-/* Output the generated parsing program for bison,
- Copyright (C) 1984, 1986, 1989, 1992, 2000, 2001, 2002
- Free Software Foundation, Inc.
+/* Output the generated parsing program for Bison.
+
+ Copyright (C) 1984, 1986, 1989, 1992, 2000-2012 Free Software
+ Foundation, Inc.
This file is part of Bison, the GNU Compiler Compiler.
- Bison is free software; you can redistribute it and/or modify it
- under the terms of the GNU General Public License as published by
- the Free Software Foundation; either version 2, or (at your option)
- any later version.
+ This program is free software: you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation, either version 3 of the License, or
+ (at your option) any later version.
- Bison is distributed in the hope that it will be useful, but
- WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- General Public License for more details.
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
You should have received a copy of the GNU General Public License
- along with Bison; see the file COPYING. If not, write to the Free
- Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
- 02111-1307, USA. */
-
-
-/* The parser tables consist of these tables.
-
- YYTRANSLATE = vector mapping yylex's token numbers into bison's
- token numbers.
-
- YYTNAME = vector of string-names indexed by bison token number.
-
- YYTOKNUM = vector of yylex token numbers corresponding to entries
- in YYTNAME.
-
- YYRLINE = vector of line-numbers of all rules. For yydebug
- printouts.
-
- YYRHS = vector of items of all rules. This is exactly what RITEMS
- contains. For yydebug and for semantic parser.
-
- YYPRHS[R] = index in YYRHS of first item for rule R.
-
- YYR1[R] = symbol number of symbol that rule R derives.
-
- YYR2[R] = number of symbols composing right hand side of rule R.
+ along with this program. If not, see <http://www.gnu.org/licenses/>. */
- YYSTOS[S] = the symbol number of the symbol that leads to state S.
-
- YYDEFACT[S] = default rule to reduce with in state s, when YYTABLE
- doesn't specify something else to do. Zero means the default is an
- error.
-
- YYDEFGOTO[I] = default state to go to after a reduction of a rule
- that generates variable NTOKENS + I, except when YYTABLE specifies
- something else to do.
-
- YYPACT[S] = index in YYTABLE of the portion describing state S.
- The lookahead token's type is used to index that portion to find
- out what to do.
-
- If the value in YYTABLE is positive, we shift the token and go to
- that state.
-
- If the value is negative, it is minus a rule number to reduce by.
-
- If the value is zero, the default action from YYDEFACT[S] is used.
-
- YYPGOTO[I] = the index in YYTABLE of the portion describing what to
- do after reducing a rule that derives variable I + NTOKENS. This
- portion is indexed by the parser state number, S, as of before the
- text for this nonterminal was read. The value from YYTABLE is the
- state to go to if the corresponding value in YYCHECK is S.
-
- YYTABLE = a vector filled with portions for different uses, found
- via YYPACT and YYPGOTO.
-
- YYCHECK = a vector indexed in parallel with YYTABLE. It indicates,
- in a roundabout way, the bounds of the portion you are trying to
- examine.
-
- Suppose that the portion of YYTABLE starts at index P and the index
- to be examined within the portion is I. Then if YYCHECK[P+I] != I,
- I is outside the bounds of what is actually allocated, and the
- default (from YYDEFACT or YYDEFGOTO) should be used. Otherwise,
- YYTABLE[P+I] should be used.
+#include <config.h>
+#include "system.h"
- YYFINAL = the state number of the termination state. YYFLAG = most
- negative short int. Used to flag ?? */
+#include <concat-filename.h>
+#include <configmake.h>
+#include <filename.h>
+#include <get-errno.h>
+#include <quotearg.h>
+#include <spawn-pipe.h>
+#include <timevar.h>
+#include <wait-process.h>
-#include "system.h"
-#include "quotearg.h"
-#include "error.h"
-#include "subpipe.h"
-#include "getargs.h"
+#include "complain.h"
#include "files.h"
+#include "getargs.h"
#include "gram.h"
-#include "complain.h"
+#include "muscle-tab.h"
#include "output.h"
#include "reader.h"
+#include "scan-code.h" /* max_left_semantic_context */
+#include "scan-skel.h"
#include "symtab.h"
#include "tables.h"
-#include "muscle_tab.h"
-
-/* From src/scan-skel.l. */
-void scan_skel (FILE *);
-
static struct obstack format_obstack;
-int error_verbose = 0;
-
-
/*-------------------------------------------------------------------.
