tests/regression.at

   1 # Bison Regressions.                               -*- Autotest -*-
   2
   3 # Copyright (C) 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free Software
   4 # 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 yyprhs[] =
 757 {
 758        0,     0,     3,     5,     6,     9,    14
 759 };
 760 static const yytype_int8 yyrhs[] =
 761 {
 762        8,     0,    -1,     9,    -1,    -1,    10,    11,    -1,     3,
 763        4,     5,     8,    -1,     6,     8,    -1
 764 };
 765 static const yytype_uint8 yyrline[] =
 766 {
 767        0,     2,     2,     3,     3,     4,     5
 768 };
 769 static const char *const yytname[] =
 770 {
 771   "$end", "error", "$undefined", "\"if\"", "\"const\"", "\"then\"",
 772   "\"else\"", "$accept", "statement", "struct_stat", "if", "else", 0
 773 };
 774 static const yytype_uint16 yytoknum[] =
 775 {
 776        0,   256,   257,   258,   259,   260,   261
 777 };
 778 static const yytype_uint8 yyr1[] =
 779 {
 780        0,     7,     8,     9,     9,    10,    11
 781 };
 782 static const yytype_uint8 yyr2[] =
 783 {
 784        0,     2,     1,     0,     2,     4,     2
 785 };
 786 static const yytype_uint8 yydefact[] =
 787 {
 788        3,     0,     0,     2,     0,     0,     1,     3,     4,     3,
 789        6,     5
 790 };
 791 static const yytype_int8 yydefgoto[] =
 792 {
 793       -1,     2,     3,     4,     8
 794 };
 795 static const yytype_int8 yypact[] =
 796 {
 797       -2,    -1,     4,    -8,     0,     2,    -8,    -2,    -8,    -2,
 798       -8,    -8
 799 };
 800 static const yytype_int8 yypgoto[] =
 801 {
 802       -8,    -7,    -8,    -8,    -8
 803 };
 804 static const yytype_uint8 yytable[] =
 805 {
 806       10,     1,    11,     5,     6,     0,     7,     9
 807 };
 808 static const yytype_int8 yycheck[] =
 809 {
 810        7,     3,     9,     4,     0,    -1,     6,     5
 811 };
 812 static const yytype_uint8 yystos[] =
 813 {
 814        0,     3,     8,     9,    10,     4,     0,     6,    11,     5,
 815        8,     8
 816 };
 817 ]])
 818
 819 AT_CLEANUP
 820
 821
 822 ## ------------------------- ##
 823 ## yycheck Bound Violation.  ##
 824 ## ------------------------- ##
 825
 826
 827 # _AT_DATA_DANCER_Y(BISON-OPTIONS)
 828 # --------------------------------
 829 # The following grammar, taken from Andrew Suffield's GPL'd implementation
 830 # of DGMTP, the Dancer Generic Message Transport Protocol, used to violate
 831 # yycheck's bounds where issuing a verbose error message.  Keep this test
 832 # so that possible bound checking compilers could check all the skeletons.
