[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
01-reader-reads-1.patch
|
From: |
Akim Demaille |
|
Subject: |
01-reader-reads-1.patch |
|
Date: |
Sun, 09 Jun 2002 19:15:32 +0200 |
Index: ChangeLog
from Akim Demaille <address@hidden>
Move symbols handling code out of the reader.
* src/reader.h, src/reader.c (errtoken, undeftoken, eoftoken)
(axiom): Move to...
* src/symtab.h, src/symtab.c: here.
* src/gram.c (start_symbol): Remove: use startsymbol->number.
* src/reader.c (startval): Rename as...
* src/symtab.h, src/symtab.c (startsymbol): this.
* src/reader.c: Adjust.
* src/reader.c (symbol_check_defined, symbol_make_alias)
(symbol_check_alias_consistence, symbol_pack, symbol_translation)
(token_translations_init)
Move to...
* src/symtab.c: here.
* src/reader.c (packsymbols): Move to...
* src/symtab.h, src/symtab.c (symbols_pack): here.
* src/symtab.h, src/symtab.c (symbol_make_alias): Takes SYMVAL as
argument.
Index: src/gram.c
--- src/gram.c Sun, 26 May 2002 20:36:39 +0200 akim
+++ src/gram.c Sun, 09 Jun 2002 15:32:26 +0200 akim
@@ -41,8 +41,6 @@
symbol_number_t *token_translations = NULL;
-symbol_number_t start_symbol = 0;
-
int max_user_token_number = 256;
int pure_parser = 0;
Index: src/reader.c
--- src/reader.c Sun, 26 May 2002 20:36:39 +0200 akim
+++ src/reader.c Sun, 09 Jun 2002 15:29:26 +0200 akim
@@ -51,7 +51,6 @@
int lineno;
static symbol_list *grammar = NULL;
static int start_flag = 0;
-static symbol_t *startval = NULL;
/* Nonzero if components of semantic values are used, implying
they must be unions. */
@@ -63,11 +62,6 @@
/* Incremented for each %left, %right or %nonassoc seen */
static int lastprec = 0;
-symbol_t *errtoken = NULL;
-symbol_t *undeftoken = NULL;
-symbol_t *eoftoken = NULL;
-symbol_t *axiom = NULL;
-
static symbol_list *
symbol_list_new (symbol_t *sym)
{
@@ -81,169 +75,6 @@
return res;
}
-/*------------------------.
-| Operations on symbols. |
-`------------------------*/
-
-
-/*-----------------------------------------------------------.
-| If THIS is not defined, report an error, and consider it a |
-| nonterminal. |
-`-----------------------------------------------------------*/
-
-static bool
-symbol_check_defined (symbol_t *this)
-{
- if (this->class == unknown_sym)
- {
- complain
- (_("symbol %s is used, but is not defined as a token and has no rules"),
- this->tag);
- this->class = nterm_sym;
- this->number = nvars++;
- }
-
- return TRUE;
-}
-
-
-/*-------------------------------------------------------------------.
-| Assign a symbol number, and write the definition of the token name |
-| into FDEFINES. Put in SYMBOLS. |
-`-------------------------------------------------------------------*/
-
-static bool
-symbol_make_alias (symbol_t *symbol, char *typename)
-{
- if (symval->alias)
- warn (_("symbol `%s' used more than once as a literal string"),
- symval->tag);
- else if (symbol->alias)
- warn (_("symbol `%s' given more than one literal string"),
- symbol->tag);
- else
- {
- symval->class = token_sym;
- symval->type_name = typename;
- symval->user_token_number = symbol->user_token_number;
- symbol->user_token_number = USER_NUMBER_ALIAS;
- symval->alias = symbol;
- symbol->alias = symval;
- /* symbol and symval combined are only one symbol */
- nsyms--;
- ntokens--;
- assert (ntokens == symbol->number || ntokens == symval->number);
- symbol->number = symval->number =
- (symval->number < symbol->number) ? symval->number : symbol->number;
- }
-
- return TRUE;
-}
-
-/*---------------------------------------------------------.