| Create a function NAME which associates to the muscle NAME the |
`-------------------------------------------------------------------*/
-#define GENERATE_MUSCLE_INSERT_TABLE(Name, Type) \
- \
-static void \
-Name (const char *name, \
- Type *table_data, \
- Type first, \
- int begin, \
- int end) \
-{ \
- Type min = first; \
- Type max = first; \
- int i; \
- int j = 1; \
- \
- obstack_fgrow1 (&format_obstack, "%6d", first); \
- for (i = begin; i < end; ++i) \
- { \
- obstack_1grow (&format_obstack, ','); \
- if (j >= 10) \
- { \
- obstack_sgrow (&format_obstack, "\n "); \
- j = 1; \
- } \
- else \
- ++j; \
- obstack_fgrow1 (&format_obstack, "%6d", table_data[i]); \
- if (table_data[i] < min) \
- min = table_data[i]; \
- if (max < table_data[i]) \
- max = table_data[i]; \
- } \
- obstack_1grow (&format_obstack, 0); \
- muscle_insert (name, obstack_finish (&format_obstack)); \
- \
- /* Build `NAME_min' and `NAME_max' in the obstack. */ \
- obstack_fgrow1 (&format_obstack, "%s_min", name); \
- obstack_1grow (&format_obstack, 0); \
- MUSCLE_INSERT_LONG_INT (obstack_finish (&format_obstack), \
- (long int) min); \
- obstack_fgrow1 (&format_obstack, "%s_max", name); \
- obstack_1grow (&format_obstack, 0); \
- MUSCLE_INSERT_LONG_INT (obstack_finish (&format_obstack), \
- (long int) max); \
+#define GENERATE_MUSCLE_INSERT_TABLE(Name, Type) \
+ \
+static void \
+Name (char const *name, \
+ Type *table_data, \
+ Type first, \
+ int begin, \
+ int end) \
+{ \
+ Type min = first; \
+ Type max = first; \
+ long int lmin; \
+ long int lmax; \
+ int i; \
+ int j = 1; \
+ \
+ obstack_fgrow1 (&format_obstack, "%6d", first); \
+ for (i = begin; i < end; ++i) \
+ { \
+ obstack_1grow (&format_obstack, ','); \
+ if (j >= 10) \
+ { \
+ obstack_sgrow (&format_obstack, "\n "); \
+ j = 1; \
+ } \
+ else \
+ ++j; \
+ obstack_fgrow1 (&format_obstack, "%6d", table_data[i]); \
+ if (table_data[i] < min) \
+ min = table_data[i]; \
+ if (max < table_data[i]) \
+ max = table_data[i]; \
+ } \
+ obstack_1grow (&format_obstack, 0); \
+ muscle_insert (name, obstack_finish (&format_obstack)); \
+ \
+ lmin = min; \
+ lmax = max; \
+ /* Build `NAME_min' and `NAME_max' in the obstack. */ \
+ obstack_fgrow1 (&format_obstack, "%s_min", name); \
+ obstack_1grow (&format_obstack, 0); \
+ MUSCLE_INSERT_LONG_INT (obstack_finish (&format_obstack), lmin); \
+ obstack_fgrow1 (&format_obstack, "%s_max", name); \
+ obstack_1grow (&format_obstack, 0); \
+ MUSCLE_INSERT_LONG_INT (obstack_finish (&format_obstack), lmax); \
}
GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_unsigned_int_table, unsigned int)
GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_int_table, int)
-GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_short_table, short)
-GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_base_table, base_t)
-GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_rule_number_table, rule_number_t)
-GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_symbol_number_table, symbol_number_t)
-GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_item_number_table, item_number_t)
-GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_state_number_table, state_number_t)
+GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_base_table, base_number)
+GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_rule_number_table, rule_number)
+GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_symbol_number_table, symbol_number)
+GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_state_number_table, state_number)
+
+
+/*--------------------------------------------------------------------.
+| Print to OUT a representation of STRING escaped both for C and M4. |
+`--------------------------------------------------------------------*/
+static void
+escaped_output (FILE *out, char const *string)
+{
+ char const *p;
+ fprintf (out, "[[");
-/*-----------------------------------------------------------------.
-| Prepare the muscles related to the tokens: translate, tname, and |
-| toknum. |
-`-----------------------------------------------------------------*/
+ for (p = quotearg_style (c_quoting_style, string); *p; p++)
+ switch (*p)
+ {
+ case '$': fputs ("$][", out); break;
+ case '@': fputs ("@@", out); break;
+ case '[': fputs ("@{", out); break;
+ case ']': fputs ("@}", out); break;
+ default: fputc (*p, out); break;
+ }
+
+ fprintf (out, "]]");
+}
+
+
+/*------------------------------------------------------------------.
