| 1 | GNU Bison NEWS |
| 2 | |
| 3 | * Noteworthy changes in release ?.? (????-??-??) [?] |
| 4 | |
| 5 | ** Incompatible changes |
| 6 | |
| 7 | *** Obsolete features |
| 8 | |
| 9 | Support for YYFAIL is removed (deprecated in Bison 2.4.2). |
| 10 | Support for yystype and yyltype (instead of YYSTYPE and YYLTYPE) |
| 11 | is removed (deprecated in Bison 1.875). |
| 12 | Support for YYPARSE_PARAM is removed (deprecated in Bison 1.875). |
| 13 | |
| 14 | ** Warnings |
| 15 | |
| 16 | *** Deprecated constructs |
| 17 | |
| 18 | The new 'deprecated' warning category flags obsolete constructs whose |
| 19 | support will be discontinued. It is enabled by default. These warnings |
| 20 | used to be reported as 'other' warnings. |
| 21 | |
| 22 | *** Warning categories are now displayed |
| 23 | |
| 24 | For instance: |
| 25 | |
| 26 | foo.y:4.6: warning: type clash on default action: <foo> != <bar> [-Wother] |
| 27 | |
| 28 | *** Useless semantic types |
| 29 | |
| 30 | Bison now warns about useless (uninhabited) semantic types. Since |
| 31 | semantic types are not declared to Bison (they are defined in the opaque |
| 32 | %union structure), it is %printer/%destructor directives about useless |
| 33 | types that trigger the warning: |
| 34 | |
| 35 | %token <type1> term |
| 36 | %type <type2> nterm |
| 37 | %printer {} <type1> <type3> |
| 38 | %destructor {} <type2> <type4> |
| 39 | %% |
| 40 | nterm: term { $$ = $1; }; |
| 41 | |
| 42 | 3.28-34: warning: type <type3> is used, but is not associated to any symbol |
| 43 | 4.28-34: warning: type <type4> is used, but is not associated to any symbol |
| 44 | |
| 45 | *** Undefined but unused symbols |
| 46 | |
| 47 | Bison used to raise an error for undefined symbols that are not used in |
| 48 | the grammar. This is now only a warning. |
| 49 | |
| 50 | %printer {} symbol1 |
| 51 | %destructor {} symbol2 |
| 52 | %type <type> symbol3 |
| 53 | %% |
| 54 | exp: "a"; |
| 55 | |
| 56 | *** Useless destructors or printers |
| 57 | |
| 58 | Bison now warns about useless destructors or printers. In the following |
| 59 | example, the printer for <type1>, and the destructor for <type2> are |
| 60 | useless: all symbols of <type1> (token1) already have a printer, and all |
| 61 | symbols of type <type2> (token2) already have a destructor. |
| 62 | |
| 63 | %token <type1> token1 |
| 64 | <type2> token2 |
| 65 | <type3> token3 |
| 66 | <type4> token4 |
| 67 | %printer {} token1 <type1> <type3> |
| 68 | %destructor {} token2 <type2> <type4> |
| 69 | |
| 70 | *** Conflicts |
| 71 | |
| 72 | The warnings and error messages about shift/reduce and reduce/reduce |
| 73 | conflicts have been normalized. For instance on the following foo.y file: |
| 74 | |
| 75 | %glr-parser |
| 76 | %% |
| 77 | exp: exp '+' exp | '0' | '0'; |
| 78 | |
| 79 | compare the previous version of bison: |
| 80 | |
| 81 | $ bison foo.y |
| 82 | foo.y: conflicts: 1 shift/reduce, 2 reduce/reduce |
| 83 | $ bison -Werror foo.y |
| 84 | bison: warnings being treated as errors |
| 85 | foo.y: conflicts: 1 shift/reduce, 2 reduce/reduce |
| 86 | |
| 87 | with the new behavior: |
| 88 | |
| 89 | $ bison foo.y |
| 90 | foo.y: warning: 1 shift/reduce conflict [-Wconflicts-sr] |
| 91 | foo.y: warning: 2 reduce/reduce conflicts [-Wconflicts-rr] |
| 92 | $ bison -Werror foo.y |
| 93 | foo.y: error: 1 shift/reduce conflict [-Werror=conflicts-sr] |
| 94 | foo.y: error: 2 reduce/reduce conflicts [-Werror=conflicts-rr] |
| 95 | |
| 96 | When %expect or %expect-rr is used, such as with bar.y: |
| 97 | |
| 98 | %expect 0 |
| 99 | %glr-parser |
| 100 | %% |
| 101 | exp: exp '+' exp | '0' | '0'; |
| 102 | |
| 103 | Former behavior: |
| 104 | |
| 105 | $ bison bar.y |
| 106 | bar.y: conflicts: 1 shift/reduce, 2 reduce/reduce |
| 107 | bar.y: expected 0 shift/reduce conflicts |
| 108 | bar.y: expected 0 reduce/reduce conflicts |
| 109 | |
| 110 | New one: |
| 111 | |
| 112 | $ bison bar.y |
| 113 | bar.y: error: shift/reduce conflicts: 1 found, 0 expected |
| 114 | bar.y: error: reduce/reduce conflicts: 2 found, 0 expected |
| 115 | |
| 116 | ** Additional yylex/yyparse arguments |
| 117 | |
| 118 | The new directive %param declares additional arguments to both yylex and |
| 119 | yyparse. The %lex-param, %parse-param, and %param directives support one |
| 120 | or more arguments. Instead of |
| 121 | |
| 122 | %lex-param {arg1_type *arg1} |
| 123 | %lex-param {arg2_type *arg2} |
| 124 | %parse-param {arg1_type *arg1} |
| 125 | %parse-param {arg2_type *arg2} |
| 126 | |
| 127 | one may now declare |
| 128 | |
| 129 | %param {arg1_type *arg1} {arg2_type *arg2} |
| 130 | |
| 131 | ** Java skeleton improvements |
| 132 | |
| 133 | The constants for token names were moved to the Lexer interface. Also, it |
| 134 | is possible to add code to the parser's constructors using "%code init" |
| 135 | and "%define init_throws". |
| 136 | |
| 137 | ** C++ skeletons improvements |
| 138 | |
| 139 | *** The parser header is no longer mandatory (lalr1.cc, glr.cc) |
| 140 | |
| 141 | Using %defines is now optional. Without it, the needed support classes |
| 142 | are defined in the generated parser, instead of additional files (such as |
| 143 | location.hh, position.hh and stack.hh). |
| 144 | |
| 145 | *** Locations are no longer mandatory (lalr1.cc, glr.cc) |
| 146 | |
| 147 | Both lalr1.cc and glr.cc no longer require %location. |
| 148 | |
| 149 | *** syntax_error exception (lalr1.cc) |
| 150 | |
| 151 | The C++ parser features a syntax_error exception, which can be |
| 152 | thrown from the scanner or from user rules to raise syntax errors. |
| 153 | This facilitates reporting errors caught in sub-functions (e.g., |
| 154 | rejecting too large integral literals from a conversion function |
| 155 | used by the scanner, or rejecting invalid combinations from a |
| 156 | factory invoked by the user actions). |
| 157 | |
| 158 | ** Variable api.tokens.prefix |
| 159 | |
| 160 | The variable api.tokens.prefix changes the way tokens are identified in |
| 161 | the generated files. This is especially useful to avoid collisions |
| 162 | with identifiers in the target language. For instance |
| 163 | |
| 164 | %token FILE for ERROR |
| 165 | %define api.tokens.prefix "TOK_" |
| 166 | %% |
| 167 | start: FILE for ERROR; |
| 168 | |
| 169 | will generate the definition of the symbols TOK_FILE, TOK_for, and |
| 170 | TOK_ERROR in the generated sources. In particular, the scanner must |
| 171 | use these prefixed token names, although the grammar itself still |
| 172 | uses the short names (as in the sample rule given above). |
| 173 | |
| 174 | ** Variable api.namespace |
| 175 | |
| 176 | The 'namespace' variable is renamed 'api.namespace'. Backward |
| 177 | compatibility is ensured, but upgrading is recommended. |
| 178 | |
| 179 | ** Variable parse.error |
| 180 | |
| 181 | This variable controls the verbosity of error messages. The use of the |
| 182 | %error-verbose directive is deprecated in favor of "%define parse.error |
| 183 | verbose". |
| 184 | |
| 185 | ** Semantic predicates |
| 186 | |
| 187 | The new, experimental, semantic-predicate feature allows actions of the |
| 188 | form "%?{ BOOLEAN-EXPRESSION }", which cause syntax errors (as for |
| 189 | YYERROR) if the expression evaluates to 0, and are evaluated immediately |
| 190 | in GLR parsers, rather than being deferred. The result is that they allow |
| 191 | the programmer to prune possible parses based on the values of run-time |
| 192 | expressions. |
| 193 | |
| 194 | ** The directive %expect-rr is now an error in non GLR mode |
| 195 | |
| 196 | It used to be an error only if used in non GLR mode, _and_ if there are |
| 197 | reduce/reduce conflicts. |
| 198 | |
| 199 | * Noteworthy changes in release ?.? (????-??-??) [?] |
| 200 | |
| 201 | ** Bug fixes |
| 202 | |
| 203 | Bugs in the test suite have been fixed. |
| 204 | |
| 205 | Some errors in translations have been addressed, and --help now directs |
| 206 | users to the appropriate place to report them. |
| 207 | |
| 208 | Stray Info files shipped by accident are removed. |
| 209 | |
| 210 | Incorrect definitions of YY_, issued by yacc.c when no parser header is |
| 211 | generated, are removed. |
| 212 | |
| 213 | ** Changes in the format of errors and exceptions output |
| 214 | |
| 215 | This used to be the format of many error reports: |
| 216 | |
| 217 | foo.y:5.10-24: result type clash on merge function 'merge': <t3> != <t2> |
| 218 | foo.y:4.13-27: previous declaration |
| 219 | |
| 220 | It is now: |
| 221 | |
| 222 | foo.y:5.10-25: result type clash on merge function 'merge': <t3> != <t2> |
| 223 | foo.y:4.13-27: previous declaration |
| 224 | |
| 225 | * Noteworthy changes in release 2.6.2 (2012-08-03) [stable] |
| 226 | |
| 227 | ** Bug fixes |
| 228 | |
| 229 | Buffer overruns, complaints from Flex, and portability issues in the test |
| 230 | suite have been fixed. |
| 231 | |
| 232 | ** Spaces in %lex- and %parse-param (lalr1.cc, glr.cc) |
| 233 | |
| 234 | Trailing end-of-lines in %parse-param or %lex-param would result in |
| 235 | invalid C++. This is fixed. |
| 236 | |
| 237 | ** Spurious spaces and end-of-lines |
| 238 | |
| 239 | The generated files no longer end (nor start) with empty lines. |
| 240 | |
| 241 | * Noteworthy changes in release 2.6.1 (2012-07-30) [stable] |
| 242 | |
| 243 | Bison no longer executes user-specified M4 code when processing a grammar. |
| 244 | |
| 245 | ** Future Changes |
| 246 | |
| 247 | In addition to the removal of the features announced in Bison 2.6, the |
| 248 | next major release will remove the "Temporary hack for adding a semicolon |
| 249 | to the user action", as announced in the release 2.5. Instead of: |
| 250 | |
| 251 | exp: exp "+" exp { $$ = $1 + $3 }; |
| 252 | |
| 253 | write: |
| 254 | |
| 255 | exp: exp "+" exp { $$ = $1 + $3; }; |
| 256 | |
| 257 | ** Bug fixes |
| 258 | |
| 259 | *** Type names are now properly escaped. |
| 260 | |
| 261 | *** glr.cc: set_debug_level and debug_level work as expected. |
| 262 | |
| 263 | *** Stray @ or $ in actions |
| 264 | |
| 265 | While Bison used to warn about stray $ or @ in action rules, it did not |
| 266 | for other actions such as printers, destructors, or initial actions. It |
| 267 | now does. |
| 268 | |
| 269 | ** Type names in actions |
| 270 | |
| 271 | For consistency with rule actions, it is now possible to qualify $$ by a |
| 272 | type-name in destructors, printers, and initial actions. For instance: |
| 273 | |
| 274 | %printer { fprintf (yyo, "(%d, %f)", $<ival>$, $<fval>$); } <*> <>; |
| 275 | |
| 276 | will display two values for each typed and untyped symbol (provided |
| 277 | that YYSTYPE has both "ival" and "fval" fields). |
| 278 | |
| 279 | * Noteworthy changes in release 2.6 (2012-07-19) [stable] |
| 280 | |
| 281 | ** Future changes: |
| 282 | |
| 283 | The next major release of Bison will drop support for the following |
| 284 | deprecated features. Please report disagreements to bug-bison@gnu.org. |
| 285 | |
| 286 | *** K&R C parsers |
| 287 | |
| 288 | Support for generating parsers in K&R C will be removed. Parsers |
| 289 | generated for C support ISO C90, and are tested with ISO C99 and ISO C11 |
| 290 | compilers. |
| 291 | |
| 292 | *** Features deprecated since Bison 1.875 |
| 293 | |
| 294 | The definitions of yystype and yyltype will be removed; use YYSTYPE and |
| 295 | YYLTYPE. |
| 296 | |
| 297 | YYPARSE_PARAM and YYLEX_PARAM, deprecated in favor of %parse-param and |
| 298 | %lex-param, will no longer be supported. |
| 299 | |
| 300 | Support for the preprocessor symbol YYERROR_VERBOSE will be removed, use |
| 301 | %error-verbose. |
| 302 | |
| 303 | *** The generated header will be included (yacc.c) |
