tests/regression.at

   1 # Bison Regressions.                               -*- Autotest -*-
   2 # Copyright (C) 2001, 2002, 2003 Free Software Foundation, Inc.
   3
   4 # This program is free software; you can redistribute it and/or modify
   5 # it under the terms of the GNU General Public License as published by
   6 # the Free Software Foundation; either version 2, or (at your option)
   7 # any later version.
   8
   9 # This program is distributed in the hope that it will be useful,
  10 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12 # GNU General Public License for more details.
  13
  14 # You should have received a copy of the GNU General Public License
  15 # along with this program; if not, write to the Free Software
  16 # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
  17 # 02111-1307, USA.
  18
  19 AT_BANNER([[Regression tests.]])
  20
  21
  22 ## ------------------------- ##
  23 ## Early token definitions.  ##
  24 ## ------------------------- ##
  25
  26
  27 AT_SETUP([Early token definitions])
  28
  29 # Found in GCJ: they expect the tokens to be defined before the user
  30 # prologue, so that they can use the token definitions in it.
  31
  32 AT_DATA_GRAMMAR([input.y],
  33 [[%{
  34 void yyerror (const char *s);
  35 int yylex (void);
  36 %}
  37
  38 %union
  39 {
  40   int val;
  41 };
  42 %{
  43 #ifndef MY_TOKEN
  44 # error "MY_TOKEN not defined."
  45 #endif
  46 %}
  47 %token MY_TOKEN
  48 %%
  49 exp: MY_TOKEN;
  50 %%
  51 ]])
  52
  53 AT_CHECK([bison -o input.c input.y])
  54 AT_COMPILE([input.o], [-c input.c])
  55
  56 AT_CLEANUP
  57
  58
  59
  60 ## ---------------- ##
  61 ## Braces parsing.  ##
  62 ## ---------------- ##
  63
  64
  65 AT_SETUP([Braces parsing])
  66
  67 AT_DATA([input.y],
  68 [[/* Bison used to swallow the character after `}'. */
  69
  70 %%
  71 exp: { tests = {{{{{{{{{{}}}}}}}}}}; };
  72 %%
  73 ]])
  74
  75 AT_CHECK([bison -v -o input.c input.y])
  76
  77 AT_CHECK([grep 'tests = {{{{{{{{{{}}}}}}}}}};' input.c], 0, [ignore])
  78
  79 AT_CLEANUP
  80
  81
  82 ## ------------------ ##
  83 ## Duplicate string.  ##
  84 ## ------------------ ##
  85
  86
  87 AT_SETUP([Duplicate string])
  88
  89 AT_DATA([input.y],
  90 [[/* `Bison -v' used to dump core when two tokens are defined with the same
  91    string, as LE and GE below. */
  92
  93 %token NUM
  94 %token LE "<="
  95 %token GE "<="
  96
  97 %%
  98 exp: '(' exp ')' | NUM ;
  99 %%
 100 ]])
 101
 102 AT_CHECK([bison -v -o input.c input.y], 0, [],
 103 [[input.y:6.8-14: warning: symbol `"<="' used more than once as a literal string
 104 ]])
 105
 106 AT_CLEANUP
 107
 108
 109 ## ------------------- ##
 110 ## Rule Line Numbers.  ##
 111 ## ------------------- ##
 112
 113 AT_SETUP([Rule Line Numbers])
 114
 115 AT_KEYWORDS([report])
 116
 117 AT_DATA([input.y],
 118 [[%%
 119 expr:
 120 'a'
 121
 122 {
 123
 124 }
 125
 126 'b'
 127
 128 {
 129
 130 }
 131
 132 |
 133
 134
 135 {
 136
 137
 138 }
 139
 140 'c'
 141
 142 {
 143
 144 };
 145 ]])
 146
 147 AT_CHECK([bison -o input.c -v input.y])
 148
 149 # Check the contents of the report.
