/* Symbol table manager for Bison.
- Copyright (C) 1984, 1989, 2000, 2001, 2002, 2004, 2005, 2006, 2007 Free
- Software Foundation, Inc.
+ Copyright (C) 1984, 1989, 2000-2002, 2004-2012 Free Software
+ Foundation, Inc.
This file is part of Bison, the GNU Compiler Compiler.
#include "system.h"
#include <hash.h>
-#include <quotearg.h>
#include "complain.h"
#include "gram.h"
#include "symtab.h"
+/*-------------------------------------------------------------------.
+| Symbols sorted by tag. Allocated by the first invocation of |
+| symbols_do, after which no more symbols should be created. |
+`-------------------------------------------------------------------*/
+
+static symbol **symbols_sorted = NULL;
+
/*------------------------.
| Distinguished symbols. |
`------------------------*/
symbol *res = xmalloc (sizeof *res);
uniqstr_assert (tag);
+
+ /* If the tag is not a string (starts with a double quote), check
+ that it is valid for Yacc. */
+ if (tag[0] != '\"' && tag[0] != '\'' && strchr (tag, '-'))
+ yacc_at (loc, _("POSIX Yacc forbids dashes in symbol names: %s"),
+ tag);
+
res->tag = tag;
res->location = loc;
if (nsyms == SYMBOL_NUMBER_MAXIMUM)
fatal (_("too many symbols in input grammar (limit is %d)"),
- SYMBOL_NUMBER_MAXIMUM);
+ SYMBOL_NUMBER_MAXIMUM);
nsyms++;
return res;
}
| Print a symbol. |
`-----------------*/
-#define SYMBOL_ATTR_PRINT(Attr) \
- if (s->Attr) \
+#define SYMBOL_ATTR_PRINT(Attr) \
+ if (s->Attr) \
fprintf (f, " %s { %s }", #Attr, s->Attr)
#define SYMBOL_CODE_PRINT(Attr) \
#undef SYMBOL_ATTR_PRINT
#undef SYMBOL_CODE_PRINT
+
+/*----------------------------------.
+| Whether S is a valid identifier. |
+`----------------------------------*/
+
+static bool
+is_identifier (uniqstr s)
+{
+ static char const alphanum[26 + 26 + 1 + 10] =
+ "abcdefghijklmnopqrstuvwxyz"
+ "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+ "_"
+ "0123456789";
+ if (!s || ! memchr (alphanum, *s, sizeof alphanum - 10))
+ return false;
+ for (++s; *s; ++s)
+ if (! memchr (alphanum, *s, sizeof alphanum))
+ return false;
+ return true;
+}
+
+
+/*-----------------------------------------------.
+| Get the identifier associated to this symbol. |
+`-----------------------------------------------*/
+uniqstr
+symbol_id_get (symbol const *sym)
+{
+ aver (sym->user_token_number != USER_NUMBER_HAS_STRING_ALIAS);
+ if (sym->alias)
+ sym = sym->alias;
+ return is_identifier (sym->tag) ? sym->tag : 0;
+}
+
+
/*------------------------------------------------------------------.
| Complain that S's WHAT is redeclared at SECOND, and was first set |
| at FIRST. |
}
+
/*-----------------------------------------------------------------.
| Set the TYPE_NAME associated with SYM. Does nothing if passed 0 |
| as TYPE_NAME. |
if (type_name)
{
if (sym->type_name)
- symbol_redeclaration (sym, "%type", sym->type_location, loc);
+ symbol_redeclaration (sym, "%type", sym->type_location, loc);
uniqstr_assert (type_name);
sym->type_name = type_name;
sym->type_location = loc;
}
}
-/*-----------------------------------.
-| Get the CLASS associated with SYM. |
-`-----------------------------------*/
-
-const char *
-symbol_class_get_string (symbol *sym)
-{
- if (sym->class)
- {
- if (sym->class == token_sym)
- return "terminal";
- else if (sym->class == nterm_sym)
- return "nonterminal";
- }
- return "unknown";
-}
-
-
/*-----------------------------------------.