| 304 | |
| 305 | Instead of duplicating the content of the generated header (definition of |
| 306 | YYSTYPE, yyparse declaration etc.), the generated parser will include it, |
| 307 | as is already the case for GLR or C++ parsers. This change is deferred |
| 308 | because existing versions of ylwrap (e.g., Automake 1.12.1) do not support |
| 309 | it. |
| 310 | |
| 311 | ** Generated Parser Headers |
| 312 | |
| 313 | *** Guards (yacc.c, glr.c, glr.cc) |
| 314 | |
| 315 | The generated headers are now guarded, as is already the case for C++ |
| 316 | parsers (lalr1.cc). For instance, with --defines=foo.h: |
| 317 | |
| 318 | #ifndef YY_FOO_H |
| 319 | # define YY_FOO_H |
| 320 | ... |
| 321 | #endif /* !YY_FOO_H */ |
| 322 | |
| 323 | *** New declarations (yacc.c, glr.c) |
| 324 | |
| 325 | The generated header now declares yydebug and yyparse. Both honor |
| 326 | --name-prefix=bar_, and yield |
| 327 | |
| 328 | int bar_parse (void); |
| 329 | |
| 330 | rather than |
| 331 | |
| 332 | #define yyparse bar_parse |
| 333 | int yyparse (void); |
| 334 | |
| 335 | in order to facilitate the inclusion of several parser headers inside a |
| 336 | single compilation unit. |
| 337 | |
| 338 | *** Exported symbols in C++ |
| 339 | |
| 340 | The symbols YYTOKEN_TABLE and YYERROR_VERBOSE, which were defined in the |
| 341 | header, are removed, as they prevent the possibility of including several |
| 342 | generated headers from a single compilation unit. |
| 343 | |
| 344 | *** YYLSP_NEEDED |
| 345 | |
| 346 | For the same reasons, the undocumented and unused macro YYLSP_NEEDED is no |
| 347 | longer defined. |
| 348 | |
| 349 | ** New %define variable: api.prefix |
| 350 | |
| 351 | Now that the generated headers are more complete and properly protected |
| 352 | against multiple inclusions, constant names, such as YYSTYPE are a |
| 353 | problem. While yyparse and others are properly renamed by %name-prefix, |
| 354 | YYSTYPE, YYDEBUG and others have never been affected by it. Because it |
| 355 | would introduce backward compatibility issues in projects not expecting |
| 356 | YYSTYPE to be renamed, instead of changing the behavior of %name-prefix, |
| 357 | it is deprecated in favor of a new %define variable: api.prefix. |
| 358 | |
| 359 | The following examples compares both: |
| 360 | |
| 361 | %name-prefix "bar_" | %define api.prefix "bar_" |
| 362 | %token <ival> FOO %token <ival> FOO |
| 363 | %union { int ival; } %union { int ival; } |
| 364 | %% %% |
| 365 | exp: 'a'; exp: 'a'; |
| 366 | |
| 367 | bison generates: |
| 368 | |
| 369 | #ifndef BAR_FOO_H #ifndef BAR_FOO_H |
| 370 | # define BAR_FOO_H # define BAR_FOO_H |
| 371 | |
| 372 | /* Enabling traces. */ /* Enabling traces. */ |
| 373 | # ifndef YYDEBUG | # ifndef BAR_DEBUG |
| 374 | > # if defined YYDEBUG |
| 375 | > # if YYDEBUG |
| 376 | > # define BAR_DEBUG 1 |
| 377 | > # else |
| 378 | > # define BAR_DEBUG 0 |
| 379 | > # endif |
| 380 | > # else |
| 381 | # define YYDEBUG 0 | # define BAR_DEBUG 0 |
| 382 | > # endif |
| 383 | # endif | # endif |
| 384 | |
| 385 | # if YYDEBUG | # if BAR_DEBUG |
| 386 | extern int bar_debug; extern int bar_debug; |
| 387 | # endif # endif |
| 388 | |
| 389 | /* Tokens. */ /* Tokens. */ |
| 390 | # ifndef YYTOKENTYPE | # ifndef BAR_TOKENTYPE |
| 391 | # define YYTOKENTYPE | # define BAR_TOKENTYPE |
| 392 | enum yytokentype { | enum bar_tokentype { |
| 393 | FOO = 258 FOO = 258 |
| 394 | }; }; |
| 395 | # endif # endif |
| 396 | |
| 397 | #if ! defined YYSTYPE \ | #if ! defined BAR_STYPE \ |
| 398 | && ! defined YYSTYPE_IS_DECLARED | && ! defined BAR_STYPE_IS_DECLARED |
| 399 | typedef union YYSTYPE | typedef union BAR_STYPE |
| 400 | { { |
| 401 | int ival; int ival; |
| 402 | } YYSTYPE; | } BAR_STYPE; |
| 403 | # define YYSTYPE_IS_DECLARED 1 | # define BAR_STYPE_IS_DECLARED 1 |
| 404 | #endif #endif |
| 405 | |
| 406 | extern YYSTYPE bar_lval; | extern BAR_STYPE bar_lval; |
| 407 | |
| 408 | int bar_parse (void); int bar_parse (void); |
| 409 | |
| 410 | #endif /* !BAR_FOO_H */ #endif /* !BAR_FOO_H */ |
| 411 | |
| 412 | * Noteworthy changes in release 2.5.1 (2012-06-05) [stable] |
| 413 | |
| 414 | ** Future changes: |
| 415 | |
| 416 | The next major release will drop support for generating parsers in K&R C. |
| 417 | |
| 418 | ** yacc.c: YYBACKUP works as expected. |
| 419 | |
| 420 | ** glr.c improvements: |
| 421 | |
| 422 | *** Location support is eliminated when not requested: |
| 423 | |
| 424 | GLR parsers used to include location-related code even when locations were |
| 425 | not requested, and therefore not even usable. |
| 426 | |
| 427 | *** __attribute__ is preserved: |
| 428 | |
| 429 | __attribute__ is no longer disabled when __STRICT_ANSI__ is defined (i.e., |
| 430 | when -std is passed to GCC). |
| 431 | |
| 432 | ** lalr1.java: several fixes: |
| 433 | |
| 434 | The Java parser no longer throws ArrayIndexOutOfBoundsException if the |
| 435 | first token leads to a syntax error. Some minor clean ups. |
| 436 | |
| 437 | ** Changes for C++: |
| 438 | |
| 439 | *** C++11 compatibility: |
| 440 | |
| 441 | C and C++ parsers use "nullptr" instead of "0" when __cplusplus is 201103L |
| 442 | or higher. |
| 443 | |
| 444 | *** Header guards |
| 445 | |
| 446 | The header files such as "parser.hh", "location.hh", etc. used a constant |
| 447 | name for preprocessor guards, for instance: |
| 448 | |
| 449 | #ifndef BISON_LOCATION_HH |
| 450 | # define BISON_LOCATION_HH |
| 451 | ... |
| 452 | #endif // !BISON_LOCATION_HH |
| 453 | |
| 454 | The inclusion guard is now computed from "PREFIX/FILE-NAME", where lower |
| 455 | case characters are converted to upper case, and series of |
| 456 | non-alphanumerical characters are converted to an underscore. |
| 457 | |
| 458 | With "bison -o lang++/parser.cc", "location.hh" would now include: |
| 459 | |
| 460 | #ifndef YY_LANG_LOCATION_HH |
| 461 | # define YY_LANG_LOCATION_HH |
| 462 | ... |
| 463 | #endif // !YY_LANG_LOCATION_HH |
| 464 | |
| 465 | *** C++ locations: |
| 466 | |
| 467 | The position and location constructors (and their initialize methods) |
| 468 | accept new arguments for line and column. Several issues in the |
| 469 | documentation were fixed. |
| 470 | |
| 471 | ** liby is no longer asking for "rpl_fprintf" on some platforms. |
| 472 | |
| 473 | ** Changes in the manual: |
| 474 | |
| 475 | *** %printer is documented |
| 476 | |
| 477 | The "%printer" directive, supported since at least Bison 1.50, is finally |
| 478 | documented. The "mfcalc" example is extended to demonstrate it. |
| 479 | |
| 480 | For consistency with the C skeletons, the C++ parsers now also support |
| 481 | "yyoutput" (as an alias to "debug_stream ()"). |
| 482 | |
| 483 | *** Several improvements have been made: |
| 484 | |
| 485 | The layout for grammar excerpts was changed to a more compact scheme. |
| 486 | Named references are motivated. The description of the automaton |
| 487 | description file (*.output) is updated to the current format. Incorrect |
| 488 | index entries were fixed. Some other errors were fixed. |
| 489 | |
| 490 | ** Building bison: |
| 491 | |
| 492 | *** Conflicting prototypes with recent/modified Flex. |
| 493 | |
| 494 | Fixed build problems with the current, unreleased, version of Flex, and |
| 495 | some modified versions of 2.5.35, which have modified function prototypes. |
| 496 | |
| 497 | *** Warnings during the build procedure have been eliminated. |
| 498 | |
| 499 | *** Several portability problems in the test suite have been fixed: |
| 500 | |
| 501 | This includes warnings with some compilers, unexpected behavior of tools |
| 502 | such as diff, warning messages from the test suite itself, etc. |
| 503 | |
| 504 | *** The install-pdf target works properly: |
| 505 | |
| 506 | Running "make install-pdf" (or -dvi, -html, -info, and -ps) no longer |
| 507 | halts in the middle of its course. |
| 508 | |
| 509 | * Changes in version 2.5 (2011-05-14): |
| 510 | |
| 511 | ** Grammar symbol names can now contain non-initial dashes: |
| 512 | |
| 513 | Consistently with directives (such as %error-verbose) and with |
| 514 | %define variables (e.g. push-pull), grammar symbol names may contain |
| 515 | dashes in any position except the beginning. This is a GNU |
| 516 | extension over POSIX Yacc. Thus, use of this extension is reported |
| 517 | by -Wyacc and rejected in Yacc mode (--yacc). |
| 518 | |
| 519 | ** Named references: |
| 520 | |
| 521 | Historically, Yacc and Bison have supported positional references |
| 522 | ($n, $$) to allow access to symbol values from inside of semantic |
| 523 | actions code. |
| 524 | |
| 525 | Starting from this version, Bison can also accept named references. |
| 526 | When no ambiguity is possible, original symbol names may be used |
| 527 | as named references: |
| 528 | |
| 529 | if_stmt : "if" cond_expr "then" then_stmt ';' |
| 530 | { $if_stmt = mk_if_stmt($cond_expr, $then_stmt); } |
| 531 | |
| 532 | In the more common case, explicit names may be declared: |
| 533 | |
| 534 | stmt[res] : "if" expr[cond] "then" stmt[then] "else" stmt[else] ';' |
| 535 | { $res = mk_if_stmt($cond, $then, $else); } |
| 536 | |
| 537 | Location information is also accessible using @name syntax. When |
| 538 | accessing symbol names containing dots or dashes, explicit bracketing |
| 539 | ($[sym.1]) must be used. |
| 540 | |
| 541 | These features are experimental in this version. More user feedback |
| 542 | will help to stabilize them. |
| 543 | |
| 544 | ** IELR(1) and canonical LR(1): |
| 545 | |
| 546 | IELR(1) is a minimal LR(1) parser table generation algorithm. That |
| 547 | is, given any context-free grammar, IELR(1) generates parser tables |
| 548 | with the full language-recognition power of canonical LR(1) but with |
| 549 | nearly the same number of parser states as LALR(1). This reduction |
| 550 | in parser states is often an order of magnitude. More importantly, |
| 551 | because canonical LR(1)'s extra parser states may contain duplicate |
| 552 | conflicts in the case of non-LR(1) grammars, the number of conflicts |
| 553 | for IELR(1) is often an order of magnitude less as well. This can |
| 554 | significantly reduce the complexity of developing of a grammar. |
| 555 | |
| 556 | Bison can now generate IELR(1) and canonical LR(1) parser tables in |
| 557 | place of its traditional LALR(1) parser tables, which remain the |
| 558 | default. You can specify the type of parser tables in the grammar |
| 559 | file with these directives: |
| 560 | |
| 561 | %define lr.type lalr |
| 562 | %define lr.type ielr |
| 563 | %define lr.type canonical-lr |
| 564 | |
| 565 | The default-reduction optimization in the parser tables can also be |
| 566 | adjusted using "%define lr.default-reductions". For details on both |
| 567 | of these features, see the new section "Tuning LR" in the Bison |
| 568 | manual. |
| 569 | |
| 570 | These features are experimental. More user feedback will help to |
| 571 | stabilize them. |
| 572 | |
| 573 | ** LAC (Lookahead Correction) for syntax error handling: |
| 574 | |
| 575 | Canonical LR, IELR, and LALR can suffer from a couple of problems |
| 576 | upon encountering a syntax error. First, the parser might perform |
| 577 | additional parser stack reductions before discovering the syntax |
| 578 | error. Such reductions can perform user semantic actions that are |
| 579 | unexpected because they are based on an invalid token, and they |
| 580 | cause error recovery to begin in a different syntactic context than |
| 581 | the one in which the invalid token was encountered. Second, when |
| 582 | verbose error messages are enabled (with %error-verbose or the |
