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 input.y:7.1-8.0: syntax error, unexpected "%{...%}"
 302 ]])
 303
 304 AT_CLEANUP
 305
 306
 307
 308 ## ------------------- ##
 309 ## Token definitions.  ##
 310 ## ------------------- ##
 311
 312
 313 AT_SETUP([Token definitions])
 314
 315 # Bison managed, when fed with `%token 'f' "f"' to #define 'f'!
 316 AT_DATA_GRAMMAR([input.y],
 317 [%{
 318 void yyerror (const char *s);
 319 int yylex (void);
 320 %}
 321 [%token MYEOF 0 "end of file"
 322 %token 'a' "a"
 323 %token b "b"
 324 %token c 'c'
 325 %token 'd' d
 326 %%
 327 exp: "a";
 328 ]])
 329
 330 AT_CHECK([bison -o input.c input.y])
 331 AT_COMPILE([input.o], [-c input.c])
 332 AT_CLEANUP
 333
 334
 335
 336 ## -------------------- ##
 337 ## Characters Escapes.  ##
 338 ## -------------------- ##
 339
 340
 341 AT_SETUP([Characters Escapes])
 342
 343 AT_DATA_GRAMMAR([input.y],
 344 [%{
 345 void yyerror (const char *s);
 346 int yylex (void);
 347 %}
 348 [%%
 349 exp:
 350   '\'' "\'"
 351 | '\"' "\""
 352 | '"'  "'"
 353 ;
 354 ]])
 355 # Pacify font-lock-mode: "
 356
 357 AT_CHECK([bison -o input.c input.y])
 358 AT_COMPILE([input.o], [-c input.c])
 359 AT_CLEANUP
 360
 361
 362
 363 ## -------------- ##
 364 ## Web2c Report.  ##
 365 ## -------------- ##
 366
 367 # The generation of the reduction was once wrong in Bison, and made it
 368 # miss some reductions.  In the following test case, the reduction on
 369 # `undef_id_tok' in state 1 was missing.  This is stripped down from
 370 # the actual web2c.y.
 371
 372 AT_SETUP([Web2c Report])
 373
 374 AT_KEYWORDS([report])
 375
 376 AT_DATA([input.y],
 377 [[%token        undef_id_tok const_id_tok
 378
 379 %start CONST_DEC_PART
 380 \f
 381 %%
 382 CONST_DEC_PART:
 383          CONST_DEC_LIST
 384         ;
 385
 386 CONST_DEC_LIST:
 387           CONST_DEC
 388         | CONST_DEC_LIST CONST_DEC
 389         ;
 390
 391 CONST_DEC:
 392           { } undef_id_tok '=' const_id_tok ';'
 393         ;
 394 %%
 395 ]])
 396
 397 AT_CHECK([bison -v input.y])
 398 AT_CHECK([cat input.output], 0,
 399 [[Grammar
 400
 401     0 $accept: CONST_DEC_PART $end
 402
 403     1 CONST_DEC_PART: CONST_DEC_LIST
 404
 405     2 CONST_DEC_LIST: CONST_DEC
 406     3               | CONST_DEC_LIST CONST_DEC
 407
 408     4 @1: /* empty */
 409
 410     5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';'
 411
 412
 413 Terminals, with rules where they appear
 414
 415 $end (0) 0
 416 ';' (59) 5
 417 '=' (61) 5
 418 error (256)
 419 undef_id_tok (258) 5
 420 const_id_tok (259) 5
 421
 422
 423 Nonterminals, with rules where they appear
 424
 425 $accept (7)
 426     on left: 0
 427 CONST_DEC_PART (8)
 428     on left: 1, on right: 0
 429 CONST_DEC_LIST (9)
 430     on left: 2 3, on right: 1 3
 431 CONST_DEC (10)
 432     on left: 5, on right: 2 3
 433 @1 (11)
 434     on left: 4, on right: 5
 435
 436
 437 state 0
 438
 439     0 $accept: . CONST_DEC_PART $end
 440
 441     $default  reduce using rule 4 (@1)
 442
 443     CONST_DEC_PART  go to state 1
 444     CONST_DEC_LIST  go to state 2
 445     CONST_DEC       go to state 3
 446     @1              go to state 4
 447
 448
 449 state 1
 450
 451     0 $accept: CONST_DEC_PART . $end
 452
 453     $end  shift, and go to state 5
 454
 455
 456 state 2
 457
 458     1 CONST_DEC_PART: CONST_DEC_LIST .
 459     3 CONST_DEC_LIST: CONST_DEC_LIST . CONST_DEC
 460
 461     undef_id_tok  reduce using rule 4 (@1)
 462     $default      reduce using rule 1 (CONST_DEC_PART)
 463
 464     CONST_DEC  go to state 6
 465     @1         go to state 4
 466
 467
 468 state 3
 469
 470     2 CONST_DEC_LIST: CONST_DEC .
