X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/43591cec981a3e7dd87bd1aef63ea0b9ec73c01c..eb1b07409f0ccad9970c8a60ded9f2a56d205b9c:/src/output.c diff --git a/src/output.c b/src/output.c index b23ffe8e..ef7683ec 100644 --- a/src/output.c +++ b/src/output.c @@ -1,5 +1,7 @@ -/* Output the generated parsing program for bison, - Copyright (C) 1984, 1986, 1989, 1992, 2000 Free Software Foundation, Inc. +/* Output the generated parsing program for Bison. + + Copyright (C) 1984, 1986, 1989, 1992, 2000, 2001, 2002, 2003, 2004, + 2005, 2006, 2007 Free Software Foundation, Inc. This file is part of Bison, the GNU Compiler Compiler. @@ -15,1243 +17,602 @@ 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. - Starred ones needed only for the semantic parser. - Double starred are output only if switches are set. - - 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. - - * 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. + Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA + 02110-1301, USA. */ - 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 ?? - YYNTBASE = ntokens. -*/ +#include <configmake.h> +#include <error.h> +#include <get-errno.h> +#include <quotearg.h> +#include <subpipe.h> +#include <timevar.h> -#include "system.h" -#include "getargs.h" -#include "xalloc.h" +#include "complain.h" #include "files.h" +#include "getargs.h" #include "gram.h" -#include "LR0.h" -#include "complain.h" +#include "muscle_tab.h" #include "output.h" -#include "lalr.h" #include "reader.h" -#include "conflicts.h" - -extern void berror PARAMS((const char *)); - - - -static int nvectors; -static int nentries; -static short **froms; -static short **tos; -static short *tally; -static short *width; -static short *actrow; -static short *state_count; -static short *order; -static short *base; -static short *pos; -static short *table; -static short *check; -static int lowzero; -static int high; - - - -static inline void -output_short_or_char_table (FILE *out, - const char *comment, - const char *type, - const char *table_name, - short *short_table, - short first_value, - short begin, short end) -{ - int i, j; +#include "scan-code.h" /* max_left_semantic_context */ +#include "scan-skel.h" +#include "symtab.h" +#include "tables.h" - if (comment) - fprintf (out, "/* %s. */\n", comment); - fprintf (out, "static const %s %s[] =\n{\n %6d", - type, table_name, first_value); +static struct obstack format_obstack; - j = 1; - for (i = begin; i < end; i++) - { - putc (',', out); - - if (j >= 10) - { - fputs ("\n ", out); - j = 1; - } - else - { - j++; - } - - fprintf (out, "%6d", short_table[i]); - } - fprintf (out, "\n};\n"); -} +/*-------------------------------------------------------------------. +| Create a function NAME which associates to the muscle NAME the | +| result of formatting the FIRST and then TABLE_DATA[BEGIN..END[ (of | +| TYPE), and to the muscle NAME_max, the max value of the | +| TABLE_DATA. | +`-------------------------------------------------------------------*/ -static inline void -output_short_table (FILE *out, - const char *comment, - const char *table_name, - short *short_table, - short first_value, - short begin, short end) -{ - output_short_or_char_table (out, comment, "short", table_name, short_table, - first_value, begin, end); +#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_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_item_number_table, item_number) +GENERATE_MUSCLE_INSERT_TABLE(muscle_insert_state_number_table, state_number) -/*--------------------------------------------------------------. -| output_headers -- Output constant strings to the beginning of | -| certain files. | -`--------------------------------------------------------------*/ - -#define GUARDSTR \ -"\n\ -#include \"%s\"\n\ -extern int yyerror;\n\ -extern int yycost;\n\ -extern char * yymsg;\n\ -extern YYSTYPE yyval;\n\ -\n\ -yyguard(n, yyvsp, yylsp)\n\ -register int n;\n\ -register YYSTYPE *yyvsp;\n\ -register YYLTYPE *yylsp;\n\ -{\n\ - yyerror = 0;\n\ - yycost = 0;\n\ - yymsg = 0;\n\ - switch (n)\n\ - {" - -#define ACTSTR \ -"\n\ -#include \"%s\"\n\ -extern YYSTYPE yyval;\n\ -extern int yychar;\n\ -\n\ -yyaction(n, yyvsp, yylsp)\n\ -register int n;\n\ -register YYSTYPE *yyvsp;\n\ -register YYLTYPE *yylsp;\n\ -{\n\ - switch (n)\n\ - {" - -#define ACTSTR_SIMPLE "\n switch (yyn) {\n" - -void -output_headers (void) -{ - if (semantic_parser) - fprintf (fguard, GUARDSTR, attrsfile); - - if (no_parser_flag) - return; - - fprintf (faction, (semantic_parser ? ACTSTR : ACTSTR_SIMPLE), attrsfile); -/* if (semantic_parser) JF moved this below - fprintf(ftable, "#include \"%s\"\n", attrsfile); - fprintf(ftable, "#include <stdio.h>\n\n"); -*/ - - /* Rename certain symbols if -p was specified. */ - if (spec_name_prefix) - { - fprintf (ftable, "#define yyparse %sparse\n", spec_name_prefix); - fprintf (ftable, "#define yylex %slex\n", spec_name_prefix); - fprintf (ftable, "#define yyerror %serror\n", spec_name_prefix); - fprintf (ftable, "#define yylval %slval\n", spec_name_prefix); - fprintf (ftable, "#define yychar %schar\n", spec_name_prefix); - fprintf (ftable, "#define yydebug %sdebug\n", spec_name_prefix); - fprintf (ftable, "#define yynerrs %snerrs\n", spec_name_prefix); - } -} +/*--------------------------------------------------------------------. +| Print to OUT a representation of STRING escaped both for C and M4. | +`--------------------------------------------------------------------*/ -/*-------------------------------------------------------. -| Output constant strings to the ends of certain files. | -`-------------------------------------------------------*/ - -void -output_trailers (void) +static void +escaped_output (FILE *out, char const *string) { - if (semantic_parser) - fprintf (fguard, "\n }\n}\n"); + char const *p; + fprintf (out, "[["); - fprintf (faction, "\n"); - - if (no_parser_flag) - return; + 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; + } - if (semantic_parser) - fprintf (faction, " }\n"); - fprintf (faction, "}\n"); + fprintf (out, "]]"); } +/*------------------------------------------------------------------. +| Prepare the muscles related to the symbols: translate, tname, and | +| toknum. | +`------------------------------------------------------------------*/ static void -output_token_translations (void) -{ - fputs ("\ -\n\ -/* YYRANSLATE(YYLEX) -- Bison token number corresponding to YYLEX. */\n", - ftable); - if (translations) - { - fprintf (ftable, - "#define YYTRANSLATE(x) ((unsigned)(x) <= %d ? yytranslate[x] : %d)\ -\n\ -\n", - max_user_token_number, nsyms); - - output_short_or_char_table (ftable, - "YYRANSLATE[YYLEX] -- Bison token number corresponding to YYLEX", - ntokens < 127 ? "char" : "short", - "yytranslate", token_translations, - 0, 1, max_user_token_number + 1); - } - else - { - fprintf (ftable, "\n#define YYTRANSLATE(x) (x)\n"); - } -} - - -static void -output_gram (void) +prepare_symbols (void) { - /* With the ordinary parser, - yyprhs and yyrhs are needed only for yydebug. */ - /* With the no_parser option, all tables are generated */ - if (!semantic_parser && !no_parser_flag) - fprintf (ftable, "\n#if YYDEBUG != 0\n"); - - output_short_table (ftable, NULL, "yyprhs", rrhs, - 0, 1, nrules + 1); - + MUSCLE_INSERT_BOOL ("token_table", token_table_flag); + 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); + + muscle_insert_symbol_number_table ("translate", + token_translations, + token_translations[0], + 1, max_user_token_number + 1); + + /* tname -- token names. */ { - size_t yyrhs_size = 1; - short *yyrhs, *sp; int i; + /* We assume that the table will be output starting at column 2. */ + int j = 2; + for (i = 0; i < nsyms; i++) + { + char const *cp = quotearg_style (c_quoting_style, symbols[i]->tag); + /* Width of the next token, including the two quotes, the + comma and the space. */ + int width = strlen (cp) + 2; - for (sp = ritem + 1; *sp; sp++) - ++yyrhs_size; - yyrhs = XMALLOC (short, yyrhs_size); + if (j + width > 75) + { + obstack_sgrow (&format_obstack, "\n "); + j = 1; + } - for (sp = ritem + 1, i = 1; *sp; ++sp, ++i) - yyrhs[i] = *sp > 0 ? *sp : 0; + if (i) + obstack_1grow (&format_obstack, ' '); + MUSCLE_OBSTACK_SGROW (&format_obstack, cp); + obstack_1grow (&format_obstack, ','); + j += width; + } + obstack_sgrow (&format_obstack, " ]b4_null["); - output_short_table (ftable, NULL, "yyrhs", yyrhs, - ritem[0], 1, yyrhs_size); - XFREE (yyrhs); + /* Finish table and store. */ + obstack_1grow (&format_obstack, 0); + muscle_insert ("tname", obstack_finish (&format_obstack)); } - if (!semantic_parser && !no_parser_flag) - fprintf (ftable, "\n#endif\n"); -} - - -static void -output_stos (void) -{ - output_short_table (ftable, NULL, "yystos", accessing_symbol, - 0, 1, nstates); -} - - -static void -output_rule_data (void) -{ - int i; - int j; - short *short_tab = NULL; - - fputs ("\n\ -#if YYDEBUG != 