-/* Input parser for bison
+/* Input parser for Bison
+
Copyright (C) 1984, 1986, 1989, 1992, 1998, 2000, 2001, 2002
Free Software Foundation, Inc.
the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA. */
-
#include "system.h"
-#include "quotearg.h"
-#include "quote.h"
-#include "getargs.h"
+
+#include <quotearg.h>
+
+#include "complain.h"
+#include "conflicts.h"
#include "files.h"
-#include "symtab.h"
-#include "symlist.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"
-int lineno;
-static symbol_list_t *grammar = NULL;
+static symbol_list *grammar = NULL;
static int start_flag = 0;
merger_list *merge_functions;
`-----------------------*/
void
-grammar_start_symbol_set (symbol_t *s, location_t l)
+grammar_start_symbol_set (symbol *s, location loc)
{
if (start_flag)
- complain_at (l, _("multiple %s declarations"), "%start");
+ complain_at (loc, _("multiple %s declarations"), "%start");
else
{
start_flag = 1;
startsymbol = s;
- startsymbol_location = l;
+ startsymbol_location = loc;
}
}
`----------------------------------------------------------------*/
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);
}
`----------------------*/
void
-epilogue_set (const char *epilogue, location_t location)
+epilogue_augment (const char *epilogue, location loc)
{
- if (!no_lines_flag)
- {
- obstack_fgrow2 (&muscle_obstack, muscle_find ("linef"),
- location.first_line,
- quotearg_style (c_quoting_style,
- muscle_find ("filename")));
- }
+ char *extension = NULL;
+ obstack_fgrow1 (&muscle_obstack, "]b4_syncline([[%d]], [[", loc.start.line);
+ MUSCLE_OBSTACK_SGROW (&muscle_obstack,
+ quotearg_style (c_quoting_style, loc.start.file));
+ obstack_sgrow (&muscle_obstack, "]])[\n");
obstack_sgrow (&muscle_obstack, epilogue);
obstack_1grow (&muscle_obstack, 0);
- muscle_insert ("epilogue", obstack_finish (&muscle_obstack));
+ extension = obstack_finish (&muscle_obstack);
+ muscle_grow ("epilogue", extension, "");
+ obstack_free (&muscle_obstack, extension);
}
\f
- /*-------------------------------------------------------------------.
+/*-------------------------------------------------------------------.
| Return the merger index for a merging function named NAME, whose |
| arguments have type TYPE. Records the function, if new, in |
-| merger_list. |
+| MERGER_LIST. |
`-------------------------------------------------------------------*/
static int
-get_merge_function (const char* name, const char* type)
+get_merge_function (uniqstr name, uniqstr type, location loc)
{
merger_list *syms;
merger_list head;
return 0;
if (type == NULL)
- type = "";
+ type = uniqstr_new ("");
head.next = merge_functions;
- for (syms = &head, n = 1; syms->next != NULL; syms = syms->next, n += 1)
- if (strcmp (name, syms->next->name) == 0)
+ 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) {
- syms->next = XMALLOC (merger_list, 1);
- syms->next->name = strdup (name);
- syms->next->type = strdup (type);
- syms->next->next = NULL;
- merge_functions = head.next;
- } else if (strcmp (type, syms->next->type) != 0)
- warn (_("result type clash on merge function %s: `%s' vs. `%s'"),
- name, type, syms->next->type);
+ if (syms->next == NULL)
+ {
+ MALLOC (syms->next, 1);
+ syms->next->name = uniqstr_new (name);
+ syms->next->type = uniqstr_new (type);
+ syms->next->next = NULL;
+ merge_functions = head.next;
+ }
+ 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;
}
}
}
-/*-------------------------------------------------------------------.
-| Generate a dummy symbol, a nonterminal, whose name cannot conflict |
-| with the user's names. |
-`-------------------------------------------------------------------*/
-
-static symbol_t *
-gensym (location_t location)
-{
- /* 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;
-}
\f
/*-------------------------------------------------------------------.
-| 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 |
`-------------------------------------------------------------------*/
/* The (currently) last symbol of GRAMMAR. */
-symbol_list_t *grammar_end = NULL;
+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;
/* 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;
+symbol_list *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;
+ startsymbol_location = loc;
start_flag = 1;
}
++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. */
++nvars;
}
else if (lhs->class == token_sym)
- complain_at (location, _("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
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_at (current_rule->location,
- _("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_at (current_rule->location,
- _("empty rule for typed nonterminal, and no action"));
- }
+ warn_at (current_rule->location,
+ _("empty rule for typed nonterminal, and no action"));
}
`-------------------------------------*/
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 ();
}
/* 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. */
/* Set the precedence symbol of the current rule to PRECSYM. */
void
-grammar_current_rule_prec_set (symbol_t *precsym, location_t location)
+grammar_current_rule_prec_set (symbol *precsym, location loc)
{
if (current_rule->ruleprec)
- complain_at (location, _("two @prec's in a row"));
+ complain_at (loc, _("only one %s allowed per rule"), "%prec");
current_rule->ruleprec = precsym;
}
/* Attach dynamic precedence DPREC to the current rule. */
void
-grammar_current_rule_dprec_set (int dprec, location_t location)
+grammar_current_rule_dprec_set (int dprec, location loc)
{
if (! glr_parser)
- warn_at (location, _("%%dprec affects only GLR parsers"));
+ warn_at (loc, _("%s affects only GLR parsers"), "%dprec");
if (dprec <= 0)
- complain_at (location, _("%%dprec must be followed by positive number"));
- else if (current_rule->dprec != 0)
- complain_at (location, _("only one %%dprec allowed per rule"));
+ 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;
}
rule. */
void
-grammar_current_rule_merge_set (const char* name, location_t location)
+grammar_current_rule_merge_set (uniqstr name, location loc)
{
if (! glr_parser)
- warn_at (location, _("%%merge affects only GLR parsers"));
- if (current_rule->merger != 0)
- complain_at (location, _("only one %%merge allowed per rule"));
- current_rule->merger =
- get_merge_function (name, current_rule->sym->type_name);
+ 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 a SYMBOL to the current rule. If needed, move the previous
+/* 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;
}
\f
static void
packgram (void)
{
- unsigned int itemno;
- int ruleno;
- symbol_list_t *p;
-
- ritem = XCALLOC (item_number_t, nritems);
- rules = XCALLOC (rule_t, nrules) - 1;
+ unsigned int itemno = 0;
+ rule_number ruleno = 0;
+ symbol_list *p = grammar;
- itemno = 0;
- ruleno = 1;
+ CALLOC (ritem, nritems);
+ CALLOC (rules, nrules);
- 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].location = p->location;
- rules[ruleno].useful = TRUE;
+ rules[ruleno].useful = true;
rules[ruleno].action = p->action;
rules[ruleno].action_location = p->action_location;
rules[ruleno].dprec = p->dprec;
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
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);
}
\f
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++;
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)
/* 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;
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);
/* 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);
}