/* Extended regular expression matching and search library.
- Copyright (C) 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
+ Copyright (C) 2002-2014 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
- the Free Software Foundation; either version 2, or (at your option)
- any later version.
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
- This program is distributed in the hope that it will be useful,
+ The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- GNU General Public License for more details.
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
- You should have received a copy of the GNU General Public License along
- with this program; if not, write to the Free Software Foundation,
- Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
static reg_errcode_t match_ctx_init (re_match_context_t *cache, int eflags,
Idx n) internal_function;
struct re_registers *regs,
bool ret_len) internal_function;
static unsigned re_copy_regs (struct re_registers *regs, regmatch_t *pmatch,
- Idx nregs, int regs_allocated) internal_function;
+ Idx nregs, int regs_allocated) internal_function;
static reg_errcode_t prune_impossible_nodes (re_match_context_t *mctx)
internal_function;
static Idx check_matching (re_match_context_t *mctx, bool fl_longest_match,
- Idx *p_match_first)
- internal_function;
+ Idx *p_match_first) internal_function;
static Idx check_halt_state_context (const re_match_context_t *mctx,
const re_dfastate_t *state, Idx idx)
internal_function;
-static void update_regs (re_dfa_t *dfa, regmatch_t *pmatch,
+static void update_regs (const re_dfa_t *dfa, regmatch_t *pmatch,
regmatch_t *prev_idx_match, Idx cur_node,
Idx cur_idx, Idx nmatch) internal_function;
static reg_errcode_t 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) internal_function;
+ re_node_set *eps_via_nodes)
+ internal_function;
static reg_errcode_t set_regs (const regex_t *preg,
const re_match_context_t *mctx,
size_t nmatch, regmatch_t *pmatch,
bool fl_backtrack) internal_function;
-static reg_errcode_t free_fail_stack_return (struct re_fail_stack_t *fs) internal_function;
+static reg_errcode_t free_fail_stack_return (struct re_fail_stack_t *fs)
+ internal_function;
#ifdef RE_ENABLE_I18N
static int sift_states_iter_mb (const re_match_context_t *mctx,
re_sift_context_t *sctx,
- Idx node_idx, Idx str_idx, Idx max_str_idx) internal_function;
+ Idx node_idx, Idx str_idx, Idx max_str_idx)
+ internal_function;
#endif /* RE_ENABLE_I18N */
-static reg_errcode_t sift_states_backward (re_match_context_t *mctx,
- re_sift_context_t *sctx) internal_function;
-static reg_errcode_t build_sifted_states (re_match_context_t *mctx,
+static reg_errcode_t sift_states_backward (const re_match_context_t *mctx,
+ re_sift_context_t *sctx)
+ internal_function;
+static reg_errcode_t build_sifted_states (const re_match_context_t *mctx,
re_sift_context_t *sctx, Idx str_idx,
- re_node_set *cur_dest) internal_function;
-static reg_errcode_t update_cur_sifted_state (re_match_context_t *mctx,
+ re_node_set *cur_dest)
+ internal_function;
+static reg_errcode_t update_cur_sifted_state (const re_match_context_t *mctx,
re_sift_context_t *sctx,
Idx str_idx,
- re_node_set *dest_nodes) internal_function;
-static reg_errcode_t add_epsilon_src_nodes (re_dfa_t *dfa,
+ re_node_set *dest_nodes)
+ internal_function;
+static reg_errcode_t add_epsilon_src_nodes (const re_dfa_t *dfa,
re_node_set *dest_nodes,
- const re_node_set *candidates) internal_function;
+ const re_node_set *candidates)
