during the second part. | |
18d192f0 | 1895 | |
82de6b0d PE |
1896 | - Semicolon changes: |
1897 | . Stray semicolons are no longer allowed at the start of a grammar. | |
1898 | . Semicolons are now required after in-grammar declarations. | |
e342c3be | 1899 | |
82de6b0d PE |
1900 | - Unescaped newlines are no longer allowed in character constants or |
1901 | string literals. They were never portable, and GCC 3.4.0 has | |
1902 | dropped support for them. Better diagnostics are now generated if | |
1903 | forget a closing quote. | |
8dd162d3 | 1904 | |
82de6b0d | 1905 | - NUL bytes are no longer allowed in Bison string literals, unfortunately. |
f74b6f91 | 1906 | |
402b123d | 1907 | ** New features |
1452af69 | 1908 | |
82de6b0d | 1909 | - GLR grammars now support locations. |
4febdd96 | 1910 | |
82de6b0d PE |
1911 | - New directive: %initial-action. |
1912 | This directive allows the user to run arbitrary code (including | |
1913 | initializing @$) from yyparse before parsing starts. | |
1452af69 | 1914 | |
82de6b0d PE |
1915 | - A new directive "%expect-rr N" specifies the expected number of |
1916 | reduce/reduce conflicts in GLR parsers. | |
1452af69 | 1917 | |
e4ab1254 | 1918 | - %token numbers can now be hexadecimal integers, e.g., "%token FOO 0x12d". |
82de6b0d | 1919 | This is a GNU extension. |
4febdd96 | 1920 | |
e4ab1254 | 1921 | - The option "--report=lookahead" was changed to "--report=look-ahead". |
9e6e7ed2 | 1922 | [However, this was changed back after 2.3.] |
1452af69 | 1923 | |
82de6b0d | 1924 | - Experimental %destructor support has been added to lalr1.cc. |
1452af69 | 1925 | |
82de6b0d PE |
1926 | - New configure option --disable-yacc, to disable installation of the |
1927 | yacc command and -ly library introduced in 1.875 for POSIX conformance. | |
6040d338 | 1928 | |
402b123d | 1929 | ** Bug fixes |
d5a3fe37 | 1930 | |
82de6b0d PE |
1931 | - For now, %expect-count violations are now just warnings, not errors. |
1932 | This is for compatibility with Bison 1.75 and earlier (when there are | |
1933 | reduce/reduce conflicts) and with Bison 1.30 and earlier (when there | |
1934 | are too many or too few shift/reduce conflicts). However, in future | |
1935 | versions of Bison we plan to improve the %expect machinery so that | |
1936 | these violations will become errors again. | |
3473d0f8 | 1937 | |
82de6b0d PE |
1938 | - Within Bison itself, numbers (e.g., goto numbers) are no longer |
1939 | arbitrarily limited to 16-bit counts. | |
d600ee67 | 1940 | |
82de6b0d | 1941 | - Semicolons are now allowed before "|" in grammar rules, as POSIX requires. |
d600ee67 | 1942 | \f |
402b123d | 1943 | * Changes in version 1.875, 2003-01-01: |
963fcc17 | 1944 | |
402b123d | 1945 | ** The documentation license has been upgraded to version 1.2 |
dc546b0f | 1946 | of the GNU Free Documentation License. |
75eb3bc4 | 1947 | |
402b123d | 1948 | ** syntax error processing |
75eb3bc4 | 1949 | |
dc546b0f PE |
1950 | - In Yacc-style parsers YYLLOC_DEFAULT is now used to compute error |
1951 | locations too. This fixes bugs in error-location computation. | |
75eb3bc4 | 1952 | |
dc546b0f PE |
1953 | - %destructor |
1954 | It is now possible to reclaim the memory associated to symbols | |
1955 | discarded during error recovery. This feature is still experimental. | |
20daca06 | 1956 | |
dc546b0f PE |
1957 | - %error-verbose |
1958 | This new directive is preferred over YYERROR_VERBOSE. | |
74724a70 | 1959 | |
dc546b0f PE |
1960 | - #defining yyerror to steal internal variables is discouraged. |
1961 | It is not guaranteed to work forever. | |
d1de5372 | 1962 | |
402b123d | 1963 | ** POSIX conformance |
d1de5372 | 1964 | |
dc546b0f PE |
1965 | - Semicolons are once again optional at the end of grammar rules. |
1966 | This reverts to the behavior of Bison 1.33 and earlier, and improves | |
1967 | compatibility with Yacc. | |
74724a70 | 1968 | |
e4ab1254 AD |
1969 | - "parse error" -> "syntax error" |
1970 | Bison now uniformly uses the term "syntax error"; formerly, the code | |
1971 | and manual sometimes used the term "parse error" instead. POSIX | |
