tests/regression.at

   1 # Bison Regressions.                               -*- Autotest -*-
   2
   3 # Copyright (C) 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009,
   4 # 2010 Free Software Foundation, Inc.
   5
   6 # This program is free software: you can redistribute it and/or modify
   7 # it under the terms of the GNU General Public License as published by
   8 # the Free Software Foundation, either version 3 of the License, or
   9 # (at your option) any later version.
  10 #
  11 # This program is distributed in the hope that it will be useful,
  12 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 # GNU General Public License for more details.
  15 #
  16 # You should have received a copy of the GNU General Public License
  17 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
  18
  19 AT_BANNER([[Regression tests.]])
  20
  21
  22 ## ------------------ ##
  23 ## Trivial grammars.  ##
  24 ## ------------------ ##
  25
  26 AT_SETUP([Trivial grammars])
  27
  28 AT_DATA_GRAMMAR([input.y],
  29 [[%{
  30 void yyerror (char const *);
  31 int yylex (void);
  32 #define YYSTYPE int *
  33 %}
  34
  35 %error-verbose
  36
  37 %%
  38
  39 program: 'x';
  40 ]])
  41
  42 AT_BISON_CHECK([-o input.c input.y])
  43 AT_COMPILE([input.o], [-c input.c])
  44 AT_COMPILE([input.o], [-DYYDEBUG -c input.c])
  45
  46 AT_CLEANUP
  47
  48
  49
  50 ## ----------------- ##
  51 ## YYSTYPE typedef.  ##
  52 ## ----------------- ##
  53
  54 AT_SETUP([YYSTYPE typedef])
  55
  56 AT_DATA_GRAMMAR([input.y],
  57 [[%{
  58 void yyerror (char const *);
  59 int yylex (void);
  60 typedef union { char const *val; } YYSTYPE;
  61 %}
  62
  63 %type <val> program
  64
  65 %%
  66
  67 program: { $$ = ""; };
  68 ]])
  69
  70 AT_BISON_CHECK([-o input.c input.y])
  71 AT_COMPILE([input.o], [-c input.c])
  72
  73 AT_CLEANUP
  74
  75
  76
  77 ## ------------------------------------- ##
  78 ## Early token definitions with --yacc.  ##
  79 ## ------------------------------------- ##
  80
  81
  82 AT_SETUP([Early token definitions with --yacc])
  83
  84 # Found in GCJ: they expect the tokens to be defined before the user
  85 # prologue, so that they can use the token definitions in it.
  86
  87 AT_DATA_GRAMMAR([input.y],
  88 [[%{
  89 void yyerror (const char *s);
  90 int yylex (void);
  91 %}
  92
  93 %union
  94 {
  95   int val;
  96 };
  97 %{
  98 #ifndef MY_TOKEN
  99 # error "MY_TOKEN not defined."
 100 #endif
 101 %}
 102 %token MY_TOKEN
 103 %%
 104 exp: MY_TOKEN;
 105 %%
 106 ]])
 107
 108 AT_BISON_CHECK([-y -o input.c input.y])
 109 AT_COMPILE([input.o], [-c input.c])
 110
 111 AT_CLEANUP
 112
 113
 114
 115 ## ---------------------------------------- ##
 116 ## Early token definitions without --yacc.  ##
 117 ## ---------------------------------------- ##
 118
 119
 120 AT_SETUP([Early token definitions without --yacc])
 121
 122 # Found in GCJ: they expect the tokens to be defined before the user
 123 # prologue, so that they can use the token definitions in it.
 124
 125 AT_DATA_GRAMMAR([input.y],
 126 [[%{
 127 #include <stdio.h>
 128 void yyerror (const char *s);
 129 int yylex (void);
 130 void print_my_token (void);
 131 %}
 132
 133 %union
 134 {
 135   int val;
 136 };
 137 %{
 138 void
 139 print_my_token (void)
 140 {
 141   enum yytokentype my_token = MY_TOKEN;
 142   printf ("%d\n", my_token);
 143 }
 144 %}
 145 %token MY_TOKEN
 146 %%
 147 exp: MY_TOKEN;
 148 %%
 149 ]])
 150
 151 AT_BISON_CHECK([-o input.c input.y])
 152 AT_COMPILE([input.o], [-c input.c])
 153
 154 AT_CLEANUP
 155
 156
 157
 158 ## ---------------- ##
 159 ## Braces parsing.  ##
 160 ## ---------------- ##
 161
 162
 163 AT_SETUP([Braces parsing])
 164
 165 AT_DATA([input.y],
 166 [[/* Bison used to swallow the character after `}'. */
 167
 168 %%
 169 exp: { tests = {{{{{{{{{{}}}}}}}}}}; };
 170 %%
 171 ]])
 172
 173 AT_BISON_CHECK([-v -o input.c input.y])
 174
 175 AT_CHECK([grep 'tests = {{{{{{{{{{}}}}}}}}}};' input.c], 0, [ignore])
 176
 177 AT_CLEANUP
 178
 179
 180 ## ------------------ ##
 181 ## Duplicate string.  ##
 182 ## ------------------ ##
 183
 184
 185 AT_SETUP([Duplicate string])
 186
 187 AT_DATA([input.y],
 188 [[/* `Bison -v' used to dump core when two tokens are defined with the same
 189    string, as LE and GE below. */
 190
 191 %token NUM
 192 %token LE "<="
 193 %token GE "<="
 194
 195 %%
 196 exp: '(' exp ')' | NUM ;
 197 %%
 198 ]])
 199
 200 AT_BISON_CHECK([-v -o input.c input.y], 0, [],
 201 [[input.y:6.8-14: warning: symbol `"<="' used more than once as a literal string
 202 ]])
 203
 204 AT_CLEANUP
 205
 206
 207 ## ------------------- ##
 208 ## Rule Line Numbers.  ##
 209 ## ------------------- ##
 210
 211 AT_SETUP([Rule Line Numbers])
 212
 213 AT_KEYWORDS([report])
 214
 215 AT_DATA([input.y],
 216 [[%%
 217 expr:
 218 'a'
 219
 220 {
 221
 222 }
 223
 224 'b'
 225
 226 {
 227
 228 }
 229
 230 |
 231
 232
 233 {
 234
 235
 236 }
 237
 238 'c'
 239
 240 {
 241
 242 };
 243 ]])
 244
 245 AT_BISON_CHECK([-o input.c -v input.y])
 246
 247 # Check the contents of the report.