+ internal_function;
static bool check_dst_limits (const re_match_context_t *mctx,
const re_node_set *limits,
Idx dst_node, Idx dst_idx, Idx src_node,
Idx src_idx) internal_function;
static int check_dst_limits_calc_pos_1 (const re_match_context_t *mctx,
int boundaries, Idx subexp_idx,
- Idx from_node, Idx bkref_idx) internal_function;
+ Idx from_node, Idx bkref_idx)
+ internal_function;
static int check_dst_limits_calc_pos (const re_match_context_t *mctx,
Idx limit, Idx subexp_idx,
Idx node, Idx str_idx,
Idx bkref_idx) internal_function;
-static reg_errcode_t check_subexp_limits (re_dfa_t *dfa,
+static reg_errcode_t check_subexp_limits (const re_dfa_t *dfa,
re_node_set *dest_nodes,
const re_node_set *candidates,
re_node_set *limits,
struct re_backref_cache_entry *bkref_ents,
Idx str_idx) internal_function;
-static reg_errcode_t sift_states_bkref (re_match_context_t *mctx,
+static reg_errcode_t sift_states_bkref (const re_match_context_t *mctx,
re_sift_context_t *sctx,
- Idx str_idx, const re_node_set *candidates) internal_function;
-static reg_errcode_t merge_state_array (re_dfa_t *dfa, re_dfastate_t **dst,
- re_dfastate_t **src, Idx num) internal_function;
+ Idx str_idx, const re_node_set *candidates)
+ internal_function;
+static reg_errcode_t merge_state_array (const re_dfa_t *dfa,
+ re_dfastate_t **dst,
+ re_dfastate_t **src, Idx num)
+ internal_function;
static re_dfastate_t *find_recover_state (reg_errcode_t *err,
re_match_context_t *mctx) internal_function;
static re_dfastate_t *transit_state (reg_errcode_t *err,
re_dfastate_t *state) internal_function;
static re_dfastate_t *merge_state_with_log (reg_errcode_t *err,
re_match_context_t *mctx,
- re_dfastate_t *next_state) internal_function;
+ re_dfastate_t *next_state)
+ internal_function;
static reg_errcode_t check_subexp_matching_top (re_match_context_t *mctx,
re_node_set *cur_nodes,
Idx str_idx) internal_function;
#if 0
static re_dfastate_t *transit_state_sb (reg_errcode_t *err,
re_match_context_t *mctx,
- re_dfastate_t *pstate) internal_function;
+ re_dfastate_t *pstate)
+ internal_function;
#endif
#ifdef RE_ENABLE_I18N
static reg_errcode_t transit_state_mb (re_match_context_t *mctx,
- re_dfastate_t *pstate) internal_function;
+ re_dfastate_t *pstate)
+ internal_function;
#endif /* RE_ENABLE_I18N */
static reg_errcode_t transit_state_bkref (re_match_context_t *mctx,
- const re_node_set *nodes) internal_function;
+ const re_node_set *nodes)
+ internal_function;
static reg_errcode_t get_subexp (re_match_context_t *mctx,
- Idx bkref_node, Idx bkref_str_idx) internal_function;
+ Idx bkref_node, Idx bkref_str_idx)
+ internal_function;
static reg_errcode_t get_subexp_sub (re_match_context_t *mctx,
const re_sub_match_top_t *sub_top,
re_sub_match_last_t *sub_last,
- Idx bkref_node, Idx bkref_str) internal_function;
+ Idx bkref_node, Idx bkref_str)
+ internal_function;
static Idx find_subexp_node (const re_dfa_t *dfa, const re_node_set *nodes,
Idx subexp_idx, int type) internal_function;
static reg_errcode_t check_arrival (re_match_context_t *mctx,
static reg_errcode_t check_arrival_add_next_nodes (re_match_context_t *mctx,
Idx str_idx,
re_node_set *cur_nodes,
- re_node_set *next_nodes) internal_function;
-static reg_errcode_t check_arrival_expand_ecl (re_dfa_t *dfa,
+ re_node_set *next_nodes)
+ internal_function;
+static reg_errcode_t check_arrival_expand_ecl (const re_dfa_t *dfa,
re_node_set *cur_nodes,
- Idx ex_subexp, int type) internal_function;
-static reg_errcode_t check_arrival_expand_ecl_sub (re_dfa_t *dfa,