1972 | requires "syntax error" in diagnostics, and it was thought better to | |
dc546b0f | 1973 | be consistent. |
74724a70 | 1974 | |
dc546b0f PE |
1975 | - The documentation now emphasizes that yylex and yyerror must be |
1976 | declared before use. C99 requires this. | |
d1de5372 | 1977 | |
dc546b0f PE |
1978 | - Bison now parses C99 lexical constructs like UCNs and |
1979 | backslash-newline within C escape sequences, as POSIX 1003.1-2001 requires. | |
d1de5372 | 1980 | |
dc546b0f PE |
1981 | - File names are properly escaped in C output. E.g., foo\bar.y is |
1982 | output as "foo\\bar.y". | |
6780ca7a | 1983 | |
dc546b0f | 1984 | - Yacc command and library now available |
e4ab1254 | 1985 | The Bison distribution now installs a "yacc" command, as POSIX requires. |
dc546b0f PE |
1986 | Also, Bison now installs a small library liby.a containing |
1987 | implementations of Yacc-compatible yyerror and main functions. | |
1988 | This library is normally not useful, but POSIX requires it. | |
6e649e65 | 1989 | |
dc546b0f | 1990 | - Type clashes now generate warnings, not errors. |
6e649e65 | 1991 | |
dc546b0f PE |
1992 | - If the user does not define YYSTYPE as a macro, Bison now declares it |
1993 | using typedef instead of defining it as a macro. | |
1994 | For consistency, YYLTYPE is also declared instead of defined. | |
9501dc6e | 1995 | |
402b123d | 1996 | ** Other compatibility issues |
886a425c | 1997 | |
e4ab1254 AD |
1998 | - %union directives can now have a tag before the "{", e.g., the |
1999 | directive "%union foo {...}" now generates the C code | |
2000 | "typedef union foo { ... } YYSTYPE;"; this is for Yacc compatibility. | |
2001 | The default union tag is "YYSTYPE", for compatibility with Solaris 9 Yacc. | |
2002 | For consistency, YYLTYPE's struct tag is now "YYLTYPE" not "yyltype". | |
dc546b0f | 2003 | This is for compatibility with both Yacc and Bison 1.35. |
72f889cc | 2004 | |
e4ab1254 | 2005 | - ";" is output before the terminating "}" of an action, for |
dc546b0f | 2006 | compatibility with Bison 1.35. |
886a425c | 2007 | |
dc546b0f | 2008 | - Bison now uses a Yacc-style format for conflict reports, e.g., |
e4ab1254 | 2009 | "conflicts: 2 shift/reduce, 1 reduce/reduce". |
437c2d80 | 2010 | |
e4ab1254 | 2011 | - "yystype" and "yyltype" are now obsolescent macros instead of being |
dc546b0f PE |
2012 | typedefs or tags; they are no longer documented and are planned to be |
2013 | withdrawn in a future release. | |
2a8d363a | 2014 | |
402b123d | 2015 | ** GLR parser notes |
2a8d363a | 2016 | |
dc546b0f PE |
2017 | - GLR and inline |
2018 | Users of Bison have to decide how they handle the portability of the | |
e4ab1254 | 2019 | C keyword "inline". |
959e5f51 | 2020 | |
e4ab1254 AD |
2021 | - "parsing stack overflow..." -> "parser stack overflow" |
2022 | GLR parsers now report "parser stack overflow" as per the Bison manual. | |
900c5db5 | 2023 | |
18ad57b3 AD |
2024 | ** %parse-param and %lex-param |
2025 | The macros YYPARSE_PARAM and YYLEX_PARAM provide a means to pass | |
2026 | additional context to yyparse and yylex. They suffer from several | |
2027 | shortcomings: | |
2028 | ||
2029 | - a single argument only can be added, | |
2030 | - their types are weak (void *), | |
242cc08e | 2031 | - this context is not passed to ancillary functions such as yyerror, |
18ad57b3 AD |
2032 | - only yacc.c parsers support them. |
2033 | ||
2034 | The new %parse-param/%lex-param directives provide a more precise control. | |
2035 | For instance: | |
2036 | ||
2037 | %parse-param {int *nastiness} | |
2038 | %lex-param {int *nastiness} | |
2039 | %parse-param {int *randomness} | |
2040 | ||
2041 | results in the following signatures: | |
2042 | ||
2043 | int yylex (int *nastiness); | |
2044 | int yyparse (int *nastiness, int *randomness); | |
2045 | ||
2046 | or, if both %pure-parser and %locations are used: | |
2047 | ||
2048 | int yylex (YYSTYPE *lvalp, YYLTYPE *llocp, int *nastiness); | |
2049 | int yyparse (int *nastiness, int *randomness); | |
2050 | ||
402b123d | 2051 | ** Bison now warns if it detects conflicting outputs to the same file, |
