X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/56c47203421107ee019a20f3b56cfb967ad75bd2..55f0c7b1f26afabb7f356343f2650626e09fce51:/src/reader.c?ds=sidebyside diff --git a/src/reader.c b/src/reader.c index 0a3f0e25..6b247b7f 100644 --- a/src/reader.c +++ b/src/reader.c @@ -1,6 +1,7 @@ -/* Input parser for bison - Copyright (C) 1984, 1986, 1989, 1992, 1998, 2000, 2001, 2002 - Free Software Foundation, Inc. +/* Input parser for Bison + + Copyright (C) 1984, 1986, 1989, 1992, 1998, 2000, 2001, 2002, 2003, + 2005 Free Software Foundation, Inc. This file is part of Bison, the GNU Compiler Compiler. @@ -16,46 +17,48 @@ 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 Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, + Boston, MA 02110-1301, USA. */ #include "system.h" -#include "quotearg.h" -#include "quote.h" -#include "getargs.h" + +#include + +#include "complain.h" +#include "conflicts.h" #include "files.h" -#include "symtab.h" -#include "symlist.h" -#include "options.h" +#include "getargs.h" #include "gram.h" -#include "complain.h" +#include "muscle_tab.h" #include "output.h" #include "reader.h" -#include "conflicts.h" -#include "muscle_tab.h" +#include "symlist.h" +#include "symtab.h" + +static symbol_list *grammar = NULL; +static bool start_flag = false; +merger_list *merge_functions; -int lineno; -static symbol_list_t *grammar = NULL; -static int start_flag = 0; +/* Has %union been seen? */ +bool typed = false; -/* Nonzero if %union has been seen. */ -int typed = 0; +/* Should rules have a default precedence? */ +bool default_prec = true; /*-----------------------. | Set the start symbol. | `-----------------------*/ void -grammar_start_symbol_set (symbol_t *s, location_t l) +grammar_start_symbol_set (symbol *sym, location loc) { if (start_flag) - complain (_("multiple %s declarations"), "%start"); + complain_at (loc, _("multiple %s declarations"), "%start"); else { - start_flag = 1; - startsymbol = s; - startsymbol_location = l; + start_flag = true; + startsymbol = sym; + startsymbol_location = loc; } } @@ -66,69 +69,79 @@ grammar_start_symbol_set (symbol_t *s, location_t l) `----------------------------------------------------------------*/ void -prologue_augment (const char *prologue, location_t location) +prologue_augment (const char *prologue, location loc) { struct obstack *oout = !typed ? &pre_prologue_obstack : &post_prologue_obstack; - if (!no_lines_flag) - { - obstack_fgrow2 (oout, muscle_find ("linef"), - location.first_line, - quotearg_style (c_quoting_style, - muscle_find ("filename"))); - } + obstack_fgrow1 (oout, "]b4_syncline(%d, [[", loc.start.line); + MUSCLE_OBSTACK_SGROW (oout, + quotearg_style (c_quoting_style, loc.start.file)); + obstack_sgrow (oout, "]])[\n"); obstack_sgrow (oout, prologue); } + + +/*-------------------------------------------------------------------. +| Return the merger index for a merging function named NAME, whose | +| arguments have type TYPE. Records the function, if new, in | +| MERGER_LIST. | +`-------------------------------------------------------------------*/ +static int +get_merge_function (uniqstr name, uniqstr type, location loc) +{ + merger_list *syms; + merger_list head; + int n; + if (! glr_parser) + return 0; -/*----------------------. -| Handle the epilogue. | -`----------------------*/ + if (type == NULL) + type = uniqstr_new (""); -void -epilogue_set (const char *epilogue, location_t location) -{ - if (!no_lines_flag) + head.next = merge_functions; + for (syms = &head, n = 1; syms->next != NULL; syms = syms->next, n += 1) + if (UNIQSTR_EQ (name, syms->next->name)) + break; + if (syms->next == NULL) { - obstack_fgrow2 (&muscle_obstack, muscle_find ("linef"), - location.first_line, - quotearg_style (c_quoting_style, - muscle_find ("filename"))); + syms->next = xmalloc (sizeof syms->next[0]); + syms->next->name = uniqstr_new (name); + syms->next->type = uniqstr_new (type); + syms->next->next = NULL; + merge_functions = head.next; } - obstack_sgrow (&muscle_obstack, epilogue); - obstack_1grow (&muscle_obstack, 0); - muscle_insert ("epilogue", obstack_finish (&muscle_obstack)); + else if (!UNIQSTR_EQ (type, syms->next->type)) + warn_at (loc, _("result type clash on merge function %s: <%s> != <%s>"), + name, type, syms->next->type); + return n; } +/*--------------------------------------. +| Free all merge-function