+| Prepare the muscles related to the symbols: translate, tname, and |
+| toknum. |
+`------------------------------------------------------------------*/
static void
-prepare_tokens (void)
+prepare_symbols (void)
{
+ MUSCLE_INSERT_INT ("tokens_number", ntokens);
+ MUSCLE_INSERT_INT ("nterms_number", nvars);
+ MUSCLE_INSERT_INT ("symbols_number", nsyms);
+ MUSCLE_INSERT_INT ("undef_token_number", undeftoken->number);
+ MUSCLE_INSERT_INT ("user_token_number_max", max_user_token_number);
+
muscle_insert_symbol_number_table ("translate",
- token_translations,
- token_translations[0],
- 1, max_user_token_number + 1);
+ token_translations,
+ token_translations[0],
+ 1, max_user_token_number + 1);
+ /* tname -- token names. */
{
int i;
- int j = 0;
+ /* We assume that the table will be output starting at column 2. */
+ int j = 2;
+ struct quoting_options *qo = clone_quoting_options (0);
+ set_quoting_style (qo, c_quoting_style);
+ set_quoting_flags (qo, QA_SPLIT_TRIGRAPHS);
for (i = 0; i < nsyms; i++)
{
- /* Be sure not to use twice the same QUOTEARG slot:
- SYMBOL_TAG_GET uses slot 0. */
- const char *cp =
- quotearg_n_style (1, c_quoting_style,
- symbols[i]->tag);
- /* Width of the next token, including the two quotes, the coma
- and the space. */
- int strsize = strlen (cp) + 2;
-
- if (j + strsize > 75)
- {
- obstack_sgrow (&format_obstack, "\n ");
- j = 2;
- }
-
- obstack_sgrow (&format_obstack, cp);
- obstack_sgrow (&format_obstack, ", ");
- j += strsize;
+ char *cp = quotearg_alloc (symbols[i]->tag, -1, qo);
+ /* Width of the next token, including the two quotes, the
+ comma and the space. */
+ int width = strlen (cp) + 2;
+
+ if (j + width > 75)
+ {
+ obstack_sgrow (&format_obstack, "\n ");
+ j = 1;
+ }
+
+ if (i)
+ obstack_1grow (&format_obstack, ' ');
+ MUSCLE_OBSTACK_SGROW (&format_obstack, cp);
+ free (cp);
+ obstack_1grow (&format_obstack, ',');
+ j += width;
}
- /* Add a NULL entry to list of tokens (well, 0, as NULL might not be
- defined). */
- obstack_sgrow (&format_obstack, "0");
+ free (qo);
+ obstack_sgrow (&format_obstack, " ]b4_null[");
/* Finish table and store. */
obstack_1grow (&format_obstack, 0);
/* Output YYTOKNUM. */
{
int i;
- int *values = XCALLOC (int, ntokens);
+ int *values = xnmalloc (ntokens, sizeof *values);
for (i = 0; i < ntokens; ++i)
values[i] = symbols[i]->user_token_number;
muscle_insert_int_table ("toknum", values,
- values[0], 1, ntokens);
+ values[0], 1, ntokens);
free (values);
}
}
-/*-------------------------------------------------------------.
-| Prepare the muscles related to the rules: rhs, prhs, r1, r2, |
-| rline, dprec, merger |
-`-------------------------------------------------------------*/
+/*----------------------------------------------------------------.
+| Prepare the muscles related to the rules: r1, r2, rline, dprec, |
+| merger, immediate. |
+`----------------------------------------------------------------*/
static void
prepare_rules (void)
{
- rule_number_t r;
- unsigned int i = 0;
- item_number_t *rhs = XMALLOC (item_number_t, nritems);
- unsigned int *prhs = XMALLOC (unsigned int, nrules);
- unsigned int *rline = XMALLOC (unsigned int, nrules);
- symbol_number_t *r1 = XMALLOC (symbol_number_t, nrules);
- unsigned int *r2 = XMALLOC (unsigned int, nrules);
- short *dprec = XMALLOC (short, nrules);
- short *merger = XMALLOC (short, nrules);
-
+ unsigned int *rline = xnmalloc (nrules, sizeof *rline);
+ symbol_number *r1 = xnmalloc (nrules, sizeof *r1);
+ unsigned int *r2 = xnmalloc (nrules, sizeof *r2);
+ int *dprec = xnmalloc (nrules, sizeof *dprec);
+ int *merger = xnmalloc (nrules, sizeof *merger);
+ int *immediate = xnmalloc (nrules, sizeof *immediate);
+
+ rule_number r;
for (r = 0; r < nrules; ++r)
{
- item_number_t *rhsp = NULL;
- /* Index of rule R in RHS. */
- prhs[r] = i;
- /* RHS of the rule R. */
- for (rhsp = rules[r].rhs; *rhsp >= 0; ++rhsp)
- rhs[i++] = *rhsp;
/* LHS of the rule R. */
r1[r] = rules[r].lhs->number;
/* Length of rule R's RHS. */
- r2[r] = i - prhs[r];
- /* Separator in RHS. */
- rhs[i++] = -1;
+ r2[r] = rule_rhs_length(&rules[r]);
/* Line where rule was defined. */
- rline[r] = rules[r].location.first_line;
- /* Dynamic precedence (GLR) */
+ rline[r] = rules[r].location.start.line;
+ /* Dynamic precedence (GLR). */
dprec[r] = rules[r].dprec;
- /* Merger-function index (GLR) */
+ /* Merger-function index (GLR). */
merger[r] = rules[r].merger;
+ /* Immediate reduction flags (GLR). */
+ immediate[r] = rules[r].is_predicate;
}
- assert (i == nritems);
- muscle_insert_item_number_table ("rhs", rhs, ritem[0], 1, nritems);
- muscle_insert_unsigned_int_table ("prhs", prhs, 0, 0, nrules);
muscle_insert_unsigned_int_table ("rline", rline, 0, 0, nrules);
muscle_insert_symbol_number_table ("r1", r1, 0, 0, nrules);
muscle_insert_unsigned_int_table ("r2", r2, 0, 0, nrules);
- muscle_insert_short_table ("dprec", dprec, 0, 0, nrules);
- muscle_insert_short_table ("merger", merger, 0, 0, nrules);
+ muscle_insert_int_table ("dprec", dprec, 0, 0, nrules);
+ muscle_insert_int_table ("merger", merger, 0, 0, nrules);
+ muscle_insert_int_table ("immediate", immediate, 0, 0, nrules);
+
+ MUSCLE_INSERT_INT ("rules_number", nrules);
+ MUSCLE_INSERT_INT ("max_left_semantic_context", max_left_semantic_context);
- free (rhs);
- free (prhs);
free (rline);
free (r1);
free (r2);
free (dprec);
free (merger);
+ free (immediate);
}
/*--------------------------------------------.