+ Idx ex_subexp, int type)
+ internal_function;
+static reg_errcode_t check_arrival_expand_ecl_sub (const re_dfa_t *dfa,
re_node_set *dst_nodes,
Idx target, Idx ex_subexp,
int type) internal_function;
static reg_errcode_t expand_bkref_cache (re_match_context_t *mctx,
re_node_set *cur_nodes, Idx cur_str,
- Idx subexp_num, int type) internal_function;
-static bool build_trtable (re_dfa_t *dfa,
+ Idx subexp_num, int type)
+ internal_function;
+static bool build_trtable (const re_dfa_t *dfa,
re_dfastate_t *state) internal_function;
#ifdef RE_ENABLE_I18N
-static int check_node_accept_bytes (re_dfa_t *dfa, Idx node_idx,
- const re_string_t *input, Idx idx) internal_function;
+static int check_node_accept_bytes (const re_dfa_t *dfa, Idx node_idx,
+ const re_string_t *input, Idx idx)
+ internal_function;
# ifdef _LIBC
static unsigned int find_collation_sequence_value (const unsigned char *mbs,
- size_t name_len) internal_function;
+ size_t name_len)
+ internal_function;
# endif /* _LIBC */
#endif /* RE_ENABLE_I18N */
static Idx group_nodes_into_DFAstates (const re_dfa_t *dfa,
const re_dfastate_t *state,
re_node_set *states_node,
- bitset *states_ch) internal_function;
+ bitset_t *states_ch) internal_function;
static bool check_node_accept (const re_match_context_t *mctx,
const re_token_t *node, Idx idx)
internal_function;
-static reg_errcode_t extend_buffers (re_match_context_t *mctx) internal_function;
+static reg_errcode_t extend_buffers (re_match_context_t *mctx, int min_len)
+ internal_function;
\f
/* Entry point for POSIX code. */
string STRING.
If NMATCH is zero or REG_NOSUB was set in the cflags argument to
- `regcomp', we ignore PMATCH. Otherwise, we assume PMATCH has at
+ 'regcomp', we ignore PMATCH. Otherwise, we assume PMATCH has at
least NMATCH elements, and we set them to the offsets of the
corresponding matched substrings.
- EFLAGS specifies `execution flags' which affect matching: if
+ EFLAGS specifies "execution flags" which affect matching: if
REG_NOTBOL is set, then ^ does not match at the beginning of the
string; if REG_NOTEOL is set, then $ does not match at the end.
We return 0 if we find a match and REG_NOMATCH if not. */
int
-regexec (const regex_t *__restrict preg, const char *__restrict string,
- size_t nmatch, regmatch_t pmatch[], int eflags)
+regexec (preg, string, nmatch, pmatch, eflags)
+ const regex_t *_Restrict_ preg;
+ const char *_Restrict_ string;
+ size_t nmatch;
+ regmatch_t pmatch[_Restrict_arr_];
+ int eflags;
{
reg_errcode_t err;
Idx start, length;
-#ifdef _LIBC
- re_dfa_t *dfa = (re_dfa_t *) preg->re_buffer;
-#endif
+ re_dfa_t *dfa = preg->buffer;
if (eflags & ~(REG_NOTBOL | REG_NOTEOL | REG_STARTEND))
return REG_BADPAT;
length = strlen (string);
}
- __libc_lock_lock (dfa->lock);
- if (preg->re_no_sub)
+ lock_lock (dfa->lock);
+ if (preg->no_sub)
err = re_search_internal (preg, string, length, start, length,
length, 0, NULL, eflags);
else
err = re_search_internal (preg, string, length, start, length,
length, nmatch, pmatch, eflags);
- __libc_lock_unlock (dfa->lock);
+ lock_unlock (dfa->lock);
return err != REG_NOERROR;
}
int
attribute_compat_text_section
-__compat_regexec (const regex_t *__restrict preg,
- const char *__restrict string, size_t nmatch,
+__compat_regexec (const regex_t *_Restrict_ preg,
+ const char *_Restrict_ string, size_t nmatch,
regmatch_t pmatch[], int eflags)
{
return regexec (preg, string, nmatch, pmatch,
the first STOP characters of the concatenation of the strings should be
concerned.