definitions. | +`--------------------------------------*/ - - -/*-------------------------------------------------------------------. -| Generate a dummy symbol, a nonterminal, whose name cannot conflict | -| with the user's names. | -`-------------------------------------------------------------------*/ - -static symbol_t * -gensym (location_t location) +void +free_merger_functions (void) { - /* Incremented for each generated symbol */ - static int gensym_count = 0; - static char buf[256]; - - symbol_t *sym; - - sprintf (buf, "@%d", ++gensym_count); - sym = getsym (buf, location); - sym->class = nterm_sym; - sym->number = nvars++; - return sym; + merger_list *L0; + if (! glr_parser) + return; + L0 = merge_functions; + while (L0 != NULL) + { + merger_list *L1 = L0->next; + free (L0); + L0 = L1; + } } + /*-------------------------------------------------------------------. -| Parse the input grammar into a one symbol_list_t structure. Each | +| Parse the input grammar into a one symbol_list structure. Each | | rule is represented by a sequence of symbols: the left hand side | | followed by the contents of the right hand side, followed by a | | null pointer instead of a symbol to terminate the rule. The next | @@ -146,13 +159,13 @@ gensym (location_t location) `-------------------------------------------------------------------*/ /* The (currently) last symbol of GRAMMAR. */ -symbol_list_t *grammar_end = NULL; +static symbol_list *grammar_end = NULL; -/* Append S to the GRAMMAR. */ +/* Append SYM to the grammar. */ void -grammar_symbol_append (symbol_t *symbol, location_t location) +grammar_symbol_append (symbol *sym, location loc) { - symbol_list_t *p = symbol_list_new (symbol, location); + symbol_list *p = symbol_list_new (sym, loc); if (grammar_end) grammar_end->next = p; @@ -165,8 +178,8 @@ grammar_symbol_append (symbol_t *symbol, location_t location) /* The rule currently being defined, and the previous rule. CURRENT_RULE points to the first LHS of the current rule, while PREVIOUS_RULE_END points to the *end* of the previous rule (NULL). */ -symbol_list_t *current_rule = NULL; -symbol_list_t *previous_rule_end = NULL; +symbol_list *current_rule = NULL; +static symbol_list *previous_rule_end = NULL; /*----------------------------------------------. @@ -174,13 +187,13 @@ symbol_list_t *previous_rule_end = NULL; `----------------------------------------------*/ void -grammar_rule_begin (symbol_t *lhs, location_t location) +grammar_rule_begin (symbol *lhs, location loc) { if (!start_flag) { startsymbol = lhs; - startsymbol_location = location; - start_flag = 1; + startsymbol_location = loc; + start_flag = true; } /* Start a new rule and record its lhs. */ @@ -188,7 +201,7 @@ grammar_rule_begin (symbol_t *lhs, location_t location) ++nritems; previous_rule_end = grammar_end; - grammar_symbol_append (lhs, location); + grammar_symbol_append (lhs, loc); current_rule = grammar_end; /* Mark the rule's lhs as a nonterminal if not already so. */ @@ -200,7 +213,7 @@ grammar_rule_begin (symbol_t *lhs, location_t location) ++nvars; } else if (lhs->class == token_sym) - complain (_("rule given for %s, which is a token"), lhs->tag); + complain_at (loc, _("rule given for %s, which is a token"), lhs->tag); } /* Check that the last rule (CURRENT_RULE) is properly defined. For @@ -209,30 +222,33 @@ grammar_rule_begin (symbol_t *lhs, location_t location) static void grammar_current_rule_check (void) { - symbol_t *lhs = current_rule->sym; - symbol_t *first_rhs = current_rule->next->sym; + symbol *lhs = current_rule->sym; + char const *lhs_type = lhs->type_name; + symbol *first_rhs = current_rule->next->sym; /* If there is an action, then there is nothing we can do: the user - is allowed to shoot in her foot. */ + is allowed to shoot herself in the foot. */ if (current_rule->action) return; - /* If $$ is being set in default way, report if any type mismatch. - */ + /* Don't worry about the default action if $$ is untyped, since $$'s + value can't be used. */ + if (! lhs_type) + return; + + /* If $$ is being set in default way, report if any type mismatch. */ if (first_rhs) { - const char *lhs_type = lhs->type_name ? lhs->type_name : ""; const char *rhs_type = first_rhs->type_name ? first_rhs->type_name : ""; - if (strcmp (lhs_type, rhs_type)) - complain (_("type clash (`%s' `%s') on default