1 /* Output the generated parsing program for Bison.
3 Copyright (C) 1984, 1986, 1989, 1992, 2000-2012 Free Software
6 This file is part of Bison, the GNU Compiler Compiler.
8 This program is free software: you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation, either version 3 of the License, or
11 (at your option) any later version.
13 This program is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with this program. If not, see <http://www.gnu.org/licenses/>. */
24 #include <concat-filename.h>
25 #include <configmake.h>
27 #include <get-errno.h>
29 #include <spawn-pipe.h>
31 #include <wait-process.h>
37 #include "muscle-tab.h"
40 #include "scan-code.h" /* max_left_semantic_context */
41 #include "scan-skel.h"
45 static struct obstack format_obstack
;
48 /*-------------------------------------------------------------------.
49 | Create a function NAME which associates to the muscle NAME the |
50 | result of formatting the FIRST and then TABLE_DATA[BEGIN..END[ (of |
51 | TYPE), and to the muscle NAME_max, the max value of the |
53 `-------------------------------------------------------------------*/
56 #define GENERATE_MUSCLE_INSERT_TABLE(Name, Type) \
59 Name (char const *name, \
72 obstack_fgrow1 (&format_obstack, "%6d", first); \
73 for (i = begin; i < end; ++i) \
75 obstack_1grow (&format_obstack, ','); \
78 obstack_sgrow (&format_obstack, "\n "); \
83 obstack_fgrow1 (&format_obstack, "%6d", table_data[i]); \
84 if (table_data[i] < min) \
85 min = table_data[i]; \
86 if (max < table_data[i]) \
87 max = table_data[i]; \
89 obstack_1grow (&format_obstack, 0); \
90 muscle_insert (name, obstack_finish (&format_obstack)); \
94 /* Build `NAME_min' and `NAME_max' in the obstack. */ \
95 obstack_fgrow1 (&format_obstack, "%s_min", name); \
96 obstack_1grow (&format_obstack, 0); \
97 MUSCLE_INSERT_LONG_INT (obstack_finish (&format_obstack), lmin); \
98 obstack_fgrow1 (&format_obstack, "%s_max", name); \
99 obstack_1grow (&format_obstack, 0); \
100 MUSCLE_INSERT_LONG_INT (obstack_finish (&format_obstack), lmax); \
103 GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_unsigned_int_table
, unsigned int)
104 GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_int_table
, int)
105 GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_base_table
, base_number
)
106 GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_rule_number_table
, rule_number
)
107 GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_symbol_number_table
, symbol_number
)
108 GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_state_number_table
, state_number
)
111 /*--------------------------------------------------------------------.
112 | Print to OUT a representation of STRING escaped both for C and M4. |
113 `--------------------------------------------------------------------*/
116 escaped_output (FILE *out
, char const *string
)
121 for (p
= quotearg_style (c_quoting_style
, string
); *p
; p
++)
124 case '$': fputs ("$][", out
); break;
125 case '@': fputs ("@@", out
); break;
126 case '[': fputs ("@{", out
); break;
127 case ']': fputs ("@}", out
); break;
128 default: fputc (*p
, out
); break;
135 /*------------------------------------------------------------------.
136 | Prepare the muscles related to the symbols: translate, tname, and |
138 `------------------------------------------------------------------*/
141 prepare_symbols (void)
143 MUSCLE_INSERT_INT ("tokens_number", ntokens
);
144 MUSCLE_INSERT_INT ("nterms_number", nvars
);
145 MUSCLE_INSERT_INT ("symbols_number", nsyms
);
146 MUSCLE_INSERT_INT ("undef_token_number", undeftoken
->number
);
147 MUSCLE_INSERT_INT ("user_token_number_max", max_user_token_number
);
149 muscle_insert_symbol_number_table ("translate",
151 token_translations
[0],
152 1, max_user_token_number
+ 1);
154 /* tname -- token names. */
157 /* We assume that the table will be output starting at column 2. */
159 struct quoting_options
*qo
= clone_quoting_options (0);
160 set_quoting_style (qo
, c_quoting_style
);
161 set_quoting_flags (qo
, QA_SPLIT_TRIGRAPHS
);
162 for (i
= 0; i
< nsyms
; i
++)
164 char *cp
= quotearg_alloc (symbols
[i
]->tag
, -1, qo
);
165 /* Width of the next token, including the two quotes, the
166 comma and the space. */
167 int width
= strlen (cp
) + 2;
171 obstack_sgrow (&format_obstack
, "\n ");
176 obstack_1grow (&format_obstack
, ' ');
177 MUSCLE_OBSTACK_SGROW (&format_obstack
, cp
);
179 obstack_1grow (&format_obstack
, ',');
183 obstack_sgrow (&format_obstack
, " ]b4_null[");
185 /* Finish table and store. */
186 obstack_1grow (&format_obstack
, 0);
187 muscle_insert ("tname", obstack_finish (&format_obstack
));
190 /* Output YYTOKNUM. */
193 int *values
= xnmalloc (ntokens
, sizeof *values
);
194 for (i
= 0; i
< ntokens
; ++i
)
195 values
[i
] = symbols
[i
]->user_token_number
;
196 muscle_insert_int_table ("toknum", values
,
197 values
[0], 1, ntokens
);
203 /*----------------------------------------------------------------.