 248 AT_CHECK([cat input.output], [],
 249 [[Grammar
 250
 251     0 $accept: expr $end
 252
 253     1 $@1: /* empty */
 254
 255     2 expr: 'a' $@1 'b'
 256
 257     3 $@2: /* empty */
 258
 259     4 expr: $@2 'c'
 260
 261
 262 Terminals, with rules where they appear
 263
 264 $end (0) 0
 265 'a' (97) 2
 266 'b' (98) 2
 267 'c' (99) 4
 268 error (256)
 269
 270
 271 Nonterminals, with rules where they appear
 272
 273 $accept (6)
 274     on left: 0
 275 expr (7)
 276     on left: 2 4, on right: 0
 277 $@1 (8)
 278     on left: 1, on right: 2
 279 $@2 (9)
 280     on left: 3, on right: 4
 281
 282
 283 state 0
 284
 285     0 $accept: . expr $end
 286
 287     'a'  shift, and go to state 1
 288
 289     $default  reduce using rule 3 ($@2)
 290
 291     expr  go to state 2
 292     $@2   go to state 3
 293
 294
 295 state 1
 296
 297     2 expr: 'a' . $@1 'b'
 298
 299     $default  reduce using rule 1 ($@1)
 300
 301     $@1  go to state 4
 302
 303
 304 state 2
 305
 306     0 $accept: expr . $end
 307
 308     $end  shift, and go to state 5
 309
 310
 311 state 3
 312
 313     4 expr: $@2 . 'c'
 314
 315     'c'  shift, and go to state 6
 316
 317
 318 state 4
 319
 320     2 expr: 'a' $@1 . 'b'
 321
 322     'b'  shift, and go to state 7
 323
 324
 325 state 5
 326
 327     0 $accept: expr $end .
 328
 329     $default  accept
 330
 331
 332 state 6
 333
 334     4 expr: $@2 'c' .
 335
 336     $default  reduce using rule 4 (expr)
 337
 338
 339 state 7
 340
 341     2 expr: 'a' $@1 'b' .
 342
 343     $default  reduce using rule 2 (expr)
 344 ]])
 345
 346 AT_CLEANUP
 347
 348
 349
 350 ## ---------------------- ##
 351 ## Mixing %token styles.  ##
 352 ## ---------------------- ##
 353
 354
 355 AT_SETUP([Mixing %token styles])
 356
 357 # Taken from the documentation.
 358 AT_DATA([input.y],
 359 [[%token  <operator>  OR      "||"
 360 %token  <operator>  LE 134  "<="
 361 %left  OR  "<="
 362 %%
 363 exp: ;
 364 %%
 365 ]])
 366
 367 AT_BISON_CHECK([-v -o input.c input.y])
 368
 369 AT_CLEANUP
 370
 371
 372
 373 ## ---------------- ##
 374 ## Invalid inputs.  ##
 375 ## ---------------- ##
 376
 377
 378 AT_SETUP([Invalid inputs])
 379
 380 AT_DATA([input.y],
 381 [[%%
 382 ?
 383 default: 'a' }
 384 %&
 385 %a-does-not-exist
 386 %-
 387 %{
 388 ]])
 389
 390 AT_BISON_CHECK([input.y], [1], [],
 391 [[input.y:2.1: invalid character: `?'
 392 input.y:3.14: invalid character: `}'
 393 input.y:4.1: invalid character: `%'
 394 input.y:4.2: invalid character: `&'
 395 input.y:5.1-17: invalid directive: `%a-does-not-exist'
 396 input.y:6.1-2: invalid directive: `%-'
 397 input.y:7.1-8.0: missing `%}' at end of file
 398 input.y:7.1-8.0: syntax error, unexpected %{...%}
 399 ]])
 400
 401 AT_CLEANUP
 402
 403
 404 AT_SETUP([Invalid inputs with {}])
 405
 406 AT_DATA([input.y],
 407 [[
 408 %destructor
 409 %initial-action
 410 %lex-param
 411 %parse-param
 412 %printer
 413 %union
 414 ]])
 415
 416 AT_BISON_CHECK([input.y], [1], [],
 417 [[input.y:3.1-15: syntax error, unexpected %initial-action, expecting {...}
 418 ]])
 419
 420 AT_CLEANUP
 421
 422
 423
 424 ## ------------------- ##
 425 ## Token definitions.  ##
 426 ## ------------------- ##
 427
 428
 429 AT_SETUP([Token definitions])
 430
 431 # Bison managed, when fed with `%token 'f' "f"' to #define 'f'!
 432 AT_DATA_GRAMMAR([input.y],
 433 [%{
 434 #include <stdlib.h>
 435 #include <stdio.h>
 436 void yyerror (const char *s);
 437 int yylex (void);
 438 %}
 439 [%error-verbose
 440 %token MYEOF 0 "end of file"
 441 %token 'a' "a"
 442 %token B_TOKEN "b"
 443 %token C_TOKEN 'c'
 444 %token 'd' D_TOKEN
 445 %token SPECIAL "\\\'\?\"\a\b\f\n\r\t\v\001\201\x001\x000081??!"
