]> git.saurik.com Git - bison.git/blob - tests/regression.at
* src/reader.c, src/symlist.h, src/symlist.c: s/mid_rule/midrule/.
[bison.git] / 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 AT_SETUP([Invalid inputs with {}])
307
308 AT_DATA([input.y],
309 [[
310 %destructor
311 %initial-action
312 %lex-param
313 %parse-param
314 %printer
315 %union
316 ]])
317
318 AT_CHECK([bison input.y], [1], [],
319 [[input.y:3.1: missing `{' in "%destructor {...}"
320 input.y:4.1: missing `{' in "%initial-action {...}"
321 input.y:4.1: syntax error, unexpected %initial-action {...}, expecting string or identifier
322 ]])
323
324 AT_CLEANUP
325
326
327
328 ## ------------------- ##
329 ## Token definitions. ##
330 ## ------------------- ##
331
332
333 AT_SETUP([Token definitions])
334
335 # Bison managed, when fed with `%token 'f' "f"' to #define 'f'!
336 AT_DATA_GRAMMAR([input.y],
337 [%{
338 #include <stdio.h>
339 void yyerror (const char *s);
340 int yylex (void);
341 %}
342 [%error-verbose
343 %token MYEOF 0 "end of file"
344 %token 'a' "a"
345 %token B_TOKEN "b"
346 %token C_TOKEN 'c'
347 %token 'd' D_TOKEN
348 %token SPECIAL "\\\'\?\"\a\b\f\n\r\t\v\001\201\x001\x000081??!"
349 %%
350 exp: "a" "\\\'\?\"\a\b\f\n\r\t\v\001\201\x001\x000081??!";
351 %%
352 void
353 yyerror (char const *s)
354 {
355 fprintf (stderr, "%s\n", s);
356 }
357
358 int
359 yylex (void)
360 {
361 return SPECIAL;
362 }
363
364 int
365 main (void)
366 {
367 return yyparse ();
368 }
369 ]])
370
371 AT_CHECK([bison -o input.c input.y])
372 AT_COMPILE([input])
373 AT_DATA([experr],
374 [[syntax error, unexpected "\\'?\"\a\b\f\n\r\t\v\001\201\001\201?\?!", expecting a
375 ]])
376 AT_PARSER_CHECK([./input], 1, [], [experr])
377 AT_CLEANUP
378
379
380
381 ## -------------------- ##
382 ## Characters Escapes. ##
383 ## -------------------- ##
384
385
386 AT_SETUP([Characters Escapes])
387
388 AT_DATA_GRAMMAR([input.y],
389 [%{
390 void yyerror (const char *s);
391 int yylex (void);
392 %}
393 [%%
394 exp:
395 '\'' "\'"
396 | '\"' "\""
397 | '"' "'"
398 ;
399 ]])
400 # Pacify font-lock-mode: "
401
402 AT_CHECK([bison -o input.c input.y])
403 AT_COMPILE([input.o], [-c input.c])
404 AT_CLEANUP
405
406
407
408 ## -------------- ##
409 ## Web2c Report. ##
410 ## -------------- ##
411
412 # The generation of the reduction was once wrong in Bison, and made it
413 # miss some reductions. In the following test case, the reduction on
414 # `undef_id_tok' in state 1 was missing. This is stripped down from
415 # the actual web2c.y.
416
417 AT_SETUP([Web2c Report])
418
419 AT_KEYWORDS([report])
420
421 AT_DATA([input.y],
422 [[%token undef_id_tok const_id_tok
423
424 %start CONST_DEC_PART
425 \f
426 %%
427 CONST_DEC_PART:
428 CONST_DEC_LIST
429 ;
430
431 CONST_DEC_LIST:
432 CONST_DEC
433 | CONST_DEC_LIST CONST_DEC
434 ;
435
436 CONST_DEC:
437 { } undef_id_tok '=' const_id_tok ';'
438 ;
439 %%
440 ]])
441
442 AT_CHECK([bison -v input.y])
443 AT_CHECK([cat input.output], 0,
444 [[Grammar
445
446 0 $accept: CONST_DEC_PART $end
447
448 1 CONST_DEC_PART: CONST_DEC_LIST
449
450 2 CONST_DEC_LIST: CONST_DEC
451 3 | CONST_DEC_LIST CONST_DEC
452
453 4 @1: /* empty */
454
455 5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';'
456
457
458 Terminals, with rules where they appear
459
460 $end (0) 0
461 ';' (59) 5
462 '=' (61) 5
463 error (256)
464 undef_id_tok (258) 5
465 const_id_tok (259) 5
466
467
468 Nonterminals, with rules where they appear
469
470 $accept (7)
471 on left: 0
472 CONST_DEC_PART (8)
473 on left: 1, on right: 0
474 CONST_DEC_LIST (9)
475 on left: 2 3, on right: 1 3
476 CONST_DEC (10)
477 on left: 5, on right: 2 3
478 @1 (11)
479 on left: 4, on right: 5
480
481
482 state 0
483
484 0 $accept: . CONST_DEC_PART $end
485
486 $default reduce using rule 4 (@1)
487
488 CONST_DEC_PART go to state 1
489 CONST_DEC_LIST go to state 2
490 CONST_DEC go to state 3
491 @1 go to state 4
492
493
494 state 1
495
496 0 $accept: CONST_DEC_PART . $end
497
498 $end shift, and go to state 5
499
500
501 state 2
502
503 1 CONST_DEC_PART: CONST_DEC_LIST .