 833 m4_define([_AT_DATA_DANCER_Y],
 834 [AT_DATA_GRAMMAR([dancer.y],
 835 [%{
 836 static int yylex (AT_LALR1_CC_IF([int *], [void]));
 837 AT_LALR1_CC_IF([],
 838 [#include <stdlib.h>
 839 #include <stdio.h>
 840 static void yyerror (const char *);])
 841 %}
 842 $1
 843 %token ARROW INVALID NUMBER STRING DATA
 844 %defines
 845 %verbose
 846 %error-verbose
 847 /* Grammar follows */
 848 %%
 849 line: header body
 850    ;
 851
 852 header: '<' from ARROW to '>' type ':'
 853    | '<' ARROW to '>' type ':'
 854    | ARROW to type ':'
 855    | type ':'
 856    | '<' '>'
 857    ;
 858
 859 from: DATA
 860    | STRING
 861    | INVALID
 862    ;
 863
 864 to: DATA
 865    | STRING
 866    | INVALID
 867    ;
 868
 869 type: DATA
 870    | STRING
 871    | INVALID
 872    ;
 873
 874 body: /* empty */
 875    | body member
 876    ;
 877
 878 member: STRING
 879    | DATA
 880    | '+' NUMBER
 881    | '-' NUMBER
 882    | NUMBER
 883    | INVALID
 884    ;
 885 %%
 886 AT_LALR1_CC_IF(
 887 [/* A C++ error reporting function. */
 888 void
 889 yy::parser::error (const location&, const std::string& m)
 890 {
 891   std::cerr << m << std::endl;
 892 }
 893
 894 int
 895 yyparse ()
 896 {
 897   yy::parser parser;
 898 #if YYDEBUG
 899   parser.set_debug_level (YYDEBUG);
 900 #endif
 901   return parser.parse ();
 902 }
 903 ],
 904 [static void
 905 yyerror (const char *s)
 906 {
 907   fprintf (stderr, "%s\n", s);
 908 }])
 909
 910 static int
 911 yylex (AT_LALR1_CC_IF([int *lval], [void]))
 912 [{
 913   static int const tokens[] =
 914     {
 915       ':', -1
 916     };
 917   static size_t toknum;
 918   ]AT_LALR1_CC_IF([*lval = 0; /* Pacify GCC.  */])[
 919   if (! (toknum < sizeof tokens / sizeof *tokens))
 920     abort ();
 921   return tokens[toknum++];
 922 }]
 923
 924 int
 925 main (void)
 926 {
 927   return yyparse ();
 928 }
 929 ])
 930 ])# _AT_DATA_DANCER_Y
 931
 932
 933 # AT_CHECK_DANCER(BISON-OPTIONS)
 934 # ------------------------------
 935 # Generate the grammar, compile it, run it.
 936 m4_define([AT_CHECK_DANCER],
 937 [AT_SETUP([Dancer $1])
 938 AT_BISON_OPTION_PUSHDEFS([$1])
 939 _AT_DATA_DANCER_Y([$1])
 940 AT_BISON_CHECK([-o dancer.c dancer.y])
 941 AT_FULL_COMPILE([dancer])
 942 AT_PARSER_CHECK([./dancer], 1, [],
 943 [syntax error, unexpected ':'
 944 ])
 945 AT_BISON_OPTION_POPDEFS
 946 AT_CLEANUP
 947 ])
 948
 949 AT_CHECK_DANCER()
 950 AT_CHECK_DANCER([%glr-parser])
 951 AT_CHECK_DANCER([%skeleton "lalr1.cc"])
 952
 953
 954 ## ------------------------------------------ ##
 955 ## Diagnostic that expects two alternatives.  ##
 956 ## ------------------------------------------ ##
 957
 958
 959 # _AT_DATA_EXPECT2_Y(BISON-OPTIONS)
 960 # --------------------------------
 961 m4_define([_AT_DATA_EXPECT2_Y],
 962 [AT_DATA_GRAMMAR([expect2.y],
 963 [%{
 964 static int yylex (AT_LALR1_CC_IF([int *], [void]));
 965 AT_LALR1_CC_IF([],
 966 [#include <stdio.h>
 967 #include <stdlib.h>
 968 static void yyerror (const char *);])
 969 %}
 970 $1
 971 %defines
 972 %error-verbose
 973 %token A 1000
 974 %token B
 975
 976 %%
 977 program: /* empty */
 978  | program e ';'
 979  | program error ';';
 980
 981 e: e '+' t | t;
 982 t: A | B;
 983
 984 %%
 985 AT_LALR1_CC_IF(
 986 [/* A C++ error reporting function. */
 987 void
 988 yy::parser::error (const location&, const std::string& m)
 989 {
 990   std::cerr << m << std::endl;
 991 }
 992
 993 int
 994 yyparse ()
 995 {
 996   yy::parser parser;
 997   return parser.parse ();
 998 }
 999 ],
1000 [static void
1001 yyerror (const char *s)
1002 {
1003   fprintf (stderr, "%s\n", s);
1004 }])
1005
1006 static int
1007 yylex (AT_LALR1_CC_IF([int *lval], [void]))
1008 [{
1009   static int const tokens[] =
1010     {
1011       1000, '+', '+', -1
1012     };
1013   static size_t toknum;
1014   ]AT_LALR1_CC_IF([*lval = 0; /* Pacify GCC.  */])[
1015   if (! (toknum < sizeof tokens / sizeof *tokens))
1016     abort ();
1017   return tokens[toknum++];
1018 }]
1019
1020 int
1021 main (void)
1022 {
1023   return yyparse ();
1024 }
1025 ])
1026 ])# _AT_DATA_EXPECT2_Y
1027
1028
1029 # AT_CHECK_EXPECT2(BISON-OPTIONS)
1030 # ------------------------------
1031 # Generate the grammar, compile it, run it.