e4ab1254 | 2052 | e.g., it generates a warning for "bison -d -o foo.h foo.y" since |
dc546b0f | 2053 | that command outputs both code and header to foo.h. |
6e40b4eb | 2054 | |
402b123d | 2055 | ** #line in output files |
dc546b0f | 2056 | - --no-line works properly. |
6e40b4eb | 2057 | |
402b123d | 2058 | ** Bison can no longer be built by a K&R C compiler; it requires C89 or |
6e40b4eb AD |
2059 | later to be built. This change originally took place a few versions |
2060 | ago, but nobody noticed until we recently asked someone to try | |
2061 | building Bison with a K&R C compiler. | |
d600ee67 | 2062 | \f |
402b123d | 2063 | * Changes in version 1.75, 2002-10-14: |
7933f2b5 | 2064 | |
402b123d | 2065 | ** Bison should now work on 64-bit hosts. |
7933f2b5 | 2066 | |
402b123d | 2067 | ** Indonesian translation thanks to Tedi Heriyanto. |
7933f2b5 | 2068 | |
402b123d | 2069 | ** GLR parsers |
f50adbbd AD |
2070 | Fix spurious parse errors. |
2071 | ||
402b123d | 2072 | ** Pure parsers |
f50adbbd AD |
2073 | Some people redefine yyerror to steal yyparse' private variables. |
2074 | Reenable this trick until an official feature replaces it. | |
2075 | ||
402b123d | 2076 | ** Type Clashes |
d90c934c AD |
2077 | In agreement with POSIX and with other Yaccs, leaving a default |
2078 | action is valid when $$ is untyped, and $1 typed: | |
2079 | ||
e9690142 | 2080 | untyped: ... typed; |
d90c934c AD |
2081 | |
2082 | but the converse remains an error: | |
2083 | ||
e9690142 | 2084 | typed: ... untyped; |
d90c934c | 2085 | |
402b123d | 2086 | ** Values of mid-rule actions |
d90c934c AD |
2087 | The following code: |
2088 | ||
e9690142 | 2089 | foo: { ... } { $$ = $1; } ... |
d90c934c AD |
2090 | |
2091 | was incorrectly rejected: $1 is defined in the second mid-rule | |
2092 | action, and is equal to the $$ of the first mid-rule action. | |
d600ee67 | 2093 | \f |
402b123d | 2094 | * Changes in version 1.50, 2002-10-04: |
adc8c848 | 2095 | |
402b123d | 2096 | ** GLR parsing |
676385e2 PH |
2097 | The declaration |
2098 | %glr-parser | |
2099 | causes Bison to produce a Generalized LR (GLR) parser, capable of handling | |
2100 | almost any context-free grammar, ambiguous or not. The new declarations | |
e8832397 | 2101 | %dprec and %merge on grammar rules allow parse-time resolution of |
676385e2 PH |
2102 | ambiguities. Contributed by Paul Hilfinger. |
2103 | ||
7933f2b5 | 2104 | Unfortunately Bison 1.50 does not work properly on 64-bit hosts |
420f93c8 PE |
2105 | like the Alpha, so please stick to 32-bit hosts for now. |
2106 | ||
402b123d | 2107 | ** Output Directory |
8c165d89 | 2108 | When not in Yacc compatibility mode, when the output file was not |
e4ab1254 AD |
2109 | specified, running "bison foo/bar.y" created "foo/bar.c". It |
2110 | now creates "bar.c". | |
8c165d89 | 2111 | |
402b123d | 2112 | ** Undefined token |
007a50a4 | 2113 | The undefined token was systematically mapped to 2 which prevented |
e88dbdbf | 2114 | the use of 2 by the user. This is no longer the case. |
007a50a4 | 2115 | |
402b123d | 2116 | ** Unknown token numbers |
e88dbdbf | 2117 | If yylex returned an out of range value, yyparse could die. This is |
007a50a4 AD |
2118 | no longer the case. |
2119 | ||
402b123d | 2120 | ** Error token |
e88dbdbf | 2121 | According to POSIX, the error token must be 256. |
23c5a174 AD |
2122 | Bison extends this requirement by making it a preference: *if* the |
2123 | user specified that one of her tokens is numbered 256, then error | |
2124 | will be mapped onto another number. | |
2125 | ||
402b123d | 2126 | ** Verbose error messages |
e4ab1254 | 2127 | They no longer report "..., expecting error or..." for states where |
217598da AD |
2128 | error recovery is possible. |
2129 | ||
402b123d | 2130 | ** End token |
e4ab1254 | 2131 | Defaults to "$end" instead of "$". |
217598da | 2132 | |
402b123d | 2133 | ** Error recovery now conforms to documentation and to POSIX |
68cd8af3 PE |
2134 | When a Bison-generated parser encounters a syntax error, it now pops |
2135 | the stack until it finds a state that allows shifting the error | |