504 3 CONST_DEC_LIST: CONST_DEC_LIST . CONST_DEC
505
506 undef_id_tok reduce using rule 4 (@1)
507 $default reduce using rule 1 (CONST_DEC_PART)
508
509 CONST_DEC go to state 6
510 @1 go to state 4
511
512
513 state 3
514
515 2 CONST_DEC_LIST: CONST_DEC .
516
517 $default reduce using rule 2 (CONST_DEC_LIST)
518
519
520 state 4
521
522 5 CONST_DEC: @1 . undef_id_tok '=' const_id_tok ';'
523
524 undef_id_tok shift, and go to state 7
525
526
527 state 5
528
529 0 $accept: CONST_DEC_PART $end .
530
531 $default accept
532
533
534 state 6
535
536 3 CONST_DEC_LIST: CONST_DEC_LIST CONST_DEC .
537
538 $default reduce using rule 3 (CONST_DEC_LIST)
539
540
541 state 7
542
543 5 CONST_DEC: @1 undef_id_tok . '=' const_id_tok ';'
544
545 '=' shift, and go to state 8
546
547
548 state 8
549
550 5 CONST_DEC: @1 undef_id_tok '=' . const_id_tok ';'
551
552 const_id_tok shift, and go to state 9
553
554
555 state 9
556
557 5 CONST_DEC: @1 undef_id_tok '=' const_id_tok . ';'
558
559 ';' shift, and go to state 10
560
561
562 state 10
563
564 5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';' .
565
566 $default reduce using rule 5 (CONST_DEC)
567 ]])
568
569 AT_CLEANUP
570
571
572 ## --------------- ##
573 ## Web2c Actions. ##
574 ## --------------- ##
575
576 # The generation of the mapping `state -> action' was once wrong in
577 # extremely specific situations. web2c.y exhibits this situation.
578 # Below is a stripped version of the grammar. It looks like one can
579 # simplify it further, but just don't: it is tuned to exhibit a bug,
580 # which disapears when applying sane grammar transformations.
581 #
582 # It used to be wrong on yydefact only:
583 #
584 # static const short int yydefact[] =
585 # {
586 # - 2, 0, 1, 0, 0, 2, 3, 2, 5, 4,
587 # + 2, 0, 1, 0, 0, 0, 3, 2, 5, 4,
588 # 0, 0
589 # };
590 #
591 # but let's check all the tables.
592
593
594 AT_SETUP([Web2c Actions])
595
596 AT_KEYWORDS([report])
597
598 AT_DATA([input.y],
599 [[%%
600 statement: struct_stat;
601 struct_stat: /* empty. */ | if else;
602 if: "if" "const" "then" statement;
603 else: "else" statement;
604 %%
605 ]])
606
607 AT_CHECK([bison -v -o input.c input.y])
608
609 # Check only the tables. We don't use --no-parser, because it is
610 # still to be implemented in the experimental branch of Bison.
611 [sed -n 's/ *$//;/^static const.*\[\] =/,/^}/p' input.c >tables.c]
612
613 AT_CHECK([[cat tables.c]], 0,
614 [[static const unsigned char yytranslate[] =
615 {
616 0, 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, 2, 2, 2, 2,
621 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
622 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
623 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
624 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
625 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
626 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
627 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
628 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
629 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
630 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
631 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
632 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
633 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
634 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
635 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
636 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
637 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
638 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
639 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
640 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
641 2, 2, 2, 2, 2, 2, 1, 2, 3, 4,
642 5, 6
643 };
644 static const unsigned char yyprhs[] =
645 {
646 0, 0, 3, 5, 6, 9, 14
647 };
648 static const yysigned_char yyrhs[] =
649 {
650 8, 0, -1, 9, -1, -1, 10, 11, -1, 3,
651 4, 5, 8, -1, 6, 8, -1
652 };