| Set the DESTRUCTOR associated with SYM. |
`-----------------------------------------*/
if (a != undef_assoc)
{
if (sym->prec != 0)
- symbol_redeclaration (sym, assoc_to_string (a), sym->prec_location,
+ symbol_redeclaration (sym, assoc_to_string (a), sym->prec_location,
loc);
sym->prec = prec;
sym->assoc = a;
if (declaring)
{
if (sym->declared)
- warn_at (loc, _("symbol %s redeclared"), sym->tag);
+ warn_at (loc, _("symbol %s redeclared"), sym->tag);
sym->declared = true;
}
}
{
int *user_token_numberp;
- aver (sym->class == token_sym);
-
- if (sym->user_token_number != USER_NUMBER_ALIAS)
+ if (sym->user_token_number != USER_NUMBER_HAS_STRING_ALIAS)
user_token_numberp = &sym->user_token_number;
else
user_token_numberp = &sym->alias->user_token_number;
if (user_token_number == 0)
{
endtoken = sym;
- endtoken->number = 0;
/* It is always mapped to 0, so it was already counted in
- NTOKENS. */
- --ntokens;
+ NTOKENS. */
+ if (endtoken->number != NUMBER_UNDEFINED)
+ --ntokens;
+ endtoken->number = 0;
}
}
if (sym->class == unknown_sym)
{
complain_at
- (sym->location,
- _("symbol %s is used, but is not defined as a token and has no rules"),
- sym->tag);
+ (sym->location,
+ _("symbol %s is used, but is not defined as a token and has no rules"),
+ sym->tag);
sym->class = nterm_sym;
sym->number = nvars++;
}
}
-/*------------------------------------------------------------------.
-| Declare the new symbol SYM. Make it an alias of SYMVAL, and type |
-| SYMVAL with SYM's type. |
-`------------------------------------------------------------------*/
-
void
-symbol_make_alias (symbol *sym, symbol *symval, location loc)
+symbol_make_alias (symbol *sym, symbol *str, location loc)
{
- if (symval->alias)
- warn_at (loc, _("symbol `%s' used more than once as a literal string"),
- symval->tag);
+ if (str->alias)
+ warn_at (loc, _("symbol %s used more than once as a literal string"),
+ str->tag);
else if (sym->alias)
- warn_at (loc, _("symbol `%s' given more than one literal string"),
- sym->tag);
+ warn_at (loc, _("symbol %s given more than one literal string"),
+ sym->tag);
else
{
- symval->class = token_sym;
- symval->user_token_number = sym->user_token_number;
- sym->user_token_number = USER_NUMBER_ALIAS;
- symval->alias = sym;
- sym->alias = symval;
- symval->number = sym->number;
- symbol_type_set (symval, sym->type_name, loc);
+ str->class = token_sym;
+ str->user_token_number = sym->user_token_number;
+ sym->user_token_number = USER_NUMBER_HAS_STRING_ALIAS;
+ str->alias = sym;
+ sym->alias = str;
+ str->number = sym->number;
+ symbol_type_set (str, sym->type_name, loc);
}
}
static inline void
symbol_check_alias_consistency (symbol *this)
{
- symbol *alias = this;
- symbol *orig = this->alias;
+ symbol *sym = this;
+ symbol *str = this->alias;
- /* Check only those that _are_ the aliases. */
- if (!(this->alias && this->user_token_number == USER_NUMBER_ALIAS))
+ /* Check only the symbol in the symbol-string pair. */
+ if (!(this->alias
+ && this->user_token_number == USER_NUMBER_HAS_STRING_ALIAS))
return;
- if (orig->type_name != alias->type_name)
+ if (str->type_name != sym->type_name)
{
- if (orig->type_name)
- symbol_type_set (alias, orig->type_name, orig->type_location);
+ if (str->type_name)
+ symbol_type_set (sym, str->type_name, str->type_location);
else
- symbol_type_set (orig, alias->type_name, alias->type_location);
+ symbol_type_set (str, sym->type_name, sym->type_location);
}
- if (orig->destructor.code || alias->destructor.code)
+ if (str->destructor.code || sym->destructor.code)
{
- if (orig->destructor.code)
- symbol_destructor_set (alias, &orig->destructor);
+ if (str->destructor.code)
+ symbol_destructor_set (sym, &str->destructor);
else
- symbol_destructor_set (orig, &alias->destructor);
+ symbol_destructor_set (str, &sym->destructor);
}
- if (orig->printer.code || alias->printer.code)
+ if (str->printer.code || sym->printer.code)
{
- if (orig->printer.code)
- symbol_printer_set (alias, &orig->printer);
+ if (str->printer.code)
+ symbol_printer_set (sym, &str->printer);
else
- symbol_printer_set (orig, &alias->printer);
+ symbol_printer_set (str, &sym->printer);
}
- if (alias->prec || orig->prec)
+ if (sym->prec || str->prec)
{
- if (orig->prec)
- symbol_precedence_set (alias, orig->prec, orig->assoc,
- orig->prec_location);
+ if (str->prec)
+ symbol_precedence_set (sym, str->prec, str->assoc,
+ str->prec_location);
else
- symbol_precedence_set (orig, alias->prec, alias->assoc,
- alias->prec_location);
+ symbol_precedence_set (str, sym->prec, sym->assoc,
+ sym->prec_location);
}
}
static bool
symbol_check_alias_consistency_processor (void *this,
- void *null ATTRIBUTE_UNUSED)
+ void *null ATTRIBUTE_UNUSED)
{
symbol_check_alias_consistency (this);
return true;
static inline bool
symbol_pack (symbol *this)
{
+ aver (this->number != NUMBER_UNDEFINED);
if (this->class == nterm_sym)
- {
- this->number += ntokens;
- }
- else if (this->alias)
- {
- /* This symbol and its alias are a single token defn.