 150 AT_CHECK([cat input.output], [],
 151 [[Grammar
 152
 153     0 $accept: expr $end
 154
 155     1 @1: /* empty */
 156
 157     2 expr: 'a' @1 'b'
 158
 159     3 @2: /* empty */
 160
 161     4 expr: @2 'c'
 162
 163
 164 Terminals, with rules where they appear
 165
 166 $end (0) 0
 167 'a' (97) 2
 168 'b' (98) 2
 169 'c' (99) 4
 170 error (256)
 171
 172
 173 Nonterminals, with rules where they appear
 174
 175 $accept (6)
 176     on left: 0
 177 expr (7)
 178     on left: 2 4, on right: 0
 179 @1 (8)
 180     on left: 1, on right: 2
 181 @2 (9)
 182     on left: 3, on right: 4
 183
 184
 185 state 0
 186
 187     0 $accept: . expr $end
 188
 189     'a'  shift, and go to state 1
 190
 191     $default  reduce using rule 3 (@2)
 192
 193     expr  go to state 2
 194     @2    go to state 3
 195
 196
 197 state 1
 198
 199     2 expr: 'a' . @1 'b'
 200
 201     $default  reduce using rule 1 (@1)
 202
 203     @1  go to state 4
 204
 205
 206 state 2
 207
 208     0 $accept: expr . $end
 209
 210     $end  shift, and go to state 5
 211
 212
 213 state 3
 214
 215     4 expr: @2 . 'c'
 216
 217     'c'  shift, and go to state 6
 218
 219
 220 state 4
 221
 222     2 expr: 'a' @1 . 'b'
 223
 224     'b'  shift, and go to state 7
 225
 226
 227 state 5
 228
 229     0 $accept: expr $end .
 230
 231     $default  accept
 232
 233
 234 state 6
 235
 236     4 expr: @2 'c' .
 237
 238     $default  reduce using rule 4 (expr)
 239
 240
 241 state 7
 242
 243     2 expr: 'a' @1 'b' .
 244
 245     $default  reduce using rule 2 (expr)
 246 ]])
 247
 248 AT_CLEANUP
 249
 250
 251
 252 ## ---------------------- ##
 253 ## Mixing %token styles.  ##
 254 ## ---------------------- ##
 255
 256
 257 AT_SETUP([Mixing %token styles])
 258
 259 # Taken from the documentation.
 260 AT_DATA([input.y],
 261 [[%token  <operator>  OR      "||"
 262 %token  <operator>  LE 134  "<="
 263 %left  OR  "<="
 264 %%
 265 exp: ;
 266 %%
 267 ]])
 268
 269 AT_CHECK([bison -v -o input.c input.y])
 270
 271 AT_CLEANUP
 272
 273
 274
 275 ## ---------------- ##
 276 ## Invalid inputs.  ##
 277 ## ---------------- ##
 278
 279
 280 AT_SETUP([Invalid inputs])
 281
 282 AT_DATA([input.y],
 283 [[%%
 284 ?
 285 default: 'a' }
 286 %&
 287 %a-does-not-exist
 288 %-
 289 %{
 290 ]])
 291
 292 AT_CHECK([bison input.y], [1], [],
 293 [[input.y:2.1: invalid character: `?'
 294 input.y:3.14: invalid character: `}'
 295 input.y:4.1: invalid character: `%'
 296 input.y:4.2: invalid character: `&'
 297 input.y:5.1-17: invalid directive: `%a-does-not-exist'
 298 input.y:6.1: invalid character: `%'
 299 input.y:6.2: invalid character: `-'
 300 input.y:7.1-8.0: missing `%}' at end of file
 301 ]])
 302
 303 AT_CLEANUP
 304
 305
 306
 307 ## ------------------- ##
 308 ## Token definitions.  ##
 309 ## ------------------- ##
 310
 311
 312 AT_SETUP([Token definitions])
 313
 314 # Bison managed, when fed with `%token 'f' "f"' to #define 'f'!
 315 AT_DATA_GRAMMAR([input.y],
 316 [%{
 317 void yyerror (const char *s);
 318 int yylex (void);
 319 %}
 320 [%token MYEOF 0 "end of file"
 321 %token 'a' "a"
 322 %token b "b"
 323 %token c 'c'
 324 %token 'd' d
 325 %%
 326 exp: "a";
 327 ]])
 328
 329 AT_CHECK([bison -o input.c input.y])
 330 AT_COMPILE([input.o], [-c input.c])
 331 AT_CLEANUP
 332
 333
 334
 335 ## -------------------- ##
 336 ## Characters Escapes.  ##
 337 ## -------------------- ##
 338
 339
 340 AT_SETUP([Characters Escapes])
 341
 342 AT_DATA_GRAMMAR([input.y],
 343 [%{
 344 void yyerror (const char *s);
 345 int yylex (void);
 346 %}
 347 [%%
 348 exp:
 349   '\'' "\'"
 350 | '\"' "\""
 351 | '"'  "'"
 352 ;
 353 ]])
 354 # Pacify font-lock-mode: "
 355
 356 AT_CHECK([bison -o input.c input.y])
 357 AT_COMPILE([input.o], [-c input.c])
 358 AT_CLEANUP
 359
 360
 361
 362 ## -------------- ##
 363 ## Web2c Report.  ##
 364 ## -------------- ##
 365
 366 # The generation of the reduction was once wrong in Bison, and made it
 367 # miss some reductions.  In the following test case, the reduction on
 368 # `undef_id_tok' in state 1 was missing.  This is stripped down from
 369 # the actual web2c.y.