2136 | token. Formerly, it popped the stack until it found a state that | |
2137 | allowed some non-error action other than a default reduction on the | |
2138 | error token. The new behavior has long been the documented behavior, | |
2139 | and has long been required by POSIX. For more details, please see | |
337116ba PE |
2140 | Paul Eggert, "Reductions during Bison error handling" (2002-05-20) |
2141 | <http://lists.gnu.org/archive/html/bug-bison/2002-05/msg00038.html>. | |
68cd8af3 | 2142 | |
402b123d | 2143 | ** Traces |
5504898e AD |
2144 | Popped tokens and nonterminals are now reported. |
2145 | ||
402b123d | 2146 | ** Larger grammars |
a861a339 PE |
2147 | Larger grammars are now supported (larger token numbers, larger grammar |
2148 | size (= sum of the LHS and RHS lengths), larger LALR tables). | |
2149 | Formerly, many of these numbers ran afoul of 16-bit limits; | |
2150 | now these limits are 32 bits on most hosts. | |
355e7c1c | 2151 | |
402b123d | 2152 | ** Explicit initial rule |
643a5994 AD |
2153 | Bison used to play hacks with the initial rule, which the user does |
2154 | not write. It is now explicit, and visible in the reports and | |
2155 | graphs as rule 0. | |
23c5a174 | 2156 | |
402b123d | 2157 | ** Useless rules |
643a5994 | 2158 | Before, Bison reported the useless rules, but, although not used, |
77714df2 | 2159 | included them in the parsers. They are now actually removed. |
23c5a174 | 2160 | |
402b123d | 2161 | ** Useless rules, useless nonterminals |
6b98e4b5 AD |
2162 | They are now reported, as a warning, with their locations. |
2163 | ||
402b123d | 2164 | ** Rules never reduced |
e8832397 AD |
2165 | Rules that can never be reduced because of conflicts are now |
2166 | reported. | |
2167 | ||
e4ab1254 | 2168 | ** Incorrect "Token not used" |
11652ab3 AD |
2169 | On a grammar such as |
2170 | ||
e29f0771 AD |
2171 | %token useless useful |
2172 | %% | |
2173 | exp: '0' %prec useful; | |
11652ab3 AD |
2174 | |
2175 | where a token was used to set the precedence of the last rule, | |
e4ab1254 | 2176 | bison reported both "useful" and "useless" as useless tokens. |
11652ab3 | 2177 | |
402b123d | 2178 | ** Revert the C++ namespace changes introduced in 1.31 |
77714df2 | 2179 | as they caused too many portability hassles. |
0179dd65 | 2180 | |
402b123d | 2181 | ** Default locations |
b2d52318 AD |
2182 | By an accident of design, the default computation of @$ was |
2183 | performed after another default computation was performed: @$ = @1. | |
2184 | The latter is now removed: YYLLOC_DEFAULT is fully responsible of | |
2185 | the computation of @$. | |
adc8c848 | 2186 | |
402b123d | 2187 | ** Token end-of-file |
b7c49edf AD |
2188 | The token end of file may be specified by the user, in which case, |
2189 | the user symbol is used in the reports, the graphs, and the verbose | |
e4ab1254 | 2190 | error messages instead of "$end", which remains being the default. |
b7c49edf | 2191 | For instance |
e29f0771 | 2192 | %token MYEOF 0 |
b7c49edf | 2193 | or |
e29f0771 | 2194 | %token MYEOF 0 "end of file" |
fdbcd8e2 | 2195 | |
402b123d | 2196 | ** Semantic parser |
fdbcd8e2 AD |
2197 | This old option, which has been broken for ages, is removed. |
2198 | ||
402b123d | 2199 | ** New translations |
a861a339 | 2200 | Brazilian Portuguese, thanks to Alexandre Folle de Menezes. |
84614e13 AD |
2201 | Croatian, thanks to Denis Lackovic. |
2202 | ||
402b123d | 2203 | ** Incorrect token definitions |
e4ab1254 AD |
2204 | When given |
2205 | %token 'a' "A" | |
2206 | bison used to output | |
2207 | #define 'a' 65 | |
b87f8b21 | 2208 | |
402b123d | 2209 | ** Token definitions as enums |
77714df2 AD |
2210 | Tokens are output both as the traditional #define's, and, provided |
2211 | the compiler supports ANSI C or is a C++ compiler, as enums. | |
e88dbdbf | 2212 | This lets debuggers display names instead of integers. |
77714df2 | 2213 | |
402b123d | 2214 | ** Reports |
ec3bc396 AD |
2215 | In addition to --verbose, bison supports --report=THINGS, which |
2216 | produces additional information: | |
