tests/regression.at

   1 # Bison Regressions.                               -*- Autotest -*-
   2 # Copyright (C) 2001, 2002, 2003, 2004, 2005 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., 51 Franklin Street, Fifth Floor, Boston, MA
  17 # 02110-1301, 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 #include <stdio.h>
 318 void yyerror (const char *s);
 319 int yylex (void);
 320 %}
 321 [%error-verbose
 322 %token MYEOF 0 "end of file"
 323 %token 'a' "a"
 324 %token B_TOKEN "b"
 325 %token C_TOKEN 'c'
 326 %token 'd' D_TOKEN
 327 %token SPECIAL "\\\'\?\"\n\t??!"
 328 %%
 329 exp: "a" "\\\'\?\"\n\t??!";
 330 %%
 331 void
 332 yyerror (char const *s)
 333 {
 334   fprintf (stderr, "%s\n", s);
 335 }
 336
 337 int
 338 yylex (void)
 339 {
 340   return SPECIAL;
 341 }
 342
 343 int
 344 main (void)
 345 {
 346   return yyparse ();
 347 }
 348 ]])
 349
 350 AT_CHECK([bison -o input.c input.y])
 351 AT_COMPILE([input])
 352 AT_PARSER_CHECK([./input], 1, [],
 353 [syntax error, unexpected \'?"
 354         ??!, expecting a
 355 ])
 356 AT_CLEANUP
 357
 358
 359
 360 ## -------------------- ##
 361 ## Characters Escapes.  ##
 362 ## -------------------- ##
 363
 364
 365 AT_SETUP([Characters Escapes])
 366
 367 AT_DATA_GRAMMAR([input.y],
 368 [%{
 369 void yyerror (const char *s);
 370 int yylex (void);
 371 %}
 372 [%%
 373 exp:
 374   '\'' "\'"
 375 | '\"' "\""
 376 | '"'  "'"
 377 ;
 378 ]])
 379 # Pacify font-lock-mode: "
 380
 381 AT_CHECK([bison -o input.c input.y])
 382 AT_COMPILE([input.o], [-c input.c])
 383 AT_CLEANUP
 384
 385
 386
 387 ## -------------- ##
 388 ## Web2c Report.  ##
 389 ## -------------- ##
 390
 391 # The generation of the reduction was once wrong in Bison, and made it
 392 # miss some reductions.  In the following test case, the reduction on
 393 # `undef_id_tok' in state 1 was missing.  This is stripped down from
 394 # the actual web2c.y.
 395
 396 AT_SETUP([Web2c Report])
 397
 398 AT_KEYWORDS([report])
 399
 400 AT_DATA([input.y],
 401 [[%token        undef_id_tok const_id_tok
 402
 403 %start CONST_DEC_PART
 404 \f
 405 %%
 406 CONST_DEC_PART:
 407          CONST_DEC_LIST
 408         ;
 409
 410 CONST_DEC_LIST:
 411           CONST_DEC
 412         | CONST_DEC_LIST CONST_DEC
 413         ;
 414
 415 CONST_DEC:
 416           { } undef_id_tok '=' const_id_tok ';'
 417         ;
 418 %%
 419 ]])
 420
 421 AT_CHECK([bison -v input.y])
 422 AT_CHECK([cat input.output], 0,
 423 [[Grammar
 424
 425     0 $accept: CONST_DEC_PART $end
 426
 427     1 CONST_DEC_PART: CONST_DEC_LIST
 428
 429     2 CONST_DEC_LIST: CONST_DEC
 430     3               | CONST_DEC_LIST CONST_DEC
 431
 432     4 @1: /* empty */
 433
 434     5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';'
 435
 436
 437 Terminals, with rules where they appear
 438
 439 $end (0) 0
 440 ';' (59) 5
 441 '=' (61) 5
 442 error (256)
 443 undef_id_tok (258) 5
 444 const_id_tok (259) 5
 445
 446
 447 Nonterminals, with rules where they appear
 448
 449 $accept (7)
 450     on left: 0
 451 CONST_DEC_PART (8)
 452     on left: 1, on right: 0
 453 CONST_DEC_LIST (9)
 454     on left: 2 3, on right: 1 3
 455 CONST_DEC (10)
 456     on left: 5, on right: 2 3
 457 @1 (11)
 458     on left: 4, on right: 5
 459
 460
 461 state 0
 462
 463     0 $accept: . CONST_DEC_PART $end
 464
 465     $default  reduce using rule 4 (@1)
 466
 467     CONST_DEC_PART  go to state 1
 468     CONST_DEC_LIST  go to state 2
 469     CONST_DEC       go to state 3
 470     @1              go to state 4
 471
 472
 473 state 1
 474
 475     0 $accept: CONST_DEC_PART . $end
 476
 477     $end  shift, and go to state 5
 478
 479
 480 state 2
 481
 482     1 CONST_DEC_PART: CONST_DEC_LIST .