204 | Prepare the muscles related to the rules: r1, r2, rline, dprec, |
205 | merger, immediate. |
206 `----------------------------------------------------------------*/
211 unsigned int *rline
= xnmalloc (nrules
, sizeof *rline
);
212 symbol_number
*r1
= xnmalloc (nrules
, sizeof *r1
);
213 unsigned int *r2
= xnmalloc (nrules
, sizeof *r2
);
214 int *dprec
= xnmalloc (nrules
, sizeof *dprec
);
215 int *merger
= xnmalloc (nrules
, sizeof *merger
);
216 int *immediate
= xnmalloc (nrules
, sizeof *immediate
);
219 for (r
= 0; r
< nrules
; ++r
)
221 /* LHS of the rule R. */
222 r1
[r
] = rules
[r
].lhs
->number
;
223 /* Length of rule R's RHS. */
224 r2
[r
] = rule_rhs_length(&rules
[r
]);
225 /* Line where rule was defined. */
226 rline
[r
] = rules
[r
].location
.start
.line
;
227 /* Dynamic precedence (GLR). */
228 dprec
[r
] = rules
[r
].dprec
;
229 /* Merger-function index (GLR). */
230 merger
[r
] = rules
[r
].merger
;
231 /* Immediate reduction flags (GLR). */
232 immediate
[r
] = rules
[r
].is_predicate
;
235 muscle_insert_unsigned_int_table ("rline", rline
, 0, 0, nrules
);
236 muscle_insert_symbol_number_table ("r1", r1
, 0, 0, nrules
);
237 muscle_insert_unsigned_int_table ("r2", r2
, 0, 0, nrules
);
238 muscle_insert_int_table ("dprec", dprec
, 0, 0, nrules
);
239 muscle_insert_int_table ("merger", merger
, 0, 0, nrules
);
240 muscle_insert_int_table ("immediate", immediate
, 0, 0, nrules
);
242 MUSCLE_INSERT_INT ("rules_number", nrules
);
243 MUSCLE_INSERT_INT ("max_left_semantic_context", max_left_semantic_context
);
253 /*--------------------------------------------.
254 | Prepare the muscles related to the states. |
255 `--------------------------------------------*/
258 prepare_states (void)
261 symbol_number
*values
= xnmalloc (nstates
, sizeof *values
);
262 for (i
= 0; i
< nstates
; ++i
)
263 values
[i
] = states
[i
]->accessing_symbol
;
264 muscle_insert_symbol_number_table ("stos", values
,
268 MUSCLE_INSERT_INT ("last", high
);
269 MUSCLE_INSERT_INT ("final_state_number", final_state
->number
);
270 MUSCLE_INSERT_INT ("states_number", nstates
);
274 /*-------------------------------------------------------.
275 | Compare two symbols by type-name, and then by number. |
276 `-------------------------------------------------------*/
279 symbol_type_name_cmp (const symbol
**lhs
, const symbol
**rhs
)
281 int res
= UNIQSTR_CMP((*lhs
)->type_name
, (*rhs
)->type_name
);
284 return (*lhs
)->number
- (*rhs
)->number
;
288 /*----------------------------------------------------------------.
289 | Return a (malloc'ed) table of the symbols sorted by type-name. |
290 `----------------------------------------------------------------*/
293 symbols_by_type_name (void)
295 typedef int (*qcmp_type
) (const void *, const void *);
296 symbol
**res
= xmemdup (symbols
, nsyms
* sizeof *res
);
297 qsort (res
, nsyms
, sizeof *res
, (qcmp_type
) &symbol_type_name_cmp
);
302 /*------------------------------------------------------------------.