 370
 371 AT_SETUP([Web2c Report])
 372
 373 AT_KEYWORDS([report])
 374
 375 AT_DATA([input.y],
 376 [[%token        undef_id_tok const_id_tok
 377
 378 %start CONST_DEC_PART
 379 \f
 380 %%
 381 CONST_DEC_PART:
 382          CONST_DEC_LIST
 383         ;
 384
 385 CONST_DEC_LIST:
 386           CONST_DEC
 387         | CONST_DEC_LIST CONST_DEC
 388         ;
 389
 390 CONST_DEC:
 391           { } undef_id_tok '=' const_id_tok ';'
 392         ;
 393 %%
 394 ]])
 395
 396 AT_CHECK([bison -v input.y])
 397 AT_CHECK([cat input.output], 0,
 398 [[Grammar
 399
 400     0 $accept: CONST_DEC_PART $end
 401
 402     1 CONST_DEC_PART: CONST_DEC_LIST
 403
 404     2 CONST_DEC_LIST: CONST_DEC
 405     3               | CONST_DEC_LIST CONST_DEC
 406
 407     4 @1: /* empty */
 408
 409     5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';'
 410
 411
 412 Terminals, with rules where they appear
 413
 414 $end (0) 0
 415 ';' (59) 5
 416 '=' (61) 5
 417 error (256)
 418 undef_id_tok (258) 5
 419 const_id_tok (259) 5
 420
 421
 422 Nonterminals, with rules where they appear
 423
 424 $accept (7)
 425     on left: 0
 426 CONST_DEC_PART (8)
 427     on left: 1, on right: 0
 428 CONST_DEC_LIST (9)
 429     on left: 2 3, on right: 1 3
 430 CONST_DEC (10)
 431     on left: 5, on right: 2 3
 432 @1 (11)
 433     on left: 4, on right: 5
 434
 435
 436 state 0
 437
 438     0 $accept: . CONST_DEC_PART $end
 439
 440     $default  reduce using rule 4 (@1)
 441
 442     CONST_DEC_PART  go to state 1
 443     CONST_DEC_LIST  go to state 2
 444     CONST_DEC       go to state 3
 445     @1              go to state 4
 446
 447
 448 state 1
 449
 450     0 $accept: CONST_DEC_PART . $end
 451
 452     $end  shift, and go to state 5
 453
 454
 455 state 2
 456
 457     1 CONST_DEC_PART: CONST_DEC_LIST .
 458     3 CONST_DEC_LIST: CONST_DEC_LIST . CONST_DEC
 459
 460     undef_id_tok  reduce using rule 4 (@1)
 461     $default      reduce using rule 1 (CONST_DEC_PART)
 462
 463     CONST_DEC  go to state 6
 464     @1         go to state 4
 465
 466
 467 state 3
 468
 469     2 CONST_DEC_LIST: CONST_DEC .
 470
 471     $default  reduce using rule 2 (CONST_DEC_LIST)
 472
 473
 474 state 4
 475
 476     5 CONST_DEC: @1 . undef_id_tok '=' const_id_tok ';'
 477
 478     undef_id_tok  shift, and go to state 7
 479
 480
 481 state 5
 482
 483     0 $accept: CONST_DEC_PART $end .
 484
 485     $default  accept
 486
 487
 488 state 6
 489
 490     3 CONST_DEC_LIST: CONST_DEC_LIST CONST_DEC .