| 583 | obsolete "#define YYERROR_VERBOSE"), the expected token list in the |
| 584 | syntax error message can both contain invalid tokens and omit valid |
| 585 | tokens. |
| 586 | |
| 587 | The culprits for the above problems are %nonassoc, default |
| 588 | reductions in inconsistent states, and parser state merging. Thus, |
| 589 | IELR and LALR suffer the most. Canonical LR can suffer only if |
| 590 | %nonassoc is used or if default reductions are enabled for |
| 591 | inconsistent states. |
| 592 | |
| 593 | LAC is a new mechanism within the parsing algorithm that solves |
| 594 | these problems for canonical LR, IELR, and LALR without sacrificing |
| 595 | %nonassoc, default reductions, or state merging. When LAC is in |
| 596 | use, canonical LR and IELR behave almost exactly the same for both |
| 597 | syntactically acceptable and syntactically unacceptable input. |
| 598 | While LALR still does not support the full language-recognition |
| 599 | power of canonical LR and IELR, LAC at least enables LALR's syntax |
| 600 | error handling to correctly reflect LALR's language-recognition |
| 601 | power. |
| 602 | |
| 603 | Currently, LAC is only supported for deterministic parsers in C. |
| 604 | You can enable LAC with the following directive: |
| 605 | |
| 606 | %define parse.lac full |
| 607 | |
| 608 | See the new section "LAC" in the Bison manual for additional |
| 609 | details including a few caveats. |
| 610 | |
| 611 | LAC is an experimental feature. More user feedback will help to |
| 612 | stabilize it. |
| 613 | |
| 614 | ** %define improvements: |
| 615 | |
| 616 | *** Can now be invoked via the command line: |
| 617 | |
| 618 | Each of these command-line options |
| 619 | |
| 620 | -D NAME[=VALUE] |
| 621 | --define=NAME[=VALUE] |
| 622 | |
| 623 | -F NAME[=VALUE] |
| 624 | --force-define=NAME[=VALUE] |
| 625 | |
| 626 | is equivalent to this grammar file declaration |
| 627 | |
| 628 | %define NAME ["VALUE"] |
| 629 | |
| 630 | except that the manner in which Bison processes multiple definitions |
| 631 | for the same NAME differs. Most importantly, -F and --force-define |
| 632 | quietly override %define, but -D and --define do not. For further |
| 633 | details, see the section "Bison Options" in the Bison manual. |
| 634 | |
| 635 | *** Variables renamed: |
| 636 | |
| 637 | The following %define variables |
| 638 | |
| 639 | api.push_pull |
| 640 | lr.keep_unreachable_states |
| 641 | |
| 642 | have been renamed to |
| 643 | |
| 644 | api.push-pull |
| 645 | lr.keep-unreachable-states |
| 646 | |
| 647 | The old names are now deprecated but will be maintained indefinitely |
| 648 | for backward compatibility. |
| 649 | |
| 650 | *** Values no longer need to be quoted in the grammar file: |
| 651 | |
| 652 | If a %define value is an identifier, it no longer needs to be placed |
| 653 | within quotations marks. For example, |
| 654 | |
| 655 | %define api.push-pull "push" |
| 656 | |
| 657 | can be rewritten as |
| 658 | |
| 659 | %define api.push-pull push |
| 660 | |
| 661 | *** Unrecognized variables are now errors not warnings. |
| 662 | |
| 663 | *** Multiple invocations for any variable is now an error not a warning. |
| 664 | |
| 665 | ** Unrecognized %code qualifiers are now errors not warnings. |
| 666 | |
| 667 | ** Character literals not of length one: |
| 668 | |
| 669 | Previously, Bison quietly converted all character literals to length |
| 670 | one. For example, without warning, Bison interpreted the operators in |
| 671 | the following grammar to be the same token: |
| 672 | |
| 673 | exp: exp '++' |
| 674 | | exp '+' exp |
| 675 | ; |
| 676 | |
| 677 | Bison now warns when a character literal is not of length one. In |
| 678 | some future release, Bison will start reporting an error instead. |
| 679 | |
| 680 | ** Destructor calls fixed for lookaheads altered in semantic actions: |
| 681 | |
| 682 | Previously for deterministic parsers in C, if a user semantic action |
| 683 | altered yychar, the parser in some cases used the old yychar value to |
| 684 | determine which destructor to call for the lookahead upon a syntax |
| 685 | error or upon parser return. This bug has been fixed. |
| 686 | |
| 687 | ** C++ parsers use YYRHSLOC: |
| 688 | |
| 689 | Similarly to the C parsers, the C++ parsers now define the YYRHSLOC |
| 690 | macro and use it in the default YYLLOC_DEFAULT. You are encouraged |
| 691 | to use it. If, for instance, your location structure has "first" |
| 692 | and "last" members, instead of |
| 693 | |
| 694 | # define YYLLOC_DEFAULT(Current, Rhs, N) \ |
| 695 | do \ |
| 696 | if (N) \ |
| 697 | { \ |
| 698 | (Current).first = (Rhs)[1].location.first; \ |
| 699 | (Current).last = (Rhs)[N].location.last; \ |
| 700 | } \ |
| 701 | else \ |
| 702 | { \ |
| 703 | (Current).first = (Current).last = (Rhs)[0].location.last; \ |
| 704 | } \ |
| 705 | while (false) |
| 706 | |
| 707 | use: |
| 708 | |
| 709 | # define YYLLOC_DEFAULT(Current, Rhs, N) \ |
| 710 | do \ |
| 711 | if (N) \ |
| 712 | { \ |
| 713 | (Current).first = YYRHSLOC (Rhs, 1).first; \ |
| 714 | (Current).last = YYRHSLOC (Rhs, N).last; \ |
| 715 | } \ |
| 716 | else \ |
| 717 | { \ |
| 718 | (Current).first = (Current).last = YYRHSLOC (Rhs, 0).last; \ |
| 719 | } \ |
| 720 | while (false) |
| 721 | |
| 722 | ** YYLLOC_DEFAULT in C++: |
| 723 | |
| 724 | The default implementation of YYLLOC_DEFAULT used to be issued in |
| 725 | the header file. It is now output in the implementation file, after |
| 726 | the user %code sections so that its #ifndef guard does not try to |
| 727 | override the user's YYLLOC_DEFAULT if provided. |
| 728 | |
| 729 | ** YYFAIL now produces warnings and Java parsers no longer implement it: |
| 730 | |
| 731 | YYFAIL has existed for many years as an undocumented feature of |
| 732 | deterministic parsers in C generated by Bison. More recently, it was |
| 733 | a documented feature of Bison's experimental Java parsers. As |
| 734 | promised in Bison 2.4.2's NEWS entry, any appearance of YYFAIL in a |
| 735 | semantic action now produces a deprecation warning, and Java parsers |
| 736 | no longer implement YYFAIL at all. For further details, including a |
| 737 | discussion of how to suppress C preprocessor warnings about YYFAIL |
| 738 | being unused, see the Bison 2.4.2 NEWS entry. |
| 739 | |
| 740 | ** Temporary hack for adding a semicolon to the user action: |
| 741 | |
| 742 | Previously, Bison appended a semicolon to every user action for |
| 743 | reductions when the output language defaulted to C (specifically, when |
| 744 | neither %yacc, %language, %skeleton, or equivalent command-line |
| 745 | options were specified). This allowed actions such as |
| 746 | |
| 747 | exp: exp "+" exp { $$ = $1 + $3 }; |
| 748 | |
| 749 | instead of |
| 750 | |
| 751 | exp: exp "+" exp { $$ = $1 + $3; }; |
| 752 | |
| 753 | As a first step in removing this misfeature, Bison now issues a |
| 754 | warning when it appends a semicolon. Moreover, in cases where Bison |
| 755 | cannot easily determine whether a semicolon is needed (for example, an |
| 756 | action ending with a cpp directive or a braced compound initializer), |
| 757 | it no longer appends one. Thus, the C compiler might now complain |
| 758 | about a missing semicolon where it did not before. Future releases of |
| 759 | Bison will cease to append semicolons entirely. |
| 760 | |
| 761 | ** Verbose syntax error message fixes: |
| 762 | |
| 763 | When %error-verbose or the obsolete "#define YYERROR_VERBOSE" is |
| 764 | specified, syntax error messages produced by the generated parser |
| 765 | include the unexpected token as well as a list of expected tokens. |
| 766 | The effect of %nonassoc on these verbose messages has been corrected |
| 767 | in two ways, but a more complete fix requires LAC, described above: |
| 768 | |
| 769 | *** When %nonassoc is used, there can exist parser states that accept no |
| 770 | tokens, and so the parser does not always require a lookahead token |
| 771 | in order to detect a syntax error. Because no unexpected token or |
| 772 | expected tokens can then be reported, the verbose syntax error |
| 773 | message described above is suppressed, and the parser instead |
| 774 | reports the simpler message, "syntax error". Previously, this |
| 775 | suppression was sometimes erroneously triggered by %nonassoc when a |
| 776 | lookahead was actually required. Now verbose messages are |
| 777 | suppressed only when all previous lookaheads have already been |
| 778 | shifted or discarded. |
| 779 | |
| 780 | *** Previously, the list of expected tokens erroneously included tokens |
| 781 | that would actually induce a syntax error because conflicts for them |
| 782 | were resolved with %nonassoc in the current parser state. Such |
| 783 | tokens are now properly omitted from the list. |
| 784 | |
| 785 | *** Expected token lists are still often wrong due to state merging |
| 786 | (from LALR or IELR) and default reductions, which can both add |
| 787 | invalid tokens and subtract valid tokens. Canonical LR almost |
| 788 | completely fixes this problem by eliminating state merging and |
| 789 | default reductions. However, there is one minor problem left even |
| 790 | when using canonical LR and even after the fixes above. That is, |
| 791 | if the resolution of a conflict with %nonassoc appears in a later |
| 792 | parser state than the one at which some syntax error is |
| 793 | discovered, the conflicted token is still erroneously included in |
| 794 | the expected token list. Bison's new LAC implementation, |
| 795 | described above, eliminates this problem and the need for |
| 796 | canonical LR. However, LAC is still experimental and is disabled |
| 797 | by default. |
| 798 | |
| 799 | ** Java skeleton fixes: |
| 800 | |
| 801 | *** A location handling bug has been fixed. |
| 802 | |
| 803 | *** The top element of each of the value stack and location stack is now |
| 804 | cleared when popped so that it can be garbage collected. |
| 805 | |
| 806 | *** Parser traces now print the top element of the stack. |
| 807 | |
| 808 | ** -W/--warnings fixes: |
| 809 | |
| 810 | *** Bison now properly recognizes the "no-" versions of categories: |
| 811 | |
| 812 | For example, given the following command line, Bison now enables all |
| 813 | warnings except warnings for incompatibilities with POSIX Yacc: |
| 814 | |
| 815 | bison -Wall,no-yacc gram.y |
| 816 | |
| 817 | *** Bison now treats S/R and R/R conflicts like other warnings: |
| 818 | |
| 819 | Previously, conflict reports were independent of Bison's normal |
| 820 | warning system. Now, Bison recognizes the warning categories |
| 821 | "conflicts-sr" and "conflicts-rr". This change has important |
| 822 | consequences for the -W and --warnings command-line options. For |
| 823 | example: |
| 824 | |
| 825 | bison -Wno-conflicts-sr gram.y # S/R conflicts not reported |
| 826 | bison -Wno-conflicts-rr gram.y # R/R conflicts not reported |
| 827 | bison -Wnone gram.y # no conflicts are reported |
| 828 | bison -Werror gram.y # any conflict is an error |
| 829 | |
| 830 | However, as before, if the %expect or %expect-rr directive is |
| 831 | specified, an unexpected number of conflicts is an error, and an |
| 832 | expected number of conflicts is not reported, so -W and --warning |
| 833 | then have no effect on the conflict report. |
| 834 | |
| 835 | *** The "none" category no longer disables a preceding "error": |
| 836 | |
| 837 | For example, for the following command line, Bison now reports |
| 838 | errors instead of warnings for incompatibilities with POSIX Yacc: |
| 839 | |
| 840 | bison -Werror,none,yacc gram.y |