static void
prepare_states (void)
{
- state_number_t i;
- symbol_number_t *values =
- (symbol_number_t *) alloca (sizeof (symbol_number_t) * nstates);
+ state_number i;
+ symbol_number *values = xnmalloc (nstates, sizeof *values);
for (i = 0; i < nstates; ++i)
values[i] = states[i]->accessing_symbol;
muscle_insert_symbol_number_table ("stos", values,
- 0, 1, nstates);
+ 0, 1, nstates);
+ free (values);
+
+ MUSCLE_INSERT_INT ("last", high);
+ MUSCLE_INSERT_INT ("final_state_number", final_state->number);
+ MUSCLE_INSERT_INT ("states_number", nstates);
+}
+
+
+/*-------------------------------------------------------.
+| Compare two symbols by type-name, and then by number. |
+`-------------------------------------------------------*/
+
+static int
+symbol_type_name_cmp (const symbol **lhs, const symbol **rhs)
+{
+ int res = UNIQSTR_CMP((*lhs)->type_name, (*rhs)->type_name);
+ if (res)
+ return res;
+ return (*lhs)->number - (*rhs)->number;
}
+/*----------------------------------------------------------------.
+| Return a (malloc'ed) table of the symbols sorted by type-name. |
+`----------------------------------------------------------------*/
+
+static symbol **
+symbols_by_type_name (void)
+{
+ typedef int (*qcmp_type) (const void *, const void *);
+ symbol **res = xmemdup (symbols, nsyms * sizeof *res);
+ qsort (res, nsyms, sizeof *res, (qcmp_type) &symbol_type_name_cmp);
+ return res;
+}
+
+
+/*------------------------------------------------------------------.
+| Define b4_type_names, which is a list of (lists of the numbers of |
+| symbols with same type-name). |
+`------------------------------------------------------------------*/
+
+static void
+type_names_output (FILE *out)
+{
+ int i;
+ symbol **syms = symbols_by_type_name ();
+ fputs ("m4_define([b4_type_names],\n[", out);
+ for (i = 0; i < nsyms; /* nothing */)
+ {
+ // The index of the first symbol of the current type-name.
+ int i0 = i;
+ fputs (i ? ",\n[" : "[", out);
+ for (; i < nsyms && syms[i]->type_name == syms[i0]->type_name; ++i)
+ fprintf (out, "%s%d", i != i0 ? ", " : "", syms[i]->number);
+ fputs ("]", out);
+ }
+ fputs ("])\n\n", out);
+ free (syms);
+}
+
+
+/*-------------------------------------.
+| The list of all the symbol numbers. |
+`-------------------------------------*/
+
+static void
+symbol_numbers_output (FILE *out)
+{
+ int i;
+ fputs ("m4_define([b4_symbol_numbers],\n[", out);
+ for (i = 0; i < nsyms; ++i)
+ fprintf (out, "%s[%d]", i ? ", " : "", i);
+ fputs ("])\n\n", out);
+}
+
/*---------------------------------.
| Output the user actions to OUT. |
static void
user_actions_output (FILE *out)
{
- rule_number_t r;
+ rule_number r;
- fputs ("m4_define([b4_actions], \n[[", out);
+ fputs ("m4_define([b4_actions], \n[", out);
for (r = 0; r < nrules; ++r)
if (rules[r].action)
{
- fprintf (out, " case %d:\n", r + 1);
-
- if (!no_lines_flag)
- fprintf (out, muscle_find ("linef"),
- rules[r].action_location.first_line,
- quotearg_style (c_quoting_style,
- muscle_find ("filename")));
- fprintf (out, " %s\n break;\n\n",
- rules[r].action);
+ fprintf (out, "b4_%scase(%d, [b4_syncline(%d, ",
+ rules[r].is_predicate ? "predicate_" : "",
+ r + 1, rules[r].action_location.start.line);
+ escaped_output (out, rules[r].action_location.start.file);
+ fprintf (out, ")\n[ %s]])\n\n", rules[r].action);
}
- fputs ("]])\n\n", out);
+ fputs ("])\n\n", out);
}
-/*--------------------------------------.
-| Output the merge functions to OUT. |
-`--------------------------------------*/
+/*------------------------------------.
+| Output the merge functions to OUT. |
+`------------------------------------*/
static void
merger_output (FILE *out)
for (n = 1, p = merge_functions; p != NULL; n += 1, p = p->next)
{
if (p->type[0] == '\0')
- fprintf (out, " case %d: yyval = %s (*yy0, *yy1); break;\n",
- n, p->name);
+ fprintf (out, " case %d: *yy0 = %s (*yy0, *yy1); break;\n",
+ n, p->name);
else
- fprintf (out, " case %d: yyval.%s = %s (*yy0, *yy1); break;\n",
- n, p->type, p->name);
+ fprintf (out, " case %d: yy0->%s = %s (*yy0, *yy1); break;\n",
+ n, p->type, p->name);
}
fputs ("]])\n\n", out);
}
-/*--------------------------------------.