b408954b AD |
2217 | - itemset |
2218 | complete the core item sets with their closure | |
e4ab1254 | 2219 | - lookahead [changed to "look-ahead" in 1.875e through 2.3, but changed back] |
9e6e7ed2 | 2220 | explicitly associate lookahead tokens to items |
b408954b AD |
2221 | - solved |
2222 | describe shift/reduce conflicts solving. | |
2223 | Bison used to systematically output this information on top of | |
2224 | the report. Solved conflicts are now attached to their states. | |
ec3bc396 | 2225 | |
402b123d | 2226 | ** Type clashes |
9af3fbce AD |
2227 | Previous versions don't complain when there is a type clash on |
2228 | the default action if the rule has a mid-rule action, such as in: | |
2229 | ||
e29f0771 AD |
2230 | %type <foo> bar |
2231 | %% | |
2232 | bar: '0' {} '0'; | |
9af3fbce AD |
2233 | |
2234 | This is fixed. | |
a861a339 | 2235 | |
402b123d | 2236 | ** GNU M4 is now required when using Bison. |
f987e9d2 | 2237 | \f |
402b123d | 2238 | * Changes in version 1.35, 2002-03-25: |
76551463 | 2239 | |
402b123d | 2240 | ** C Skeleton |
76551463 AD |
2241 | Some projects use Bison's C parser with C++ compilers, and define |
2242 | YYSTYPE as a class. The recent adjustment of C parsers for data | |
2243 | alignment and 64 bit architectures made this impossible. | |
2244 | ||
2245 | Because for the time being no real solution for C++ parser | |
2246 | generation exists, kludges were implemented in the parser to | |
2247 | maintain this use. In the future, when Bison has C++ parsers, this | |
2248 | kludge will be disabled. | |
2249 | ||
2250 | This kludge also addresses some C++ problems when the stack was | |
2251 | extended. | |
76551463 | 2252 | \f |
402b123d | 2253 | * Changes in version 1.34, 2002-03-12: |
76551463 | 2254 | |
402b123d | 2255 | ** File name clashes are detected |
76551463 | 2256 | $ bison foo.y -d -o foo.x |
e4ab1254 | 2257 | fatal error: header and parser would both be named "foo.x" |
76551463 | 2258 | |
e4ab1254 | 2259 | ** A missing ";" at the end of a rule triggers a warning |
76551463 AD |
2260 | In accordance with POSIX, and in agreement with other |
2261 | Yacc implementations, Bison will mandate this semicolon in the near | |
2262 | future. This eases the implementation of a Bison parser of Bison | |
2263 | grammars by making this grammar LALR(1) instead of LR(2). To | |
2264 | facilitate the transition, this release introduces a warning. | |
2265 | ||
402b123d | 2266 | ** Revert the C++ namespace changes introduced in 1.31, as they caused too |
76551463 AD |
2267 | many portability hassles. |
2268 | ||
402b123d | 2269 | ** DJGPP support added. |
76551463 | 2270 | |
402b123d | 2271 | ** Fix test suite portability problems. |
76551463 | 2272 | \f |
402b123d | 2273 | * Changes in version 1.33, 2002-02-07: |
76551463 | 2274 | |
402b123d | 2275 | ** Fix C++ issues |
76551463 AD |
2276 | Groff could not be compiled for the definition of size_t was lacking |
2277 | under some conditions. | |
2278 | ||
402b123d | 2279 | ** Catch invalid @n |
76551463 AD |
2280 | As is done with $n. |
2281 | \f | |
402b123d | 2282 | * Changes in version 1.32, 2002-01-23: |
76551463 | 2283 | |
402b123d | 2284 | ** Fix Yacc output file names |
76551463 | 2285 | |
402b123d | 2286 | ** Portability fixes |
76551463 | 2287 | |
402b123d | 2288 | ** Italian, Dutch translations |
76551463 | 2289 | \f |
402b123d | 2290 | * Changes in version 1.31, 2002-01-14: |
52d1aeee | 2291 | |
402b123d | 2292 | ** Many Bug Fixes |
52d1aeee | 2293 | |
402b123d | 2294 | ** GNU Gettext and %expect |
52d1aeee MA |
2295 | GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that |
2296 | Bison dies on incorrect %expectations, we fear there will be | |
2297 | too many bug reports for Gettext, so _for the time being_, %expect | |
e4ab1254 | 2298 | does not trigger an error when the input file is named "plural.y". |
52d1aeee | 2299 | |
402b123d | 2300 | ** Use of alloca in parsers |
52d1aeee MA |
2301 | If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use |
2302 | malloc exclusively. Since 1.29, but was not NEWS'ed. | |
2303 | ||