| 841 | |
| 842 | *** The "none" category now disables all Bison warnings: |
| 843 | |
| 844 | Previously, the "none" category disabled only Bison warnings for |
| 845 | which there existed a specific -W/--warning category. However, |
| 846 | given the following command line, Bison is now guaranteed to |
| 847 | suppress all warnings: |
| 848 | |
| 849 | bison -Wnone gram.y |
| 850 | |
| 851 | ** Precedence directives can now assign token number 0: |
| 852 | |
| 853 | Since Bison 2.3b, which restored the ability of precedence |
| 854 | directives to assign token numbers, doing so for token number 0 has |
| 855 | produced an assertion failure. For example: |
| 856 | |
| 857 | %left END 0 |
| 858 | |
| 859 | This bug has been fixed. |
| 860 | |
| 861 | * Changes in version 2.4.3 (2010-08-05): |
| 862 | |
| 863 | ** Bison now obeys -Werror and --warnings=error for warnings about |
| 864 | grammar rules that are useless in the parser due to conflicts. |
| 865 | |
| 866 | ** Problems with spawning M4 on at least FreeBSD 8 and FreeBSD 9 have |
| 867 | been fixed. |
| 868 | |
| 869 | ** Failures in the test suite for GCC 4.5 have been fixed. |
| 870 | |
| 871 | ** Failures in the test suite for some versions of Sun Studio C++ have |
| 872 | been fixed. |
| 873 | |
| 874 | ** Contrary to Bison 2.4.2's NEWS entry, it has been decided that |
| 875 | warnings about undefined %prec identifiers will not be converted to |
| 876 | errors in Bison 2.5. They will remain warnings, which should be |
| 877 | sufficient for POSIX while avoiding backward compatibility issues. |
| 878 | |
| 879 | ** Minor documentation fixes. |
| 880 | |
| 881 | * Changes in version 2.4.2 (2010-03-20): |
| 882 | |
| 883 | ** Some portability problems that resulted in failures and livelocks |
| 884 | in the test suite on some versions of at least Solaris, AIX, HP-UX, |
| 885 | RHEL4, and Tru64 have been addressed. As a result, fatal Bison |
| 886 | errors should no longer cause M4 to report a broken pipe on the |
| 887 | affected platforms. |
| 888 | |
| 889 | ** "%prec IDENTIFIER" requires IDENTIFIER to be defined separately. |
| 890 | |
| 891 | POSIX specifies that an error be reported for any identifier that does |
| 892 | not appear on the LHS of a grammar rule and that is not defined by |
| 893 | %token, %left, %right, or %nonassoc. Bison 2.3b and later lost this |
| 894 | error report for the case when an identifier appears only after a |
| 895 | %prec directive. It is now restored. However, for backward |
| 896 | compatibility with recent Bison releases, it is only a warning for |
| 897 | now. In Bison 2.5 and later, it will return to being an error. |
| 898 | [Between the 2.4.2 and 2.4.3 releases, it was decided that this |
| 899 | warning will not be converted to an error in Bison 2.5.] |
| 900 | |
| 901 | ** Detection of GNU M4 1.4.6 or newer during configure is improved. |
| 902 | |
| 903 | ** Warnings from gcc's -Wundef option about undefined YYENABLE_NLS, |
| 904 | YYLTYPE_IS_TRIVIAL, and __STRICT_ANSI__ in C/C++ parsers are now |
| 905 | avoided. |
| 906 | |
| 907 | ** %code is now a permanent feature. |
| 908 | |
| 909 | A traditional Yacc prologue directive is written in the form: |
| 910 | |
| 911 | %{CODE%} |
| 912 | |
| 913 | To provide a more flexible alternative, Bison 2.3b introduced the |
| 914 | %code directive with the following forms for C/C++: |
| 915 | |
| 916 | %code {CODE} |
| 917 | %code requires {CODE} |
| 918 | %code provides {CODE} |
| 919 | %code top {CODE} |
| 920 | |
| 921 | These forms are now considered permanent features of Bison. See the |
| 922 | %code entries in the section "Bison Declaration Summary" in the Bison |
| 923 | manual for a summary of their functionality. See the section |
| 924 | "Prologue Alternatives" for a detailed discussion including the |
| 925 | advantages of %code over the traditional Yacc prologue directive. |
| 926 | |
| 927 | Bison's Java feature as a whole including its current usage of %code |
| 928 | is still considered experimental. |
| 929 | |
| 930 | ** YYFAIL is deprecated and will eventually be removed. |
| 931 | |
| 932 | YYFAIL has existed for many years as an undocumented feature of |
| 933 | deterministic parsers in C generated by Bison. Previously, it was |
| 934 | documented for Bison's experimental Java parsers. YYFAIL is no longer |
| 935 | documented for Java parsers and is formally deprecated in both cases. |
| 936 | Users are strongly encouraged to migrate to YYERROR, which is |
| 937 | specified by POSIX. |
| 938 | |
| 939 | Like YYERROR, you can invoke YYFAIL from a semantic action in order to |
| 940 | induce a syntax error. The most obvious difference from YYERROR is |
| 941 | that YYFAIL will automatically invoke yyerror to report the syntax |
| 942 | error so that you don't have to. However, there are several other |
| 943 | subtle differences between YYERROR and YYFAIL, and YYFAIL suffers from |
| 944 | inherent flaws when %error-verbose or "#define YYERROR_VERBOSE" is |
| 945 | used. For a more detailed discussion, see: |
| 946 | |
| 947 | http://lists.gnu.org/archive/html/bison-patches/2009-12/msg00024.html |
| 948 | |
| 949 | The upcoming Bison 2.5 will remove YYFAIL from Java parsers, but |
| 950 | deterministic parsers in C will continue to implement it. However, |
| 951 | because YYFAIL is already flawed, it seems futile to try to make new |
| 952 | Bison features compatible with it. Thus, during parser generation, |
| 953 | Bison 2.5 will produce a warning whenever it discovers YYFAIL in a |
| 954 | rule action. In a later release, YYFAIL will be disabled for |
| 955 | %error-verbose and "#define YYERROR_VERBOSE". Eventually, YYFAIL will |
| 956 | be removed altogether. |
| 957 | |
| 958 | There exists at least one case where Bison 2.5's YYFAIL warning will |
| 959 | be a false positive. Some projects add phony uses of YYFAIL and other |
| 960 | Bison-defined macros for the sole purpose of suppressing C |
| 961 | preprocessor warnings (from GCC cpp's -Wunused-macros, for example). |
| 962 | To avoid Bison's future warning, such YYFAIL uses can be moved to the |
| 963 | epilogue (that is, after the second "%%") in the Bison input file. In |
| 964 | this release (2.4.2), Bison already generates its own code to suppress |
| 965 | C preprocessor warnings for YYFAIL, so projects can remove their own |
| 966 | phony uses of YYFAIL if compatibility with Bison releases prior to |
| 967 | 2.4.2 is not necessary. |
| 968 | |
| 969 | ** Internationalization. |
| 970 | |
| 971 | Fix a regression introduced in Bison 2.4: Under some circumstances, |
| 972 | message translations were not installed although supported by the |
| 973 | host system. |
| 974 | |
| 975 | * Changes in version 2.4.1 (2008-12-11): |
| 976 | |
| 977 | ** In the GLR defines file, unexpanded M4 macros in the yylval and yylloc |
| 978 | declarations have been fixed. |
| 979 | |
| 980 | ** Temporary hack for adding a semicolon to the user action. |
| 981 | |
| 982 | Bison used to prepend a trailing semicolon at the end of the user |
| 983 | action for reductions. This allowed actions such as |
| 984 | |
| 985 | exp: exp "+" exp { $$ = $1 + $3 }; |
| 986 | |
| 987 | instead of |
| 988 | |
| 989 | exp: exp "+" exp { $$ = $1 + $3; }; |
| 990 | |
| 991 | Some grammars still depend on this "feature". Bison 2.4.1 restores |
| 992 | the previous behavior in the case of C output (specifically, when |
| 993 | neither %language or %skeleton or equivalent command-line options |
| 994 | are used) to leave more time for grammars depending on the old |
| 995 | behavior to be adjusted. Future releases of Bison will disable this |
| 996 | feature. |
| 997 | |
| 998 | ** A few minor improvements to the Bison manual. |
| 999 | |
| 1000 | * Changes in version 2.4 (2008-11-02): |
| 1001 | |
| 1002 | ** %language is an experimental feature. |
| 1003 | |
| 1004 | We first introduced this feature in test release 2.3b as a cleaner |
| 1005 | alternative to %skeleton. Since then, we have discussed the possibility of |
| 1006 | modifying its effect on Bison's output file names. Thus, in this release, |
| 1007 | we consider %language to be an experimental feature that will likely evolve |
| 1008 | in future releases. |
| 1009 | |
| 1010 | ** Forward compatibility with GNU M4 has been improved. |
| 1011 | |
| 1012 | ** Several bugs in the C++ skeleton and the experimental Java skeleton have been |
| 1013 | fixed. |
| 1014 | |
| 1015 | * Changes in version 2.3b (2008-05-27): |
| 1016 | |
| 1017 | ** The quotes around NAME that used to be required in the following directive |
| 1018 | are now deprecated: |
| 1019 | |
| 1020 | %define NAME "VALUE" |
| 1021 | |
| 1022 | ** The directive "%pure-parser" is now deprecated in favor of: |
| 1023 | |
| 1024 | %define api.pure |
| 1025 | |
| 1026 | which has the same effect except that Bison is more careful to warn about |
| 1027 | unreasonable usage in the latter case. |
| 1028 | |
| 1029 | ** Push Parsing |
| 1030 | |
| 1031 | Bison can now generate an LALR(1) parser in C with a push interface. That |
| 1032 | is, instead of invoking "yyparse", which pulls tokens from "yylex", you can |
| 1033 | push one token at a time to the parser using "yypush_parse", which will |
| 1034 | return to the caller after processing each token. By default, the push |
| 1035 | interface is disabled. Either of the following directives will enable it: |
| 1036 | |
| 1037 | %define api.push_pull "push" // Just push; does not require yylex. |
| 1038 | %define api.push_pull "both" // Push and pull; requires yylex. |
| 1039 | |
| 1040 | See the new section "A Push Parser" in the Bison manual for details. |
| 1041 | |
| 1042 | The current push parsing interface is experimental and may evolve. More user |
| 1043 | feedback will help to stabilize it. |
| 1044 | |
| 1045 | ** The -g and --graph options now output graphs in Graphviz DOT format, |
| 1046 | not VCG format. Like --graph, -g now also takes an optional FILE argument |
| 1047 | and thus cannot be bundled with other short options. |
| 1048 | |
| 1049 | ** Java |
| 1050 | |
| 1051 | Bison can now generate an LALR(1) parser in Java. The skeleton is |
| 1052 | "data/lalr1.java". Consider using the new %language directive instead of |
| 1053 | %skeleton to select it. |
| 1054 | |
| 1055 | See the new section "Java Parsers" in the Bison manual for details. |
| 1056 | |
| 1057 | The current Java interface is experimental and may evolve. More user |
| 1058 | feedback will help to stabilize it. |
| 1059 | |
| 1060 | ** %language |
| 1061 | |
| 1062 | This new directive specifies the programming language of the generated |
| 1063 | parser, which can be C (the default), C++, or Java. Besides the skeleton |
| 1064 | that Bison uses, the directive affects the names of the generated files if |
| 1065 | the grammar file's name ends in ".y". |
| 1066 | |
| 1067 | ** XML Automaton Report |
| 1068 | |
| 1069 | Bison can now generate an XML report of the LALR(1) automaton using the new |
| 1070 | "--xml" option. The current XML schema is experimental and may evolve. More |
| 1071 | user feedback will help to stabilize it. |
| 1072 | |
| 1073 | ** The grammar file may now specify the name of the parser header file using |
| 1074 | %defines. For example: |
| 1075 | |
| 1076 | %defines "parser.h" |
| 1077 | |
| 1078 | ** When reporting useless rules, useless nonterminals, and unused terminals, |
| 1079 | Bison now employs the terms "useless in grammar" instead of "useless", |
| 1080 | "useless in parser" instead of "never reduced", and "unused in grammar" |
| 1081 | instead of "unused". |
| 1082 | |
| 1083 | ** Unreachable State Removal |
| 1084 | |