1032 m4_define([AT_CHECK_EXPECT2],
1033 [AT_SETUP([Expecting two tokens $1])
1034 AT_BISON_OPTION_PUSHDEFS([$1])
1035 _AT_DATA_EXPECT2_Y([$1])
1036 AT_BISON_CHECK([-o expect2.c expect2.y])
1037 AT_FULL_COMPILE([expect2])
1038 AT_PARSER_CHECK([./expect2], 1, [],
1039 [syntax error, unexpected '+', expecting A or B
1040 ])
1041 AT_BISON_OPTION_POPDEFS
1042 AT_CLEANUP
1043 ])
1044
1045 AT_CHECK_EXPECT2()
1046 AT_CHECK_EXPECT2([%glr-parser])
1047 AT_CHECK_EXPECT2([%skeleton "lalr1.cc"])
1048
1049
1050
1051 ## --------------------------------------------- ##
1052 ## Braced code in declaration in rules section.  ##
1053 ## --------------------------------------------- ##
1054
1055 AT_SETUP([Braced code in declaration in rules section])
1056
1057 # Bison once mistook braced code in a declaration in the rules section to be a
1058 # rule action.
1059
1060 AT_DATA_GRAMMAR([input.y],
1061 [[%{
1062 #include <stdio.h>
1063 static void yyerror (char const *msg);
1064 static int yylex (void);
1065 %}
1066
1067 %error-verbose
1068
1069 %%
1070
1071 start:
1072   {
1073     printf ("Bison would once convert this action to a midrule because of the"
1074             " subsequent braced code.\n");
1075   }
1076   ;
1077
1078 %destructor { fprintf (stderr, "DESTRUCTOR\n"); } 'a';
1079 %printer { fprintf (yyoutput, "PRINTER"); } 'a';
1080
1081 %%
1082
1083 static void
1084 yyerror (char const *msg)
1085 {
1086   fprintf (stderr, "%s\n", msg);
1087 }
1088
1089 static int
1090 yylex (void)
1091 {
1092   return 'a';
1093 }
1094
1095 int
1096 main (void)
1097 {
1098   yydebug = 1;
1099   return !yyparse ();
1100 }
1101 ]])
1102
1103 AT_BISON_CHECK([-t -o input.c input.y])
1104 AT_COMPILE([input])
1105 AT_PARSER_CHECK([./input], 0,
1106 [[Bison would once convert this action to a midrule because of the subsequent braced code.
1107 ]],
1108 [[Starting parse
1109 Entering state 0
1110 Reducing stack by rule 1 (line 20):
1111 -> $$ = nterm start ()
1112 Stack now 0
1113 Entering state 1
1114 Reading a token: Next token is token 'a' (PRINTER)
1115 syntax error, unexpected 'a', expecting $end
1116 Error: popping nterm start ()
1117 Stack now 0
1118 Cleanup: discarding lookahead token 'a' (PRINTER)
1119 DESTRUCTOR
1120 Stack now 0
1121 ]])
1122
1123 AT_CLEANUP
1124
1125
1126
1127 ## --------------------------------- ##
1128 ## String alias declared after use.  ##
1129 ## --------------------------------- ##
1130
1131 AT_SETUP([String alias declared after use])
1132
1133 # Bison once incorrectly asserted that the symbol number for either a token or
1134 # its alias was the highest symbol number so far at the point of the alias
1135 # declaration.  That was true unless the declaration appeared after their first
1136 # uses and other tokens appeared in between.
