- Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009
- Free Software Foundation, Inc.
+ Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010 Free
+ Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
This file is part of the GNU C Library.
Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
re_search_internal (const regex_t *preg,
const char *string, Idx length,
Idx start, Idx last_start, Idx stop,
re_search_internal (const regex_t *preg,
const char *string, Idx length,
Idx start, Idx last_start, Idx stop,
- for (reg_idx = 0; reg_idx + 1 < nmatch; reg_idx++)
- if (dfa->subexp_map[reg_idx] != reg_idx)
- {
- pmatch[reg_idx + 1].rm_so
- = pmatch[dfa->subexp_map[reg_idx] + 1].rm_so;
- pmatch[reg_idx + 1].rm_eo
- = pmatch[dfa->subexp_map[reg_idx] + 1].rm_eo;
- }
+ for (reg_idx = 0; reg_idx + 1 < nmatch; reg_idx++)
+ if (dfa->subexp_map[reg_idx] != reg_idx)
+ {
+ pmatch[reg_idx + 1].rm_so
+ = pmatch[dfa->subexp_map[reg_idx] + 1].rm_so;
+ pmatch[reg_idx + 1].rm_eo
+ = pmatch[dfa->subexp_map[reg_idx] + 1].rm_eo;
+ }
{
err = transit_state_bkref (mctx, &cur_state->nodes);
if (BE (err != REG_NOERROR, 0))
{
err = transit_state_bkref (mctx, &cur_state->nodes);
if (BE (err != REG_NOERROR, 0))
Idx next_char_idx = re_string_cur_idx (&mctx->input) + 1;
if (BE (next_char_idx >= mctx->input.bufs_len, 0)
Idx next_char_idx = re_string_cur_idx (&mctx->input) + 1;
if (BE (next_char_idx >= mctx->input.bufs_len, 0)
- || (BE (next_char_idx >= mctx->input.valid_len, 0)
- && mctx->input.valid_len < mctx->input.len))
- {
- err = extend_buffers (mctx);
- if (BE (err != REG_NOERROR, 0))
+ || (BE (next_char_idx >= mctx->input.valid_len, 0)
+ && mctx->input.valid_len < mctx->input.len))
+ {
+ err = extend_buffers (mctx);
+ if (BE (err != REG_NOERROR, 0))
cur_state = transit_state (&err, mctx, cur_state);
if (mctx->state_log != NULL)
cur_state = transit_state (&err, mctx, cur_state);
if (mctx->state_log != NULL)
/* Otherwise, push the second epsilon-transition on the fail stack. */
else if (fs != NULL
&& push_fail_stack (fs, *pidx, candidate, nregs, regs,
/* Otherwise, push the second epsilon-transition on the fail stack. */
else if (fs != NULL
&& push_fail_stack (fs, *pidx, candidate, nregs, regs,
push_fail_stack (struct re_fail_stack_t *fs, Idx str_idx, Idx dest_node,
Idx nregs, regmatch_t *regs, re_node_set *eps_via_nodes)
{
push_fail_stack (struct re_fail_stack_t *fs, Idx str_idx, Idx dest_node,
Idx nregs, regmatch_t *regs, re_node_set *eps_via_nodes)
{
set_regs (const regex_t *preg, const re_match_context_t *mctx, size_t nmatch,
regmatch_t *pmatch, bool fl_backtrack)
{
set_regs (const regex_t *preg, const re_match_context_t *mctx, size_t nmatch,
regmatch_t *pmatch, bool fl_backtrack)
{
if (mctx->state_log[str_idx])
{
err = build_sifted_states (mctx, sctx, str_idx, &cur_dest);
if (mctx->state_log[str_idx])
{
err = build_sifted_states (mctx, sctx, str_idx, &cur_dest);
build_sifted_states (const re_match_context_t *mctx, re_sift_context_t *sctx,
Idx str_idx, re_node_set *cur_dest)
{
build_sifted_states (const re_match_context_t *mctx, re_sift_context_t *sctx,
Idx str_idx, re_node_set *cur_dest)
{
add_epsilon_src_nodes (const re_dfa_t *dfa, re_node_set *dest_nodes,
const re_node_set *candidates)
{
add_epsilon_src_nodes (const re_dfa_t *dfa, re_node_set *dest_nodes,
const re_node_set *candidates)
{
{
err = re_node_set_alloc (&state->inveclosure, dest_nodes->nelem);
if (BE (err != REG_NOERROR, 0))
{
err = re_node_set_alloc (&state->inveclosure, dest_nodes->nelem);
if (BE (err != REG_NOERROR, 0))
- re_node_set_merge (&state->inveclosure,
- dfa->inveclosures + dest_nodes->elems[i]);
+ {
+ err = re_node_set_merge (&state->inveclosure,
+ dfa->inveclosures + dest_nodes->elems[i]);
+ if (BE (err != REG_NOERROR, 0))
+ return REG_ESPACE;
+ }
if (subexp_idx < BITSET_WORD_BITS)
ent->eps_reachable_subexps_map
&= ~((bitset_word_t) 1 << subexp_idx);
if (subexp_idx < BITSET_WORD_BITS)
ent->eps_reachable_subexps_map
&= ~((bitset_word_t) 1 << subexp_idx);
sift_states_bkref (const re_match_context_t *mctx, re_sift_context_t *sctx,
Idx str_idx, const re_node_set *candidates)
{
sift_states_bkref (const re_match_context_t *mctx, re_sift_context_t *sctx,
Idx str_idx, const re_node_set *candidates)
{
re_node_set_remove (&local_sctx.limits, enabled_idx);