2304 | alloca is used only when compiled with GCC, to avoid portability | |
2305 | problems as on AIX. | |
2306 | ||
402b123d | 2307 | ** yyparse now returns 2 if memory is exhausted; formerly it dumped core. |
b47dbebe | 2308 | |
402b123d | 2309 | ** When the generated parser lacks debugging code, YYDEBUG is now 0 |
52d1aeee MA |
2310 | (as POSIX requires) instead of being undefined. |
2311 | ||
402b123d | 2312 | ** User Actions |
52d1aeee MA |
2313 | Bison has always permitted actions such as { $$ = $1 }: it adds the |
2314 | ending semicolon. Now if in Yacc compatibility mode, the semicolon | |
2315 | is no longer output: one has to write { $$ = $1; }. | |
2316 | ||
402b123d | 2317 | ** Better C++ compliance |
52d1aeee | 2318 | The output parsers try to respect C++ namespaces. |
76551463 | 2319 | [This turned out to be a failed experiment, and it was reverted later.] |
52d1aeee | 2320 | |
402b123d | 2321 | ** Reduced Grammars |
52d1aeee MA |
2322 | Fixed bugs when reporting useless nonterminals. |
2323 | ||
402b123d | 2324 | ** 64 bit hosts |
52d1aeee MA |
2325 | The parsers work properly on 64 bit hosts. |
2326 | ||
402b123d | 2327 | ** Error messages |
52d1aeee MA |
2328 | Some calls to strerror resulted in scrambled or missing error messages. |
2329 | ||
402b123d | 2330 | ** %expect |
52d1aeee MA |
2331 | When the number of shift/reduce conflicts is correct, don't issue |
2332 | any warning. | |
2333 | ||
402b123d | 2334 | ** The verbose report includes the rule line numbers. |
52d1aeee | 2335 | |
402b123d | 2336 | ** Rule line numbers are fixed in traces. |
52d1aeee | 2337 | |
402b123d | 2338 | ** Swedish translation |
52d1aeee | 2339 | |
402b123d | 2340 | ** Parse errors |
52d1aeee MA |
2341 | Verbose parse error messages from the parsers are better looking. |
2342 | Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'('' | |
2343 | Now: parse error: unexpected '/', expecting "number" or '-' or '(' | |
2344 | ||
402b123d | 2345 | ** Fixed parser memory leaks. |
52d1aeee MA |
2346 | When the generated parser was using malloc to extend its stacks, the |
2347 | previous allocations were not freed. | |
2348 | ||
402b123d | 2349 | ** Fixed verbose output file. |
52d1aeee MA |
2350 | Some newlines were missing. |
2351 | Some conflicts in state descriptions were missing. | |
2352 | ||
402b123d | 2353 | ** Fixed conflict report. |
52d1aeee MA |
2354 | Option -v was needed to get the result. |
2355 | ||
402b123d | 2356 | ** %expect |
52d1aeee MA |
2357 | Was not used. |
2358 | Mismatches are errors, not warnings. | |
2359 | ||
402b123d | 2360 | ** Fixed incorrect processing of some invalid input. |
52d1aeee | 2361 | |
402b123d | 2362 | ** Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H. |
52d1aeee | 2363 | |
402b123d | 2364 | ** Fixed some typos in the documentation. |
52d1aeee | 2365 | |
402b123d | 2366 | ** %token MY_EOF 0 is supported. |
52d1aeee MA |
2367 | Before, MY_EOF was silently renumbered as 257. |
2368 | ||
402b123d | 2369 | ** doc/refcard.tex is updated. |
52d1aeee | 2370 | |
402b123d | 2371 | ** %output, %file-prefix, %name-prefix. |
52d1aeee MA |
2372 | New. |
2373 | ||
402b123d | 2374 | ** --output |
e4ab1254 | 2375 | New, aliasing "--output-file". |
52d1aeee | 2376 | \f |
402b123d | 2377 | * Changes in version 1.30, 2001-10-26: |
342b8b6e | 2378 | |
e4ab1254 AD |
2379 | ** "--defines" and "--graph" have now an optional argument which is the |
2380 | output file name. "-d" and "-g" do not change; they do not take any | |
342b8b6e AD |
2381 | argument. |
2382 | ||
e4ab1254 | 2383 | ** "%source_extension" and "%header_extension" are removed, failed |
342b8b6e AD |
2384 | experiment. |
2385 | ||
402b123d | 2386 | ** Portability fixes. |
f987e9d2 | 2387 | \f |
402b123d | 2388 | * Changes in version 1.29, 2001-09-07: |
342b8b6e | 2389 | |
402b123d | 2390 | ** The output file does not define const, as this caused problems when used |
342b8b6e AD |
2391 | with common autoconfiguration schemes. If you still use ancient compilers |
2392 | that lack const, compile with the equivalent of the C compiler option | |