action"), - lhs_type, rhs_type); + if (!UNIQSTR_EQ (lhs_type, rhs_type)) + warn_at (current_rule->location, + _("type clash on default action: <%s> != <%s>"), + lhs_type, rhs_type); } /* Warn if there is no default for $$ but we need one. */ else - { - if (lhs->type_name) - complain (_("empty rule for typed nonterminal, and no action")); - } + warn_at (current_rule->location, + _("empty rule for typed nonterminal, and no action")); } @@ -241,11 +257,11 @@ grammar_current_rule_check (void) `-------------------------------------*/ void -grammar_rule_end (location_t location) +grammar_rule_end (location loc) { /* Put an empty link in the list to mark the end of this rule */ grammar_symbol_append (NULL, grammar_end->location); - current_rule->location = location; + current_rule->location = loc; grammar_current_rule_check (); } @@ -266,9 +282,9 @@ grammar_midrule_action (void) /* Make a DUMMY nonterminal, whose location is that of the midrule action. Create the MIDRULE. */ - location_t dummy_location = current_rule->action_location; - symbol_t *dummy = gensym (dummy_location); - symbol_list_t *midrule = symbol_list_new (dummy, dummy_location); + location dummy_location = current_rule->action_location; + symbol *dummy = dummy_symbol_get (dummy_location); + symbol_list *midrule = symbol_list_new (dummy, dummy_location); /* Make a new rule, whose body is empty, before the current one, so that the action just read can belong to it. */ @@ -299,36 +315,63 @@ grammar_midrule_action (void) /* Set the precedence symbol of the current rule to PRECSYM. */ void -grammar_current_rule_prec_set (symbol_t *precsym) +grammar_current_rule_prec_set (symbol *precsym, location loc) { if (current_rule->ruleprec) - complain (_("two @prec's in a row")); + complain_at (loc, _("only one %s allowed per rule"), "%prec"); current_rule->ruleprec = precsym; } -/* Attach a SYMBOL to the current rule. If needed, move the previous +/* Attach dynamic precedence DPREC to the current rule. */ + +void +grammar_current_rule_dprec_set (int dprec, location loc) +{ + if (! glr_parser) + warn_at (loc, _("%s affects only GLR parsers"), "%dprec"); + if (dprec <= 0) + complain_at (loc, _("%s must be followed by positive number"), "%dprec"); + else if (current_rule->dprec != 0) + complain_at (loc, _("only one %s allowed per rule"), "%dprec"); + current_rule->dprec = dprec; +} + +/* Attach a merge function NAME with argument type TYPE to current + rule. */ + +void +grammar_current_rule_merge_set (uniqstr name, location loc) +{ + if (! glr_parser) + warn_at (loc, _("%s affects only GLR parsers"), "%merge"); + if (current_rule->merger != 0) + complain_at (loc, _("only one %s allowed per rule"), "%merge"); + current_rule->merger = + get_merge_function (name, current_rule->sym->type_name, loc); +} + +/* Attach SYM to the current rule. If needed, move the previous action as a mid-rule action. */ void -grammar_current_rule_symbol_append (symbol_t *symbol, location_t location) +grammar_current_rule_symbol_append (symbol *sym, location loc) { if (current_rule->action) grammar_midrule_action (); ++nritems; - grammar_symbol_append (symbol, location); + grammar_symbol_append (sym, loc); } - /* Attach an ACTION to the current rule. If needed, move the previous action as a mid-rule action. */ void -grammar_current_rule_action_append (const char *action, location_t location) +grammar_current_rule_action_append (const char *action, location loc) { if (current_rule->action) grammar_midrule_action (); current_rule->action = action; - current_rule->action_location = location; + current_rule->action_location = loc; } @@ -340,38 +383,38 @@ grammar_current_rule_action_append (const char *action, location_t location) static void packgram (void) { - unsigned int itemno; - int ruleno; - symbol_list_t *p; + unsigned int itemno = 0; + rule_number ruleno = 0; + symbol_list *p = grammar; - ritem = XCALLOC (item_number_t, nritems); - rules = XCALLOC (rule_t, nrules) - 1; + ritem = xnmalloc (nritems, sizeof *ritem); + rules = xnmalloc (nrules, sizeof *rules); - itemno = 0; - ruleno = 1; - - p = grammar; while (p) { - symbol_t *ruleprec = p->ruleprec; + symbol *ruleprec = p->ruleprec; rules[ruleno].user_number = ruleno; rules[ruleno].number = ruleno; rules[ruleno].lhs = p->sym; rules[ruleno].rhs = ritem + itemno; + rules[ruleno].prec = NULL; + rules[ruleno].dprec = p->dprec; + rules[ruleno].merger = p->merger; + rules[ruleno].precsym = NULL; rules[ruleno].location = p->location; - rules[ruleno].useful = TRUE; + rules[ruleno].useful = true; rules[ruleno].action = p->action; rules[ruleno].action_location = p->action_location; p = p->next; while (p && p->sym) { - /* item_number_t = symbol_number_t. + /* item_number = symbol_number. But the former needs to contain more: negative rule numbers. */ ritem[itemno++] = symbol_number_as_item_number (p->sym->number); /* A rule gets by default the precedence and associativity of the last token in it. */ - if (p->sym->class == token_sym) + if (p->sym->class == token_sym && default_prec) rules[ruleno].prec = p->sym; if (p) p = p->next; @@ -384,16 +427,17 @@ packgram (void) rules[ruleno].precsym = ruleprec; rules[ruleno].prec = ruleprec; } - ritem[itemno++] = -ruleno; + ritem[itemno++] = rule_number_as_item_number (ruleno); ++ruleno; if (p) p = p->next; } - assert (itemno == nritems); + if (itemno != nritems) + abort (); - if (trace_flag) + if (trace_flag & trace_sets) ritem_print (stderr); } @@ -407,25 +451,22 @@ packgram (void) void reader (void) { - gram_control_t gram_control; - lineno = 1; - /* Initialize the symbol table. */ symbols_new (); - /* Construct the axiom symbol. */ - axiom = getsym ("$axiom", empty_location); - axiom->class = nterm_sym; - axiom->number = nvars++; + /* Construct the accept symbol. */ + accept = symbol_get ("$accept", empty_location); + accept->class = nterm_sym; + accept->number = nvars++; /* Construct the error token */ - errtoken = getsym ("error", empty_location); + errtoken = symbol_get ("error", empty_location); errtoken->class = token_sym; errtoken->number = ntokens++; /* Construct a token that represents all undefined literal tokens. It is always token number 2. */ - undeftoken = getsym ("$undefined.", empty_location); + undeftoken = symbol_get ("$undefined", empty_location); undeftoken->class = token_sym; undeftoken->number = ntokens++; @@ -433,13 +474,18 @@ reader (void) obstack_init (&pre_prologue_obstack); obstack_init (&post_prologue_obstack); - finput = xfopen (infile, "r"); + finput = xfopen (grammar_file, "r"); gram_in = finput; - gram_debug = !!getenv ("parse"); - gram__flex_debug = !!getenv ("scan"); + gram__flex_debug = trace_flag & trace_scan; + gram_debug = trace_flag & trace_parse; scanner_initialize (); - gram_parse (&gram_control); + gram_parse (); + + /* If something went wrong during the parsing, don't try to + continue. */ + if (complaint_issued) + return; /* Grammar has been read. Do some checking */ if (nrules == 0) @@ -448,25 +494,25 @@ reader (void) /* Report any undefined symbols and consider them nonterminals. */ symbols_check_defined (); - /* If the user did not define her EOFTOKEN, do it now. */ - if (!eoftoken) + /* If the user did not define her ENDTOKEN, do it now. */ + if (!endtoken) { - eoftoken = getsym ("$", empty_location); - eoftoken->class = token_sym; - eoftoken->number = 0; + endtoken = symbol_get ("$end", empty_location); + endtoken->class = token_sym; + endtoken->number = 0; /* Value specified by POSIX. */ - eoftoken->user_token_number = 0; + endtoken->user_token_number = 0; } /* Insert the initial rule, which line is that of the first rule (not that of the start symbol): - axiom: %start EOF. */ + accept: %start EOF. */ { - symbol_list_t *p = symbol_list_new (axiom, empty_location); + symbol_list *p = symbol_list_new (accept, empty_location); p->location = grammar->location; p->next = symbol_list_new (startsymbol, empty_location); - p->next->next = symbol_list_new (eoftoken, empty_location); + p->next->next = symbol_list_new (endtoken, empty_location); p->next->next->next = symbol_list_new (NULL, empty_location); p->next->next->next->next = grammar; nrules += 1; @@ -474,11 +520,8 @@ reader (void) grammar = p; } - if (nsyms > SHRT_MAX) - fatal (_("too many symbols (tokens plus nonterminals); maximum %d"), - SHRT_MAX); - - assert (nsyms == ntokens + nvars); + if (! (nsyms <= SYMBOL_NUMBER_MAXIMUM && nsyms == ntokens + nvars)) + abort (); xfclose (finput); @@ -489,6 +532,6 @@ reader (void) /* Convert the grammar into the format described in gram.h. */ packgram (); - /* The grammar as a symbol_list_t is no longer needed. */ - LIST_FREE (symbol_list_t, grammar); + /* The grammar as a symbol_list is no longer needed. */ + LIST_FREE (symbol_list, grammar); }