303 | Define b4_type_names, which is a list of (lists of the numbers of |
304 | symbols with same type-name). |
305 `------------------------------------------------------------------*/
308 type_names_output (FILE *out
)
311 symbol
**syms
= symbols_by_type_name ();
312 fputs ("m4_define([b4_type_names],\n[", out
);
313 for (i
= 0; i
< nsyms
; /* nothing */)
315 // The index of the first symbol of the current type-name.
317 fputs (i
? ",\n[" : "[", out
);
318 for (; i
< nsyms
&& syms
[i
]->type_name
== syms
[i0
]->type_name
; ++i
)
319 fprintf (out
, "%s%d", i
!= i0
? ", " : "", syms
[i
]->number
);
322 fputs ("])\n\n", out
);
327 /*-------------------------------------.
328 | The list of all the symbol numbers. |
329 `-------------------------------------*/
332 symbol_numbers_output (FILE *out
)
335 fputs ("m4_define([b4_symbol_numbers],\n[", out
);
336 for (i
= 0; i
< nsyms
; ++i
)
337 fprintf (out
, "%s[%d]", i
? ", " : "", i
);
338 fputs ("])\n\n", out
);
342 /*---------------------------------.
343 | Output the user actions to OUT. |
344 `---------------------------------*/
347 user_actions_output (FILE *out
)
351 fputs ("m4_define([b4_actions], \n[", out
);
352 for (r
= 0; r
< nrules
; ++r
)
355 fprintf (out
, "b4_%scase(%d, [b4_syncline(%d, ",
356 rules
[r
].is_predicate
? "predicate_" : "",
357 r
+ 1, rules
[r
].action_location
.start
.line
);
358 escaped_output (out
, rules
[r
].action_location
.start
.file
);
359 fprintf (out
, ")\n[ %s]])\n\n", rules
[r
].action
);
361 fputs ("])\n\n", out
);
364 /*------------------------------------.
365 | Output the merge functions to OUT. |
366 `------------------------------------*/
369 merger_output (FILE *out
)
374 fputs ("m4_define([b4_mergers], \n[[", out
);
375 for (n
= 1, p
= merge_functions
; p
!= NULL
; n
+= 1, p
= p
->next
)
377 if (p
->type
[0] == '\0')
378 fprintf (out
, " case %d: *yy0 = %s (*yy0, *yy1); break;\n",
381 fprintf (out
, " case %d: yy0->%s = %s (*yy0, *yy1); break;\n",
382 n
, p
->type
, p
->name
);
384 fputs ("]])\n\n", out
);
388 /*---------------------------------------------.
389 | Prepare the muscles for symbol definitions. |
390 `---------------------------------------------*/
393 prepare_symbol_definitions (void)
396 for (i
= 0; i
< nsyms
; ++i
)
398 symbol
*sym
= symbols
[i
];
402 #define SET_KEY(Entry) \
403 obstack_fgrow2 (&format_obstack, "symbol(%d, %s)", \
405 obstack_1grow (&format_obstack, 0); \
406 key = obstack_finish (&format_obstack);
408 #define SET_KEY2(Entry, Suffix) \
409 obstack_fgrow3 (&format_obstack, "symbol(%d, %s_%s)", \
411 obstack_1grow (&format_obstack, 0); \
412 key = obstack_finish (&format_obstack);
414 // Whether the symbol has an identifier.
415 value
= symbol_id_get (sym
);
417 MUSCLE_INSERT_INT (key
, !!value
);
421 MUSCLE_INSERT_STRING (key
, value
? value
: "");
423 // Its tag. Typically for documentation purpose.
425 MUSCLE_INSERT_STRING (key
, sym
->tag
);
427 SET_KEY("user_number");
428 MUSCLE_INSERT_INT (key
, sym
->user_token_number
);
431 MUSCLE_INSERT_INT (key
,
432 i
< ntokens
&& sym
!= errtoken
&& sym
!= undeftoken
);
435 MUSCLE_INSERT_INT (key
, sym
->number
);
438 MUSCLE_INSERT_INT (key
, !!sym
->type_name
);
441 MUSCLE_INSERT_STRING (key
, sym
->type_name
? sym
->type_name
: "");
445 for (j
= 0; j
< CODE_PROPS_SIZE
; ++j
)
447 /* "printer", not "%printer". */
448 char const *pname
= code_props_type_string (j
) + 1;
449 code_props
const *p
= symbol_code_props_get (sym
, j
);
450 SET_KEY2("has", pname
);
451 MUSCLE_INSERT_INT (key
, !!p
->code
);
455 SET_KEY2(pname
, "file");
456 MUSCLE_INSERT_STRING (key
, p
->location
.start
.file
);
458 SET_KEY2(pname
, "line");
459 MUSCLE_INSERT_INT (key
, p
->location
.start
.line
);
462 MUSCLE_INSERT_STRING_RAW (key
, p
->code
);
472 /*--------------------------------------.