 491
 492     $default  reduce using rule 3 (CONST_DEC_LIST)
 493
 494
 495 state 7
 496
 497     5 CONST_DEC: @1 undef_id_tok . '=' const_id_tok ';'
 498
 499     '='  shift, and go to state 8
 500
 501
 502 state 8
 503
 504     5 CONST_DEC: @1 undef_id_tok '=' . const_id_tok ';'
 505
 506     const_id_tok  shift, and go to state 9
 507
 508
 509 state 9
 510
 511     5 CONST_DEC: @1 undef_id_tok '=' const_id_tok . ';'
 512
 513     ';'  shift, and go to state 10
 514
 515
 516 state 10
 517
 518     5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';' .
 519
 520     $default  reduce using rule 5 (CONST_DEC)
 521 ]])
 522
 523 AT_CLEANUP
 524
 525
 526 ## --------------- ##
 527 ## Web2c Actions.  ##
 528 ## --------------- ##
 529
 530 # The generation of the mapping `state -> action' was once wrong in
 531 # extremely specific situations.  web2c.y exhibits this situation.
 532 # Below is a stripped version of the grammar.  It looks like one can
 533 # simplify it further, but just don't: it is tuned to exhibit a bug,
 534 # which disapears when applying sane grammar transformations.
 535 #
 536 # It used to be wrong on yydefact only:
 537 #
 538 # static const short yydefact[] =
 539 #  {
 540 # -       2,     0,     1,     0,     0,     2,     3,     2,     5,     4,
 541 # +       2,     0,     1,     0,     0,     0,     3,     2,     5,     4,
 542 #         0,     0
 543 #  };
 544 #
 545 # but let's check all the tables.
 546
 547
 548 AT_SETUP([Web2c Actions])
 549
 550 AT_KEYWORDS([report])
 551
 552 AT_DATA([input.y],
 553 [[%%
 554 statement:  struct_stat;
 555 struct_stat:  /* empty. */ | if else;
 556 if: "if" "const" "then" statement;
 557 else: "else" statement;
 558 %%
 559 ]])
 560
 561 AT_CHECK([bison -v -o input.c input.y])
 562
 563 # Check only the tables.  We don't use --no-parser, because it is
 564 # still to be implemented in the experimental branch of Bison.
 565 [sed -n 's/  *$//;/^static const.*\[\] =/,/^}/p' input.c >tables.c]
 566
 567 AT_CHECK([[cat tables.c]], 0,
 568 [[static const unsigned char yytranslate[] =
 569 {
 570        0,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 571        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 572        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 573        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 574        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 575        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 576        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 577        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 578        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 579        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 580        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 581        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 582        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 583        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 584        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 585        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 586        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 587        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 588        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 589        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 590        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 591        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 592        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 593        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 594        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 595        2,     2,     2,     2,     2,     2,     1,     2,     3,     4,
 596        5,     6
 597 };
 598 static const unsigned char yyprhs[] =
 599 {
 600        0,     0,     3,     5,     6,     9,    14
 601 };
 602 static const yysigned_char yyrhs[] =
 603 {
 604        8,     0,    -1,     9,    -1,    -1,    10,    11,    -1,     3,
 605        4,     5,     8,    -1,     6,     8,    -1
 606 };
 607 static const unsigned char yyrline[] =
 608 {
 609        0,     2,     2,     3,     3,     4,     5
 610 };
 611 static const char *const yytname[] =
 612 {
 613   "$end", "error", "$undefined", "\"if\"", "\"const\"", "\"then\"",
 614   "\"else\"", "$accept", "statement", "struct_stat", "if", "else", 0
 615 };
 616 static const unsigned short yytoknum[] =
 617 {
 618        0,   256,   257,   258,   259,   260,   261
 619 };
 620 static const unsigned char yyr1[] =
 621 {
 622        0,     7,     8,     9,     9,    10,    11
 623 };
 624 static const unsigned char yyr2[] =
 625 {
 626        0,     2,     1,     0,     2,     4,     2
 627 };
 628 static const unsigned char yydefact[] =
 629 {
 630        3,     0,     0,     2,     0,     0,     1,     3,     4,     3,
 631        6,     5
 632 };
 633 static const yysigned_char yydefgoto[] =
 634 {
 635       -1,     2,     3,     4,     8
 636 };
 637 static const yysigned_char yypact[] =
 638 {
 639       -2,    -1,     4,    -8,     0,     2,    -8,    -2,    -8,    -2,
 640       -8,    -8
 641 };
 642 static const yysigned_char yypgoto[] =
 643 {
 644       -8,    -7,    -8,    -8,    -8
 645 };
 646 static const unsigned char yytable[] =
 647 {
 648       10,     1,    11,     5,     6,     0,     7,     9
 649 };
 650 static const yysigned_char yycheck[] =
 651 {
 652        7,     3,     9,     4,     0,    -1,     6,     5
 653 };
 654 static const unsigned char yystos[] =
 655 {
 656        0,     3,     8,     9,    10,     4,     0,     6,    11,     5,
 657        8,     8
 658 };
 659 ]])
 660
 661 AT_CLEANUP
 662
 663
 664 ## ------------------------- ##
 665 ## yycheck Bound Violation.  ##
 666 ## ------------------------- ##
 667
 668
 669 # _AT_DATA_DANCER_Y(BISON-OPTIONS)
 670 # --------------------------------
 671 # The following grammar, taken from Andrew Suffield's GPL'd implementation
 672 # of DGMTP, the Dancer Generic Message Transport Protocol, used to violate
 673 # yycheck's bounds where issuing a verbose error message.  Keep this test
 674 # so that possible bound checking compilers could check all the skeletons.