 446 %token SPECIAL "\\\'\?\"\a\b\f\n\r\t\v\001\201\x001\x000081??!"
 447 %%
 448 exp: "a" "\\\'\?\"\a\b\f\n\r\t\v\001\201\x001\x000081??!";
 449 %%
 450 void
 451 yyerror (char const *s)
 452 {
 453   fprintf (stderr, "%s\n", s);
 454 }
 455
 456 int
 457 yylex (void)
 458 {
 459   static int called;
 460   if (called++)
 461     abort ();
 462   return SPECIAL;
 463 }
 464
 465 int
 466 main (void)
 467 {
 468   return yyparse ();
 469 }
 470 ]])
 471
 472 # Checking the warning message guarantees that the trigraph "??!" isn't
 473 # unnecessarily escaped here even though it would need to be if encoded in a
 474 # C-string literal.  Also notice that unnecessary escaping, such as "\?", from
 475 # the user specification is eliminated.
 476 AT_BISON_CHECK([-o input.c input.y], [[0]], [[]],
 477 [[input.y:22.8-14: warning: symbol SPECIAL redeclared
 478 input.y:22.8-63: warning: symbol `"\\'?\"\a\b\f\n\r\t\v\001\201\001\201??!"' used more than once as a literal string
 479 ]])
 480 AT_COMPILE([input])
 481
 482 # Checking the error message here guarantees that yytname, which does contain
 483 # C-string literals, does have the trigraph escaped correctly.  Thus, the
 484 # symbol name reported by the parser is exactly the same as that reported by
 485 # Bison itself.
 486 AT_DATA([experr],
 487 [[syntax error, unexpected "\\'?\"\a\b\f\n\r\t\v\001\201\001\201??!", expecting a
 488 ]])
 489 AT_PARSER_CHECK([./input], 1, [], [experr])
 490 AT_CLEANUP
 491
 492
 493
 494 ## -------------------- ##
 495 ## Characters Escapes.  ##
 496 ## -------------------- ##
 497
 498
 499 AT_SETUP([Characters Escapes])
 500
 501 AT_DATA_GRAMMAR([input.y],
 502 [%{
 503 void yyerror (const char *s);
 504 int yylex (void);
 505 %}
 506 [%%
 507 exp:
 508   '\'' "\'"
 509 | '\"' "\""
 510 | '"'  "'"
 511 ;
 512 ]])
 513 # Pacify font-lock-mode: "
 514
 515 AT_BISON_CHECK([-o input.c input.y])
 516 AT_COMPILE([input.o], [-c input.c])
 517 AT_CLEANUP
 518
 519
 520
 521 ## -------------- ##
 522 ## Web2c Report.  ##
 523 ## -------------- ##
 524
 525 # The generation of the reduction was once wrong in Bison, and made it
 526 # miss some reductions.  In the following test case, the reduction on
 527 # `undef_id_tok' in state 1 was missing.  This is stripped down from
 528 # the actual web2c.y.
 529
 530 AT_SETUP([Web2c Report])
 531
 532 AT_KEYWORDS([report])
 533
 534 AT_DATA([input.y],
 535 [[%token        undef_id_tok const_id_tok
 536
 537 %start CONST_DEC_PART
 538 \f
 539 %%
 540 CONST_DEC_PART:
 541          CONST_DEC_LIST
 542         ;
 543
 544 CONST_DEC_LIST:
 545           CONST_DEC
 546         | CONST_DEC_LIST CONST_DEC
 547         ;
 548
 549 CONST_DEC:
 550           { } undef_id_tok '=' const_id_tok ';'
 551         ;
 552 %%
 553 ]])
 554
 555 AT_BISON_CHECK([-v input.y])
 556 AT_CHECK([cat input.output], 0,
 557 [[Grammar
 558
 559     0 $accept: CONST_DEC_PART $end
 560
 561     1 CONST_DEC_PART: CONST_DEC_LIST
 562
 563     2 CONST_DEC_LIST: CONST_DEC
 564     3               | CONST_DEC_LIST CONST_DEC
 565
 566     4 $@1: /* empty */
 567
 568     5 CONST_DEC: $@1 undef_id_tok '=' const_id_tok ';'
 569
 570
 571 Terminals, with rules where they appear
 572
 573 $end (0) 0
 574 ';' (59) 5
 575 '=' (61) 5
 576 error (256)
 577 undef_id_tok (258) 5
 578 const_id_tok (259) 5
 579
 580
 581 Nonterminals, with rules where they appear
 582
 583 $accept (7)
 584     on left: 0
 585 CONST_DEC_PART (8)
 586     on left: 1, on right: 0
 587 CONST_DEC_LIST (9)
 588     on left: 2 3, on right: 1 3
 589 CONST_DEC (10)
 590     on left: 5, on right: 2 3
 591 $@1 (11)
 592     on left: 4, on right: 5
 593
 594
 595 state 0
 596
 597     0 $accept: . CONST_DEC_PART $end
 598
 599     $default  reduce using rule 4 ($@1)
 600
 601     CONST_DEC_PART  go to state 1
 602     CONST_DEC_LIST  go to state 2
 603     CONST_DEC       go to state 3
 604     $@1             go to state 4
 605
 606
 607 state 1
 608
 609     0 $accept: CONST_DEC_PART . $end
 610
 611     $end  shift, and go to state 5
 612
 613
 614 state 2
 615
 616     1 CONST_DEC_PART: CONST_DEC_LIST .