473 | Output the tokens definition to OUT. |
474 `--------------------------------------*/
477 token_definitions_output (FILE *out
)
480 char const *sep
= "";
482 fputs ("m4_define([b4_tokens], \n[", out
);
483 for (i
= 0; i
< ntokens
; ++i
)
485 symbol
*sym
= symbols
[i
];
486 int number
= sym
->user_token_number
;
487 uniqstr id
= symbol_id_get (sym
);
489 /* At this stage, if there are literal string aliases, they are
490 part of SYMBOLS, so we should not find their aliased symbols
492 aver (number
!= USER_NUMBER_HAS_STRING_ALIAS
);
494 /* Skip error token and tokens without identifier. */
495 if (sym
!= errtoken
&& id
)
497 fprintf (out
, "%s[[[%s]], %d]",
502 fputs ("])\n\n", out
);
507 prepare_actions (void)
509 /* Figure out the actions for the specified state, indexed by
510 lookahead token type. */
512 muscle_insert_rule_number_table ("defact", yydefact
,
513 yydefact
[0], 1, nstates
);
515 /* Figure out what to do after reducing with each rule, depending on
516 the saved state from before the beginning of parsing the data
517 that matched this rule. */
518 muscle_insert_state_number_table ("defgoto", yydefgoto
,
519 yydefgoto
[0], 1, nsyms
- ntokens
);
523 muscle_insert_base_table ("pact", base
,
524 base
[0], 1, nstates
);
525 MUSCLE_INSERT_INT ("pact_ninf", base_ninf
);
528 muscle_insert_base_table ("pgoto", base
,
529 base
[nstates
], nstates
+ 1, nvectors
);
531 muscle_insert_base_table ("table", table
,
532 table
[0], 1, high
+ 1);
533 MUSCLE_INSERT_INT ("table_ninf", table_ninf
);
535 muscle_insert_base_table ("check", check
,
536 check
[0], 1, high
+ 1);
538 /* GLR parsing slightly modifies YYTABLE and YYCHECK (and thus
539 YYPACT) so that in states with unresolved conflicts, the default
540 reduction is not used in the conflicted entries, so that there is
541 a place to put a conflict pointer.
543 This means that YYCONFLP and YYCONFL are nonsense for a non-GLR
544 parser, so we could avoid accidents by not writing them out in
545 that case. Nevertheless, it seems even better to be able to use
546 the GLR skeletons even without the non-deterministic tables. */
547 muscle_insert_unsigned_int_table ("conflict_list_heads", conflict_table
,
548 conflict_table
[0], 1, high
+ 1);
549 muscle_insert_unsigned_int_table ("conflicting_rules", conflict_list
,
550 0, 1, conflict_list_cnt
);
554 /*--------------------------------------------.
555 | Output the definitions of all the muscles. |
556 `--------------------------------------------*/
559 muscles_output (FILE *out
)
561 fputs ("m4_init()\n", out
);
563 symbol_numbers_output (out
);
564 token_definitions_output (out
);
565 type_names_output (out
);
566 user_actions_output (out
);
568 muscles_m4_output (out
);
571 /*---------------------------.
572 | Call the skeleton parser. |
573 `---------------------------*/
576 output_skeleton (void)
581 /* Compute the names of the package data dir and skeleton files. */
582 char const *m4
= (m4
= getenv ("M4")) ? m4
: M4
;
583 char const *datadir
= pkgdatadir ();
584 char *m4sugar
= xconcatenated_filename (datadir
, "m4sugar/m4sugar.m4", NULL
);
585 char *m4bison
= xconcatenated_filename (datadir
, "bison.m4", NULL
);
586 char *skel
= (IS_PATH_WITH_DIR (skeleton
)
588 : xconcatenated_filename (datadir
, skeleton
, NULL
));
590 /* Test whether m4sugar.m4 is readable, to check for proper
591 installation. A faulty installation can cause deadlock, so a
592 cheap sanity check is worthwhile. */
593 xfclose (xfopen (m4sugar
, "r"));
595 /* Create an m4 subprocess connected to us via two pipes. */
597 if (trace_flag
& trace_tools
)
598 fprintf (stderr
, "running: %s %s - %s %s\n",
599 m4
, m4sugar
, m4bison
, skel
);
601 /* Some future version of GNU M4 (most likely 1.6) may treat the -dV in a
602 position-dependent manner. Keep it as the first argument so that all
605 See the thread starting at
606 <http://lists.gnu.org/archive/html/bug-bison/2008-07/msg00000.html>
609 char const *argv
[10];
613 /* When POSIXLY_CORRECT is set, GNU M4 1.6 and later disable GNU
614 extensions, which Bison's skeletons depend on. With older M4,
615 it has no effect. M4 1.4.12 added a -g/--gnu command-line
616 option to make it explicit that a program wants GNU M4
617 extensions even when POSIXLY_CORRECT is set.