e4ab1254 | 2393 | "-Dconst=". Autoconf's AC_C_CONST macro provides one way to do this. |
342b8b6e | 2394 | |
e4ab1254 | 2395 | ** Added "-g" and "--graph". |
f87a2205 | 2396 | |
402b123d | 2397 | ** The Bison manual is now distributed under the terms of the GNU FDL. |
f2b5126e | 2398 | |
402b123d | 2399 | ** The input and the output files has automatically a similar extension. |
234a3be3 | 2400 | |
402b123d | 2401 | ** Russian translation added. |
f87a2205 | 2402 | |
402b123d | 2403 | ** NLS support updated; should hopefully be less troublesome. |
f87a2205 | 2404 | |
402b123d | 2405 | ** Added the old Bison reference card. |
c33638bb | 2406 | |
e4ab1254 | 2407 | ** Added "--locations" and "%locations". |
6deb4447 | 2408 | |
e4ab1254 | 2409 | ** Added "-S" and "--skeleton". |
cd5bd6ac | 2410 | |
e4ab1254 | 2411 | ** "%raw", "-r", "--raw" is disabled. |
62ab6972 | 2412 | |
402b123d | 2413 | ** Special characters are escaped when output. This solves the problems |
cd5bd6ac AD |
2414 | of the #line lines with path names including backslashes. |
2415 | ||
402b123d | 2416 | ** New directives. |
e4ab1254 AD |
2417 | "%yacc", "%fixed_output_files", "%defines", "%no_parser", "%verbose", |
2418 | "%debug", "%source_extension" and "%header_extension". | |
f987e9d2 | 2419 | |
402b123d | 2420 | ** @$ |
f987e9d2 | 2421 | Automatic location tracking. |
f87a2205 | 2422 | \f |
402b123d | 2423 | * Changes in version 1.28, 1999-07-06: |
d2e00347 | 2424 | |
402b123d | 2425 | ** Should compile better now with K&R compilers. |
d2e00347 | 2426 | |
402b123d | 2427 | ** Added NLS. |
d2e00347 | 2428 | |
402b123d | 2429 | ** Fixed a problem with escaping the double quote character. |
d2e00347 | 2430 | |
402b123d | 2431 | ** There is now a FAQ. |
d2e00347 | 2432 | \f |
402b123d | 2433 | * Changes in version 1.27: |
5c31c3c2 | 2434 | |
402b123d | 2435 | ** The make rule which prevented bison.simple from being created on |
5c31c3c2 JT |
2436 | some systems has been fixed. |
2437 | \f | |
402b123d | 2438 | * Changes in version 1.26: |
4be07551 | 2439 | |
7e508a2b | 2440 | ** Bison now uses Automake. |
4be07551 | 2441 | |
402b123d | 2442 | ** New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>. |
4be07551 | 2443 | |
402b123d | 2444 | ** Token numbers now start at 257 as previously documented, not 258. |
4be07551 | 2445 | |
402b123d | 2446 | ** Bison honors the TMPDIR environment variable. |
4be07551 | 2447 | |
402b123d | 2448 | ** A couple of buffer overruns have been fixed. |
f51dbca1 | 2449 | |
402b123d | 2450 | ** Problems when closing files should now be reported. |
f51dbca1 | 2451 | |
402b123d | 2452 | ** Generated parsers should now work even on operating systems which do |
f51dbca1 | 2453 | not provide alloca(). |
4be07551 | 2454 | \f |
402b123d | 2455 | * Changes in version 1.25, 1995-10-16: |
df8878c5 | 2456 | |
402b123d | 2457 | ** Errors in the input grammar are not fatal; Bison keeps reading |
df8878c5 | 2458 | the grammar file, and reports all the errors found in it. |
8c44d3ec | 2459 | |
402b123d | 2460 | ** Tokens can now be specified as multiple-character strings: for |
df8878c5 | 2461 | example, you could use "<=" for a token which looks like <=, instead |
7e508a2b | 2462 | of choosing a name like LESSEQ. |
df8878c5 | 2463 | |
402b123d | 2464 | ** The %token_table declaration says to write a table of tokens (names |
df8878c5 RS |
2465 | and numbers) into the parser file. The yylex function can use this |
2466 | table to recognize multiple-character string tokens, or for other | |
2467 | purposes. | |
2468 | ||
402b123d | 2469 | ** The %no_lines declaration says not to generate any #line preprocessor |
df8878c5 RS |
2470 | directives in the parser file. |
2471 | ||
402b123d | 2472 | ** The %raw declaration says to use internal Bison token numbers, not |
df8878c5 RS |
2473 | Yacc-compatible token numbers, when token names are defined as macros. |
2474 | ||
402b123d | 2475 | ** The --no-parser option produces the parser tables without including |
df8878c5 RS |
2476 | the parser engine; a project can now use its own parser engine. |
2477 | The actions go into a separate file called NAME.act, in the form of | |