 617     3 CONST_DEC_LIST: CONST_DEC_LIST . CONST_DEC
 618
 619     undef_id_tok  reduce using rule 4 ($@1)
 620     $default      reduce using rule 1 (CONST_DEC_PART)
 621
 622     CONST_DEC  go to state 6
 623     $@1        go to state 4
 624
 625
 626 state 3
 627
 628     2 CONST_DEC_LIST: CONST_DEC .
 629
 630     $default  reduce using rule 2 (CONST_DEC_LIST)
 631
 632
 633 state 4
 634
 635     5 CONST_DEC: $@1 . undef_id_tok '=' const_id_tok ';'
 636
 637     undef_id_tok  shift, and go to state 7
 638
 639
 640 state 5
 641
 642     0 $accept: CONST_DEC_PART $end .
 643
 644     $default  accept
 645
 646
 647 state 6
 648
 649     3 CONST_DEC_LIST: CONST_DEC_LIST CONST_DEC .
 650
 651     $default  reduce using rule 3 (CONST_DEC_LIST)
 652
 653
 654 state 7
 655
 656     5 CONST_DEC: $@1 undef_id_tok . '=' const_id_tok ';'
 657
 658     '='  shift, and go to state 8
 659
 660
 661 state 8
 662
 663     5 CONST_DEC: $@1 undef_id_tok '=' . const_id_tok ';'
 664
 665     const_id_tok  shift, and go to state 9
 666
 667
 668 state 9
 669
 670     5 CONST_DEC: $@1 undef_id_tok '=' const_id_tok . ';'
 671
 672     ';'  shift, and go to state 10
 673
 674
 675 state 10
 676
 677     5 CONST_DEC: $@1 undef_id_tok '=' const_id_tok ';' .
 678
 679     $default  reduce using rule 5 (CONST_DEC)
 680 ]])
 681
 682 AT_CLEANUP
 683
 684
 685 ## --------------- ##
 686 ## Web2c Actions.  ##
 687 ## --------------- ##
 688
 689 # The generation of the mapping `state -> action' was once wrong in
 690 # extremely specific situations.  web2c.y exhibits this situation.
 691 # Below is a stripped version of the grammar.  It looks like one can
 692 # simplify it further, but just don't: it is tuned to exhibit a bug,
 693 # which disapears when applying sane grammar transformations.
 694 #
 695 # It used to be wrong on yydefact only:
 696 #
 697 # static const yytype_uint8 yydefact[] =
 698 #  {
 699 # -       2,     0,     1,     0,     0,     2,     3,     2,     5,     4,
 700 # +       2,     0,     1,     0,     0,     0,     3,     2,     5,     4,
 701 #         0,     0
 702 #  };
 703 #
 704 # but let's check all the tables.
 705
 706
 707 AT_SETUP([Web2c Actions])
 708
 709 AT_KEYWORDS([report])
 710
 711 AT_DATA([input.y],
 712 [[%%
 713 statement:  struct_stat;
 714 struct_stat:  /* empty. */ | if else;
 715 if: "if" "const" "then" statement;
 716 else: "else" statement;
 717 %%
 718 ]])
 719
 720 AT_BISON_CHECK([-v -o input.c input.y])
 721
 722 # Check only the tables.
 723 [sed -n 's/  *$//;/^static const.*\[\] =/,/^}/p' input.c >tables.c]
 724
 725 AT_CHECK([[cat tables.c]], 0,
 726 [[static const yytype_uint8 yytranslate[] =
 727 {
 728        0,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 729        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 730        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 731        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 732        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 733        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 734        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 735        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 736        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 737        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 738        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 739        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 740        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 741        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 742        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 743        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 744        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 745        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 746        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 747        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 748        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 749        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 750        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 751        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 752        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 753        2,     2,     2,     2,     2,     2,     1,     2,     3,     4,
 754        5,     6
 755 };
 756 static const yytype_uint8 yyrline[] =
 757 {
 758        0,     2,     2,     3,     3,     4,     5
 759 };
 760 static const char *const yytname[] =
 761 {
 762   "$end", "error", "$undefined", "\"if\"", "\"const\"", "\"then\"",
 763   "\"else\"", "$accept", "statement", "struct_stat", "if", "else", 0
 764 };
 765 static const yytype_uint16 yytoknum[] =
 766 {
 767        0,   256,   257,   258,   259,   260,   261
 768 };
 769 static const yytype_int8 yypact[] =
 770 {
 771       -2,    -1,     4,    -8,     0,     2,    -8,    -2,    -8,    -2,
 772       -8,    -8
 773 };
 774 static const yytype_uint8 yydefact[] =
 775 {
 776        3,     0,     0,     2,     0,     0,     1,     3,     4,     3,
 777        6,     5
 778 };
 779 static const yytype_int8 yypgoto[] =
 780 {
 781       -8,    -7,    -8,    -8,    -8
 782 };
 783 static const yytype_int8 yydefgoto[] =
 784 {
 785       -1,     2,     3,     4,     8
 786 };
 787 static const yytype_uint8 yytable[] =
 788 {
 789       10,     1,    11,     5,     6,     0,     7,     9
 790 };
 791 static const yytype_int8 yycheck[] =
 792 {
 793        7,     3,     9,     4,     0,    -1,     6,     5
 794 };
 795 static const yytype_uint8 yystos[] =
 796 {
 797        0,     3,     8,     9,    10,     4,     0,     6,    11,     5,
 798        8,     8
 799 };
 800 static const yytype_uint8 yyr1[] =
 801 {
 802        0,     7,     8,     9,     9,    10,    11
 803 };
 804 static const yytype_uint8 yyr2[] =
 805 {
 806        0,     2,     1,     0,     2,     4,     2
 807 };
 808 ]])
 809
 810 AT_CLEANUP
 811
 812
 813 ## ------------------------- ##
 814 ## yycheck Bound Violation.  ##
 815 ## ------------------------- ##
 816
 817
 818 # _AT_DATA_DANCER_Y(BISON-OPTIONS)
 819 # --------------------------------
 820 # The following grammar, taken from Andrew Suffield's GPL'd implementation
 821 # of DGMTP, the Dancer Generic Message Transport Protocol, used to violate
 822 # yycheck's bounds where issuing a verbose error message.  Keep this test
 823 # so that possible bound checking compilers could check all the skeletons.