0\n", - ftable); - - output_short_table (ftable, - "YYRLINE[YYN] -- source line where rule number YYN was defined", - "yyrline", rline, - 0, 1, nrules + 1); - - fputs ("#endif\n\n", ftable); - - if (token_table_flag || no_parser_flag) - { - fprintf (ftable, "#define YYNTOKENS %d\n", ntokens); - fprintf (ftable, "#define YYNNTS %d\n", nvars); - fprintf (ftable, "#define YYNRULES %d\n", nrules); - fprintf (ftable, "#define YYNSTATES %d\n", nstates); - fprintf (ftable, "#define YYMAXUTOK %d\n\n", max_user_token_number); - } - - /* Output the table of symbol names. */ - if (!token_table_flag && !no_parser_flag) - fputs ("\n#if YYDEBUG != 0 || defined YYERROR_VERBOSE\n\n", ftable); - fputs ("\ -/* YYTNAME[TOKEN_NUM] -- String name of the token TOKEN_NUM. */\n", - ftable); - fprintf (ftable, - "static const char *const yytname[] =\n{\n "); - - j = 0; - for (i = 0; i < nsyms; i++) - /* this used to be i<=nsyms, but that output a final "" symbol - almost by accident */ - { - /* Width of the next token, including the two quotes, the coma - and the space. */ - int strsize = 4; - char *p; - - for (p = tags[i]; p && *p; p++) - if (*p == '"' || *p == '\\' || *p == '\n' || *p == '\t' - || *p == '\b') - strsize += 2; - else if (*p < 040 || *p >= 0177) - strsize += 4; - else - strsize++; - - if (j + strsize > 75) - { - fputs ("\n ", ftable); - j = 2; - } - - putc ('\"', ftable); - for (p = tags[i]; p && *p; p++) - { - if (*p == '"' || *p == '\\') - { - fprintf (ftable, "\\%c", *p); - } - else if (*p == '\n') - { - fprintf (ftable, "\\n"); - } - else if (*p == '\t') - { - fprintf (ftable, "\\t"); - } - else if (*p == '\b') - { - fprintf (ftable, "\\b"); - } - else if (*p < 040 || *p >= 0177) - { - fprintf (ftable, "\\%03o", *p); - } - else - { - putc (*p, ftable); - } - } - - fputs ("\", ", ftable); - j += strsize; - } - /* add a NULL entry to list of tokens */ - fprintf (ftable, "NULL\n};\n"); - - if (!token_table_flag && !no_parser_flag) - fprintf (ftable, "#endif\n\n"); - /* Output YYTOKNUM. */ - if (token_table_flag) - { - output_short_table (ftable, - "YYTOKNUM[YYLEX] -- Index in YYTNAME corresponding to YYLEX", - "yytoknum", user_toknums, - 0, 1, ntokens + 1); - } - - /* Output YYR1. */ - output_short_table (ftable, - "YYR1[YYN] -- Symbol number of symbol that rule YYN derives", - "yyr1", rlhs, - 0, 1, nrules + 1); - XFREE (rlhs + 1); - - putc ('\n', ftable); - - /* Output YYR2. */ - short_tab = XMALLOC (short, nrules + 1); - for (i = 1; i < nrules; i++) - short_tab[i] = rrhs[i + 1] - rrhs[i] - 1; - short_tab[nrules] = nitems - rrhs[nrules] - 1; - output_short_table (ftable, - "YYR2[YYN] -- Number of symbols composing right hand side of rule YYN", - "yyr2", short_tab, - 0, 1, nrules + 1); - putc ('\n', ftable); - - XFREE (short_tab); - - XFREE (rrhs + 1); -} - - -static void -output_defines (void) -{ - fprintf (ftable, "\n\n#define\tYYFINAL\t\t%d\n", final_state); - fprintf (ftable, "#define\tYYFLAG\t\t%d\n", MINSHORT); - fprintf (ftable, "#define\tYYNTBASE\t%d\n", ntokens); + { + int i; + 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); + free (values); + } } -/*------------------------------------------------------------------. -| Decide what to do for each type of token if seen as the lookahead | -| token in specified state. The value returned is used as the | -| default action (yydefact) for the state. In addition, actrow is | -| filled with what to do for each kind of token, index by symbol | -| number, with zero meaning do the default action. The value | -| MINSHORT, a very negative number, means this situation is an | -| error. The parser recognizes this value specially. | -| | -| This is where conflicts are resolved. The loop over lookahead | -| rules considered lower-numbered rules last, and the last rule | -| considered that likes a token gets to handle it. | -`------------------------------------------------------------------*/ - -static int -action_row (int state) -{ - int i; - int j; - int k; - int m = 0; - int n = 0; - int count; - int default_rule; - int nreds; - int max; - int rule; - int shift_state; - int symbol; - unsigned mask; - unsigned *wordp; - reductions *redp; - shifts *shiftp; - errs *errp; - int nodefault = 0; /* set nonzero to inhibit having any default reduction */ - - for (i = 0; i < ntokens; i++) - actrow[i] = 0; - - default_rule = 0; - nreds = 0; - redp = reduction_table[state]; - - if (redp) - { - nreds = redp->nreds; - - if (nreds >= 1) - { - /* loop over all the rules