 471
 472     $default  reduce using rule 2 (CONST_DEC_LIST)
 473
 474
 475 state 4
 476
 477     5 CONST_DEC: @1 . undef_id_tok '=' const_id_tok ';'
 478
 479     undef_id_tok  shift, and go to state 7
 480
 481
 482 state 5
 483
 484     0 $accept: CONST_DEC_PART $end .
 485
 486     $default  accept
 487
 488
 489 state 6
 490
 491     3 CONST_DEC_LIST: CONST_DEC_LIST CONST_DEC .
 492
 493     $default  reduce using rule 3 (CONST_DEC_LIST)
 494
 495
 496 state 7
 497
 498     5 CONST_DEC: @1 undef_id_tok . '=' const_id_tok ';'
 499
 500     '='  shift, and go to state 8
 501
 502
 503 state 8
 504
 505     5 CONST_DEC: @1 undef_id_tok '=' . const_id_tok ';'
 506
 507     const_id_tok  shift, and go to state 9
 508
 509
 510 state 9
 511
 512     5 CONST_DEC: @1 undef_id_tok '=' const_id_tok . ';'
 513
 514     ';'  shift, and go to state 10
 515
 516
 517 state 10
 518
 519     5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';' .
 520
 521     $default  reduce using rule 5 (CONST_DEC)
 522 ]])
 523
 524 AT_CLEANUP
 525
 526
 527 ## --------------- ##
 528 ## Web2c Actions.  ##
 529 ## --------------- ##
 530
 531 # The generation of the mapping `state -> action' was once wrong in
 532 # extremely specific situations.  web2c.y exhibits this situation.
 533 # Below is a stripped version of the grammar.  It looks like one can
 534 # simplify it further, but just don't: it is tuned to exhibit a bug,
 535 # which disapears when applying sane grammar transformations.
 536 #
 537 # It used to be wrong on yydefact only:
 538 #
 539 # static const short yydefact[] =
 540 #  {
 541 # -       2,     0,     1,     0,     0,     2,     3,     2,     5,     4,
 542 # +       2,     0,     1,     0,     0,     0,     3,     2,     5,     4,
 543 #         0,     0
 544 #  };
 545 #
 546 # but let's check all the tables.
 547
 548
 549 AT_SETUP([Web2c Actions])
 550
 551 AT_KEYWORDS([report])
 552
 553 AT_DATA([input.y],
 554 [[%%
 555 statement:  struct_stat;
 556 struct_stat:  /* empty. */ | if else;
 557 if: "if" "const" "then" statement;
 558 else: "else" statement;
 559 %%
 560 ]])
 561
 562 AT_CHECK([bison -v -o input.c input.y])
 563
 564 # Check only the tables.  We don't use --no-parser, because it is
 565 # still to be implemented in the experimental branch of Bison.
 566 [sed -n 's/  *$//;/^static const.*\[\] =/,/^}/p' input.c >tables.c]
 567
 568 AT_CHECK([[cat tables.c]], 0,
 569 [[static const unsigned char yytranslate[] =
 570 {
 571        0,     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,     2,     2,     2,     2,
 596        2,     2,     2,     2,     2,     2,     1,     2,     3,     4,
 597        5,     6
 598 };
 599 static const unsigned char yyprhs[] =
 600 {
 601        0,     0,     3,     5,     6,     9,    14
 602 };
 603 static const yysigned_char yyrhs[] =
 604 {
 605        8,     0,    -1,     9,    -1,    -1,    10,    11,    -1,     3,
 606        4,     5,     8,    -1,     6,     8,    -1
 607 };
 608 static const unsigned char yyrline[] =
 609 {
 610        0,     2,     2,     3,     3,     4,     5
 611 };
 612 static const char *const yytname[] =
 613 {
 614   "$end", "error", "$undefined", "\"if\"", "\"const\"", "\"then\"",
 615   "\"else\"", "$accept", "statement", "struct_stat", "if", "else", 0
 616 };
 617 static const unsigned short yytoknum[] =
 618 {
 619        0,   256,   257,   258,   259,   260,   261
 620 };
 621 static const unsigned char yyr1[] =
 622 {
 623        0,     7,     8,     9,     9,    10,    11
 624 };
 625 static const unsigned char yyr2[] =
 626 {
 627        0,     2,     1,     0,     2,     4,     2
 628 };
 629 static const unsigned char yydefact[] =
 630 {
 631        3,     0,     0,     2,     0,     0,     1,     3,     4,     3,
 632        6,     5
 633 };
 634 static const yysigned_char yydefgoto[] =
 635 {
 636       -1,     2,     3,     4,     8
 637 };
 638 static const yysigned_char yypact[] =
 639 {
 640       -2,    -1,     4,    -8,     0,     2,    -8,    -2,    -8,    -2,
 641       -8,    -8
 642 };
 643 static const yysigned_char yypgoto[] =
 644 {
 645       -8,    -7,    -8,    -8,    -8
 646 };
 647 static const unsigned char yytable[] =
 648 {
 649       10,     1,    11,     5,     6,     0,     7,     9
 650 };
 651 static const yysigned_char yycheck[] =
 652 {
 653        7,     3,     9,     4,     0,    -1,     6,     5
 654 };
 655 static const unsigned char yystos[] =
 656 {
 657        0,     3,     8,     9,    10,     4,     0,     6,    11,     5,
 658        8,     8
 659 };
 660 ]])
 661
 662 AT_CLEANUP
 663
 664
 665 ## ------------------------- ##
 666 ## yycheck Bound Violation.  ##
 667 ## ------------------------- ##
 668
 669
 670 # _AT_DATA_DANCER_Y(BISON-OPTIONS)
 671 # --------------------------------
 672 # The following grammar, taken from Andrew Suffield's GPL'd implementation
 673 # of DGMTP, the Dancer Generic Message Transport Protocol, used to violate
 674 # yycheck's bounds where issuing a verbose error message.  Keep this test
 675 # so that possible bound checking compilers could check all the skeletons.