| 1085 | Previously, Bison sometimes generated parser tables containing unreachable |
| 1086 | states. A state can become unreachable during conflict resolution if Bison |
| 1087 | disables a shift action leading to it from a predecessor state. Bison now: |
| 1088 | |
| 1089 | 1. Removes unreachable states. |
| 1090 | |
| 1091 | 2. Does not report any conflicts that appeared in unreachable states. |
| 1092 | WARNING: As a result, you may need to update %expect and %expect-rr |
| 1093 | directives in existing grammar files. |
| 1094 | |
| 1095 | 3. For any rule used only in such states, Bison now reports the rule as |
| 1096 | "useless in parser due to conflicts". |
| 1097 | |
| 1098 | This feature can be disabled with the following directive: |
| 1099 | |
| 1100 | %define lr.keep_unreachable_states |
| 1101 | |
| 1102 | See the %define entry in the "Bison Declaration Summary" in the Bison manual |
| 1103 | for further discussion. |
| 1104 | |
| 1105 | ** Lookahead Set Correction in the ".output" Report |
| 1106 | |
| 1107 | When instructed to generate a ".output" file including lookahead sets |
| 1108 | (using "--report=lookahead", for example), Bison now prints each reduction's |
| 1109 | lookahead set only next to the associated state's one item that (1) is |
| 1110 | associated with the same rule as the reduction and (2) has its dot at the end |
| 1111 | of its RHS. Previously, Bison also erroneously printed the lookahead set |
| 1112 | next to all of the state's other items associated with the same rule. This |
| 1113 | bug affected only the ".output" file and not the generated parser source |
| 1114 | code. |
| 1115 | |
| 1116 | ** --report-file=FILE is a new option to override the default ".output" file |
| 1117 | name. |
| 1118 | |
| 1119 | ** The "=" that used to be required in the following directives is now |
| 1120 | deprecated: |
| 1121 | |
| 1122 | %file-prefix "parser" |
| 1123 | %name-prefix "c_" |
| 1124 | %output "parser.c" |
| 1125 | |
| 1126 | ** An Alternative to "%{...%}" -- "%code QUALIFIER {CODE}" |
| 1127 | |
| 1128 | Bison 2.3a provided a new set of directives as a more flexible alternative to |
| 1129 | the traditional Yacc prologue blocks. Those have now been consolidated into |
| 1130 | a single %code directive with an optional qualifier field, which identifies |
| 1131 | the purpose of the code and thus the location(s) where Bison should generate |
| 1132 | it: |
| 1133 | |
| 1134 | 1. "%code {CODE}" replaces "%after-header {CODE}" |
| 1135 | 2. "%code requires {CODE}" replaces "%start-header {CODE}" |
| 1136 | 3. "%code provides {CODE}" replaces "%end-header {CODE}" |
| 1137 | 4. "%code top {CODE}" replaces "%before-header {CODE}" |
| 1138 | |
| 1139 | See the %code entries in section "Bison Declaration Summary" in the Bison |
| 1140 | manual for a summary of the new functionality. See the new section "Prologue |
| 1141 | Alternatives" for a detailed discussion including the advantages of %code |
| 1142 | over the traditional Yacc prologues. |
| 1143 | |
| 1144 | The prologue alternatives are experimental. More user feedback will help to |
| 1145 | determine whether they should become permanent features. |
| 1146 | |
| 1147 | ** Revised warning: unset or unused mid-rule values |
| 1148 | |
| 1149 | Since Bison 2.2, Bison has warned about mid-rule values that are set but not |
| 1150 | used within any of the actions of the parent rule. For example, Bison warns |
| 1151 | about unused $2 in: |
| 1152 | |
| 1153 | exp: '1' { $$ = 1; } '+' exp { $$ = $1 + $4; }; |
| 1154 | |
| 1155 | Now, Bison also warns about mid-rule values that are used but not set. For |
| 1156 | example, Bison warns about unset $$ in the mid-rule action in: |
| 1157 | |
| 1158 | exp: '1' { $1 = 1; } '+' exp { $$ = $2 + $4; }; |
| 1159 | |
| 1160 | However, Bison now disables both of these warnings by default since they |
| 1161 | sometimes prove to be false alarms in existing grammars employing the Yacc |
| 1162 | constructs $0 or $-N (where N is some positive integer). |
| 1163 | |
| 1164 | To enable these warnings, specify the option "--warnings=midrule-values" or |
| 1165 | "-W", which is a synonym for "--warnings=all". |
| 1166 | |
| 1167 | ** Default %destructor or %printer with "<*>" or "<>" |
| 1168 | |
| 1169 | Bison now recognizes two separate kinds of default %destructor's and |
| 1170 | %printer's: |
| 1171 | |
| 1172 | 1. Place "<*>" in a %destructor/%printer symbol list to define a default |
| 1173 | %destructor/%printer for all grammar symbols for which you have formally |
| 1174 | declared semantic type tags. |
| 1175 | |
| 1176 | 2. Place "<>" in a %destructor/%printer symbol list to define a default |
| 1177 | %destructor/%printer for all grammar symbols without declared semantic |
| 1178 | type tags. |
| 1179 | |
| 1180 | Bison no longer supports the "%symbol-default" notation from Bison 2.3a. |
| 1181 | "<*>" and "<>" combined achieve the same effect with one exception: Bison no |
| 1182 | longer applies any %destructor to a mid-rule value if that mid-rule value is |
| 1183 | not actually ever referenced using either $$ or $n in a semantic action. |
| 1184 | |
| 1185 | The default %destructor's and %printer's are experimental. More user |
| 1186 | feedback will help to determine whether they should become permanent |
| 1187 | features. |
| 1188 | |
| 1189 | See the section "Freeing Discarded Symbols" in the Bison manual for further |
| 1190 | details. |
| 1191 | |
| 1192 | ** %left, %right, and %nonassoc can now declare token numbers. This is required |
| 1193 | by POSIX. However, see the end of section "Operator Precedence" in the Bison |
| 1194 | manual for a caveat concerning the treatment of literal strings. |
| 1195 | |
| 1196 | ** The nonfunctional --no-parser, -n, and %no-parser options have been |
| 1197 | completely removed from Bison. |
| 1198 | |
| 1199 | * Changes in version 2.3a, 2006-09-13: |
| 1200 | |
| 1201 | ** Instead of %union, you can define and use your own union type |
| 1202 | YYSTYPE if your grammar contains at least one <type> tag. |
| 1203 | Your YYSTYPE need not be a macro; it can be a typedef. |
| 1204 | This change is for compatibility with other Yacc implementations, |
| 1205 | and is required by POSIX. |
| 1206 | |
| 1207 | ** Locations columns and lines start at 1. |
| 1208 | In accordance with the GNU Coding Standards and Emacs. |
| 1209 | |
| 1210 | ** You may now declare per-type and default %destructor's and %printer's: |
| 1211 | |
| 1212 | For example: |
| 1213 | |
| 1214 | %union { char *string; } |
| 1215 | %token <string> STRING1 |
| 1216 | %token <string> STRING2 |
| 1217 | %type <string> string1 |
| 1218 | %type <string> string2 |
| 1219 | %union { char character; } |
| 1220 | %token <character> CHR |
| 1221 | %type <character> chr |
| 1222 | %destructor { free ($$); } %symbol-default |
| 1223 | %destructor { free ($$); printf ("%d", @$.first_line); } STRING1 string1 |
| 1224 | %destructor { } <character> |
| 1225 | |
| 1226 | guarantees that, when the parser discards any user-defined symbol that has a |
| 1227 | semantic type tag other than "<character>", it passes its semantic value to |
| 1228 | "free". However, when the parser discards a "STRING1" or a "string1", it |
| 1229 | also prints its line number to "stdout". It performs only the second |
| 1230 | "%destructor" in this case, so it invokes "free" only once. |
| 1231 | |
| 1232 | [Although we failed to mention this here in the 2.3a release, the default |
| 1233 | %destructor's and %printer's were experimental, and they were rewritten in |
| 1234 | future versions.] |
| 1235 | |
| 1236 | ** Except for LALR(1) parsers in C with POSIX Yacc emulation enabled (with "-y", |
| 1237 | "--yacc", or "%yacc"), Bison no longer generates #define statements for |
| 1238 | associating token numbers with token names. Removing the #define statements |
| 1239 | helps to sanitize the global namespace during preprocessing, but POSIX Yacc |
| 1240 | requires them. Bison still generates an enum for token names in all cases. |
| 1241 | |
| 1242 | ** Handling of traditional Yacc prologue blocks is now more consistent but |
| 1243 | potentially incompatible with previous releases of Bison. |
| 1244 | |
| 1245 | As before, you declare prologue blocks in your grammar file with the |
| 1246 | "%{ ... %}" syntax. To generate the pre-prologue, Bison concatenates all |
| 1247 | prologue blocks that you've declared before the first %union. To generate |
| 1248 | the post-prologue, Bison concatenates all prologue blocks that you've |
| 1249 | declared after the first %union. |
| 1250 | |
| 1251 | Previous releases of Bison inserted the pre-prologue into both the header |
| 1252 | file and the code file in all cases except for LALR(1) parsers in C. In the |
| 1253 | latter case, Bison inserted it only into the code file. For parsers in C++, |
| 1254 | the point of insertion was before any token definitions (which associate |
| 1255 | token numbers with names). For parsers in C, the point of insertion was |
| 1256 | after the token definitions. |
| 1257 | |
| 1258 | Now, Bison never inserts the pre-prologue into the header file. In the code |
| 1259 | file, it always inserts it before the token definitions. |
| 1260 | |
| 1261 | ** Bison now provides a more flexible alternative to the traditional Yacc |
| 1262 | prologue blocks: %before-header, %start-header, %end-header, and |
| 1263 | %after-header. |
| 1264 | |
| 1265 | For example, the following declaration order in the grammar file reflects the |
| 1266 | order in which Bison will output these code blocks. However, you are free to |
| 1267 | declare these code blocks in your grammar file in whatever order is most |
| 1268 | convenient for you: |
| 1269 | |
| 1270 | %before-header { |
| 1271 | /* Bison treats this block like a pre-prologue block: it inserts it into |
| 1272 | * the code file before the contents of the header file. It does *not* |
| 1273 | * insert it into the header file. This is a good place to put |
| 1274 | * #include's that you want at the top of your code file. A common |
| 1275 | * example is '#include "system.h"'. */ |
| 1276 | } |
| 1277 | %start-header { |
| 1278 | /* Bison inserts this block into both the header file and the code file. |
| 1279 | * In both files, the point of insertion is before any Bison-generated |
| 1280 | * token, semantic type, location type, and class definitions. This is a |
| 1281 | * good place to define %union dependencies, for example. */ |
| 1282 | } |
| 1283 | %union { |
| 1284 | /* Unlike the traditional Yacc prologue blocks, the output order for the |
| 1285 | * new %*-header blocks is not affected by their declaration position |
| 1286 | * relative to any %union in the grammar file. */ |
| 1287 | } |
| 1288 | %end-header { |
| 1289 | /* Bison inserts this block into both the header file and the code file. |
| 1290 | * In both files, the point of insertion is after the Bison-generated |
| 1291 | * definitions. This is a good place to declare or define public |
| 1292 | * functions or data structures that depend on the Bison-generated |
| 1293 | * definitions. */ |
| 1294 | } |
| 1295 | %after-header { |
| 1296 | /* Bison treats this block like a post-prologue block: it inserts it into |
| 1297 | * the code file after the contents of the header file. It does *not* |
| 1298 | * insert it into the header file. This is a good place to declare or |
| 1299 | * define internal functions or data structures that depend on the |
| 1300 | * Bison-generated definitions. */ |
| 1301 | } |
| 1302 | |
| 1303 | If you have multiple occurrences of any one of the above declarations, Bison |
| 1304 | will concatenate the contents in declaration order. |
| 1305 | |