-| Check that THIS, and its alias, have same precedence and |
-| associativity. |
-`---------------------------------------------------------*/
-
-static bool
-symbol_check_alias_consistence (symbol_t *this)
-{
- /* Check only those who _are_ the aliases. */
- if (this->alias && this->user_token_number == USER_NUMBER_ALIAS)
- {
- if (this->prec != this->alias->prec)
- {
- if (this->prec != 0 && this->alias->prec != 0)
- complain (_("conflicting precedences for %s and %s"),
- this->tag, this->alias->tag);
- if (this->prec != 0)
- this->alias->prec = this->prec;
- else
- this->prec = this->alias->prec;
- }
-
- if (this->assoc != this->alias->assoc)
- {
- if (this->assoc != 0 && this->alias->assoc != 0)
- complain (_("conflicting assoc values for %s and %s"),
- this->tag, this->alias->tag);
- if (this->assoc != 0)
- this->alias->assoc = this->assoc;
- else
- this->assoc = this->alias->assoc;
- }
- }
- return TRUE;
-}
-
-
-/*-------------------------------------------------------------------.
-| Assign a symbol number, and write the definition of the token name |
-| into FDEFINES. Put in SYMBOLS. |
-`-------------------------------------------------------------------*/
-
-static bool
-symbol_pack (symbol_t *this)
-{
- 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 == eoftoken || this->alias == eoftoken)
- this->number = this->alias->number = 0;
- else
- {
- assert (this->alias->number != NUMBER_UNDEFINED);
- this->number = this->alias->number;
- }
- }
- /* Do not do processing below for USER_NUMBER_ALIASs. */
- if (this->user_token_number == USER_NUMBER_ALIAS)
- return TRUE;
- }
- else /* this->class == token_sym */
- {
- assert (this->number != NUMBER_UNDEFINED);
- }
-
- symbols[this->number] = this;
- return TRUE;
-}
-
-
-
-
-/*--------------------------------------------------.
-| Put THIS in TOKEN_TRANSLATIONS if it is a token. |
-`--------------------------------------------------*/
-
-static bool
-symbol_translation (symbol_t *this)
-{
- /* Non-terminal? */
- if (this->class == token_sym
- && this->user_token_number != USER_NUMBER_ALIAS)
- {
- /* A token which translation has already been set? */
- if (token_translations[this->user_token_number] != undeftoken->number)
- complain (_("tokens %s and %s both assigned number %d"),
- symbols[token_translations[this->user_token_number]]->tag,
- this->tag, this->user_token_number);
-
- token_translations[this->user_token_number] = this->number;
- }
-
- return TRUE;
-}
-
-
/*===================\
| Low level lexing. |
\===================*/
@@ -699,7 +530,7 @@
}
else if (token == tok_identifier && *symval->tag == '\"' && symbol)
{
- symbol_make_alias (symbol, typename);
+ symbol_make_alias (symbol, symval, typename);
symbol = NULL;
}
else if (token == tok_identifier)
@@ -762,7 +593,7 @@
else
{
start_flag = 1;
- startval = symval;
+ startsymbol = symval;
}
}
@@ -1401,7 +1232,7 @@
if (!start_flag)
{
- startval = lhs;
+ startsymbol = lhs;
start_flag = 1;
}
@@ -1599,7 +1430,7 @@
fatal (_("no rules in the input grammar"));
/* Report any undefined symbols and consider them nonterminals. */
- symbols_do (symbol_check_defined, NULL);
+ symbols_check_defined ();
/* Insert the initial rule, which line is that of the first rule
(not that of the start symbol):
@@ -1607,14 +1438,13 @@
axiom: %start EOF. */
p = symbol_list_new (axiom);
p->line = grammar->line;
- p->next = symbol_list_new (startval);
+ p->next = symbol_list_new (startsymbol);
p->next->next = symbol_list_new (eoftoken);
p->next->next->next = symbol_list_new (NULL);
p->next->next->next->next = grammar;
nrules += 1;
nritems += 3;
grammar = p;
- startval = axiom;
if (nsyms > SHRT_MAX)
fatal (_("too many symbols (tokens plus nonterminals); maximum %d"),
@@ -1649,85 +1479,6 @@
}
-/*------------------------------------------------------------------.
-| Set TOKEN_TRANSLATIONS. Check that no two symbols share the same |
-| number. |
-`------------------------------------------------------------------*/
-
-static void
-token_translations_init (void)
-{
- int num_256_available_p = TRUE;
- int i;
-
- /* Find the highest user token number, and whether 256, the POSIX
- preferred user token number for the error token, is used. */
- max_user_token_number = 0;
- for (i = 0; i < ntokens; ++i)
- {
- symbol_t *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 256 is not used, assign it to error, to follow POSIX. */
- if (num_256_available_p
- && errtoken->user_token_number == USER_NUMBER_UNDEFINED)
- errtoken->user_token_number = 256;
-
- /* Set the missing user numbers. */
- if (max_user_token_number < 256)
- max_user_token_number = 256;
-
- for (i = 0; i < ntokens; ++i)
- {
- symbol_t *this = symbols[i];
- if (this->user_token_number == USER_NUMBER_UNDEFINED)
- 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;
- }
-
- token_translations = XCALLOC (symbol_number_t, max_user_token_number + 1);
-
- /* Initialize all entries for literal tokens to 2, 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, NULL);
-}
-
-
-/*----------------------------------------------------------------.