1137
1138 AT_DATA([input.y],
1139 [[%%
1140 start: 'a' "A" 'b';
1141 %token 'a' "A";
1142 ]])
1143
1144 AT_BISON_CHECK([-t -o input.c input.y])
1145
1146 AT_CLEANUP
1147
1148
1149
1150 ## -------------------------------- ##
1151 ## Extra lookahead sets in report.  ##
1152 ## -------------------------------- ##
1153
1154 AT_SETUP([[Extra lookahead sets in report]])
1155
1156 # Bison prints each reduction's lookahead set only next to the associated
1157 # state's one item that (1) is associated with the same rule as the reduction
1158 # and (2) has its dot at the end of its RHS.  Previously, Bison also
1159 # erroneously printed the lookahead set next to all of the state's other items
1160 # associated with the same rule.  This bug affected only the `.output' file and
1161 # not the generated parser source code.
1162
1163 AT_DATA([[input.y]],
1164 [[%%
1165 start: a | 'a' a 'a' ;
1166 a: 'a' ;
1167 ]])
1168
1169 AT_BISON_CHECK([[--report=all input.y]])
1170 AT_CHECK([[sed -n '/^state 1$/,/^state 2$/p' input.output]], [[0]],
1171 [[state 1
1172
1173     2 start: 'a' . a 'a'
1174     3 a: . 'a'
1175     3  | 'a' .  [$end]
1176
1177     'a'  shift, and go to state 4
1178
1179     $default  reduce using rule 3 (a)
1180
1181     a  go to state 5
1182
1183
1184 state 2
1185 ]])
1186
1187 AT_CLEANUP
1188
1189
1190
1191 ## ---------------------------------------- ##
1192 ## Token number in precedence declaration.  ##
1193 ## ---------------------------------------- ##
1194
1195 AT_SETUP([[Token number in precedence declaration]])
1196
1197 # POSIX says token numbers can be declared in %left, %right, and %nonassoc, but
1198 # we lost this in Bison 1.50.
1199
1200 AT_DATA_GRAMMAR([input.y],
1201 [[%{
1202   #include <stdio.h>
1203   void yyerror (char const *);
1204   int yylex (void);
1205 %}
1206
1207 %error-verbose
1208 %left TK1 1 TK2 2 "tok alias" 3
1209
1210 %%
1211
1212 start: TK1 sr_conflict "tok alias" ;
1213
1214 sr_conflict:
1215   TK2
1216   | TK2 "tok alias"
1217   ;
1218
1219 %%
1220
1221 void
1222 yyerror (char const *msg)
1223 {
1224   fprintf (stderr, "%s\n", msg);
1225 }
1226
1227 int
1228 yylex (void)
1229 {
1230   static int const input[] = { 1, 2, 3, 0 };
1231   static int const *inputp = input;
1232   return *inputp++;
1233 }
1234
1235 int
1236 main (void)
1237 {
1238   return yyparse ();
1239 }
1240 ]])
1241
1242 AT_BISON_CHECK([[-o input.c input.y]], [[0]],,
1243 [[input.y:24.5-19: warning: rule useless in parser due to conflicts: sr_conflict: TK2 "tok alias"
1244 ]])
1245 AT_COMPILE([[input]])
1246 AT_PARSER_CHECK([[./input]])
1247
1248 AT_CLEANUP
1249
1250
1251
1252 ## --------------------------- ##
1253 ## parse-gram.y: LALR = IELR.  ##
1254 ## --------------------------- ##
1255
1256 # If parse-gram.y's LALR and IELR parser tables ever begin to differ, we
1257 # need to fix parse-gram.y or start using IELR.
1258
1259 AT_SETUP([[parse-gram.y: LALR = IELR]])
1260
1261 # Avoid differences in synclines by telling bison that the output files
1262 # have the same name.