| 1306 | [Although we failed to mention this here in the 2.3a release, the prologue |
| 1307 | alternatives were experimental, and they were rewritten in future versions.] |
| 1308 | |
| 1309 | ** The option "--report=look-ahead" has been changed to "--report=lookahead". |
| 1310 | The old spelling still works, but is not documented and may be removed |
| 1311 | in a future release. |
| 1312 | |
| 1313 | * Changes in version 2.3, 2006-06-05: |
| 1314 | |
| 1315 | ** GLR grammars should now use "YYRECOVERING ()" instead of "YYRECOVERING", |
| 1316 | for compatibility with LALR(1) grammars. |
| 1317 | |
| 1318 | ** It is now documented that any definition of YYSTYPE or YYLTYPE should |
| 1319 | be to a type name that does not contain parentheses or brackets. |
| 1320 | |
| 1321 | * Changes in version 2.2, 2006-05-19: |
| 1322 | |
| 1323 | ** The distribution terms for all Bison-generated parsers now permit |
| 1324 | using the parsers in nonfree programs. Previously, this permission |
| 1325 | was granted only for Bison-generated LALR(1) parsers in C. |
| 1326 | |
| 1327 | ** %name-prefix changes the namespace name in C++ outputs. |
| 1328 | |
| 1329 | ** The C++ parsers export their token_type. |
| 1330 | |
| 1331 | ** Bison now allows multiple %union declarations, and concatenates |
| 1332 | their contents together. |
| 1333 | |
| 1334 | ** New warning: unused values |
| 1335 | Right-hand side symbols whose values are not used are reported, |
| 1336 | if the symbols have destructors. For instance: |
| 1337 | |
| 1338 | exp: exp "?" exp ":" exp { $1 ? $1 : $3; } |
| 1339 | | exp "+" exp |
| 1340 | ; |
| 1341 | |
| 1342 | will trigger a warning about $$ and $5 in the first rule, and $3 in |
| 1343 | the second ($1 is copied to $$ by the default rule). This example |
| 1344 | most likely contains three errors, and could be rewritten as: |
| 1345 | |
| 1346 | exp: exp "?" exp ":" exp |
| 1347 | { $$ = $1 ? $3 : $5; free ($1 ? $5 : $3); free ($1); } |
| 1348 | | exp "+" exp |
| 1349 | { $$ = $1 ? $1 : $3; if ($1) free ($3); } |
| 1350 | ; |
| 1351 | |
| 1352 | However, if the original actions were really intended, memory leaks |
| 1353 | and all, the warnings can be suppressed by letting Bison believe the |
| 1354 | values are used, e.g.: |
| 1355 | |
| 1356 | exp: exp "?" exp ":" exp { $1 ? $1 : $3; (void) ($$, $5); } |
| 1357 | | exp "+" exp { $$ = $1; (void) $3; } |
| 1358 | ; |
| 1359 | |
| 1360 | If there are mid-rule actions, the warning is issued if no action |
| 1361 | uses it. The following triggers no warning: $1 and $3 are used. |
| 1362 | |
| 1363 | exp: exp { push ($1); } '+' exp { push ($3); sum (); }; |
| 1364 | |
| 1365 | The warning is intended to help catching lost values and memory leaks. |
| 1366 | If a value is ignored, its associated memory typically is not reclaimed. |
| 1367 | |
| 1368 | ** %destructor vs. YYABORT, YYACCEPT, and YYERROR. |
| 1369 | Destructors are now called when user code invokes YYABORT, YYACCEPT, |
| 1370 | and YYERROR, for all objects on the stack, other than objects |
| 1371 | corresponding to the right-hand side of the current rule. |
| 1372 | |
| 1373 | ** %expect, %expect-rr |
| 1374 | Incorrect numbers of expected conflicts are now actual errors, |
| 1375 | instead of warnings. |
| 1376 | |
| 1377 | ** GLR, YACC parsers. |
| 1378 | The %parse-params are available in the destructors (and the |
| 1379 | experimental printers) as per the documentation. |
| 1380 | |
| 1381 | ** Bison now warns if it finds a stray "$" or "@" in an action. |
| 1382 | |
| 1383 | ** %require "VERSION" |
| 1384 | This specifies that the grammar file depends on features implemented |
| 1385 | in Bison version VERSION or higher. |
| 1386 | |
| 1387 | ** lalr1.cc: The token and value types are now class members. |
| 1388 | The tokens were defined as free form enums and cpp macros. YYSTYPE |
| 1389 | was defined as a free form union. They are now class members: |
| 1390 | tokens are enumerations of the "yy::parser::token" struct, and the |
| 1391 | semantic values have the "yy::parser::semantic_type" type. |
| 1392 | |
| 1393 | If you do not want or can update to this scheme, the directive |
| 1394 | '%define "global_tokens_and_yystype" "1"' triggers the global |
| 1395 | definition of tokens and YYSTYPE. This change is suitable both |
| 1396 | for previous releases of Bison, and this one. |
| 1397 | |
| 1398 | If you wish to update, then make sure older version of Bison will |
| 1399 | fail using '%require "2.2"'. |
| 1400 | |
| 1401 | ** DJGPP support added. |
| 1402 | \f |
| 1403 | * Changes in version 2.1, 2005-09-16: |
| 1404 | |
| 1405 | ** The C++ lalr1.cc skeleton supports %lex-param. |
| 1406 | |
| 1407 | ** Bison-generated parsers now support the translation of diagnostics like |
| 1408 | "syntax error" into languages other than English. The default |
| 1409 | language is still English. For details, please see the new |
| 1410 | Internationalization section of the Bison manual. Software |
| 1411 | distributors should also see the new PACKAGING file. Thanks to |
| 1412 | Bruno Haible for this new feature. |
| 1413 | |
| 1414 | ** Wording in the Bison-generated parsers has been changed slightly to |
| 1415 | simplify translation. In particular, the message "memory exhausted" |
| 1416 | has replaced "parser stack overflow", as the old message was not |
| 1417 | always accurate for modern Bison-generated parsers. |
| 1418 | |
| 1419 | ** Destructors are now called when the parser aborts, for all symbols left |
| 1420 | behind on the stack. Also, the start symbol is now destroyed after a |
| 1421 | successful parse. In both cases, the behavior was formerly inconsistent. |
| 1422 | |
| 1423 | ** When generating verbose diagnostics, Bison-generated parsers no longer |
| 1424 | quote the literal strings associated with tokens. For example, for |
| 1425 | a syntax error associated with '%token NUM "number"' they might |
| 1426 | print 'syntax error, unexpected number' instead of 'syntax error, |
| 1427 | unexpected "number"'. |
| 1428 | \f |
| 1429 | * Changes in version 2.0, 2004-12-25: |
| 1430 | |
| 1431 | ** Possibly-incompatible changes |
| 1432 | |
| 1433 | - Bison-generated parsers no longer default to using the alloca function |
| 1434 | (when available) to extend the parser stack, due to widespread |
| 1435 | problems in unchecked stack-overflow detection. You can "#define |
| 1436 | YYSTACK_USE_ALLOCA 1" to require the use of alloca, but please read |
| 1437 | the manual to determine safe values for YYMAXDEPTH in that case. |
| 1438 | |
| 1439 | - Error token location. |
| 1440 | During error recovery, the location of the syntax error is updated |
| 1441 | to cover the whole sequence covered by the error token: it includes |
| 1442 | the shifted symbols thrown away during the first part of the error |
| 1443 | recovery, and the lookahead rejected during the second part. |
| 1444 | |
| 1445 | - Semicolon changes: |
| 1446 | . Stray semicolons are no longer allowed at the start of a grammar. |
| 1447 | . Semicolons are now required after in-grammar declarations. |
| 1448 | |
| 1449 | - Unescaped newlines are no longer allowed in character constants or |
| 1450 | string literals. They were never portable, and GCC 3.4.0 has |
| 1451 | dropped support for them. Better diagnostics are now generated if |
| 1452 | forget a closing quote. |
| 1453 | |
| 1454 | - NUL bytes are no longer allowed in Bison string literals, unfortunately. |
| 1455 | |
| 1456 | ** New features |
| 1457 | |
| 1458 | - GLR grammars now support locations. |
| 1459 | |
| 1460 | - New directive: %initial-action. |
| 1461 | This directive allows the user to run arbitrary code (including |
| 1462 | initializing @$) from yyparse before parsing starts. |
| 1463 | |
| 1464 | - A new directive "%expect-rr N" specifies the expected number of |
| 1465 | reduce/reduce conflicts in GLR parsers. |
| 1466 | |
| 1467 | - %token numbers can now be hexadecimal integers, e.g., "%token FOO 0x12d". |
| 1468 | This is a GNU extension. |
| 1469 | |
| 1470 | - The option "--report=lookahead" was changed to "--report=look-ahead". |
| 1471 | [However, this was changed back after 2.3.] |
| 1472 | |
| 1473 | - Experimental %destructor support has been added to lalr1.cc. |
| 1474 | |
| 1475 | - New configure option --disable-yacc, to disable installation of the |
| 1476 | yacc command and -ly library introduced in 1.875 for POSIX conformance. |
| 1477 | |
| 1478 | ** Bug fixes |
| 1479 | |
| 1480 | - For now, %expect-count violations are now just warnings, not errors. |
| 1481 | This is for compatibility with Bison 1.75 and earlier (when there are |
| 1482 | reduce/reduce conflicts) and with Bison 1.30 and earlier (when there |
| 1483 | are too many or too few shift/reduce conflicts). However, in future |
| 1484 | versions of Bison we plan to improve the %expect machinery so that |
| 1485 | these violations will become errors again. |
| 1486 | |
| 1487 | - Within Bison itself, numbers (e.g., goto numbers) are no longer |
| 1488 | arbitrarily limited to 16-bit counts. |
| 1489 | |
| 1490 | - Semicolons are now allowed before "|" in grammar rules, as POSIX requires. |
| 1491 | \f |
| 1492 | * Changes in version 1.875, 2003-01-01: |
| 1493 | |
| 1494 | ** The documentation license has been upgraded to version 1.2 |
| 1495 | of the GNU Free Documentation License. |
| 1496 | |
| 1497 | ** syntax error processing |
| 1498 | |
| 1499 | - In Yacc-style parsers YYLLOC_DEFAULT is now used to compute error |
| 1500 | locations too. This fixes bugs in error-location computation. |
| 1501 | |
| 1502 | - %destructor |
| 1503 | It is now possible to reclaim the memory associated to symbols |
| 1504 | discarded during error recovery. This feature is still experimental. |
| 1505 | |
| 1506 | - %error-verbose |
| 1507 | This new directive is preferred over YYERROR_VERBOSE. |
| 1508 | |
| 1509 | - #defining yyerror to steal internal variables is discouraged. |
| 1510 | It is not guaranteed to work forever. |
| 1511 | |
| 1512 | ** POSIX conformance |
| 1513 | |
| 1514 | - Semicolons are once again optional at the end of grammar rules. |
| 1515 | This reverts to the behavior of Bison 1.33 and earlier, and improves |
| 1516 | compatibility with Yacc. |
| 1517 | |
| 1518 | - "parse error" -> "syntax error" |
| 1519 | Bison now uniformly uses the term "syntax error"; formerly, the code |
| 1520 | and manual sometimes used the term "parse error" instead. POSIX |
| 1521 | requires "syntax error" in diagnostics, and it was thought better to |
| 1522 | be consistent. |
| 1523 | |
| 1524 | - The documentation now emphasizes that yylex and yyerror must be |
| 1525 | declared before use. C99 requires this. |
| 1526 | |
| 1527 | - Bison now parses C99 lexical constructs like UCNs and |
| 1528 | backslash-newline within C escape sequences, as POSIX 1003.1-2001 requires. |
| 1529 | |
| 1530 | - File names are properly escaped in C output. E.g., foo\bar.y is |
| 1531 | output as "foo\\bar.y". |
| 1532 | |
| 1533 | - Yacc command and library now available |
| 1534 | The Bison distribution now installs a "yacc" command, as POSIX requires. |
| 1535 | Also, Bison now installs a small library liby.a containing |
| 1536 | implementations of Yacc-compatible yyerror and main functions. |
| 1537 | This library is normally not useful, but POSIX requires it. |
| 1538 | |
| 1539 | - Type clashes now generate warnings, not errors. |
| 1540 | |
| 1541 | - If the user does not define YYSTYPE as a macro, Bison now declares it |
| 1542 | using typedef instead of defining it as a macro. |
| 1543 | For consistency, YYLTYPE is also declared instead of defined. |
| 1544 | |
| 1545 | ** Other compatibility issues |