-| Assign symbol numbers, and write definition of token names into |
-| FDEFINES. Set up vectors SYMBOL_TABLE, TAGS of symbols. |
-`----------------------------------------------------------------*/
-
-static void
-packsymbols (void)
-{
- symbols = XCALLOC (symbol_t *, nsyms);
-
- symbols_do (symbol_check_alias_consistence, NULL);
- symbols_do (symbol_pack, NULL);
-
- token_translations_init ();
-
- if (startval->class == unknown_sym)
- fatal (_("the start symbol %s is undefined"), startval->tag);
- else if (startval->class == token_sym)
- fatal (_("the start symbol %s is a token"), startval->tag);
-
- start_symbol = startval->number;
-}
-
-
/*---------------------------------------------------------------.
| Convert the rules into the representation using RRHS, RLHS and |
| RITEM. |
@@ -1862,7 +1613,7 @@
/* Assign the symbols their symbol numbers. Write #defines for the
token symbols into FDEFINES if requested. */
- packsymbols ();
+ symbols_pack ();
/* Convert the grammar into the format described in gram.h. */
packgram ();
Index: src/reader.h
--- src/reader.h Sat, 04 May 2002 21:05:50 +0200 akim
+++ src/reader.h Sun, 09 Jun 2002 15:25:55 +0200 akim
@@ -30,9 +30,4 @@
extern int lineno;
-extern symbol_t *errtoken;
-extern symbol_t *undeftoken;
-extern symbol_t *eoftoken;
-extern symbol_t *axiom;
-
#endif /* !READER_H_ */
Index: src/reduce.c
--- src/reduce.c Sun, 26 May 2002 20:36:39 +0200 akim
+++ src/reduce.c Sun, 09 Jun 2002 15:31:27 +0200 akim
@@ -165,9 +165,9 @@
Pp = bitset_create (nrules + 1, BITSET_FIXED);
/* If the start symbol isn't useful, then nothing will be useful. */
- if (bitset_test (N, start_symbol - ntokens))
+ if (bitset_test (N, axiom->number - ntokens))
{
- bitset_set (V, start_symbol);
+ bitset_set (V, axiom->number);
while (1)
{
@@ -315,7 +315,7 @@
if (ISVAR (*rhsp))
*rhsp = symbol_number_as_item_number (nontermmap[*rhsp]);
}
- start_symbol = nontermmap[start_symbol];
+ axiom->number = nontermmap[axiom->number];
}
nsyms -= nuseless_nonterminals;
@@ -435,9 +435,9 @@
reduce_print ();
- if (!bitset_test (N, start_symbol - ntokens))
+ if (!bitset_test (N, axiom->number - ntokens))
fatal (_("Start symbol %s does not derive any sentence"),
- quotearg_style (escape_quoting_style, symbols[start_symbol]->tag));
+ quotearg_style (escape_quoting_style, symbols[axiom->number]->tag));
/* First reduce the nonterminals, as they renumber themselves in the
whole grammar. If you change the order, nonterms would be
Index: src/symtab.c
--- src/symtab.c Sat, 04 May 2002 10:44:42 +0200 akim
+++ src/symtab.c Sun, 09 Jun 2002 15:32:34 +0200 akim
@@ -24,6 +24,16 @@
#include "symtab.h"
#include "gram.h"
+/*------------------------.
+| Distinguished symbols. |
+`------------------------*/
+
+symbol_t *errtoken = NULL;
+symbol_t *undeftoken = NULL;
+symbol_t *eoftoken = NULL;
+symbol_t *axiom = NULL;
+symbol_t *startsymbol = NULL;
+
/*---------------------------------.
| Create a new symbol, named TAG. |
`---------------------------------*/
@@ -65,9 +75,165 @@
}
+/*-----------------------------------------------------------.
+| If THIS is not defined, report an error, and consider it a |
+| nonterminal. |
+`-----------------------------------------------------------*/
+
+static bool
+symbol_check_defined (symbol_t *this)
+{
+ if (this->class == unknown_sym)
+ {
+ complain
+ (_("symbol %s is used, but is not defined as a token and has no rules"),
+ this->tag);
+ this->class = nterm_sym;
+ this->number = nvars++;
+ }
+
+ return TRUE;
+}
+
+
+/*-------------------------------------------------------------------.