653 static const unsigned char yyrline[] =
654 {
655 0, 2, 2, 3, 3, 4, 5
656 };
657 static const char *const yytname[] =
658 {
659 "$end", "error", "$undefined", "\"if\"", "\"const\"", "\"then\"",
660 "\"else\"", "$accept", "statement", "struct_stat", "if", "else", 0
661 };
662 static const unsigned short int yytoknum[] =
663 {
664 0, 256, 257, 258, 259, 260, 261
665 };
666 static const unsigned char yyr1[] =
667 {
668 0, 7, 8, 9, 9, 10, 11
669 };
670 static const unsigned char yyr2[] =
671 {
672 0, 2, 1, 0, 2, 4, 2
673 };
674 static const unsigned char yydefact[] =
675 {
676 3, 0, 0, 2, 0, 0, 1, 3, 4, 3,
677 6, 5
678 };
679 static const yysigned_char yydefgoto[] =
680 {
681 -1, 2, 3, 4, 8
682 };
683 static const yysigned_char yypact[] =
684 {
685 -2, -1, 4, -8, 0, 2, -8, -2, -8, -2,
686 -8, -8
687 };
688 static const yysigned_char yypgoto[] =
689 {
690 -8, -7, -8, -8, -8
691 };
692 static const unsigned char yytable[] =
693 {
694 10, 1, 11, 5, 6, 0, 7, 9
695 };
696 static const yysigned_char yycheck[] =
697 {
698 7, 3, 9, 4, 0, -1, 6, 5
699 };
700 static const unsigned char yystos[] =
701 {
702 0, 3, 8, 9, 10, 4, 0, 6, 11, 5,
703 8, 8
704 };
705 ]])
706
707 AT_CLEANUP
708
709
710 ## ------------------------- ##
711 ## yycheck Bound Violation. ##
712 ## ------------------------- ##
713
714
715 # _AT_DATA_DANCER_Y(BISON-OPTIONS)
716 # --------------------------------
717 # The following grammar, taken from Andrew Suffield's GPL'd implementation
718 # of DGMTP, the Dancer Generic Message Transport Protocol, used to violate
719 # yycheck's bounds where issuing a verbose error message. Keep this test
720 # so that possible bound checking compilers could check all the skeletons.
721 m4_define([_AT_DATA_DANCER_Y],
722 [AT_DATA_GRAMMAR([dancer.y],
723 [%{
724 static int yylex (AT_LALR1_CC_IF([int *], [void]));
725 AT_LALR1_CC_IF([],
726 [#include <stdio.h>
727 static void yyerror (const char *);])
728 %}
729 $1
730 %token ARROW INVALID NUMBER STRING DATA
731 %defines
732 %verbose
733 %error-verbose
734 /* Grammar follows */
735 %%
736 line: header body
737 ;
738
739 header: '<' from ARROW to '>' type ':'
740 | '<' ARROW to '>' type ':'
741 | ARROW to type ':'
742 | type ':'
743 | '<' '>'
744 ;
745
746 from: DATA
747 | STRING
748 | INVALID
749 ;
750
751 to: DATA
752 | STRING
753 | INVALID
754 ;
755
756 type: DATA
757 | STRING
758 | INVALID
759 ;
760
761 body: /* empty */
762 | body member
763 ;
764
765 member: STRING
766 | DATA
767 | '+' NUMBER
768 | '-' NUMBER
769 | NUMBER
770 | INVALID
771 ;
772 %%
773 AT_LALR1_CC_IF(
774 [/* A C++ error reporting function. */
775 void
776 yy::parser::error (const location&, const std::string& m)
777 {
778 std::cerr << m << std::endl;
779 }
780
781 int
782 yyparse ()
783 {
784 yy::parser parser;
785 parser.set_debug_level (!!YYDEBUG);
786 return parser.parse ();
787 }
788 ],
789 [static void
790 yyerror (const char *s)
791 {
792 fprintf (stderr, "%s\n", s);
793 }])
794
795 static int
796 yylex (AT_LALR1_CC_IF([int *lval], [void]))
797 [{
798 static int toknum = 0;
799 int tokens[] =
800 {
801 ':', -1
802 };
803 ]AT_LALR1_CC_IF([*lval = 0; /* Pacify GCC. */])[
804 return tokens[toknum++];
805 }]
806
807 int
808 main (void)
809 {
810 return yyparse ();
811 }
812 ])
813 ])# _AT_DATA_DANCER_Y
814
815
816 # AT_CHECK_DANCER(BISON-OPTIONS)
817 # ------------------------------
818 # Generate the grammar, compile it, run it.
819 m4_define([AT_CHECK_DANCER],
820 [AT_SETUP([Dancer $1])
821 AT_BISON_OPTION_PUSHDEFS([$1])
822 _AT_DATA_DANCER_Y([$1])
823 AT_CHECK([bison -o dancer.c dancer.y])
824 AT_LALR1_CC_IF(
825 [AT_CHECK([bison -o dancer.cc dancer.y])
826 AT_COMPILE_CXX([dancer])],
827 [AT_CHECK([bison -o dancer.c dancer.y])
828 AT_COMPILE([dancer])])
829 AT_PARSER_CHECK([./dancer], 1, [],
830 [syntax error, unexpected ':'
831 ])
832 AT_BISON_OPTION_POPDEFS
833 AT_CLEANUP
834 ])
835
836 AT_CHECK_DANCER()
837 AT_CHECK_DANCER([%glr-parser])
838 AT_CHECK_DANCER([%skeleton "lalr1.cc"])