-| Output the tokens definition to OUT. |
-`--------------------------------------*/
+
+/*---------------------------------------------.
+| Prepare the muscles for symbol definitions. |
+`---------------------------------------------*/
static void
-token_definitions_output (FILE *out)
+prepare_symbol_definitions (void)
{
int i;
- int first = 1;
-
- fputs ("m4_define([b4_tokens], \n[", out);
- for (i = 0; i < ntokens; ++i)
+ for (i = 0; i < nsyms; ++i)
{
- symbol_t *symbol = symbols[i];
- int number = symbol->user_token_number;
-
- /* At this stage, if there are literal aliases, they are part of
- SYMBOLS, so we should not find symbols which are the aliases
- here. */
- assert (number != USER_NUMBER_ALIAS);
-
- /* Skip error token. */
- if (symbol == errtoken)
- continue;
-
- /* If this string has an alias, then it is necessarily the alias
- which is to be output. */
- if (symbol->alias)
- symbol = symbol->alias;
-
- /* Don't output literal chars or strings (when defined only as a
- string). Note that must be done after the alias resolution:
- think about `%token 'f' "f"'. */
- if (symbol->tag[0] == '\'' || symbol->tag[0] == '\"')
- continue;
-
- /* Don't #define nonliteral tokens whose names contain periods
- or '$' (as does the default value of the EOF token). */
- if (strchr (symbol->tag, '.') || strchr (symbol->tag, '$'))
- continue;
-
- fprintf (out, "%s[[[%s]], [%d]]",
- first ? "" : ",\n", symbol->tag, number);
-
- first = 0;
- }
- fputs ("])\n\n", out);
-}
+ symbol *sym = symbols[i];
+ const char *key;
+ const char *value;
+#define SET_KEY(Entry) \
+ obstack_fgrow2 (&format_obstack, "symbol(%d, %s)", \
+ i, Entry); \
+ obstack_1grow (&format_obstack, 0); \
+ key = obstack_finish (&format_obstack);
-/*---------------------------------------.
-| Output the symbol destructors to OUT. |
-`---------------------------------------*/
+#define SET_KEY2(Entry, Suffix) \
+ obstack_fgrow3 (&format_obstack, "symbol(%d, %s_%s)", \
+ i, Entry, Suffix); \
+ obstack_1grow (&format_obstack, 0); \
+ key = obstack_finish (&format_obstack);
-static void
-symbol_destructors_output (FILE *out)
-{
- int i;
- int first = 1;
+ // Whether the symbol has an identifier.
+ value = symbol_id_get (sym);
+ SET_KEY("has_id");
+ MUSCLE_INSERT_INT (key, !!value);
+
+ // Its identifier.
+ SET_KEY("id");
+ MUSCLE_INSERT_STRING (key, value ? value : "");
+
+ // Its tag. Typically for documentation purpose.
+ SET_KEY("tag");
+ MUSCLE_INSERT_STRING (key, sym->tag);
+
+ SET_KEY("user_number");
+ MUSCLE_INSERT_INT (key, sym->user_token_number);
+
+ SET_KEY("is_token");
+ MUSCLE_INSERT_INT (key,
+ i < ntokens && sym != errtoken && sym != undeftoken);
+
+ SET_KEY("number");
+ MUSCLE_INSERT_INT (key, sym->number);
+
+ SET_KEY("has_type");
+ MUSCLE_INSERT_INT (key, !!sym->type_name);
+
+ SET_KEY("type");
+ MUSCLE_INSERT_STRING (key, sym->type_name ? sym->type_name : "");
- fputs ("m4_define([b4_symbol_destructors], \n[", out);
- for (i = 0; i < nsyms; ++i)
- if (symbols[i]->destructor)
{
- symbol_t *symbol = symbols[i];
-
- /* Filename, lineno,
- Symbol-name, Symbol-number,
- destructor, typename. */
- fprintf (out, "%s[[[%s]], [[%d]], [[%s]], [[%d]], [[%s]], [[%s]]]",
- first ? "" : ",\n",
- infile, symbol->destructor_location.first_line,
- symbol->tag,
- symbol->number,
- symbol->destructor,
- symbol->type_name);
-
- first = 0;
+ int j;
+ for (j = 0; j < CODE_PROPS_SIZE; ++j)
+ {
+ /* "printer", not "%printer". */
+ char const *pname = code_props_type_string (j) + 1;
+ code_props const *p = symbol_code_props_get (sym, j);
+ SET_KEY2("has", pname);
+ MUSCLE_INSERT_INT (key, !!p->code);
+
+ if (p->code)
+ {
+ SET_KEY2(pname, "file");
+ MUSCLE_INSERT_STRING (key, p->location.start.file);
+
+ SET_KEY2(pname, "line");
+ MUSCLE_INSERT_INT (key, p->location.start.line);
+
+ SET_KEY(pname);
+ MUSCLE_INSERT_STRING_RAW (key, p->code);
+ }
+ }
}
- fputs ("])\n\n", out);
+#undef SET_KEY2
+#undef SET_KEY
+ }
}
-/*------------------------------------.