- Allocate a tokno, and assign to both check agreement of
- prec and assoc fields and make both the same */
- if (this->number == NUMBER_UNDEFINED)
- {
- if (this == endtoken || this->alias == endtoken)
- this->number = this->alias->number = 0;
- else
- {
- aver (this->alias->number != NUMBER_UNDEFINED);
- this->number = this->alias->number;
- }
- }
- /* Do not do processing below for USER_NUMBER_ALIASes. */
- if (this->user_token_number == USER_NUMBER_ALIAS)
- return true;
- }
- else /* this->class == token_sym */
- aver (this->number != NUMBER_UNDEFINED);
+ this->number += ntokens;
+ else if (this->user_token_number == USER_NUMBER_HAS_STRING_ALIAS)
+ return true;
symbols[this->number] = this;
return true;
}
+static void
+user_token_number_redeclaration (int num, symbol *first, symbol *second)
+{
+ /* User token numbers are not assigned during the parsing, but in a
+ second step, via a traversal of the symbol table sorted on tag.
+ However, error messages make more sense if we keep the first
+ declaration first. */
+ if (location_cmp (first->location, second->location) > 0)
+ {
+ symbol* tmp = first;
+ first = second;
+ second = tmp;
+ }
+ complain_at (second->location,
+ _("user token number %d redeclaration for %s"),
+ num, second->tag);
+ complain_at (first->location, _("previous declaration for %s"),
+ first->tag);
+}
/*--------------------------------------------------.
| Put THIS in TOKEN_TRANSLATIONS if it is a token. |
{
/* Non-terminal? */
if (this->class == token_sym
- && this->user_token_number != USER_NUMBER_ALIAS)
+ && this->user_token_number != USER_NUMBER_HAS_STRING_ALIAS)
{
/* A token which translation has already been set? */
if (token_translations[this->user_token_number] != undeftoken->number)
- complain_at (this->location,
- _("tokens %s and %s both assigned number %d"),
- symbols[token_translations[this->user_token_number]]->tag,
- this->tag, this->user_token_number);
+ user_token_number_redeclaration
+ (this->user_token_number,
+ symbols[token_translations[this->user_token_number]],
+ this);
token_translations[this->user_token_number] = this->number;
}
symbols_new (void)
{
symbol_table = hash_initialize (HT_INITIAL_CAPACITY,
- NULL,
- hash_symbol_hasher,
- hash_symbol_comparator,
- free);
+ NULL,
+ hash_symbol_hasher,
+ hash_symbol_comparator,
+ free);
semantic_type_table = hash_initialize (HT_INITIAL_CAPACITY,
- NULL,
- hash_semantic_type_hasher,
- hash_semantic_type_comparator,
- free);
+ NULL,
+ hash_semantic_type_hasher,
+ hash_semantic_type_comparator,
+ free);
}
if (!entry)
{
/* First insertion in the hash. */
+ aver (!symbols_sorted);
entry = symbol_new (key, loc);
- hash_insert (symbol_table, entry);
+ if (!hash_insert (symbol_table, entry))
+ xalloc_die ();
}
return entry;
}
{
/* First insertion in the hash. */
entry = semantic_type_new (key);
- hash_insert (semantic_type_table, entry);
+ if (!hash_insert (semantic_type_table, entry))
+ xalloc_die ();
}
return entry;
}
hash_free (symbol_table);
hash_free (semantic_type_table);
free (symbols);
+ free (symbols_sorted);
}
| terminals. |
`---------------------------------------------------------------*/
+static int
+symbols_cmp (symbol const *a, symbol const *b)
+{
+ return strcmp (a->tag, b->tag);
+}
+
+static int
+symbols_cmp_qsort (void const *a, void const *b)
+{
+ return symbols_cmp (*(symbol * const *)a, *(symbol * const *)b);
+}
+
static void
symbols_do (Hash_processor processor, void *processor_data)