 483     3 CONST_DEC_LIST: CONST_DEC_LIST . CONST_DEC
 484
 485     undef_id_tok  reduce using rule 4 (@1)
 486     $default      reduce using rule 1 (CONST_DEC_PART)
 487
 488     CONST_DEC  go to state 6
 489     @1         go to state 4
 490
 491
 492 state 3
 493
 494     2 CONST_DEC_LIST: CONST_DEC .
 495
 496     $default  reduce using rule 2 (CONST_DEC_LIST)
 497
 498
 499 state 4
 500
 501     5 CONST_DEC: @1 . undef_id_tok '=' const_id_tok ';'
 502
 503     undef_id_tok  shift, and go to state 7
 504
 505
 506 state 5
 507
 508     0 $accept: CONST_DEC_PART $end .
 509
 510     $default  accept
 511
 512
 513 state 6
 514
 515     3 CONST_DEC_LIST: CONST_DEC_LIST CONST_DEC .
 516
 517     $default  reduce using rule 3 (CONST_DEC_LIST)
 518
 519
 520 state 7
 521
 522     5 CONST_DEC: @1 undef_id_tok . '=' const_id_tok ';'
 523
 524     '='  shift, and go to state 8
 525
 526
 527 state 8
 528
 529     5 CONST_DEC: @1 undef_id_tok '=' . const_id_tok ';'
 530
 531     const_id_tok  shift, and go to state 9
 532
 533
 534 state 9
 535
 536     5 CONST_DEC: @1 undef_id_tok '=' const_id_tok . ';'
 537
 538     ';'  shift, and go to state 10
 539
 540
 541 state 10
 542
 543     5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';' .
 544
 545     $default  reduce using rule 5 (CONST_DEC)
 546 ]])
 547
 548 AT_CLEANUP
 549
 550
 551 ## --------------- ##
 552 ## Web2c Actions.  ##
 553 ## --------------- ##
 554
 555 # The generation of the mapping `state -> action' was once wrong in
 556 # extremely specific situations.  web2c.y exhibits this situation.
 557 # Below is a stripped version of the grammar.  It looks like one can
 558 # simplify it further, but just don't: it is tuned to exhibit a bug,
 559 # which disapears when applying sane grammar transformations.
 560 #
 561 # It used to be wrong on yydefact only:
 562 #
 563 # static const short int yydefact[] =
 564 #  {
 565 # -       2,     0,     1,     0,     0,     2,     3,     2,     5,     4,
 566 # +       2,     0,     1,     0,     0,     0,     3,     2,     5,     4,
 567 #         0,     0
 568 #  };
 569 #
 570 # but let's check all the tables.
 571
 572
 573 AT_SETUP([Web2c Actions])
 574
 575 AT_KEYWORDS([report])
 576
 577 AT_DATA([input.y],
 578 [[%%
 579 statement:  struct_stat;
 580 struct_stat:  /* empty. */ | if else;
 581 if: "if" "const" "then" statement;
 582 else: "else" statement;
 583 %%
 584 ]])
 585
 586 AT_CHECK([bison -v -o input.c input.y])
 587
 588 # Check only the tables.  We don't use --no-parser, because it is
 589 # still to be implemented in the experimental branch of Bison.
 590 [sed -n 's/  *$//;/^static const.*\[\] =/,/^}/p' input.c >tables.c]
 591
 592 AT_CHECK([[cat tables.c]], 0,
 593 [[static const unsigned char yytranslate[] =
 594 {
 595        0,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 596        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 597        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 598        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 599        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 600        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 601        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 602        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 603        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 604        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 605        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 606        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 607        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 608        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 609        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 610        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 611        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 612        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 613        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 614        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 615        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 616        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 617        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 618        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 619        2,     2,     2,     2,     2,     2,     2,     2,     2,     2,
 620        2,     2,     2,     2,     2,     2,     1,     2,     3,     4,
 621        5,     6
 622 };
 623 static const unsigned char yyprhs[] =
 624 {
 625        0,     0,     3,     5,     6,     9,    14
 626 };
 627 static const yysigned_char yyrhs[] =
 628 {
 629        8,     0,    -1,     9,    -1,    -1,    10,    11,    -1,     3,
 630        4,     5,     8,    -1,     6,     8,    -1
 631 };
 632 static const unsigned char yyrline[] =
 633 {
 634        0,     2,     2,     3,     3,     4,     5
 635 };
 636 static const char *const yytname[] =
 637 {
 638   "$end", "error", "$undefined", "if", "const", "then", "else", "$accept",
 639   "statement", "struct_stat", "if", "else", 0
 640 };
 641 static const unsigned short int yytoknum[] =
 642 {
 643        0,   256,   257,   258,   259,   260,   261
 644 };
 645 static const unsigned char yyr1[] =
 646 {
 647        0,     7,     8,     9,     9,    10,    11
 648 };
 649 static const unsigned char yyr2[] =
 650 {
 651        0,     2,     1,     0,     2,     4,     2
 652 };
 653 static const unsigned char yydefact[] =
 654 {
 655        3,     0,     0,     2,     0,     0,     1,     3,     4,     3,
 656        6,     5
 657 };
 658 static const yysigned_char yydefgoto[] =
 659 {
 660       -1,     2,     3,     4,     8
 661 };
 662 static const yysigned_char yypact[] =
 663 {
 664       -2,    -1,     4,    -8,     0,     2,    -8,    -2,    -8,    -2,
 665       -8,    -8
 666 };
 667 static const yysigned_char yypgoto[] =
 668 {
 669       -8,    -7,    -8,    -8,    -8
 670 };
 671 static const unsigned char yytable[] =
 672 {
 673       10,     1,    11,     5,     6,     0,     7,     9
 674 };
 675 static const yysigned_char yycheck[] =
 676 {
 677        7,     3,     9,     4,     0,    -1,     6,     5
 678 };
 679 static const unsigned char yystos[] =
 680 {
 681        0,     3,     8,     9,    10,     4,     0,     6,    11,     5,
 682        8,     8
 683 };
 684 ]])
 685
 686 AT_CLEANUP
 687
 688
 689 ## ------------------------- ##
 690 ## yycheck Bound Violation.  ##
 691 ## ------------------------- ##
 692
 693
 694 # _AT_DATA_DANCER_Y(BISON-OPTIONS)
 695 # --------------------------------
 696 # The following grammar, taken from Andrew Suffield's GPL'd implementation
 697 # of DGMTP, the Dancer Generic Message Transport Protocol, used to violate
 698 # yycheck's bounds where issuing a verbose error message.  Keep this test
 699 # so that possible bound checking compilers could check all the skeletons.