 824 m4_define([_AT_DATA_DANCER_Y],
 825 [AT_DATA_GRAMMAR([dancer.y],
 826 [%{
 827 static int yylex (AT_LALR1_CC_IF([int *], [void]));
 828 AT_LALR1_CC_IF([#include <cstdlib>],
 829 [#include <stdlib.h>
 830 #include <stdio.h>
 831 static void yyerror (const char *);])
 832 %}
 833 $1
 834 %token ARROW INVALID NUMBER STRING DATA
 835 %defines
 836 %verbose
 837 %error-verbose
 838 /* Grammar follows */
 839 %%
 840 line: header body
 841    ;
 842
 843 header: '<' from ARROW to '>' type ':'
 844    | '<' ARROW to '>' type ':'
 845    | ARROW to type ':'
 846    | type ':'
 847    | '<' '>'
 848    ;
 849
 850 from: DATA
 851    | STRING
 852    | INVALID
 853    ;
 854
 855 to: DATA
 856    | STRING
 857    | INVALID
 858    ;
 859
 860 type: DATA
 861    | STRING
 862    | INVALID
 863    ;
 864
 865 body: /* empty */
 866    | body member
 867    ;
 868
 869 member: STRING
 870    | DATA
 871    | '+' NUMBER
 872    | '-' NUMBER
 873    | NUMBER
 874    | INVALID
 875    ;
 876 %%
 877 AT_LALR1_CC_IF(
 878 [/* A C++ error reporting function. */
 879 void
 880 yy::parser::error (const std::string& m)
 881 {
 882   std::cerr << m << std::endl;
 883 }
 884
 885 int
 886 yyparse ()
 887 {
 888   yy::parser parser;
 889 #if YYDEBUG
 890   parser.set_debug_level (YYDEBUG);
 891 #endif
 892   return parser.parse ();
 893 }
 894 ],
 895 [static void
 896 yyerror (const char *s)
 897 {
 898   fprintf (stderr, "%s\n", s);
 899 }])
 900
 901 static int
 902 yylex (AT_LALR1_CC_IF([int *lval], [void]))
 903 [{
 904   static int const tokens[] =
 905     {
 906       ':', -1
 907     };
 908   static size_t toknum;
 909   ]AT_LALR1_CC_IF([*lval = 0; /* Pacify GCC.  */])[
 910   if (! (toknum < sizeof tokens / sizeof *tokens))
 911     abort ();
 912   return tokens[toknum++];
 913 }]
 914
 915 int
 916 main (void)
 917 {
 918   return yyparse ();
 919 }
 920 ])
 921 ])# _AT_DATA_DANCER_Y
 922
 923
 924 # AT_CHECK_DANCER(BISON-OPTIONS)
 925 # ------------------------------
 926 # Generate the grammar, compile it, run it.
 927 m4_define([AT_CHECK_DANCER],
 928 [AT_SETUP([Dancer $1])
 929 AT_BISON_OPTION_PUSHDEFS([$1])
 930 _AT_DATA_DANCER_Y([$1])
 931 AT_BISON_CHECK([-o dancer.c dancer.y])
 932 AT_FULL_COMPILE([dancer])
 933 AT_PARSER_CHECK([./dancer], 1, [],
 934 [syntax error, unexpected ':'
 935 ])
 936 AT_BISON_OPTION_POPDEFS
 937 AT_CLEANUP
 938 ])
 939
 940 AT_CHECK_DANCER()
 941 AT_CHECK_DANCER([%glr-parser])
 942 AT_CHECK_DANCER([%skeleton "lalr1.cc"])
 943
 944
 945 ## ------------------------------------------ ##
 946 ## Diagnostic that expects two alternatives.  ##
 947 ## ------------------------------------------ ##
 948
 949
 950 # _AT_DATA_EXPECT2_Y(BISON-OPTIONS)
 951 # --------------------------------
 952 m4_define([_AT_DATA_EXPECT2_Y],
 953 [AT_DATA_GRAMMAR([expect2.y],
 954 [%{
 955 static int yylex (AT_LALR1_CC_IF([int *], [void]));
 956 AT_LALR1_CC_IF([#include <cstdlib>],
 957 [#include <stdio.h>
 958 #include <stdlib.h>
 959 static void yyerror (const char *);])
 960 %}
 961 $1
 962 %defines
 963 %error-verbose
 964 %token A 1000
 965 %token B
 966
 967 %%
 968 program: /* empty */
 969  | program e ';'
 970  | program error ';';
 971
 972 e: e '+' t | t;
 973 t: A | B;
 974
 975 %%
 976 AT_LALR1_CC_IF(
 977 [/* A C++ error reporting function. */
 978 void
 979 yy::parser::error (const std::string& m)
 980 {
 981   std::cerr << m << std::endl;
 982 }
 983
 984 int
 985 yyparse ()
 986 {
 987   yy::parser parser;
 988   return parser.parse ();
 989 }
 990 ],
 991 [static void
 992 yyerror (const char *s)
 993 {
 994   fprintf (stderr, "%s\n", s);
 995 }])
 996
 997 static int
 998 yylex (AT_LALR1_CC_IF([int *lval], [void]))
 999 [{
1000   static int const tokens[] =
1001     {
1002       1000, '+', '+', -1
1003     };
1004   static size_t toknum;
1005   ]AT_LALR1_CC_IF([*lval = 0; /* Pacify GCC.  */])[
1006   if (! (toknum < sizeof tokens / sizeof *tokens))
1007     abort ();
1008   return tokens[toknum++];
1009 }]
1010
1011 int
1012 main (void)
1013 {
1014   return yyparse ();
1015 }
1016 ])
1017 ])# _AT_DATA_EXPECT2_Y
1018
1019
1020 # AT_CHECK_EXPECT2(BISON-OPTIONS)
1021 # ------------------------------
1022 # Generate the grammar, compile it, run it.