+| Declare the new SYMBOL. Make it an alias of SYMVAL, and type them |
+| with TYPENAME. |
+`-------------------------------------------------------------------*/
+
+void
+symbol_make_alias (symbol_t *symbol, symbol_t *symval, char *typename)
+{
+ if (symval->alias)
+ warn (_("symbol `%s' used more than once as a literal string"),
+ symval->tag);
+ else if (symbol->alias)
+ warn (_("symbol `%s' given more than one literal string"),
+ symbol->tag);
+ else
+ {
+ symval->class = token_sym;
+ symval->type_name = typename;
+ symval->user_token_number = symbol->user_token_number;
+ symbol->user_token_number = USER_NUMBER_ALIAS;
+ symval->alias = symbol;
+ symbol->alias = symval;
+ /* symbol and symval combined are only one symbol */
+ nsyms--;
+ ntokens--;
+ assert (ntokens == symbol->number || ntokens == symval->number);
+ symbol->number = symval->number =
+ (symval->number < symbol->number) ? symval->number : symbol->number;
+ }
+}
+
+
+/*---------------------------------------------------------.
+| Check that THIS, and its alias, have same precedence and |
+| associativity. |
+`---------------------------------------------------------*/
+
+static bool
+symbol_check_alias_consistence (symbol_t *this)
+{
+ /* Check only those who _are_ the aliases. */
+ if (this->alias && this->user_token_number == USER_NUMBER_ALIAS)
+ {
+ if (this->prec != this->alias->prec)
+ {
+ if (this->prec != 0 && this->alias->prec != 0)
+ complain (_("conflicting precedences for %s and %s"),
+ this->tag, this->alias->tag);
+ if (this->prec != 0)
+ this->alias->prec = this->prec;
+ else
+ this->prec = this->alias->prec;
+ }
+
+ if (this->assoc != this->alias->assoc)
+ {
+ if (this->assoc != 0 && this->alias->assoc != 0)
+ complain (_("conflicting assoc values for %s and %s"),
+ this->tag, this->alias->tag);
+ if (this->assoc != 0)
+ this->alias->assoc = this->assoc;
+ else
+ this->assoc = this->alias->assoc;
+ }
+ }
+ return TRUE;
+}
+
+
+/*-------------------------------------------------------------------.
+| Assign a symbol number, and write the definition of the token name |
+| into FDEFINES. Put in SYMBOLS. |
+`-------------------------------------------------------------------*/
+
+static bool
+symbol_pack (symbol_t *this)
+{
+ 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 == eoftoken || this->alias == eoftoken)
+ this->number = this->alias->number = 0;
+ else
+ {
+ assert (this->alias->number != NUMBER_UNDEFINED);
+ this->number = this->alias->number;
+ }
+ }
+ /* Do not do processing below for USER_NUMBER_ALIASs. */
+ if (this->user_token_number == USER_NUMBER_ALIAS)
+ return TRUE;
+ }
+ else /* this->class == token_sym */
+ {
+ assert (this->number != NUMBER_UNDEFINED);
+ }
+
+ symbols[this->number] = this;
+ return TRUE;
+}
+
+
+
+
+/*--------------------------------------------------.
+| Put THIS in TOKEN_TRANSLATIONS if it is a token. |
+`--------------------------------------------------*/
+
+static bool
+symbol_translation (symbol_t *this)
+{
+ /* Non-terminal? */
+ if (this->class == token_sym
+ && this->user_token_number != USER_NUMBER_ALIAS)
+ {
+ /* A token which translation has already been set? */
+ if (token_translations[this->user_token_number] != undeftoken->number)
+ complain (_("tokens %s and %s both assigned number %d"),
+ symbols[token_translations[this->user_token_number]]->tag,
+ this->tag, this->user_token_number);
+
+ token_translations[this->user_token_number] = this->number;
+ }
+
+ return TRUE;
+}
+
/*----------------------.
-| A symbol_t hash table. |
+| A symbol hash table. |
`----------------------*/
/* Initial capacity of symbols hash table. */
@@ -89,7 +255,7 @@
/*-------------------------------.
-| Create the symbol_t hash table. |
+| Create the symbol hash table. |
`-------------------------------*/
void
@@ -149,4 +315,92 @@
hash_do_for_each (symbol_table,
(Hash_processor) processor,
processor_data);
+}
+
+
+/*--------------------------------------------------------------.