-| Output the symbol printers to OUT. |
-`------------------------------------*/
+/*--------------------------------------.
+| Output the tokens definition to OUT. |
+`--------------------------------------*/
static void
-symbol_printers_output (FILE *out)
+token_definitions_output (FILE *out)
{
int i;
- int first = 1;
+ char const *sep = "";
- fputs ("m4_define([b4_symbol_printers], \n[", out);
- for (i = 0; i < nsyms; ++i)
- if (symbols[i]->destructor)
- {
- symbol_t *symbol = symbols[i];
-
- /* Filename, lineno,
- Symbol-name, Symbol-number,
- printer, typename. */
- fprintf (out, "%s[[[%s]], [[%d]], [[%s]], [[%d]], [[%s]], [[%s]]]",
- first ? "" : ",\n",
- infile, symbol->printer_location.first_line,
- symbol->tag,
- symbol->number,
- symbol->printer,
- symbol->type_name);
-
- first = 0;
- }
+ fputs ("m4_define([b4_tokens], \n[", out);
+ for (i = 0; i < ntokens; ++i)
+ {
+ symbol *sym = symbols[i];
+ int number = sym->user_token_number;
+ uniqstr id = symbol_id_get (sym);
+
+ /* At this stage, if there are literal string aliases, they are
+ part of SYMBOLS, so we should not find their aliased symbols
+ here. */
+ aver (number != USER_NUMBER_HAS_STRING_ALIAS);
+
+ /* Skip error token and tokens without identifier. */
+ if (sym != errtoken && id)
+ {
+ fprintf (out, "%s[[[%s]], %d]",
+ sep, id, number);
+ sep = ",\n";
+ }
+ }
fputs ("])\n\n", out);
}
lookahead token type. */
muscle_insert_rule_number_table ("defact", yydefact,
- yydefact[0], 1, nstates);
+ yydefact[0], 1, nstates);
/* Figure out what to do after reducing with each rule, depending on
the saved state from before the beginning of parsing the data
that matched this rule. */
muscle_insert_state_number_table ("defgoto", yydefgoto,
- yydefgoto[0], 1, nsyms - ntokens);
+ yydefgoto[0], 1, nsyms - ntokens);
/* Output PACT. */
muscle_insert_base_table ("pact", base,
- base[0], 1, nstates);
+ base[0], 1, nstates);
MUSCLE_INSERT_INT ("pact_ninf", base_ninf);
/* Output PGOTO. */
muscle_insert_base_table ("pgoto", base,
- base[nstates], nstates + 1, nvectors);
+ base[nstates], nstates + 1, nvectors);
muscle_insert_base_table ("table", table,
- table[0], 1, high + 1);
+ table[0], 1, high + 1);
MUSCLE_INSERT_INT ("table_ninf", table_ninf);
muscle_insert_base_table ("check", check,
- check[0], 1, high + 1);
+ check[0], 1, high + 1);
/* GLR parsing slightly modifies YYTABLE and YYCHECK (and thus
YYPACT) so that in states with unresolved conflicts, the default
that case. Nevertheless, it seems even better to be able to use
the GLR skeletons even without the non-deterministic tables. */
muscle_insert_unsigned_int_table ("conflict_list_heads", conflict_table,
- conflict_table[0], 1, high+1);
+ conflict_table[0], 1, high + 1);
muscle_insert_unsigned_int_table ("conflicting_rules", conflict_list,
- conflict_list[0], 1, conflict_list_cnt);
+ 0, 1, conflict_list_cnt);
}
+
+/*--------------------------------------------.
+| Output the definitions of all the muscles. |
+`--------------------------------------------*/
+
+static void
+muscles_output (FILE *out)
+{
+ fputs ("m4_init()\n", out);
+ merger_output (out);
+ symbol_numbers_output (out);
+ token_definitions_output (out);
+ type_names_output (out);
+ user_actions_output (out);
+ // Must be last.
+ muscles_m4_output (out);
+}
\f
/*---------------------------.
| Call the skeleton parser. |
static void
output_skeleton (void)
{
- FILE *in;
- FILE *out;
int filter_fd[2];
- char const *argv[7];
pid_t pid;
- /* Compute the names of the package data dir and skeleton file.