619 See the thread starting at
620 <http://lists.gnu.org/archive/html/bug-bison/2008-07/msg00000.html>
623 argv
[i
++] = M4_GNU_OPTION
;
627 if (trace_flag
& trace_m4
)
634 aver (i
<= ARRAY_CARDINALITY (argv
));
636 /* The ugly cast is because gnulib gets the const-ness wrong. */
637 pid
= create_pipe_bidi ("m4", m4
, (char **)(void*)argv
, false, true,
645 if (trace_flag
& trace_muscles
)
646 muscles_output (stderr
);
648 FILE *out
= xfdopen (filter_fd
[1], "w");
649 muscles_output (out
);
653 /* Read and process m4's output. */
654 timevar_push (TV_M4
);
656 FILE *in
= xfdopen (filter_fd
[0], "r");
658 /* scan_skel should have read all of M4's output. Otherwise, when we
659 close the pipe, we risk letting M4 report a broken-pipe to the
664 wait_subprocess (pid
, "m4", false, false, true, true, NULL
);
671 /* BISON_USE_PUSH_FOR_PULL is for the test suite and should not be
672 documented for the user. */
673 char const *cp
= getenv ("BISON_USE_PUSH_FOR_PULL");
674 bool use_push_for_pull_flag
= cp
&& *cp
&& strtol (cp
, 0, 10);
677 MUSCLE_INSERT_BOOL ("defines_flag", defines_flag
);
678 MUSCLE_INSERT_BOOL ("glr_flag", glr_parser
);
679 MUSCLE_INSERT_BOOL ("nondeterministic_flag", nondeterministic_parser
);
680 MUSCLE_INSERT_BOOL ("synclines_flag", !no_lines_flag
);
681 MUSCLE_INSERT_BOOL ("tag_seen_flag", tag_seen
);
682 MUSCLE_INSERT_BOOL ("token_table_flag", token_table_flag
);
683 MUSCLE_INSERT_BOOL ("use_push_for_pull_flag", use_push_for_pull_flag
);
684 MUSCLE_INSERT_BOOL ("yacc_flag", yacc_flag
);
687 if (spec_name_prefix
)
688 MUSCLE_INSERT_STRING ("prefix", spec_name_prefix
);
690 MUSCLE_INSERT_STRING ("file_name_all_but_ext", all_but_ext
);
692 #define DEFINE(Name) MUSCLE_INSERT_STRING (#Name, Name ? Name : "")
694 DEFINE (parser_file_name
);
695 DEFINE (spec_defines_file
);
696 DEFINE (spec_file_prefix
);
697 DEFINE (spec_graph_file
);
698 DEFINE (spec_name_prefix
);
699 DEFINE (spec_outfile
);
700 DEFINE (spec_verbose_file
);
703 /* Find the right skeleton file, and add muscles about the skeletons. */
705 MUSCLE_INSERT_C_STRING ("skeleton", skeleton
);
707 skeleton
= language
->skeleton
;
709 /* About the skeletons. */
711 /* b4_pkgdatadir is used inside m4_include in the skeletons, so digraphs
712 would never be expanded. Hopefully no one has M4-special characters in
713 his Bison installation path. */
714 MUSCLE_INSERT_STRING_RAW ("pkgdatadir", pkgdatadir ());
719 /*----------------------------------------------------------.
720 | Output the parsing tables and the parser code to ftable. |
721 `----------------------------------------------------------*/
726 obstack_init (&format_obstack
);
732 prepare_symbol_definitions ();
736 /* Process the selected skeleton file. */
739 obstack_free (&format_obstack
, NULL
);
745 char const *cp
= getenv ("BISON_PKGDATADIR");
746 return cp
? cp
: PKGDATADIR
;