2478 | a switch statement body. | |
2479 | \f | |
402b123d | 2480 | * Changes in version 1.23: |
6780ca7a | 2481 | |
4d019228 DM |
2482 | The user can define YYPARSE_PARAM as the name of an argument to be |
2483 | passed into yyparse. The argument should have type void *. It should | |
2484 | actually point to an object. Grammar actions can access the variable | |
2485 | by casting it to the proper pointer type. | |
6780ca7a | 2486 | |
6780ca7a | 2487 | Line numbers in output file corrected. |
6780ca7a | 2488 | \f |
402b123d | 2489 | * Changes in version 1.22: |
6780ca7a DM |
2490 | |
2491 | --help option added. | |
6780ca7a | 2492 | \f |
402b123d | 2493 | * Changes in version 1.20: |
6780ca7a DM |
2494 | |
2495 | Output file does not redefine const for C++. | |
9f4503d6 | 2496 | |
76551463 AD |
2497 | ----- |
2498 | ||
7d6bad19 | 2499 | Copyright (C) 1995-2013 Free Software Foundation, Inc. |
76551463 | 2500 | |
74553c98 | 2501 | This file is part of Bison, the GNU Parser Generator. |
76551463 | 2502 | |
f16b0819 | 2503 | This program is free software: you can redistribute it and/or modify |
76551463 | 2504 | it under the terms of the GNU General Public License as published by |
f16b0819 PE |
2505 | the Free Software Foundation, either version 3 of the License, or |
2506 | (at your option) any later version. | |
76551463 | 2507 | |
f16b0819 | 2508 | This program is distributed in the hope that it will be useful, |
76551463 AD |
2509 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
2510 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
2511 | GNU General Public License for more details. | |
2512 | ||
2513 | You should have received a copy of the GNU General Public License | |
f16b0819 | 2514 | along with this program. If not, see <http://www.gnu.org/licenses/>. |
7e508a2b AD |
2515 | |
2516 | LocalWords: yacc YYBACKUP glr GCC lalr ArrayIndexOutOfBoundsException nullptr | |
2517 | LocalWords: cplusplus liby rpl fprintf mfcalc Wyacc stmt cond expr mk sym lr | |
2518 | LocalWords: IELR ielr Lookahead YYERROR nonassoc LALR's api lookaheads yychar | |
2519 | LocalWords: destructor lookahead YYRHSLOC YYLLOC Rhs ifndef YYFAIL cpp sr rr | |
2520 | LocalWords: preprocessor initializer Wno Wnone Werror FreeBSD prec livelocks | |
2521 | LocalWords: Solaris AIX UX RHEL Tru LHS gcc's Wundef YYENABLE NLS YYLTYPE VCG | |
2522 | LocalWords: yyerror cpp's Wunused yylval yylloc prepend yyparse yylex yypush | |
2523 | LocalWords: Graphviz xml nonterminals midrule destructor's YYSTYPE typedef ly | |
2524 | LocalWords: CHR chr printf stdout namespace preprocessing enum pre include's | |
2525 | LocalWords: YYRECOVERING nonfree destructors YYABORT YYACCEPT params enums de | |
2526 | LocalWords: struct yystype DJGPP lex param Haible NUM alloca YYSTACK NUL goto | |
2527 | LocalWords: YYMAXDEPTH Unescaped UCNs YYLTYPE's yyltype typedefs inline Yaccs | |
2528 | LocalWords: Heriyanto Reenable dprec Hilfinger Eggert MYEOF Folle Menezes EOF | |
242cc08e | 2529 | LocalWords: Lackovic define's itemset Groff Gettext malloc NEWS'ed YYDEBUG YY |
7e508a2b | 2530 | LocalWords: namespaces strerror const autoconfiguration Dconst Autoconf's FDL |
242cc08e AD |
2531 | LocalWords: Automake TMPDIR LESSEQ ylwrap endif yydebug YYTOKEN YYLSP ival hh |
2532 | LocalWords: extern YYTOKENTYPE TOKENTYPE yytokentype tokentype STYPE lval pdf | |
dcb366b1 | 2533 | LocalWords: lang yyoutput dvi html ps POSIX lvalp llocp Wother nterm arg init |
1282c124 AD |
2534 | LocalWords: TOK calc yyo fval Wconflicts parsers yystackp yyval yynerrs |
2535 | LocalWords: Théophile Ranquet Santet fno fnone stype associativity Tolmer | |
2536 | LocalWords: Wprecedence Rassoul Wempty Paolo Bonzini parser's Michiel loc | |
2537 | LocalWords: redeclaration sval fcaret reentrant XSLT xsl Wmaybe yyvsp Tedi | |
2538 | LocalWords: pragmas noreturn untyped Rozenman unexpanded Wojciech Polak | |
2539 | LocalWords: Alexandre MERCHANTABILITY | |
7e508a2b AD |
2540 | |
2541 | Local Variables: | |
2542 | mode: outline | |
e4ab1254 | 2543 | fill-column: 76 |
7e508a2b | 2544 | End: |