1023 m4_define([AT_CHECK_EXPECT2],
1024 [AT_SETUP([Expecting two tokens $1])
1025 AT_BISON_OPTION_PUSHDEFS([$1])
1026 _AT_DATA_EXPECT2_Y([$1])
1027 AT_BISON_CHECK([-o expect2.c expect2.y])
1028 AT_FULL_COMPILE([expect2])
1029 AT_PARSER_CHECK([./expect2], 1, [],
1030 [syntax error, unexpected '+', expecting A or B
1031 ])
1032 AT_BISON_OPTION_POPDEFS
1033 AT_CLEANUP
1034 ])
1035
1036 AT_CHECK_EXPECT2()
1037 AT_CHECK_EXPECT2([%glr-parser])
1038 AT_CHECK_EXPECT2([%skeleton "lalr1.cc"])
1039
1040
1041
1042 ## --------------------------------------------- ##
1043 ## Braced code in declaration in rules section.  ##
1044 ## --------------------------------------------- ##
1045
1046 AT_SETUP([Braced code in declaration in rules section])
1047
1048 # Bison once mistook braced code in a declaration in the rules section to be a
1049 # rule action.
1050
1051 AT_DATA_GRAMMAR([input.y],
1052 [[%{
1053 #include <stdio.h>
1054 static void yyerror (char const *msg);
1055 static int yylex (void);
1056 %}
1057
1058 %error-verbose
1059
1060 %%
1061
1062 start:
1063   {
1064     printf ("Bison would once convert this action to a midrule because of the"
1065             " subsequent braced code.\n");
1066   }
1067   ;
1068
1069 %destructor { fprintf (stderr, "DESTRUCTOR\n"); } 'a';
1070 %printer { fprintf (yyoutput, "PRINTER"); } 'a';
1071
1072 %%
1073
1074 static void
1075 yyerror (char const *msg)
1076 {
1077   fprintf (stderr, "%s\n", msg);
1078 }
1079
1080 static int
1081 yylex (void)
1082 {
1083   return 'a';
1084 }
1085
1086 int
1087 main (void)
1088 {
1089   yydebug = 1;
1090   return !yyparse ();
1091 }
1092 ]])
1093
1094 AT_BISON_CHECK([-t -o input.c input.y])
1095 AT_COMPILE([input])
1096 AT_PARSER_CHECK([./input], 0,
1097 [[Bison would once convert this action to a midrule because of the subsequent braced code.
1098 ]],
1099 [[Starting parse
1100 Entering state 0
1101 Reducing stack by rule 1 (line 20):
1102 -> $$ = nterm start ()
1103 Stack now 0
1104 Entering state 1
1105 Reading a token: Next token is token 'a' (PRINTER)
1106 syntax error, unexpected 'a', expecting $end
1107 Error: popping nterm start ()
1108 Stack now 0
1109 Cleanup: discarding lookahead token 'a' (PRINTER)
1110 DESTRUCTOR
1111 Stack now 0
1112 ]])
1113
1114 AT_CLEANUP
1115
1116
1117
1118 ## --------------------------------- ##
1119 ## String alias declared after use.  ##
1120 ## --------------------------------- ##
1121
1122 AT_SETUP([String alias declared after use])
1123
1124 # Bison once incorrectly asserted that the symbol number for either a token or
1125 # its alias was the highest symbol number so far at the point of the alias
1126 # declaration.  That was true unless the declaration appeared after their first
1127 # uses and other tokens appeared in between.
1128
1129 AT_DATA([input.y],
1130 [[%%
1131 start: 'a' "A" 'b';
1132 %token 'a' "A";
1133 ]])
1134
1135 AT_BISON_CHECK([-t -o input.c input.y])
1136
1137 AT_CLEANUP
1138
1139
1140
1141 ## -------------------------------- ##
1142 ## Extra lookahead sets in report.  ##
1143 ## -------------------------------- ##
1144
1145 AT_SETUP([[Extra lookahead sets in report]])
1146
1147 # Bison prints each reduction's lookahead set only next to the associated
1148 # state's one item that (1) is associated with the same rule as the reduction
1149 # and (2) has its dot at the end of its RHS.  Previously, Bison also
1150 # erroneously printed the lookahead set next to all of the state's other items
1151 # associated with the same rule.  This bug affected only the `.output' file and
1152 # not the generated parser source code.