| 1546 | |
| 1547 | - %union directives can now have a tag before the "{", e.g., the |
| 1548 | directive "%union foo {...}" now generates the C code |
| 1549 | "typedef union foo { ... } YYSTYPE;"; this is for Yacc compatibility. |
| 1550 | The default union tag is "YYSTYPE", for compatibility with Solaris 9 Yacc. |
| 1551 | For consistency, YYLTYPE's struct tag is now "YYLTYPE" not "yyltype". |
| 1552 | This is for compatibility with both Yacc and Bison 1.35. |
| 1553 | |
| 1554 | - ";" is output before the terminating "}" of an action, for |
| 1555 | compatibility with Bison 1.35. |
| 1556 | |
| 1557 | - Bison now uses a Yacc-style format for conflict reports, e.g., |
| 1558 | "conflicts: 2 shift/reduce, 1 reduce/reduce". |
| 1559 | |
| 1560 | - "yystype" and "yyltype" are now obsolescent macros instead of being |
| 1561 | typedefs or tags; they are no longer documented and are planned to be |
| 1562 | withdrawn in a future release. |
| 1563 | |
| 1564 | ** GLR parser notes |
| 1565 | |
| 1566 | - GLR and inline |
| 1567 | Users of Bison have to decide how they handle the portability of the |
| 1568 | C keyword "inline". |
| 1569 | |
| 1570 | - "parsing stack overflow..." -> "parser stack overflow" |
| 1571 | GLR parsers now report "parser stack overflow" as per the Bison manual. |
| 1572 | |
| 1573 | ** %parse-param and %lex-param |
| 1574 | The macros YYPARSE_PARAM and YYLEX_PARAM provide a means to pass |
| 1575 | additional context to yyparse and yylex. They suffer from several |
| 1576 | shortcomings: |
| 1577 | |
| 1578 | - a single argument only can be added, |
| 1579 | - their types are weak (void *), |
| 1580 | - this context is not passed to ancillary functions such as yyerror, |
| 1581 | - only yacc.c parsers support them. |
| 1582 | |
| 1583 | The new %parse-param/%lex-param directives provide a more precise control. |
| 1584 | For instance: |
| 1585 | |
| 1586 | %parse-param {int *nastiness} |
| 1587 | %lex-param {int *nastiness} |
| 1588 | %parse-param {int *randomness} |
| 1589 | |
| 1590 | results in the following signatures: |
| 1591 | |
| 1592 | int yylex (int *nastiness); |
| 1593 | int yyparse (int *nastiness, int *randomness); |
| 1594 | |
| 1595 | or, if both %pure-parser and %locations are used: |
| 1596 | |
| 1597 | int yylex (YYSTYPE *lvalp, YYLTYPE *llocp, int *nastiness); |
| 1598 | int yyparse (int *nastiness, int *randomness); |
| 1599 | |
| 1600 | ** Bison now warns if it detects conflicting outputs to the same file, |
| 1601 | e.g., it generates a warning for "bison -d -o foo.h foo.y" since |
| 1602 | that command outputs both code and header to foo.h. |
| 1603 | |
| 1604 | ** #line in output files |
| 1605 | - --no-line works properly. |
| 1606 | |
| 1607 | ** Bison can no longer be built by a K&R C compiler; it requires C89 or |
| 1608 | later to be built. This change originally took place a few versions |
| 1609 | ago, but nobody noticed until we recently asked someone to try |
| 1610 | building Bison with a K&R C compiler. |
| 1611 | \f |
| 1612 | * Changes in version 1.75, 2002-10-14: |
| 1613 | |
| 1614 | ** Bison should now work on 64-bit hosts. |
| 1615 | |
| 1616 | ** Indonesian translation thanks to Tedi Heriyanto. |
| 1617 | |
| 1618 | ** GLR parsers |
| 1619 | Fix spurious parse errors. |
| 1620 | |
| 1621 | ** Pure parsers |
| 1622 | Some people redefine yyerror to steal yyparse' private variables. |
| 1623 | Reenable this trick until an official feature replaces it. |
| 1624 | |
| 1625 | ** Type Clashes |
| 1626 | In agreement with POSIX and with other Yaccs, leaving a default |
| 1627 | action is valid when $$ is untyped, and $1 typed: |
| 1628 | |
| 1629 | untyped: ... typed; |
| 1630 | |
| 1631 | but the converse remains an error: |
| 1632 | |
| 1633 | typed: ... untyped; |
| 1634 | |
| 1635 | ** Values of mid-rule actions |
| 1636 | The following code: |
| 1637 | |
| 1638 | foo: { ... } { $$ = $1; } ... |
| 1639 | |
| 1640 | was incorrectly rejected: $1 is defined in the second mid-rule |
| 1641 | action, and is equal to the $$ of the first mid-rule action. |
| 1642 | \f |
| 1643 | * Changes in version 1.50, 2002-10-04: |
| 1644 | |
| 1645 | ** GLR parsing |
| 1646 | The declaration |
| 1647 | %glr-parser |
| 1648 | causes Bison to produce a Generalized LR (GLR) parser, capable of handling |
| 1649 | almost any context-free grammar, ambiguous or not. The new declarations |
| 1650 | %dprec and %merge on grammar rules allow parse-time resolution of |
| 1651 | ambiguities. Contributed by Paul Hilfinger. |
| 1652 | |
| 1653 | Unfortunately Bison 1.50 does not work properly on 64-bit hosts |
| 1654 | like the Alpha, so please stick to 32-bit hosts for now. |
| 1655 | |
| 1656 | ** Output Directory |
| 1657 | When not in Yacc compatibility mode, when the output file was not |
| 1658 | specified, running "bison foo/bar.y" created "foo/bar.c". It |
| 1659 | now creates "bar.c". |
| 1660 | |
| 1661 | ** Undefined token |
| 1662 | The undefined token was systematically mapped to 2 which prevented |
| 1663 | the use of 2 by the user. This is no longer the case. |
| 1664 | |
| 1665 | ** Unknown token numbers |
| 1666 | If yylex returned an out of range value, yyparse could die. This is |
| 1667 | no longer the case. |
| 1668 | |
| 1669 | ** Error token |
| 1670 | According to POSIX, the error token must be 256. |
| 1671 | Bison extends this requirement by making it a preference: *if* the |
| 1672 | user specified that one of her tokens is numbered 256, then error |
| 1673 | will be mapped onto another number. |
| 1674 | |
| 1675 | ** Verbose error messages |
| 1676 | They no longer report "..., expecting error or..." for states where |
| 1677 | error recovery is possible. |
| 1678 | |
| 1679 | ** End token |
| 1680 | Defaults to "$end" instead of "$". |
| 1681 | |
| 1682 | ** Error recovery now conforms to documentation and to POSIX |
| 1683 | When a Bison-generated parser encounters a syntax error, it now pops |
| 1684 | the stack until it finds a state that allows shifting the error |
| 1685 | token. Formerly, it popped the stack until it found a state that |
| 1686 | allowed some non-error action other than a default reduction on the |
| 1687 | error token. The new behavior has long been the documented behavior, |
| 1688 | and has long been required by POSIX. For more details, please see |
| 1689 | Paul Eggert, "Reductions during Bison error handling" (2002-05-20) |
| 1690 | <http://lists.gnu.org/archive/html/bug-bison/2002-05/msg00038.html>. |
| 1691 | |
| 1692 | ** Traces |
| 1693 | Popped tokens and nonterminals are now reported. |
| 1694 | |
| 1695 | ** Larger grammars |
| 1696 | Larger grammars are now supported (larger token numbers, larger grammar |
| 1697 | size (= sum of the LHS and RHS lengths), larger LALR tables). |
| 1698 | Formerly, many of these numbers ran afoul of 16-bit limits; |
| 1699 | now these limits are 32 bits on most hosts. |
| 1700 | |
| 1701 | ** Explicit initial rule |
| 1702 | Bison used to play hacks with the initial rule, which the user does |
| 1703 | not write. It is now explicit, and visible in the reports and |
| 1704 | graphs as rule 0. |
| 1705 | |
| 1706 | ** Useless rules |
| 1707 | Before, Bison reported the useless rules, but, although not used, |
| 1708 | included them in the parsers. They are now actually removed. |
| 1709 | |
| 1710 | ** Useless rules, useless nonterminals |
| 1711 | They are now reported, as a warning, with their locations. |
| 1712 | |
| 1713 | ** Rules never reduced |
| 1714 | Rules that can never be reduced because of conflicts are now |
| 1715 | reported. |
| 1716 | |
| 1717 | ** Incorrect "Token not used" |
| 1718 | On a grammar such as |
| 1719 | |
| 1720 | %token useless useful |
| 1721 | %% |
| 1722 | exp: '0' %prec useful; |
| 1723 | |
| 1724 | where a token was used to set the precedence of the last rule, |
| 1725 | bison reported both "useful" and "useless" as useless tokens. |
| 1726 | |
| 1727 | ** Revert the C++ namespace changes introduced in 1.31 |
| 1728 | as they caused too many portability hassles. |
| 1729 | |
| 1730 | ** Default locations |
| 1731 | By an accident of design, the default computation of @$ was |
| 1732 | performed after another default computation was performed: @$ = @1. |
| 1733 | The latter is now removed: YYLLOC_DEFAULT is fully responsible of |
| 1734 | the computation of @$. |
| 1735 | |
| 1736 | ** Token end-of-file |
| 1737 | The token end of file may be specified by the user, in which case, |
| 1738 | the user symbol is used in the reports, the graphs, and the verbose |
| 1739 | error messages instead of "$end", which remains being the default. |
| 1740 | For instance |
| 1741 | %token MYEOF 0 |
| 1742 | or |
| 1743 | %token MYEOF 0 "end of file" |
| 1744 | |
| 1745 | ** Semantic parser |
| 1746 | This old option, which has been broken for ages, is removed. |
| 1747 | |
| 1748 | ** New translations |
| 1749 | Brazilian Portuguese, thanks to Alexandre Folle de Menezes. |
| 1750 | Croatian, thanks to Denis Lackovic. |
| 1751 | |
| 1752 | ** Incorrect token definitions |
| 1753 | When given |
| 1754 | %token 'a' "A" |
| 1755 | bison used to output |
| 1756 | #define 'a' 65 |
| 1757 | |
| 1758 | ** Token definitions as enums |
| 1759 | Tokens are output both as the traditional #define's, and, provided |
| 1760 | the compiler supports ANSI C or is a C++ compiler, as enums. |
| 1761 | This lets debuggers display names instead of integers. |
| 1762 | |
| 1763 | ** Reports |
| 1764 | In addition to --verbose, bison supports --report=THINGS, which |
| 1765 | produces additional information: |
| 1766 | - itemset |
| 1767 | complete the core item sets with their closure |
| 1768 | - lookahead [changed to "look-ahead" in 1.875e through 2.3, but changed back] |
| 1769 | explicitly associate lookahead tokens to items |
| 1770 | - solved |
| 1771 | describe shift/reduce conflicts solving. |
| 1772 | Bison used to systematically output this information on top of |
| 1773 | the report. Solved conflicts are now attached to their states. |
| 1774 | |
| 1775 | ** Type clashes |
| 1776 | Previous versions don't complain when there is a type clash on |
| 1777 | the default action if the rule has a mid-rule action, such as in: |
| 1778 | |
| 1779 | %type <foo> bar |
| 1780 | %% |
| 1781 | bar: '0' {} '0'; |
| 1782 | |
| 1783 | This is fixed. |
| 1784 | |
| 1785 | ** GNU M4 is now required when using Bison. |
| 1786 | \f |
| 1787 | * Changes in version 1.35, 2002-03-25: |
| 1788 | |
| 1789 | ** C Skeleton |
| 1790 | Some projects use Bison's C parser with C++ compilers, and define |
| 1791 | YYSTYPE as a class. The recent adjustment of C parsers for data |
| 1792 | alignment and 64 bit architectures made this impossible. |
| 1793 | |
| 1794 | Because for the time being no real solution for C++ parser |
| 1795 | generation exists, kludges were implemented in the parser to |
| 1796 | maintain this use. In the future, when Bison has C++ parsers, this |
| 1797 | kludge will be disabled. |
| 1798 | |
| 1799 | This kludge also addresses some C++ problems when the stack was |
| 1800 | extended. |
| 1801 | \f |
| 1802 | * Changes in version 1.34, 2002-03-12: |
| 1803 | |
| 1804 | ** File name clashes are detected |
| 1805 | $ bison foo.y -d -o foo.x |
| 1806 | fatal error: header and parser would both be named "foo.x" |
| 1807 | |
| 1808 | ** A missing ";" at the end of a rule triggers a warning |
| 1809 | In accordance with POSIX, and in agreement with other |
| 1810 | Yacc implementations, Bison will mandate this semicolon in the near |
| 1811 | future. This eases the implementation of a Bison parser of Bison |
| 1812 | grammars by making this grammar LALR(1) instead of LR(2). To |