1263 [cp $abs_top_srcdir/src/parse-gram.y input.y]
1264 AT_BISON_CHECK([[-o input.c -Dlr.type=lalr input.y]])
1265 [mv input.c lalr.c]
1266 AT_BISON_CHECK([[-o input.c -Dlr.type=ielr input.y]])
1267 [mv input.c ielr.c]
1268 AT_CHECK([[diff -u lalr.c ielr.c]])
1269
1270 AT_CLEANUP
1271
1272
1273
1274 ## --------------------------------------- ##
1275 ## %error-verbose and YYSTACK_USE_ALLOCA.  ##
1276 ## --------------------------------------- ##
1277
1278 AT_SETUP([[%error-verbose and YYSTACK_USE_ALLOCA]])
1279
1280 AT_DATA_GRAMMAR([input.y],
1281 [[%code {
1282   #include <stdio.h>
1283   void yyerror (char const *);
1284   int yylex (void);
1285   #define YYSTACK_USE_ALLOCA 1
1286 }
1287
1288 %error-verbose
1289
1290 %%
1291
1292 start: check syntax_error syntax_error ;
1293
1294 check:
1295 {
1296   if (128 < sizeof yymsgbuf)
1297     {
1298       fprintf (stderr,
1299                "The initial size of yymsgbuf in yyparse has increased\n"
1300                "since this test group was last updated.  As a result,\n"
1301                "this test group may no longer manage to induce a\n"
1302                "reallocation of the syntax error message buffer.\n"
1303                "This test group must be adjusted to produce a longer\n"
1304                "error message.\n");
1305       YYABORT;
1306     }
1307 }
1308 ;
1309
1310 // Induce a syntax error message whose total length is more than
1311 // sizeof yymsgbuf in yyparse.  Each token here is 64 bytes.
1312 syntax_error:
1313   "123456789112345678921234567893123456789412345678951234567896123A"
1314 | "123456789112345678921234567893123456789412345678951234567896123B"
1315 | error 'a' 'b' 'c'
1316 ;
1317
1318 %%
1319
1320 void
1321 yyerror (char const *msg)
1322 {
1323   fprintf (stderr, "%s\n", msg);
1324 }
1325
1326 int
1327 yylex (void)
1328 {
1329   /* Induce two syntax error messages (which requires full error
1330      recovery by shifting 3 tokens) in order to detect any loss of the
1331      reallocated buffer.  */
1332   static char const *input = "abc";
1333   return *input++;
1334 }
1335
1336 int
1337 main (void)
1338 {
1339   return yyparse ();
1340 }
1341 ]])
1342
1343 AT_BISON_CHECK([[-o input.c input.y]])
1344 AT_COMPILE([[input]])
1345 AT_PARSER_CHECK([[./input]], [[1]], [],
1346 [[syntax error, unexpected 'a', expecting 123456789112345678921234567893123456789412345678951234567896123A or 123456789112345678921234567893123456789412345678951234567896123B
1347 syntax error, unexpected $end, expecting 123456789112345678921234567893123456789412345678951234567896123A or 123456789112345678921234567893123456789412345678951234567896123B
1348 ]])
1349
1350 AT_CLEANUP
1351
1352
1353
1354 ## ------------------------- ##
1355 ## %error-verbose overflow.  ##
1356 ## ------------------------- ##
1357
1358 # Imagine the case where YYSTACK_ALLOC_MAXIMUM = YYSIZE_MAXIMUM and an
1359 # invocation of yysyntax_error has caused yymsg_alloc to grow to exactly
1360 # YYSTACK_ALLOC_MAXIMUM (perhaps because the normal doubling of size had
1361 # to be clipped to YYSTACK_ALLOC_MAXIMUM).  In an old version of yacc.c,
1362 # a subsequent invocation of yysyntax_error that overflows during its
1363 # size calculation would return YYSIZE_MAXIMUM to yyparse.  Then,
1364 # yyparse would invoke yyerror using the old contents of yymsg.