1153
1154 AT_DATA([[input.y]],
1155 [[%%
1156 start: a | 'a' a 'a' ;
1157 a: 'a' ;
1158 ]])
1159
1160 AT_BISON_CHECK([[--report=all input.y]])
1161 AT_CHECK([[sed -n '/^state 1$/,/^state 2$/p' input.output]], [[0]],
1162 [[state 1
1163
1164     2 start: 'a' . a 'a'
1165     3 a: . 'a'
1166     3  | 'a' .  [$end]
1167
1168     'a'  shift, and go to state 4
1169
1170     $default  reduce using rule 3 (a)
1171
1172     a  go to state 5
1173
1174
1175 state 2
1176 ]])
1177
1178 AT_CLEANUP
1179
1180
1181
1182 ## ---------------------------------------- ##
1183 ## Token number in precedence declaration.  ##
1184 ## ---------------------------------------- ##
1185
1186 AT_SETUP([[Token number in precedence declaration]])
1187
1188 # POSIX says token numbers can be declared in %left, %right, and %nonassoc, but
1189 # we lost this in Bison 1.50.
1190
1191 AT_DATA_GRAMMAR([input.y],
1192 [[%{
1193   #include <stdio.h>
1194   void yyerror (char const *);
1195   int yylex (void);
1196 %}
1197
1198 %error-verbose
1199 %left TK1 1 TK2 2 "tok alias" 3
1200
1201 %%
1202
1203 start: TK1 sr_conflict "tok alias" ;
1204
1205 sr_conflict:
1206   TK2
1207   | TK2 "tok alias"
1208   ;
1209
1210 %%
1211
1212 void
1213 yyerror (char const *msg)
1214 {
1215   fprintf (stderr, "%s\n", msg);
1216 }
1217
1218 int
1219 yylex (void)
1220 {
1221   static int const input[] = { 1, 2, 3, 0 };
1222   static int const *inputp = input;
1223   return *inputp++;
1224 }
1225
1226 int
1227 main (void)
1228 {
1229   return yyparse ();
1230 }
1231 ]])
1232
1233 AT_BISON_CHECK([[-o input.c input.y]], [[0]],,
1234 [[input.y:24.5-19: warning: rule useless in parser due to conflicts: sr_conflict: TK2 "tok alias"
1235 ]])
1236 AT_COMPILE([[input]])
1237 AT_PARSER_CHECK([[./input]])
1238
1239 AT_CLEANUP
1240
1241
1242
1243 ## --------------------------- ##
1244 ## parse-gram.y: LALR = IELR.  ##
1245 ## --------------------------- ##
1246
1247 # If parse-gram.y's LALR and IELR parser tables ever begin to differ, we
1248 # need to fix parse-gram.y or start using IELR.
1249
1250 AT_SETUP([[parse-gram.y: LALR = IELR]])
1251
1252 # Avoid tests/bison's dark magic by processing a local copy of the
1253 # grammar.  Avoid differences in synclines by telling bison that the
1254 # output files have the same name.
1255 [cp $abs_top_srcdir/src/parse-gram.y input.y]
1256 AT_BISON_CHECK([[-o input.c -Dlr.type=lalr input.y]])
1257 [mv input.c lalr.c]
1258 AT_BISON_CHECK([[-o input.c -Dlr.type=ielr input.y]])
1259 [mv input.c ielr.c]
1260 AT_CHECK([[diff -u lalr.c ielr.c]])
1261
1262 AT_CLEANUP
1263
1264
1265
1266 ## -------------------------------------------- ##
1267 ## parse.error=verbose and YYSTACK_USE_ALLOCA.  ##
1268 ## -------------------------------------------- ##
1269
1270 AT_SETUP([[parse.error=verbose and YYSTACK_USE_ALLOCA]])
1271
1272 AT_DATA_GRAMMAR([input.y],
1273 [[%code {
1274   #include <stdio.h>
1275   void yyerror (char const *);
1276   int yylex (void);
1277   #define YYSTACK_USE_ALLOCA 1
1278 }
1279
1280 %define parse.error verbose
1281
1282 %%
1283
1284 start: check syntax_error syntax_error ;
1285
1286 check:
1287 {
1288   if (128 < sizeof yymsgbuf)
1289     {
1290       fprintf (stderr,
1291                "The initial size of yymsgbuf in yyparse has increased\n"
1292                "since this test group was last updated.  As a result,\n"
1293                "this test group may no longer manage to induce a\n"
1294                "reallocation of the syntax error message buffer.\n"
1295                "This test group must be adjusted to produce a longer\n"
1296                "error message.\n");
1297       YYABORT;
1298     }
1299 }
1300 ;
1301
1302 // Induce a syntax error message whose total length is more than
1303 // sizeof yymsgbuf in yyparse.  Each token here is 64 bytes.
1304 syntax_error:
1305   "123456789112345678921234567893123456789412345678951234567896123A"
1306 | "123456789112345678921234567893123456789412345678951234567896123B"
1307 | error 'a' 'b' 'c'
1308 ;
1309
1310 %%
1311
1312 void
1313 yyerror (char const *msg)
1314 {
1315   fprintf (stderr, "%s\n", msg);
1316 }
1317
1318 int
1319 yylex (void)
1320 {
1321   /* Induce two syntax error messages (which requires full error
1322      recovery by shifting 3 tokens) in order to detect any loss of the
1323      reallocated buffer.  */
1324   static char const *input = "abc";
1325   return *input++;
1326 }
1327
1328 int
1329 main (void)
1330 {
1331   return yyparse ();
1332 }
1333 ]])
1334
1335 AT_BISON_CHECK([[-o input.c input.y]])
1336 AT_COMPILE([[input]])
1337 AT_PARSER_CHECK([[./input]], [[1]], [],
1338 [[syntax error, unexpected 'a', expecting 123456789112345678921234567893123456789412345678951234567896123A or 123456789112345678921234567893123456789412345678951234567896123B
1339 syntax error, unexpected $end, expecting 123456789112345678921234567893123456789412345678951234567896123A or 123456789112345678921234567893123456789412345678951234567896123B
1340 ]])
1341
1342 AT_CLEANUP
1343
1344
1345
1346 ## ------------------------------ ##
1347 ## parse.error=verbose overflow.  ##
1348 ## ------------------------------ ##
1349
1350 # Imagine the case where YYSTACK_ALLOC_MAXIMUM = YYSIZE_MAXIMUM and an
1351 # invocation of yysyntax_error has caused yymsg_alloc to grow to exactly
1352 # YYSTACK_ALLOC_MAXIMUM (perhaps because the normal doubling of size had
1353 # to be clipped to YYSTACK_ALLOC_MAXIMUM).  In an old version of yacc.c,
1354 # a subsequent invocation of yysyntax_error that overflows during its
1355 # size calculation would return YYSIZE_MAXIMUM to yyparse.  Then,
1356 # yyparse would invoke yyerror using the old contents of yymsg.