- Test whether m4sugar.m4 is readable, to check for proper
+ /* Compute the names of the package data dir and skeleton files. */
+ char const *m4 = (m4 = getenv ("M4")) ? m4 : M4;
+ char const *datadir = pkgdatadir ();
+ char *m4sugar = xconcatenated_filename (datadir, "m4sugar/m4sugar.m4", NULL);
+ char *m4bison = xconcatenated_filename (datadir, "bison.m4", NULL);
+ char *skel = (IS_PATH_WITH_DIR (skeleton)
+ ? xstrdup (skeleton)
+ : xconcatenated_filename (datadir, skeleton, NULL));
+
+ /* Test whether m4sugar.m4 is readable, to check for proper
installation. A faulty installation can cause deadlock, so a
cheap sanity check is worthwhile. */
- char const m4sugar[] = "m4sugar/m4sugar.m4";
- char *full_path;
- char const *p;
- char const *m4 = (p = getenv ("M4")) ? p : M4;
- char const *pkgdatadir = (p = getenv ("BISON_PKGDATADIR")) ? p : PKGDATADIR;
- size_t skeleton_size = strlen (skeleton) + 1;
- size_t pkgdatadirlen = strlen (pkgdatadir);
- while (pkgdatadirlen && pkgdatadir[pkgdatadirlen - 1] == '/')
- pkgdatadirlen--;
- full_path = xmalloc (pkgdatadirlen + 1
- + (skeleton_size < sizeof m4sugar
- ? sizeof m4sugar : skeleton_size));
- strcpy (full_path, pkgdatadir);
- full_path[pkgdatadirlen] = '/';
- strcpy (full_path + pkgdatadirlen + 1, m4sugar);
- in = fopen (full_path, "r");
- if (! in || fclose (in) != 0)
- error (EXIT_FAILURE, errno, "%s", full_path);
- strcpy (full_path + pkgdatadirlen + 1, skeleton);
+ xfclose (xfopen (m4sugar, "r"));
/* Create an m4 subprocess connected to us via two pipes. */
if (trace_flag & trace_tools)
- fprintf (stderr, "running: %s -I %s %s - %s\n",
- m4, pkgdatadir, m4sugar, full_path);
-
- argv[0] = m4;
- argv[1] = "-I";
- argv[2] = pkgdatadir;
- argv[3] = m4sugar;
- argv[4] = "-";
- argv[5] = full_path;
- argv[6] = NULL;
-
- init_subpipe ();
- pid = create_subpipe (argv, filter_fd);
- free (full_path);
-
- out = fdopen (filter_fd[0], "w");
- if (! out)
- error (EXIT_FAILURE, errno, "fdopen");
-
- /* Output the definitions of all the muscles. */
-
- /* There are no comments, especially not `#': we do want M4 expansion
- after `#': think of CPP macros! */
- fputs ("m4_changecom()\n", out);
- fputs ("m4_init()\n", out);
+ fprintf (stderr, "running: %s %s - %s %s\n",
+ m4, m4sugar, m4bison, skel);
- user_actions_output (out);
- merger_output (out);
- token_definitions_output (out);
- symbol_destructors_output (out);
- symbol_printers_output (out);
+ /* Some future version of GNU M4 (most likely 1.6) may treat the -dV in a
+ position-dependent manner. Keep it as the first argument so that all
+ files are traced.
- muscles_m4_output (out);
+ See the thread starting at
+ <http://lists.gnu.org/archive/html/bug-bison/2008-07/msg00000.html>
+ for details. */
+ {
+ char const *argv[10];
+ int i = 0;
+ argv[i++] = m4;
+
+ /* When POSIXLY_CORRECT is set, GNU M4 1.6 and later disable GNU
+ extensions, which Bison's skeletons depend on. With older M4,
+ it has no effect. M4 1.4.12 added a -g/--gnu command-line
+ option to make it explicit that a program wants GNU M4
+ extensions even when POSIXLY_CORRECT is set.
+
+ See the thread starting at
+ <http://lists.gnu.org/archive/html/bug-bison/2008-07/msg00000.html>
+ for details. */
+ if (*M4_GNU_OPTION)
+ argv[i++] = M4_GNU_OPTION;
+
+ argv[i++] = "-I";
+ argv[i++] = datadir;
+ if (trace_flag & trace_m4)
+ argv[i++] = "-dV";
+ argv[i++] = m4sugar;
+ argv[i++] = "-";
+ argv[i++] = m4bison;
+ argv[i++] = skel;
+ argv[i++] = NULL;
+ aver (i <= ARRAY_CARDINALITY (argv));
+
+ /* The ugly cast is because gnulib gets the const-ness wrong. */
+ pid = create_pipe_bidi ("m4", m4, (char **)(void*)argv, false, true,
+ true, filter_fd);
+ }
- fputs ("m4_wrap([m4_divert_pop(0)])\n", out);
- fputs ("m4_divert_push(0)dnl\n", out);
- if (ferror (out))
- error (EXIT_FAILURE, 0, "pipe output error");
- xfclose (out);
+ free (m4sugar);
+ free (m4bison);
+ free (skel);
+
+ if (trace_flag & trace_muscles)
+ muscles_output (stderr);
+ {
+ FILE *out = xfdopen (filter_fd[1], "w");
+ muscles_output (out);
+ xfclose (out);
+ }
/* Read and process m4's output. */
timevar_push (TV_M4);
- in = fdopen (filter_fd[1], "r");
- if (! in)
- error (EXIT_FAILURE, errno, "fdopen");
- scan_skel (in);
- if (ferror (in))