{
- hash_do_for_each (symbol_table, processor, processor_data);
+ size_t count = hash_get_n_entries (symbol_table);
+ if (!symbols_sorted)
+ {
+ symbols_sorted = xnmalloc (count, sizeof *symbols_sorted);
+ hash_get_entries (symbol_table, (void**)symbols_sorted, count);
+ qsort (symbols_sorted, count, sizeof *symbols_sorted,
+ symbols_cmp_qsort);
+ }
+ {
+ size_t i;
+ for (i = 0; i < count; ++i)
+ processor (symbols_sorted[i], processor_data);
+ }
}
-
/*--------------------------------------------------------------.
| Check that all the symbols are defined. Report any undefined |
| symbols and consider them nonterminals. |
{
symbol *this = symbols[i];
if (this->user_token_number != USER_NUMBER_UNDEFINED)
- {
- if (this->user_token_number > max_user_token_number)
- max_user_token_number = this->user_token_number;
- if (this->user_token_number == 256)
- num_256_available_p = false;
- }
+ {
+ if (this->user_token_number > max_user_token_number)
+ max_user_token_number = this->user_token_number;
+ if (this->user_token_number == 256)
+ num_256_available_p = false;
+ }
}
/* If 256 is not used, assign it to error, to follow POSIX. */
{
symbol *this = symbols[i];
if (this->user_token_number == USER_NUMBER_UNDEFINED)
- this->user_token_number = ++max_user_token_number;
+ this->user_token_number = ++max_user_token_number;
if (this->user_token_number > max_user_token_number)
- max_user_token_number = this->user_token_number;
+ max_user_token_number = this->user_token_number;
}
token_translations = xnmalloc (max_user_token_number + 1,
- sizeof *token_translations);
+ sizeof *token_translations);
- /* Initialize all entries for literal tokens to 2, the internal
- token number for $undefined, which represents all invalid inputs.
- */
+ /* Initialize all entries for literal tokens to the internal token
+ number for $undefined, which represents all invalid inputs. */
for (i = 0; i < max_user_token_number + 1; i++)
token_translations[i] = undeftoken->number;
symbols_do (symbol_translation_processor, NULL);
if (startsymbol->class == unknown_sym)
fatal_at (startsymbol_location,
- _("the start symbol %s is undefined"),
- startsymbol->tag);
+ _("the start symbol %s is undefined"),
+ startsymbol->tag);
else if (startsymbol->class == token_sym)
fatal_at (startsymbol_location,
- _("the start symbol %s is a token"),
- startsymbol->tag);
+ _("the start symbol %s is a token"),
+ startsymbol->tag);
}
complain_at (destructor->location,
_("redeclaration for default tagged %%destructor"));
complain_at (default_tagged_destructor.location,
- _("previous declaration"));
+ _("previous declaration"));
}
default_tagged_destructor = *destructor;
}
complain_at (destructor->location,
_("redeclaration for default tagless %%destructor"));
complain_at (default_tagless_destructor.location,
- _("previous declaration"));
+ _("previous declaration"));
}
default_tagless_destructor = *destructor;
}
complain_at (printer->location,
_("redeclaration for default tagged %%printer"));
complain_at (default_tagged_printer.location,
- _("previous declaration"));
+ _("previous declaration"));
}
default_tagged_printer = *printer;
}
complain_at (printer->location,
_("redeclaration for default tagless %%printer"));
complain_at (default_tagless_printer.location,
- _("previous declaration"));
+ _("previous declaration"));
}
default_tagless_printer = *printer;
}
projects / bison.git / blobdiff