+| Check that all the symbols are defined. Report any undefined |
+| symbols and consider them nonterminals. |
+`--------------------------------------------------------------*/
+
+void
+symbols_check_defined (void)
+{
+ symbols_do (symbol_check_defined, NULL);
+}
+
+/*------------------------------------------------------------------.
+| Set TOKEN_TRANSLATIONS. Check that no two symbols share the same |
+| number. |
+`------------------------------------------------------------------*/
+
+static void
+symbols_token_translations_init (void)
+{
+ int num_256_available_p = TRUE;
+ int i;
+
+ /* Find the highest user token number, and whether 256, the POSIX
+ preferred user token number for the error token, is used. */
+ max_user_token_number = 0;
+ for (i = 0; i < ntokens; ++i)
+ {
+ symbol_t *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 256 is not used, assign it to error, to follow POSIX. */
+ if (num_256_available_p
+ && errtoken->user_token_number == USER_NUMBER_UNDEFINED)
+ errtoken->user_token_number = 256;
+
+ /* Set the missing user numbers. */
+ if (max_user_token_number < 256)
+ max_user_token_number = 256;
+
+ for (i = 0; i < ntokens; ++i)
+ {
+ symbol_t *this = symbols[i];
+ if (this->user_token_number == USER_NUMBER_UNDEFINED)
+ 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;
+ }
+
+ token_translations = XCALLOC (symbol_number_t, max_user_token_number + 1);
+
+ /* Initialize all entries for literal tokens to 2, 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, NULL);
+}
+
+
+/*----------------------------------------------------------------.
+| Assign symbol numbers, and write definition of token names into |
+| FDEFINES. Set up vectors SYMBOL_TABLE, TAGS of symbols. |
+`----------------------------------------------------------------*/
+
+void
+symbols_pack (void)
+{
+ symbols = XCALLOC (symbol_t *, nsyms);
+
+ symbols_do (symbol_check_alias_consistence, NULL);
+ symbols_do (symbol_pack, NULL);
+
+ symbols_token_translations_init ();
+
+ if (startsymbol->class == unknown_sym)
+ fatal (_("the start symbol %s is undefined"), startsymbol->tag);
+ else if (startsymbol->class == token_sym)
+ fatal (_("the start symbol %s is a token"), startsymbol->tag);
}
Index: src/symtab.h
--- src/symtab.h Sun, 26 May 2002 20:36:39 +0200 akim
+++ src/symtab.h Sun, 09 Jun 2002 15:28:37 +0200 akim
@@ -22,7 +22,9 @@
#ifndef SYMTAB_H_
# define SYMTAB_H_
-#define TABSIZE 1009
+/*----------.
+| Symbols. |
+`----------*/
/* Associativity values for tokens and rules. */
typedef enum
@@ -79,13 +81,39 @@
#define NUMBER_UNDEFINED ((symbol_number_t) -1)
-/* A function to apply to each symbol. */
-typedef bool (*symbol_processor) PARAMS ((symbol_t *));
+/* Fetch (or create) the symbol associated to KEY. */
+symbol_t *getsym PARAMS ((const char *key));
+
+/* Distinguished symbols. AXIOM is the real start symbol, that used
+ by the automaton. STARTSYMBOL is the one specified by the user.
+ */
+extern symbol_t *errtoken;
+extern symbol_t *undeftoken;
+extern symbol_t *eoftoken;
+extern symbol_t *axiom;
+extern symbol_t *startsymbol;
-symbol_t *getsym PARAMS ((const char *));
+/*---------------.
+| Symbol table. |
+`---------------*/
+
+
+/* Create the symbol table. */
void symbols_new PARAMS ((void));
+
+/* A function to apply to each symbol. */
+typedef bool (*symbol_processor) PARAMS ((symbol_t *));
+
+/* Apply PROCESSOR to all the symbols. PROCESSOR must return TRUE: on
+ FALSE, the processing stops. */
void symbols_do PARAMS ((symbol_processor processor, void *processor_data));
+
+/* Free all the memory allocated for symbols. */
void symbols_free PARAMS ((void));
+
+/* Perform various sanity checks, assign symbol numbers, and set up
+ TOKEN_TRANSLATIONS. */
+void symbols_pack PARAMS ((void));
#endif /* !SYMTAB_H_ */
| [Prev in Thread] |
Current Thread |
[Next in Thread] |
- 01-reader-reads-1.patch,
Akim Demaille <=