/* mctx->bkref_ents may have changed, reload the pointer. */
re_node_set_remove (&local_sctx.limits, enabled_idx);
/* mctx->bkref_ents may have changed, reload the pointer. */
trtable = state->word_trtable;
if (BE (trtable != NULL, 1))
trtable = state->word_trtable;
if (BE (trtable != NULL, 1))
unsigned int context;
re_node_set next_nodes, *log_nodes, *table_nodes = NULL;
/* If (state_log[cur_idx] != 0), it implies that cur_idx is
unsigned int context;
re_node_set next_nodes, *log_nodes, *table_nodes = NULL;
/* If (state_log[cur_idx] != 0), it implies that cur_idx is
- the destination of a multibyte char/collating element/
- back reference. Then the next state is the union set of
- these destinations and the results of the transition table. */
+ the destination of a multibyte char/collating element/
+ back reference. Then the next state is the union set of
+ these destinations and the results of the transition table. */
- {
- table_nodes = next_state->entrance_nodes;
- *err = re_node_set_init_union (&next_nodes, table_nodes,
+ {
+ table_nodes = next_state->entrance_nodes;
+ *err = re_node_set_init_union (&next_nodes, table_nodes,
/* Note: We already add the nodes of the initial state,
then we don't need to add them here. */
/* Note: We already add the nodes of the initial state,
then we don't need to add them here. */
re_string_cur_idx (&mctx->input) - 1,
mctx->eflags);
next_state = mctx->state_log[cur_idx]
re_string_cur_idx (&mctx->input) - 1,
mctx->eflags);
next_state = mctx->state_log[cur_idx]
- = re_acquire_state_context (err, dfa, &next_nodes, context);
+ = re_acquire_state_context (err, dfa, &next_nodes, context);
- if (++cur_str_idx > max)
- return NULL;
- re_string_skip_bytes (&mctx->input, 1);
+ if (++cur_str_idx > max)
+ return NULL;
+ re_string_skip_bytes (&mctx->input, 1);
get_subexp (re_match_context_t *mctx, Idx bkref_node, Idx bkref_str_idx)
{
const re_dfa_t *const dfa = mctx->dfa;
get_subexp (re_match_context_t *mctx, Idx bkref_node, Idx bkref_str_idx)
{
const re_dfa_t *const dfa = mctx->dfa;
check_arrival (re_match_context_t *mctx, state_array_t *path, Idx top_node,
Idx top_str, Idx last_node, Idx last_str, int type)
{
check_arrival (re_match_context_t *mctx, state_array_t *path, Idx top_node,
Idx top_str, Idx last_node, Idx last_str, int type)
{
check_arrival_add_next_nodes (re_match_context_t *mctx, Idx str_idx,
re_node_set *cur_nodes, re_node_set *next_nodes)
{
check_arrival_add_next_nodes (re_match_context_t *mctx, Idx str_idx,
re_node_set *cur_nodes, re_node_set *next_nodes)
{
check_arrival_expand_ecl_sub (const re_dfa_t *dfa, re_node_set *dst_nodes,
Idx target, Idx ex_subexp, int type)
{
check_arrival_expand_ecl_sub (const re_dfa_t *dfa, re_node_set *dst_nodes,
Idx target, Idx ex_subexp, int type)
{
expand_bkref_cache (re_match_context_t *mctx, re_node_set *cur_nodes,
Idx cur_str, Idx subexp_num, int type)
{
expand_bkref_cache (re_match_context_t *mctx, re_node_set *cur_nodes,
Idx cur_str, Idx subexp_num, int type)
{
if (ASCII_CHARS % BITSET_WORD_BITS == 0)
memset (accepts, -1, ASCII_CHARS / CHAR_BIT);
else
if (ASCII_CHARS % BITSET_WORD_BITS == 0)
memset (accepts, -1, ASCII_CHARS / CHAR_BIT);
else
bitset_clear (accepts, '\n');
if (dfa->syntax & RE_DOT_NOT_NULL)
bitset_clear (accepts, '\0');
bitset_clear (accepts, '\n');
if (dfa->syntax & RE_DOT_NOT_NULL)
bitset_clear (accepts, '\0');
/* FIXME: I don't think this if is needed, as both '\n'
and '\0' are char_len == 1. */
/* '.' accepts any one character except the following two cases. */
/* FIXME: I don't think this if is needed, as both '\n'
and '\0' are char_len == 1. */
/* '.' accepts any one character except the following two cases. */
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
indirect = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
indirect = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
- size_t weight_len = weights[idx];
- if (weight_len == weights[equiv_class_idx])
+ size_t weight_len = weights[idx & 0xffffff];
+ if (weight_len == weights[equiv_class_idx & 0xffffff]
+ && (idx >> 24) == (equiv_class_idx >> 24))
match_ctx_add_entry (re_match_context_t *mctx, Idx node, Idx str_idx, Idx from,
Idx to)
{
match_ctx_add_entry (re_match_context_t *mctx, Idx node, Idx str_idx, Idx from,
Idx to)
{