 676 m4_define([_AT_DATA_DANCER_Y],
 677 [AT_DATA_GRAMMAR([dancer.y],
 678 [%{
 679 AT_LALR1_CC_IF(
 680 [static int yylex (int *lval);],
 681 [#include <stdio.h>
 682 static void yyerror (const char *s);
 683 static int yylex ();])
 684 %}
 685 $1
 686 %token ARROW INVALID NUMBER STRING DATA
 687 %defines
 688 %verbose
 689 %error-verbose
 690 /* Grammar follows */
 691 %%
 692 line: header body
 693    ;
 694
 695 header: '<' from ARROW to '>' type ':'
 696    | '<' ARROW to '>' type ':'
 697    | ARROW to type ':'
 698    | type ':'
 699    | '<' '>'
 700    ;
 701
 702 from: DATA
 703    | STRING
 704    | INVALID
 705    ;
 706
 707 to: DATA
 708    | STRING
 709    | INVALID
 710    ;
 711
 712 type: DATA
 713    | STRING
 714    | INVALID
 715    ;
 716
 717 body: /* empty */
 718    | body member
 719    ;
 720
 721 member: STRING
 722    | DATA
 723    | '+' NUMBER
 724    | '-' NUMBER
 725    | NUMBER
 726    | INVALID
 727    ;
 728 %%
 729 AT_LALR1_CC_IF(
 730 [/* Currently, print_ is required in C++. */
 731 void
 732 yy::Parser::print_ ()
 733 {
 734 }
 735
 736 /* A C++ error reporting function. */
 737 void
 738 yy::Parser::error_ ()
 739 {
 740   std::cerr << message << std::endl;
 741 }
 742
 743 int
 744 yyparse (void)
 745 {
 746   yy::Parser parser (!!YYDEBUG);
 747   return parser.parse ();
 748 }
 749 ],
 750 [static void
 751 yyerror (const char *s)
 752 {
 753   fprintf (stderr, "%s\n", s);
 754 }])
 755
 756 static int
 757 yylex (AT_LALR1_CC_IF([int *lval]))
 758 [{
 759   static int toknum = 0;
 760   int tokens[] =
 761     {
 762       ':', -1
 763     };
 764   return tokens[toknum++];
 765 }]
 766
 767 int
 768 main (void)
 769 {
 770   return yyparse ();
 771 }
 772 ])
 773 ])# _AT_DATA_DANCER_Y
 774
 775
 776 # AT_CHECK_DANCER(BISON-OPTIONS)
 777 # ------------------------------
 778 # Generate the grammar, compile it, run it.
 779 m4_define([AT_CHECK_DANCER],
 780 [AT_SETUP([Dancer $1])
 781 AT_BISON_OPTION_PUSHDEFS([$1])
 782 _AT_DATA_DANCER_Y([$1])
 783 AT_CHECK([bison -o dancer.c dancer.y])
 784 AT_LALR1_CC_IF([AT_COMPILE_CXX([dancer])],
 785                [AT_COMPILE([dancer])])
 786 AT_PARSER_CHECK([./dancer], 1, [],
 787 [syntax error, unexpected ':'
 788 ])
 789 AT_BISON_OPTION_POPDEFS
 790 AT_CLEANUP
 791 ])
 792
 793 AT_CHECK_DANCER()
 794 AT_CHECK_DANCER([%glr-parser])
 795 AT_CHECK_DANCER([%skeleton "lalr1.cc"])