 675 m4_define([_AT_DATA_DANCER_Y],
 676 [AT_DATA_GRAMMAR([dancer.y],
 677 [%{
 678 AT_LALR1_CC_IF(
 679 [static int yylex (int *lval);],
 680 [#include <stdio.h>
 681 static void yyerror (const char *s);
 682 static int yylex ();])
 683 %}
 684 $1
 685 %token ARROW INVALID NUMBER STRING DATA
 686 %defines
 687 %verbose
 688 %error-verbose
 689 /* Grammar follows */
 690 %%
 691 line: header body
 692    ;
 693
 694 header: '<' from ARROW to '>' type ':'
 695    | '<' ARROW to '>' type ':'
 696    | ARROW to type ':'
 697    | type ':'
 698    | '<' '>'
 699    ;
 700
 701 from: DATA
 702    | STRING
 703    | INVALID
 704    ;
 705
 706 to: DATA
 707    | STRING
 708    | INVALID
 709    ;
 710
 711 type: DATA
 712    | STRING
 713    | INVALID
 714    ;
 715
 716 body: /* empty */
 717    | body member
 718    ;
 719
 720 member: STRING
 721    | DATA
 722    | '+' NUMBER
 723    | '-' NUMBER
 724    | NUMBER
 725    | INVALID
 726    ;
 727 %%
 728 AT_LALR1_CC_IF(
 729 [/* Currently, print_ is required in C++. */
 730 void
 731 yy::Parser::print_ ()
 732 {
 733 }
 734
 735 /* A C++ error reporting function. */
 736 void
 737 yy::Parser::error_ ()
 738 {
 739   std::cerr << message << std::endl;
 740 }
 741
 742 int
 743 yyparse (void)
 744 {
 745   yy::Parser parser (!!YYDEBUG);
 746   return parser.parse ();
 747 }
 748 ],
 749 [static void
 750 yyerror (const char *s)
 751 {
 752   fprintf (stderr, "%s\n", s);
 753 }])
 754
 755 static int
 756 yylex (AT_LALR1_CC_IF([int *lval]))
 757 [{
 758   static int toknum = 0;
 759   int tokens[] =
 760     {
 761       ':', -1
 762     };
 763   return tokens[toknum++];
 764 }]
 765
 766 int
 767 main (void)
 768 {
 769   return yyparse ();
 770 }
 771 ])
 772 ])# _AT_DATA_DANCER_Y
 773
 774
 775 # AT_CHECK_DANCER(BISON-OPTIONS)
 776 # ------------------------------
 777 # Generate the grammar, compile it, run it.
 778 m4_define([AT_CHECK_DANCER],
 779 [AT_SETUP([Dancer $1])
 780 AT_BISON_OPTION_PUSHDEFS([$1])
 781 _AT_DATA_DANCER_Y([$1])
 782 AT_CHECK([bison -o dancer.c dancer.y])
 783 AT_LALR1_CC_IF([AT_COMPILE_CXX([dancer])],
 784                [AT_COMPILE([dancer])])
 785 AT_PARSER_CHECK([./dancer], 1, [],
 786 [syntax error, unexpected ':'
 787 ])
 788 AT_BISON_OPTION_POPDEFS
 789 AT_CLEANUP
 790 ])
 791
 792 AT_CHECK_DANCER()
 793 AT_CHECK_DANCER([%glr-parser])
 794 AT_CHECK_DANCER([%skeleton "lalr1.cc"])