 700 m4_define([_AT_DATA_DANCER_Y],
 701 [AT_DATA_GRAMMAR([dancer.y],
 702 [%{
 703 static int yylex (AT_LALR1_CC_IF([int *], [void]));
 704 AT_LALR1_CC_IF([],
 705 [#include <stdio.h>
 706 static void yyerror (const char *);])
 707 %}
 708 $1
 709 %token ARROW INVALID NUMBER STRING DATA
 710 %defines
 711 %verbose
 712 %error-verbose
 713 /* Grammar follows */
 714 %%
 715 line: header body
 716    ;
 717
 718 header: '<' from ARROW to '>' type ':'
 719    | '<' ARROW to '>' type ':'
 720    | ARROW to type ':'
 721    | type ':'
 722    | '<' '>'
 723    ;
 724
 725 from: DATA
 726    | STRING
 727    | INVALID
 728    ;
 729
 730 to: DATA
 731    | STRING
 732    | INVALID
 733    ;
 734
 735 type: DATA
 736    | STRING
 737    | INVALID
 738    ;
 739
 740 body: /* empty */
 741    | body member
 742    ;
 743
 744 member: STRING
 745    | DATA
 746    | '+' NUMBER
 747    | '-' NUMBER
 748    | NUMBER
 749    | INVALID
 750    ;
 751 %%
 752 AT_LALR1_CC_IF(
 753 [/* A C++ error reporting function. */
 754 void
 755 yy::parser::error (const location&, const std::string& m)
 756 {
 757   std::cerr << m << std::endl;
 758 }
 759
 760 int
 761 yyparse ()
 762 {
 763   yy::parser parser;
 764   parser.set_debug_level (!!YYDEBUG);
 765   return parser.parse ();
 766 }
 767 ],
 768 [static void
 769 yyerror (const char *s)
 770 {
 771   fprintf (stderr, "%s\n", s);
 772 }])
 773
 774 static int
 775 yylex (AT_LALR1_CC_IF([int *lval], [void]))
 776 [{
 777   static int toknum = 0;
 778   int tokens[] =
 779     {
 780       ':', -1
 781     };
 782   ]AT_LALR1_CC_IF([*lval = 0; /* Pacify GCC.  */])[
 783   return tokens[toknum++];
 784 }]
 785
 786 int
 787 main (void)
 788 {
 789   return yyparse ();
 790 }
 791 ])
 792 ])# _AT_DATA_DANCER_Y
 793
 794
 795 # AT_CHECK_DANCER(BISON-OPTIONS)
 796 # ------------------------------
 797 # Generate the grammar, compile it, run it.
 798 m4_define([AT_CHECK_DANCER],
 799 [AT_SETUP([Dancer $1])
 800 AT_BISON_OPTION_PUSHDEFS([$1])
 801 _AT_DATA_DANCER_Y([$1])
 802 AT_CHECK([bison -o dancer.c dancer.y])
 803 AT_LALR1_CC_IF(
 804   [AT_CHECK([bison -o dancer.cc dancer.y])
 805    AT_COMPILE_CXX([dancer])],
 806   [AT_CHECK([bison -o dancer.c dancer.y])
 807    AT_COMPILE([dancer])])
 808 AT_PARSER_CHECK([./dancer], 1, [],
 809 [syntax error, unexpected ':'
 810 ])
 811 AT_BISON_OPTION_POPDEFS
 812 AT_CLEANUP
 813 ])
 814
 815 AT_CHECK_DANCER()
 816 AT_CHECK_DANCER([%glr-parser])
 817 AT_CHECK_DANCER([%skeleton "lalr1.cc"])