| 1813 | facilitate the transition, this release introduces a warning. |
| 1814 | |
| 1815 | ** Revert the C++ namespace changes introduced in 1.31, as they caused too |
| 1816 | many portability hassles. |
| 1817 | |
| 1818 | ** DJGPP support added. |
| 1819 | |
| 1820 | ** Fix test suite portability problems. |
| 1821 | \f |
| 1822 | * Changes in version 1.33, 2002-02-07: |
| 1823 | |
| 1824 | ** Fix C++ issues |
| 1825 | Groff could not be compiled for the definition of size_t was lacking |
| 1826 | under some conditions. |
| 1827 | |
| 1828 | ** Catch invalid @n |
| 1829 | As is done with $n. |
| 1830 | \f |
| 1831 | * Changes in version 1.32, 2002-01-23: |
| 1832 | |
| 1833 | ** Fix Yacc output file names |
| 1834 | |
| 1835 | ** Portability fixes |
| 1836 | |
| 1837 | ** Italian, Dutch translations |
| 1838 | \f |
| 1839 | * Changes in version 1.31, 2002-01-14: |
| 1840 | |
| 1841 | ** Many Bug Fixes |
| 1842 | |
| 1843 | ** GNU Gettext and %expect |
| 1844 | GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that |
| 1845 | Bison dies on incorrect %expectations, we fear there will be |
| 1846 | too many bug reports for Gettext, so _for the time being_, %expect |
| 1847 | does not trigger an error when the input file is named "plural.y". |
| 1848 | |
| 1849 | ** Use of alloca in parsers |
| 1850 | If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use |
| 1851 | malloc exclusively. Since 1.29, but was not NEWS'ed. |
| 1852 | |
| 1853 | alloca is used only when compiled with GCC, to avoid portability |
| 1854 | problems as on AIX. |
| 1855 | |
| 1856 | ** yyparse now returns 2 if memory is exhausted; formerly it dumped core. |
| 1857 | |
| 1858 | ** When the generated parser lacks debugging code, YYDEBUG is now 0 |
| 1859 | (as POSIX requires) instead of being undefined. |
| 1860 | |
| 1861 | ** User Actions |
| 1862 | Bison has always permitted actions such as { $$ = $1 }: it adds the |
| 1863 | ending semicolon. Now if in Yacc compatibility mode, the semicolon |
| 1864 | is no longer output: one has to write { $$ = $1; }. |
| 1865 | |
| 1866 | ** Better C++ compliance |
| 1867 | The output parsers try to respect C++ namespaces. |
| 1868 | [This turned out to be a failed experiment, and it was reverted later.] |
| 1869 | |
| 1870 | ** Reduced Grammars |
| 1871 | Fixed bugs when reporting useless nonterminals. |
| 1872 | |
| 1873 | ** 64 bit hosts |
| 1874 | The parsers work properly on 64 bit hosts. |
| 1875 | |
| 1876 | ** Error messages |
| 1877 | Some calls to strerror resulted in scrambled or missing error messages. |
| 1878 | |
| 1879 | ** %expect |
| 1880 | When the number of shift/reduce conflicts is correct, don't issue |
| 1881 | any warning. |
| 1882 | |
| 1883 | ** The verbose report includes the rule line numbers. |
| 1884 | |
| 1885 | ** Rule line numbers are fixed in traces. |
| 1886 | |
| 1887 | ** Swedish translation |
| 1888 | |
| 1889 | ** Parse errors |
| 1890 | Verbose parse error messages from the parsers are better looking. |
| 1891 | Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'('' |
| 1892 | Now: parse error: unexpected '/', expecting "number" or '-' or '(' |
| 1893 | |
| 1894 | ** Fixed parser memory leaks. |
| 1895 | When the generated parser was using malloc to extend its stacks, the |
| 1896 | previous allocations were not freed. |
| 1897 | |
| 1898 | ** Fixed verbose output file. |
| 1899 | Some newlines were missing. |
| 1900 | Some conflicts in state descriptions were missing. |
| 1901 | |
| 1902 | ** Fixed conflict report. |
| 1903 | Option -v was needed to get the result. |
| 1904 | |
| 1905 | ** %expect |
| 1906 | Was not used. |
| 1907 | Mismatches are errors, not warnings. |
| 1908 | |
| 1909 | ** Fixed incorrect processing of some invalid input. |
| 1910 | |
| 1911 | ** Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H. |
| 1912 | |
| 1913 | ** Fixed some typos in the documentation. |
| 1914 | |
| 1915 | ** %token MY_EOF 0 is supported. |
| 1916 | Before, MY_EOF was silently renumbered as 257. |
| 1917 | |
| 1918 | ** doc/refcard.tex is updated. |
| 1919 | |
| 1920 | ** %output, %file-prefix, %name-prefix. |
| 1921 | New. |
| 1922 | |
| 1923 | ** --output |
| 1924 | New, aliasing "--output-file". |
| 1925 | \f |
| 1926 | * Changes in version 1.30, 2001-10-26: |
| 1927 | |
| 1928 | ** "--defines" and "--graph" have now an optional argument which is the |
| 1929 | output file name. "-d" and "-g" do not change; they do not take any |
| 1930 | argument. |
| 1931 | |
| 1932 | ** "%source_extension" and "%header_extension" are removed, failed |
| 1933 | experiment. |
| 1934 | |
| 1935 | ** Portability fixes. |
| 1936 | \f |
| 1937 | * Changes in version 1.29, 2001-09-07: |
| 1938 | |
| 1939 | ** The output file does not define const, as this caused problems when used |
| 1940 | with common autoconfiguration schemes. If you still use ancient compilers |
| 1941 | that lack const, compile with the equivalent of the C compiler option |
| 1942 | "-Dconst=". Autoconf's AC_C_CONST macro provides one way to do this. |
| 1943 | |
| 1944 | ** Added "-g" and "--graph". |
| 1945 | |
| 1946 | ** The Bison manual is now distributed under the terms of the GNU FDL. |
| 1947 | |
| 1948 | ** The input and the output files has automatically a similar extension. |
| 1949 | |
| 1950 | ** Russian translation added. |
| 1951 | |
| 1952 | ** NLS support updated; should hopefully be less troublesome. |
| 1953 | |
| 1954 | ** Added the old Bison reference card. |
| 1955 | |
| 1956 | ** Added "--locations" and "%locations". |
| 1957 | |
| 1958 | ** Added "-S" and "--skeleton". |
| 1959 | |
| 1960 | ** "%raw", "-r", "--raw" is disabled. |
| 1961 | |
| 1962 | ** Special characters are escaped when output. This solves the problems |
| 1963 | of the #line lines with path names including backslashes. |
| 1964 | |
| 1965 | ** New directives. |
| 1966 | "%yacc", "%fixed_output_files", "%defines", "%no_parser", "%verbose", |
| 1967 | "%debug", "%source_extension" and "%header_extension". |
| 1968 | |
| 1969 | ** @$ |
| 1970 | Automatic location tracking. |
| 1971 | \f |
| 1972 | * Changes in version 1.28, 1999-07-06: |
| 1973 | |
| 1974 | ** Should compile better now with K&R compilers. |
| 1975 | |
| 1976 | ** Added NLS. |
| 1977 | |
| 1978 | ** Fixed a problem with escaping the double quote character. |
| 1979 | |
| 1980 | ** There is now a FAQ. |
| 1981 | \f |
| 1982 | * Changes in version 1.27: |
| 1983 | |
| 1984 | ** The make rule which prevented bison.simple from being created on |
| 1985 | some systems has been fixed. |
| 1986 | \f |
| 1987 | * Changes in version 1.26: |
| 1988 | |
| 1989 | ** Bison now uses Automake. |
| 1990 | |
| 1991 | ** New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>. |
| 1992 | |
| 1993 | ** Token numbers now start at 257 as previously documented, not 258. |
| 1994 | |
| 1995 | ** Bison honors the TMPDIR environment variable. |
| 1996 | |
| 1997 | ** A couple of buffer overruns have been fixed. |
| 1998 | |
| 1999 | ** Problems when closing files should now be reported. |
| 2000 | |
| 2001 | ** Generated parsers should now work even on operating systems which do |
| 2002 | not provide alloca(). |
| 2003 | \f |
| 2004 | * Changes in version 1.25, 1995-10-16: |
| 2005 | |
| 2006 | ** Errors in the input grammar are not fatal; Bison keeps reading |
| 2007 | the grammar file, and reports all the errors found in it. |
| 2008 | |
| 2009 | ** Tokens can now be specified as multiple-character strings: for |
| 2010 | example, you could use "<=" for a token which looks like <=, instead |
| 2011 | of choosing a name like LESSEQ. |
| 2012 | |
| 2013 | ** The %token_table declaration says to write a table of tokens (names |
| 2014 | and numbers) into the parser file. The yylex function can use this |
| 2015 | table to recognize multiple-character string tokens, or for other |
| 2016 | purposes. |
| 2017 | |
| 2018 | ** The %no_lines declaration says not to generate any #line preprocessor |
| 2019 | directives in the parser file. |
| 2020 | |
| 2021 | ** The %raw declaration says to use internal Bison token numbers, not |
| 2022 | Yacc-compatible token numbers, when token names are defined as macros. |
| 2023 | |
| 2024 | ** The --no-parser option produces the parser tables without including |
| 2025 | the parser engine; a project can now use its own parser engine. |
| 2026 | The actions go into a separate file called NAME.act, in the form of |
| 2027 | a switch statement body. |
| 2028 | \f |
| 2029 | * Changes in version 1.23: |
| 2030 | |
| 2031 | The user can define YYPARSE_PARAM as the name of an argument to be |
| 2032 | passed into yyparse. The argument should have type void *. It should |
| 2033 | actually point to an object. Grammar actions can access the variable |
| 2034 | by casting it to the proper pointer type. |
| 2035 | |
| 2036 | Line numbers in output file corrected. |
| 2037 | \f |
| 2038 | * Changes in version 1.22: |
| 2039 | |
| 2040 | --help option added. |
| 2041 | \f |
| 2042 | * Changes in version 1.20: |
| 2043 | |
| 2044 | Output file does not redefine const for C++. |
| 2045 | |
| 2046 | ----- |
| 2047 | |
| 2048 | Copyright (C) 1995-2012 Free Software Foundation, Inc. |
| 2049 | |
| 2050 | This file is part of Bison, the GNU Parser Generator. |
| 2051 | |
| 2052 | This program is free software: you can redistribute it and/or modify |
| 2053 | it under the terms of the GNU General Public License as published by |
| 2054 | the Free Software Foundation, either version 3 of the License, or |
| 2055 | (at your option) any later version. |
| 2056 | |
| 2057 | This program is distributed in the hope that it will be useful, |
| 2058 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 2059 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 2060 | GNU General Public License for more details. |
| 2061 | |
| 2062 | You should have received a copy of the GNU General Public License |
| 2063 | along with this program. If not, see <http://www.gnu.org/licenses/>. |
| 2064 | |
| 2065 | LocalWords: yacc YYBACKUP glr GCC lalr ArrayIndexOutOfBoundsException nullptr |
| 2066 | LocalWords: cplusplus liby rpl fprintf mfcalc Wyacc stmt cond expr mk sym lr |
| 2067 | LocalWords: IELR ielr Lookahead YYERROR nonassoc LALR's api lookaheads yychar |
| 2068 | LocalWords: destructor lookahead YYRHSLOC YYLLOC Rhs ifndef YYFAIL cpp sr rr |
| 2069 | LocalWords: preprocessor initializer Wno Wnone Werror FreeBSD prec livelocks |
| 2070 | LocalWords: Solaris AIX UX RHEL Tru LHS gcc's Wundef YYENABLE NLS YYLTYPE VCG |
| 2071 | LocalWords: yyerror cpp's Wunused yylval yylloc prepend yyparse yylex yypush |
| 2072 | LocalWords: Graphviz xml nonterminals midrule destructor's YYSTYPE typedef ly |
| 2073 | LocalWords: CHR chr printf stdout namespace preprocessing enum pre include's |
| 2074 | LocalWords: YYRECOVERING nonfree destructors YYABORT YYACCEPT params enums de |
| 2075 | LocalWords: struct yystype DJGPP lex param Haible NUM alloca YYSTACK NUL goto |
| 2076 | LocalWords: YYMAXDEPTH Unescaped UCNs YYLTYPE's yyltype typedefs inline Yaccs |
| 2077 | LocalWords: Heriyanto Reenable dprec Hilfinger Eggert MYEOF Folle Menezes EOF |
| 2078 | LocalWords: Lackovic define's itemset Groff Gettext malloc NEWS'ed YYDEBUG YY |
| 2079 | LocalWords: namespaces strerror const autoconfiguration Dconst Autoconf's FDL |
| 2080 | LocalWords: Automake TMPDIR LESSEQ ylwrap endif yydebug YYTOKEN YYLSP ival hh |
| 2081 | LocalWords: extern YYTOKENTYPE TOKENTYPE yytokentype tokentype STYPE lval pdf |
| 2082 | LocalWords: lang yyoutput dvi html ps POSIX lvalp llocp Wother nterm arg init |
| 2083 | LocalWords: TOK |
| 2084 | |
| 2085 | Local Variables: |
| 2086 | mode: outline |
| 2087 | fill-column: 76 |
| 2088 | End: |