- error (EXIT_FAILURE, 0, "pipe input error");
- xfclose (in);
- reap_subpipe (pid, m4);
+ {
+ FILE *in = xfdopen (filter_fd[0], "r");
+ scan_skel (in);
+ /* scan_skel should have read all of M4's output. Otherwise, when we
+ close the pipe, we risk letting M4 report a broken-pipe to the
+ Bison user. */
+ aver (feof (in));
+ xfclose (in);
+ }
+ wait_subprocess (pid, "m4", false, false, true, true, NULL);
timevar_pop (TV_M4);
}
static void
prepare (void)
{
- /* Flags. */
- MUSCLE_INSERT_INT ("locations_flag", locations_flag);
- MUSCLE_INSERT_INT ("defines_flag", defines_flag);
- MUSCLE_INSERT_INT ("error_verbose", error_verbose);
- MUSCLE_INSERT_INT ("pure", pure_parser);
- MUSCLE_INSERT_INT ("debug", debug_flag);
-
- /* FIXME: This is wrong: the muscles should decide whether they hold
- a copy or not, but the situation is too obscure currently. */
- MUSCLE_INSERT_STRING ("prefix", spec_name_prefix ? spec_name_prefix : "yy");
- MUSCLE_INSERT_STRING ("output_infix", output_infix ? output_infix : "");
- MUSCLE_INSERT_STRING ("output_prefix", short_base_name);
- MUSCLE_INSERT_STRING ("output_parser_name", parser_file_name);
- MUSCLE_INSERT_STRING ("output_header_name", spec_defines_file);
-
- /* Symbols. */
- MUSCLE_INSERT_INT ("tokens_number", ntokens);
- MUSCLE_INSERT_INT ("nterms_number", nvars);
- MUSCLE_INSERT_INT ("undef_token_number", undeftoken->number);
- MUSCLE_INSERT_INT ("user_token_number_max", max_user_token_number);
-
- /* Rules. */
- MUSCLE_INSERT_INT ("rules_number", nrules);
-
- /* States. */
- MUSCLE_INSERT_INT ("last", high);
- MUSCLE_INSERT_INT ("final_state_number", final_state->number);
- MUSCLE_INSERT_INT ("states_number", nstates);
-
- /* User Code. */
- obstack_1grow (&pre_prologue_obstack, 0);
- obstack_1grow (&post_prologue_obstack, 0);
- muscle_insert ("pre_prologue", obstack_finish (&pre_prologue_obstack));
- muscle_insert ("post_prologue", obstack_finish (&post_prologue_obstack));
+ /* BISON_USE_PUSH_FOR_PULL is for the test suite and should not be
+ documented for the user. */
+ char const *cp = getenv ("BISON_USE_PUSH_FOR_PULL");
+ bool use_push_for_pull_flag = cp && *cp && strtol (cp, 0, 10);
- /* Find the right skeleton file. */
- if (!skeleton)
- {
- if (glr_parser)
- skeleton = "glr.c";
- else
- skeleton = "yacc.c";
- }
-
- /* Parse the skeleton file and output the needed parsers. */
- MUSCLE_INSERT_STRING ("skeleton", skeleton);
+ /* Flags. */
+ MUSCLE_INSERT_BOOL ("defines_flag", defines_flag);
+ MUSCLE_INSERT_BOOL ("glr_flag", glr_parser);
+ MUSCLE_INSERT_BOOL ("nondeterministic_flag", nondeterministic_parser);
+ MUSCLE_INSERT_BOOL ("synclines_flag", !no_lines_flag);
+ MUSCLE_INSERT_BOOL ("tag_seen_flag", tag_seen);
+ MUSCLE_INSERT_BOOL ("token_table_flag", token_table_flag);
+ MUSCLE_INSERT_BOOL ("use_push_for_pull_flag", use_push_for_pull_flag);
+ MUSCLE_INSERT_BOOL ("yacc_flag", yacc_flag);
+
+ /* File names. */
+ if (spec_name_prefix)
+ MUSCLE_INSERT_STRING ("prefix", spec_name_prefix);
+
+ MUSCLE_INSERT_STRING ("file_name_all_but_ext", all_but_ext);
+
+#define DEFINE(Name) MUSCLE_INSERT_STRING (#Name, Name ? Name : "")
+ DEFINE (dir_prefix);
+ DEFINE (parser_file_name);
+ DEFINE (spec_defines_file);
+ DEFINE (spec_file_prefix);
+ DEFINE (spec_graph_file);
+ DEFINE (spec_name_prefix);
+ DEFINE (spec_outfile);
+ DEFINE (spec_verbose_file);
+#undef DEFINE
+
+ /* Find the right skeleton file, and add muscles about the skeletons. */
+ if (skeleton)
+ MUSCLE_INSERT_C_STRING ("skeleton", skeleton);
+ else
+ skeleton = language->skeleton;
+
+ /* About the skeletons. */
+ {
+ /* b4_pkgdatadir is used inside m4_include in the skeletons, so digraphs
+ would never be expanded. Hopefully no one has M4-special characters in
+ his Bison installation path. */
+ MUSCLE_INSERT_STRING_RAW ("pkgdatadir", pkgdatadir ());
+ }
}
{
obstack_init (&format_obstack);
- prepare_tokens ();
+ prepare_symbols ();
prepare_rules ();
prepare_states ();
prepare_actions ();
+ prepare_symbol_definitions ();
prepare ();
output_skeleton ();
obstack_free (&format_obstack, NULL);
- obstack_free (&pre_prologue_obstack, NULL);
- obstack_free (&post_prologue_obstack, NULL);
+}
+
+char const *
+pkgdatadir (void)
+{
+ char const *cp = getenv ("BISON_PKGDATADIR");
+ return cp ? cp : PKGDATADIR;
}