1365
1366 AT_SETUP([[%error-verbose overflow]])
1367
1368 AT_DATA_GRAMMAR([input.y],
1369 [[%code {
1370   #include <stdio.h>
1371   void yyerror (char const *);
1372   int yylex (void);
1373
1374   /* This prevents this test case from having to induce error messages
1375      large enough to overflow size_t.  */
1376   #define YYSIZE_T unsigned char
1377
1378   /* Bring in malloc so yacc.c doesn't try to provide a malloc prototype
1379      using our YYSIZE_T.  */
1380   #include <stdlib.h>
1381
1382   /* Max depth is usually much smaller than YYSTACK_ALLOC_MAXIMUM, and
1383      we don't want gcc to warn everywhere this constant would be too big
1384      to make sense for our YYSIZE_T.  */
1385   #define YYMAXDEPTH 100
1386 }
1387
1388 %error-verbose
1389
1390 %%
1391
1392 start: syntax_error1 check syntax_error2 ;
1393
1394 // Induce a syntax error message whose total length causes yymsg in
1395 // yyparse to be reallocated to size YYSTACK_ALLOC_MAXIMUM, which
1396 // should be 255.  Each token here is 64 bytes.
1397 syntax_error1:
1398   "123456789112345678921234567893123456789412345678951234567896123A"
1399 | "123456789112345678921234567893123456789412345678951234567896123B"
1400 | "123456789112345678921234567893123456789412345678951234567896123C"
1401 | error 'a' 'b' 'c'
1402 ;
1403
1404 check:
1405 {
1406   if (yymsg_alloc != YYSTACK_ALLOC_MAXIMUM
1407       || YYSTACK_ALLOC_MAXIMUM != YYSIZE_MAXIMUM
1408       || YYSIZE_MAXIMUM != 255)
1409     {
1410       fprintf (stderr,
1411                "The assumptions of this test group are no longer\n"
1412                "valid, so it may no longer catch the error it was\n"
1413                "designed to catch.  Specifically, the following\n"
1414                "values should all be 255:\n\n");
1415       fprintf (stderr, "  yymsg_alloc = %d\n", yymsg_alloc);
1416       fprintf (stderr, "  YYSTACK_ALLOC_MAXIMUM = %d\n",
1417                YYSTACK_ALLOC_MAXIMUM);
1418       fprintf (stderr, "  YYSIZE_MAXIMUM = %d\n", YYSIZE_MAXIMUM);
1419       YYABORT;
1420     }
1421 }
1422 ;
1423
1424 // Now overflow.
1425 syntax_error2:
1426   "123456789112345678921234567893123456789412345678951234567896123A"
1427 | "123456789112345678921234567893123456789412345678951234567896123B"
1428 | "123456789112345678921234567893123456789412345678951234567896123C"
1429 | "123456789112345678921234567893123456789412345678951234567896123D"
1430 | "123456789112345678921234567893123456789412345678951234567896123E"
1431 ;
1432
1433 %%
1434
1435 void
1436 yyerror (char const *msg)
1437 {
1438   fprintf (stderr, "%s\n", msg);
1439 }
1440
1441 int
1442 yylex (void)
1443 {
1444   /* Induce two syntax error messages (which requires full error
1445      recovery by shifting 3 tokens).  */
1446   static char const *input = "abc";
1447   return *input++;
1448 }
1449
1450 int
1451 main (void)
1452 {
1453   /* Push parsers throw away the message buffer between tokens, so skip
1454      this test under maintainer-push-check.  */
1455   if (YYPUSH)
1456     return 77;
1457   return yyparse ();
1458 }
1459 ]])
1460
1461 AT_BISON_CHECK([[-o input.c input.y]])
1462
1463 # gcc warns about tautologies and fallacies involving comparisons for
1464 # unsigned char.  However, it doesn't produce these same warnings for
1465 # size_t and many other types when the warnings would seem to make just
1466 # as much sense.  We ignore the warnings.
1467 [CFLAGS="$NO_WERROR_CFLAGS"]
1468 AT_COMPILE([[input]])
1469
1470 AT_PARSER_CHECK([[./input]], [[2]], [],
1471 [[syntax error, unexpected 'a', expecting 123456789112345678921234567893123456789412345678951234567896123A or 123456789112345678921234567893123456789412345678951234567896123B or 123456789112345678921234567893123456789412345678951234567896123C
1472 syntax error
1473 memory exhausted
1474 ]])
1475
1476 AT_CLEANUP