1357
1358 AT_SETUP([[parse.error=verbose overflow]])
1359
1360 AT_DATA_GRAMMAR([input.y],
1361 [[%code {
1362   #include <stdio.h>
1363   void yyerror (char const *);
1364   int yylex (void);
1365
1366   /* This prevents this test case from having to induce error messages
1367      large enough to overflow size_t.  */
1368   #define YYSIZE_T unsigned char
1369
1370   /* Bring in malloc and set EXIT_SUCCESS so yacc.c doesn't try to
1371      provide a malloc prototype using our YYSIZE_T.  */
1372   #include <stdlib.h>
1373   #ifndef EXIT_SUCCESS
1374   # define EXIT_SUCCESS 0
1375   #endif
1376
1377   /* Max depth is usually much smaller than YYSTACK_ALLOC_MAXIMUM, and
1378      we don't want gcc to warn everywhere this constant would be too big
1379      to make sense for our YYSIZE_T.  */
1380   #define YYMAXDEPTH 100
1381 }
1382
1383 %define parse.error verbose
1384
1385 %%
1386
1387 start: syntax_error1 check syntax_error2 ;
1388
1389 // Induce a syntax error message whose total length causes yymsg in
1390 // yyparse to be reallocated to size YYSTACK_ALLOC_MAXIMUM, which
1391 // should be 255.  Each token here is 64 bytes.
1392 syntax_error1:
1393   "123456789112345678921234567893123456789412345678951234567896123A"
1394 | "123456789112345678921234567893123456789412345678951234567896123B"
1395 | "123456789112345678921234567893123456789412345678951234567896123C"
1396 | error 'a' 'b' 'c'
1397 ;
1398
1399 check:
1400 {
1401   if (yymsg_alloc != YYSTACK_ALLOC_MAXIMUM
1402       || YYSTACK_ALLOC_MAXIMUM != YYSIZE_MAXIMUM
1403       || YYSIZE_MAXIMUM != 255)
1404     {
1405       fprintf (stderr,
1406                "The assumptions of this test group are no longer\n"
1407                "valid, so it may no longer catch the error it was\n"
1408                "designed to catch.  Specifically, the following\n"
1409                "values should all be 255:\n\n");
1410       fprintf (stderr, "  yymsg_alloc = %d\n", yymsg_alloc);
1411       fprintf (stderr, "  YYSTACK_ALLOC_MAXIMUM = %d\n",
1412                YYSTACK_ALLOC_MAXIMUM);
1413       fprintf (stderr, "  YYSIZE_MAXIMUM = %d\n", YYSIZE_MAXIMUM);
1414       YYABORT;
1415     }
1416 }
1417 ;
1418
1419 // Now overflow.
1420 syntax_error2:
1421   "123456789112345678921234567893123456789412345678951234567896123A"
1422 | "123456789112345678921234567893123456789412345678951234567896123B"
1423 | "123456789112345678921234567893123456789412345678951234567896123C"
1424 | "123456789112345678921234567893123456789412345678951234567896123D"
1425 | "123456789112345678921234567893123456789412345678951234567896123E"
1426 ;
1427
1428 %%
1429
1430 void
1431 yyerror (char const *msg)
1432 {
1433   fprintf (stderr, "%s\n", msg);
1434 }
1435
1436 int
1437 yylex (void)
1438 {
1439   /* Induce two syntax error messages (which requires full error
1440      recovery by shifting 3 tokens).  */
1441   static char const *input = "abc";
1442   return *input++;
1443 }
1444
1445 int
1446 main (void)
1447 {
1448   /* Push parsers throw away the message buffer between tokens, so skip
1449      this test under maintainer-push-check.  */
1450   if (YYPUSH)
1451     return 77;
1452   return yyparse ();
1453 }
1454 ]])
1455
1456 AT_BISON_CHECK([[-o input.c input.y]])
1457
1458 # gcc warns about tautologies and fallacies involving comparisons for
1459 # unsigned char.  However, it doesn't produce these same warnings for
1460 # size_t and many other types when the warnings would seem to make just
1461 # as much sense.  We ignore the warnings.
1462 [CFLAGS="$NO_WERROR_CFLAGS"]
1463 AT_COMPILE([[input]])
1464
1465 AT_PARSER_CHECK([[./input]], [[2]], [],
1466 [[syntax error, unexpected 'a', expecting 123456789112345678921234567893123456789412345678951234567896123A or 123456789112345678921234567893123456789412345678951234567896123B or 123456789112345678921234567893123456789412345678951234567896123C
1467 syntax error
1468 memory exhausted
1469 ]])
1470
1471 AT_CLEANUP