available here which require - lookahead */ - m = lookaheads[state]; - n = lookaheads[state + 1]; - - for (i = n - 1; i >= m; i--) - { - rule = -LAruleno[i]; - wordp = LA + i * tokensetsize; - mask = 1; - - /* and find each token which the rule finds acceptable - to come next */ - for (j = 0; j < ntokens; j++) - { - /* and record this rule as the rule to use if that - token follows. */ - if (mask & *wordp) - actrow[j] = rule; - - mask <<= 1; - if (mask == 0) - { - mask = 1; - wordp++; - } - } - } - } - } - - shiftp = shift_table[state]; - - /* Now see which tokens are allowed for shifts in this state. For - them, record the shift as the thing to do. So shift is preferred - to reduce. */ - - if (shiftp) - { - k = shiftp->nshifts; - - for (i = 0; i < k; i++) - { - shift_state = shiftp->shifts[i]; - if (!shift_state) - continue; - - symbol = accessing_symbol[shift_state]; - - if (ISVAR (symbol)) - break; - - actrow[symbol] = shift_state; - - /* Do not use any default reduction if there is a shift for - error */ - if (symbol == error_token_number) - nodefault = 1; - } - } - - errp = err_table[state]; - - /* See which tokens are an explicit error in this state (due to - %nonassoc). For them, record MINSHORT as the action. */ - - if (errp) - { - k = errp->nerrs; - - for (i = 0; i < k; i++) - { - symbol = errp->errs[i]; - actrow[symbol] = MINSHORT; - } - } - - /* Now find the most common reduction and make it the default action - for this state. */ - - if (nreds >= 1 && !nodefault) - { - if (consistent[state]) - default_rule = redp->rules[0]; - else - { - max = 0; - for (i = m; i < n; i++) - { - count = 0; - rule = -LAruleno[i]; - - for (j = 0; j < ntokens; j++) - { - if (actrow[j] == rule) - count++; - } - - if (count > max) - { - max = count; - default_rule = rule; - } - } - - /* actions which match the default are replaced with zero, - which means "use the default" */ - - if (max > 0) - { - for (j = 0; j < ntokens; j++) - { - if (actrow[j] == default_rule) - actrow[j] = 0; - } - - default_rule = -default_rule; - } - } - } - - /* If have no default rule, the default is an error. - So replace any action which says "error" with "use default". */ - - if (default_rule == 0) - for (j = 0; j < ntokens; j++) - { - if (actrow[j] == MINSHORT) - actrow[j] = 0; - } - - return default_rule; -} - +/*-------------------------------------------------------------. +| Prepare the muscles related to the rules: rhs, prhs, r1, r2, | +| rline, dprec, merger. | +`-------------------------------------------------------------*/ static void -save_row (int state) +prepare_rules (void) { - int i; - int count; - short *sp; - short *sp1; - short *sp2; - - count = 0; - for (i = 0; i < ntokens; i++) + rule_number r; + unsigned int i = 0; + item_number *rhs = xnmalloc (nritems, sizeof *rhs); + unsigned int *prhs = xnmalloc (nrules, sizeof *prhs); + 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); + + for (r = 0; r < nrules; ++r) { - if (actrow[i] != 0) - count++; + item_number *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; + /* Line where rule was defined. */ + rline[r] = rules[r].location.start.line; + /* Dynamic precedence (GLR). */ + dprec[r] = rules[r].dprec; + /* Merger-function index (GLR). */ + merger[r] = rules[r].merger; } - - if (count == 0) - return; - - froms[state] = sp1 = sp = XCALLOC (short, count); - tos[state] = sp2 = XCALLOC (short, count); - - for (i = 0; i < ntokens; i++) - { - if (actrow[i] != 0) - { - *sp1++ = i; - *sp2++ = actrow[i]; - } - } - - tally[state] = count; - width[state] = sp1[-1] - sp[0] + 1; + aver (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_int_table ("dprec", dprec, 0, 0, nrules); + muscle_insert_int_table ("merger", merger, 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); } - -/*------------------------------------------------------------------. -| Figure out the actions for the specified state, indexed by | -| lookahead token type. | -| | -| The YYDEFACT table is output now. The detailed info is saved for | -| putting into YYTABLE later. | -`------------------------------------------------------------------*/ +/*--------------------------------------------. +| Prepare the muscles related to the states. | +`--------------------------------------------*/ static void -token_actions (void) +prepare_states (void) { - int i; - short *yydefact = XCALLOC (short, nstates); - - actrow = XCALLOC (short, ntokens); + state_number i; + symbol_number *values = xnmalloc (nstates, sizeof *values); for (i = 0; i < nstates; ++i) - { - yydefact[i] = action_row (i); - save_row (i); - } - XFREE (actrow); - - output_short_table (ftable, - "YYDEFACT[S] -- default rule to reduce with in state S when YYTABLE\n\ - doesn't specify something else to do. Zero means the default is an\n\ - error", - "yydefact", yydefact, - yydefact[0], 1, nstates); - putc ('\n', ftable); - XFREE (yydefact); + values[i] = states[i]->accessing_symbol; + muscle_insert_symbol_number_table ("stos", values, + 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); } -static void -free_shifts (void) -{ - shifts *sp, *sptmp; /* JF derefrenced freed ptr */ - - XFREE (shift_table); - - for (sp = first_shift; sp; sp = sptmp) - { - sptmp = sp->next; - XFREE (sp); - } -} +/*---------------------------------. +| Output the user actions to OUT. | +`---------------------------------*/ static void -free_reductions (void) +user_actions_output (FILE *out) { - reductions *rp, *rptmp; /* JF fixed freed ptr */ - - XFREE (reduction_table); + rule_number r; - for (rp = first_reduction; rp; rp = rptmp) - { - rptmp = rp->next; - XFREE (rp); - } + fputs ("m4_define([b4_actions], \n[", out); + for (r = 0; r < nrules; ++r) + if (rules[r].action) + { + fprintf (out, "b4_case(%d, [b4_syncline(%d, ", 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); } - +/*--------------------------------------. +| Output the merge functions to OUT. | +`--------------------------------------*/ static void -save_column (int symbol, int default_state) -{ - int i; - short *sp; - short *sp1; - short *sp2; - int count; - int symno; - - short begin = goto_map[symbol]; - short end = goto_map[symbol + 1]; - - count = 0; - for (i = begin; i < end; i++) - { - if (to_state[i] != default_state) - count++; - } - - if (count == 0) - return; - - symno = symbol - ntokens + nstates; - - froms[symno] = sp1 = sp = XCALLOC (short, count); - tos[symno] = sp2 = XCALLOC (short, count); - - for (i = begin; i < end; i++) - { - if (to_state[i] != default_state) - { - *sp1++ = from_state[i]; - *sp2++ = to_state[i]; - } - } - - tally[symno] = count; - width[symno] = sp1[-1] - sp[0] + 1; -} - -static int -default_goto (int symbol) +merger_output (FILE *out) { - int i; - int m; int n; - int default_state; - int max; - - m = goto_map[symbol]; - n = goto_map[symbol + 1]; + merger_list* p; - if (m == n) - return -1; - - for (i = 0; i < nstates; i++) - state_count[i] = 0; - - for (i = m; i < n; i++) - state_count[to_state[i]]++; - - max = 0; - default_state = -1; - - for (i = 0; i < nstates; i++) + fputs ("m4_define([b4_mergers], \n[[", out); + for (n = 1, p = merge_functions; p != NULL; n += 1, p = p->next) { - if (state_count[i] > max) - { - max = state_count[i]; - default_state = i; - } + if (p->type[0] == '\0') + fprintf (out, " case %d: *yy0 = %s (*yy0, *yy1); break;\n", + n, p->name); + else + fprintf (out, " case %d: yy0->%s = %s (*yy0, *yy1); break;\n", + n, p->type, p->name); } - - return default_state; + fputs ("]])\n\n", out); } - -/*-------------------------------------------------------------------. -| 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. | -| | -| The YYDEFGOTO table is output now. The detailed info is saved for | -| putting into YYTABLE later. | -`-------------------------------------------------------------------*/ +/*--------------------------------------. +| Output the tokens definition to OUT. | +`--------------------------------------*/ static void -goto_actions (void) +token_definitions_output (FILE *out) { int i; + char const *sep = ""; - short *yydefgoto = XMALLOC (short, nsyms - ntokens); - state_count = XCALLOC (short, nstates); - - for (i = ntokens; i < nsyms; ++i) + fputs ("m4_define([b4_tokens], \n[", out); + for (i = 0; i < ntokens; ++i) { - int default_state = default_goto (i); - save_column (i, default_state); - yydefgoto[i - ntokens] = default_state; + symbol *sym = symbols[i]; + int number = sym->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. */ + aver (number != USER_NUMBER_ALIAS); + + /* Skip error token. */ + if (sym == errtoken) + continue; + + /* If this string has an alias, then it is necessarily the alias + which is to be output. */ + if (sym->alias) + sym = sym->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 (sym->tag[0] == '\'' || sym->tag[0] == '\"') + continue; + + /* Don't #define nonliteral tokens whose names contain periods + or '$' (as does the default value of the EOF token). */ + if (strchr (sym->tag, '.') || strchr (sym->tag, '$')) + continue; + + fprintf (out, "%s[[[%s]], %d]", + sep, sym->tag, number); + sep = ",\n"; } - - output_short_table (ftable, NULL, "yydefgoto", yydefgoto, - yydefgoto[0], 1, nsyms - ntokens); - - XFREE (state_count); - XFREE (yydefgoto); + fputs ("])\n\n", out); } -/* The next few functions decide how to pack the actions and gotos - information into yytable. */ +/*---------------------------------------------------. +| Output the symbol destructors or printers to OUT. | +`---------------------------------------------------*/ static void -sort_actions (void) +symbol_code_props_output (FILE *out, char const *what, + code_props const *(*get)(symbol const *)) { int i; - int j; - int k; - int t; - int w; - - order = XCALLOC (short, nvectors); - nentries = 0; + char const *sep = ""; - for (i = 0; i < nvectors; i++) + fputs ("m4_define([b4_symbol_", out); + fputs (what, out); + fputs ("], \n[", out); + for (i = 0; i < nsyms; ++i) { - if (tally[i] > 0) - { - t = tally[i]; - w = width[i]; - j = nentries - 1; - - while (j >= 0 && (width[order[j]] < w)) - j--; - - while (j >= 0 && (width[order[j]] == w) && (tally[order[j]] < t)) - j--; - - for (k = nentries - 1; k > j; k--) - order[k + 1] = order[k]; - - order[j + 1] = i; - nentries++; - } + symbol *sym = symbols[i]; + char const *code = (*get) (sym)->code; + if (code) + { + location loc = (*get) (sym)->location; + /* Filename, lineno, + Symbol-name, Symbol-number, + code, optional typename. */ + fprintf (out, "%s[", sep); + sep = ",\n"; + escaped_output (out, loc.start.file); + fprintf (out, ", %d, ", loc.start.line); + escaped_output (out, sym->tag); + fprintf (out, ", %d, [[%s]]", sym->number, code); + if (sym->type_name) + fprintf (out, ", [[%s]]", sym->type_name); + fputc (']', out); + } } -} - - -static int -matching_state (int vector) -{ - int i; - int j; - int k; - int t; - int w; - int match; - int prev; - - i = order[vector]; - if (i >= nstates) - return -1; - - t = tally[i]; - w = width[i]; - - for (prev = vector - 1; prev >= 0; prev--) - { - j = order[prev]; - if (width[j] != w || tally[j] != t) - return -1; - - match = 1; - for (k = 0; match && k < t; k++) - { - if (tos[j][k] != tos[i][k] || froms[j][k] != froms[i][k]) - match = 0; - } - - if (match) - return j; - } - - return -1; -} - - -static int -pack_vector (int vector) -{ - int i; - int j; - int k; - int t; - int loc = 0; - int ok; - short *from; - short *to; - - i = order[vector]; - t = tally[i]; - - assert (t); - - from = froms[i]; - to = tos[i]; - - for (j = lowzero - from[0]; j < MAXTABLE; j++) - { - ok = 1; - - for (k = 0; ok && k < t; k++) - { - loc = j + from[k]; - if (loc > MAXTABLE) - fatal (_("maximum table size (%d) exceeded"), MAXTABLE); - - if (table[loc] != 0) - ok = 0; - } - - for (k = 0; ok && k < vector; k++) - { - if (pos[k] == j) - ok = 0; - } - - if (ok) - { - for (k = 0; k < t; k++) - { - loc = j + from[k]; - table[loc] = to[k]; - check[loc] = from[k]; - } - - while (table[lowzero] != 0) - lowzero++; - - if (loc > high) - high = loc; - - return j; - } - } - - berror ("pack_vector"); - return 0; /* JF keep lint happy */ + fputs ("])\n\n", out); } static void -pack_table (void) +prepare_actions (void) { - int i; - int place; - int state; - - base = XCALLOC (short, nvectors); - pos = XCALLOC (short, nentries); - table = XCALLOC (short, MAXTABLE); - check = XCALLOC (short, MAXTABLE); - - lowzero = 0; - high = 0; - - for (i = 0; i < nvectors; i++) - base[i] = MINSHORT; - - for (i = 0; i < MAXTABLE; i++) - check[i] = -1; - - for (i = 0; i < nentries; i++) - { - state = matching_state (i); - - if (state < 0) - place = pack_vector (i); - else - place = base[state]; - - pos[i] = place; - base[order[i]] = place; - } - - for (i = 0; i < nvectors; i++) - { - if (froms[i]) - XFREE (froms[i]); - if (tos[i]) - XFREE (tos[i]); - } - - XFREE (froms); - XFREE (tos); - XFREE (pos); + /* Figure out the actions for the specified state, indexed by + lookahead token type. */ + + muscle_insert_rule_number_table ("defact", yydefact, + 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); + + + /* Output PACT. */ + muscle_insert_base_table ("pact", base, + base[0], 1, nstates); + MUSCLE_INSERT_INT ("pact_ninf", base_ninf); + + /* Output PGOTO. */ + muscle_insert_base_table ("pgoto", base, + base[nstates], nstates + 1, nvectors); + + muscle_insert_base_table ("table", table, + table[0], 1, high + 1); + MUSCLE_INSERT_INT ("table_ninf", table_ninf); + + muscle_insert_base_table ("check", check, + check[0], 1, high + 1); + + /* GLR parsing slightly modifies YYTABLE and YYCHECK (and thus + YYPACT) so that in states with unresolved conflicts, the default + reduction is not used in the conflicted entries, so that there is + a place to put a conflict pointer. + + This means that YYCONFLP and YYCONFL are nonsense for a non-GLR + parser, so we could avoid accidents by not writing them out in + 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); + muscle_insert_unsigned_int_table ("conflicting_rules", conflict_list, + 0, 1, conflict_list_cnt); } -/* the following functions output yytable, yycheck - and the vectors whose elements index the portion starts */ + +/*---------------------------. +| Call the skeleton parser. | +`---------------------------*/ static void -output_base (void) +output_skeleton (void) { - output_short_table (ftable, NULL, "yypact", base, - base[0], 1, nstates); - - putc ('\n', ftable); - - output_short_table (ftable, NULL, "yypgoto", base, - base[nstates], nstates + 1, nvectors); - - XFREE (base); -} - - -static void -output_table (void) -{ - fprintf (ftable, "\n\n#define\tYYLAST\t\t%d\n\n\n", high); - output_short_table (ftable, NULL, "yytable", table, - table[0], 1, high + 1); - XFREE (table); -} - - -static void -output_check (void) -{ - output_short_table (ftable, NULL, "yycheck", check, - check[0], 1, high + 1); - XFREE (check); -} - -/* compute and output yydefact, yydefgoto, yypact, yypgoto, yytable - and yycheck. */ - -static void -output_actions (void) -{ - nvectors = nstates + nvars; - - froms = XCALLOC (short *, nvectors); - tos = XCALLOC (short *, nvectors); - tally = XCALLOC (short, nvectors); - width = XCALLOC (short, nvectors); - - token_actions (); - free_shifts (); - free_reductions (); - XFREE (lookaheads); - XFREE (LA); - XFREE (LAruleno); - XFREE (accessing_symbol); - - goto_actions (); - XFREE (goto_map + ntokens); - XFREE (from_state); - XFREE (to_state); - - sort_actions (); - pack_table (); - putc ('\n', ftable); - output_base (); - output_table (); - putc ('\n', ftable); - output_check (); -} - -/* copy the parser code into the ftable file at the end. */ - -static void -output_parser (void) -{ - int c; - static int number_of_dollar_signs = 0; -#ifdef DONTDEF - FILE *fpars; -#else -#define fpars fparser -#endif - - if (pure_parser) - fprintf (ftable, "#define YYPURE 1\n\n"); - -#ifdef DONTDEF - /* JF no longer needed 'cuz open_extra_files changes the currently - open parser from bison.simple to bison.hairy */ - if (semantic_parser) - fpars = fparser; + 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 + installation. A faulty installation can cause deadlock, so a + cheap sanity check is worthwhile. */ + char const m4sugar[] = "m4sugar/m4sugar.m4"; + char const m4bison[] = "bison.m4"; + char *full_m4sugar; + char *full_m4bison; + char *full_skeleton; + 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_skeleton = xmalloc (pkgdatadirlen + 1 + + (skeleton_size < sizeof m4sugar + ? sizeof m4sugar : skeleton_size)); + strncpy (full_skeleton, pkgdatadir, pkgdatadirlen); + full_skeleton[pkgdatadirlen] = '/'; + strcpy (full_skeleton + pkgdatadirlen + 1, m4sugar); + full_m4sugar = xstrdup (full_skeleton); + strcpy (full_skeleton + pkgdatadirlen + 1, m4bison); + full_m4bison = xstrdup (full_skeleton); + if (strchr (skeleton, '/')) + strcpy (full_skeleton, skeleton); else - fpars = fparser1; -#endif - - /* Loop over lines in the standard parser file. */ - - while (1) - { - int write_line = 1; - - c = getc (fpars); - - /* See if the line starts with `#line. - If so, set write_line to 0. */ - if (no_lines_flag) - if (c == '#') - { - c = getc (fpars); - if (c == 'l') - { - c = getc (fpars); - if (c == 'i') - { - c = getc (fpars); - if (c == 'n') - { - c = getc (fpars); - if (c == 'e') - write_line = 0; - else - fprintf (ftable, "#lin"); - } - else - fprintf (ftable, "#li"); - } - else - fprintf (ftable, "#l"); - } - else - fprintf (ftable, "#"); - } - - /* now write out the line... */ - for (; c != '\n' && c != EOF; c = getc (fpars)) - if (write_line) - { - if (c == '$') - { - number_of_dollar_signs++; - assert (number_of_dollar_signs == 1); - /* `$' in the parser file indicates where to put the actions. - Copy them in at this point. */ - rewind (faction); - for (c = getc (faction); c != EOF; c = getc (faction)) - putc (c, ftable); - /* Skip the end of the line containing `$'. */ - write_line = 0; - } - else - putc (c, ftable); - } - if (c == EOF) - break; - putc (c, ftable); - } - assert (number_of_dollar_signs == 1); -} - -static void -output_program (void) -{ - int c; - - if (!no_lines_flag) - fprintf (ftable, "#line %d \"%s\"\n", lineno, infile); - - c = getc (finput); - while (c != EOF) - { - putc (c, ftable); - c = getc (finput); - } + strcpy (full_skeleton + pkgdatadirlen + 1, skeleton); + xfclose (xfopen (full_m4sugar, "r")); + + /* Create an m4 subprocess connected to us via two pipes. */ + + if (trace_flag & trace_tools) + fprintf (stderr, "running: %s %s - %s %s\n", + m4, full_m4sugar, full_m4bison, full_skeleton); + + argv[0] = m4; + argv[1] = full_m4sugar; + argv[2] = "-"; + argv[3] = full_m4bison; + argv[4] = full_skeleton; + argv[5] = trace_flag & trace_m4 ? "-dV" : NULL; + argv[6] = NULL; + + init_subpipe (); + pid = create_subpipe (argv, filter_fd); + free (full_m4bison); + free (full_m4sugar); + free (full_skeleton); + + out = fdopen (filter_fd[0], "w"); + if (! out) + error (EXIT_FAILURE, get_errno (), + "fdopen"); + + /* Output the definitions of all the muscles. */ + fputs ("m4_init()\n", out); + + user_actions_output (out); + merger_output (out); + token_definitions_output (out); + symbol_code_props_output (out, "destructors", &symbol_destructor_get); + symbol_code_props_output (out, "printers", &symbol_printer_get); + + muscles_m4_output (out); + xfclose (out); + + /* Read and process m4's output. */ + timevar_push (TV_M4); + end_of_output_subpipe (pid, filter_fd); + in = fdopen (filter_fd[1], "r"); + if (! in) + error (EXIT_FAILURE, get_errno (), + "fdopen"); + scan_skel (in); + xfclose (in); + reap_subpipe (pid, m4); + timevar_pop (TV_M4); } - static void -free_itemsets (void) +prepare (void) { - core *cp, *cptmp; - - XFREE (state_table); + /* BISON_USE_PUSH_FOR_PULL is for the test suite and should not be documented + for the user. */ + char const *use_push_for_pull_env = getenv ("BISON_USE_PUSH_FOR_PULL"); + bool use_push_for_pull_flag = false; + if (use_push_for_pull_env != NULL + && use_push_for_pull_env[0] != '\0' + && 0 != strcmp (use_push_for_pull_env, "0")) + use_push_for_pull_flag = true; + + /* Flags. */ + MUSCLE_INSERT_BOOL ("debug_flag", debug_flag); + MUSCLE_INSERT_BOOL ("defines_flag", defines_flag); + MUSCLE_INSERT_BOOL ("error_verbose_flag", error_verbose); + MUSCLE_INSERT_BOOL ("glr_flag", glr_parser); + MUSCLE_INSERT_BOOL ("locations_flag", locations_flag); + MUSCLE_INSERT_BOOL ("nondeterministic_flag", nondeterministic_parser); + MUSCLE_INSERT_BOOL ("pull_flag", pull_parser); + MUSCLE_INSERT_BOOL ("pure_flag", pure_parser); + MUSCLE_INSERT_BOOL ("push_flag", push_parser); + MUSCLE_INSERT_BOOL ("synclines_flag", !no_lines_flag); + MUSCLE_INSERT_BOOL ("tag_seen_flag", tag_seen); + 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; - for (cp = first_state; cp; cp = cptmp) - { - cptmp = cp->next; - XFREE (cp); - } + /* About the skeletons. */ + { + char const *pkgdatadir = getenv ("BISON_PKGDATADIR"); + /* 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 ? pkgdatadir : PKGDATADIR); + } } @@ -1262,52 +623,17 @@ free_itemsets (void) void output (void) { - int c; + obstack_init (&format_obstack); - /* output_token_defines(ftable); / * JF put out token defines FIRST */ - if (!semantic_parser) /* JF Put out other stuff */ - { - rewind (fattrs); - while ((c = getc (fattrs)) != EOF) - putc (c, ftable); - } - reader_output_yylsp (ftable); - if (debug_flag) - fputs ("\ -#ifndef YYDEBUG\n\ -# define YYDEBUG 1\n\ -#endif\n\ -\n", - ftable); - - if (semantic_parser) - fprintf (ftable, "#include \"%s\"\n", attrsfile); - - if (!no_parser_flag) - fprintf (ftable, "#include <stdio.h>\n\n"); - - /* Make "const" do nothing if not in ANSI C. */ - fputs ("\ -#ifndef __cplusplus\n\ -# ifndef __STDC__\n\ -# define const\n\ -# endif\n\ -#endif\n\ -\n", - ftable); - - free_itemsets (); - output_defines (); - output_token_translations (); -/* if (semantic_parser) */ - /* This is now unconditional because debugging printouts can use it. */ - output_gram (); - XFREE (ritem); - if (semantic_parser) - output_stos (); - output_rule_data (); - output_actions (); - if (!no_parser_flag) - output_parser (); - output_program (); + prepare_symbols (); + prepare_rules (); + prepare_states (); + prepare_actions (); + + prepare (); + + /* Process the selected skeleton file. */ + output_skeleton (); + + obstack_free (&format_obstack, NULL); }