- If REGS is not NULL, and BUFP->re_no_sub is not set, the offsets of the match
- and all groups is stroed in REGS. (For the "_2" variants, the offsets are
+ If REGS is not NULL, and BUFP->no_sub is not set, the offsets of the match
+ and all groups is stored in REGS. (For the "_2" variants, the offsets are
computed relative to the concatenation, not relative to the individual
strings.)
match was found and -2 indicates an internal error. */
regoff_t
-re_match (struct re_pattern_buffer *bufp, const char *string,
- Idx length, Idx start, struct re_registers *regs)
+re_match (bufp, string, length, start, regs)
+ struct re_pattern_buffer *bufp;
+ const char *string;
+ Idx length, start;
+ struct re_registers *regs;
{
return re_search_stub (bufp, string, length, start, 0, length, regs, true);
}
#endif
regoff_t
-re_search (struct re_pattern_buffer *bufp, const char *string,
- Idx length, Idx start, regoff_t range, struct re_registers *regs)
+re_search (bufp, string, length, start, range, regs)
+ struct re_pattern_buffer *bufp;
+ const char *string;
+ Idx length, start;
+ regoff_t range;
+ struct re_registers *regs;
{
return re_search_stub (bufp, string, length, start, range, length, regs,
false);
#endif
regoff_t
-re_match_2 (struct re_pattern_buffer *bufp,
- const char *string1, Idx length1,
- const char *string2, Idx length2,
- Idx start, struct re_registers *regs, Idx stop)
+re_match_2 (bufp, string1, length1, string2, length2, start, regs, stop)
+ struct re_pattern_buffer *bufp;
+ const char *string1, *string2;
+ Idx length1, length2, start, stop;
+ struct re_registers *regs;
{
return re_search_2_stub (bufp, string1, length1, string2, length2,
start, 0, regs, stop, true);
#endif
regoff_t
-re_search_2 (struct re_pattern_buffer *bufp,
- const char *string1, Idx length1,
- const char *string2, Idx length2,
- Idx start, regoff_t range, struct re_registers *regs, Idx stop)
+re_search_2 (bufp, string1, length1, string2, length2, start, range, regs, stop)
+ struct re_pattern_buffer *bufp;
+ const char *string1, *string2;
+ Idx length1, length2, start, stop;
+ regoff_t range;
+ struct re_registers *regs;
{
return re_search_2_stub (bufp, string1, length1, string2, length2,
start, range, regs, stop, false);
#endif
static regoff_t
-internal_function
re_search_2_stub (struct re_pattern_buffer *bufp,
const char *string1, Idx length1,
const char *string2, Idx length2,
if (BE (s == NULL, 0))
return -2;
+#ifdef _LIBC
+ memcpy (__mempcpy (s, string1, length1), string2, length2);
+#else
memcpy (s, string1, length1);
memcpy (s + length1, string2, length2);
+#endif
str = s;
}
else
otherwise the position of the match is returned. */
static regoff_t
-internal_function
re_search_stub (struct re_pattern_buffer *bufp,
const char *string, Idx length,
Idx start, regoff_t range, Idx stop, struct re_registers *regs,
Idx nregs;
regoff_t rval;
int eflags = 0;
-#ifdef _LIBC
- re_dfa_t *dfa = (re_dfa_t *) bufp->re_buffer;
-#endif
+ re_dfa_t *dfa = bufp->buffer;
Idx last_start = start + range;
/* Check for out-of-range. */
if (BE (start < 0 || start > length, 0))
return -1;
- if (sizeof start < sizeof range)
- {
- regoff_t length_offset = length;
- regoff_t start_offset = start;
- if (BE (length_offset - start_offset < range, 0))
- last_start = length;
- else if (BE (range < - start_offset, 0))
- last_start = 0;
- }
- else
- {
- if (BE ((last_start < start) != (range < 0), 0))
- {
- /* Overflow occurred when computing last_start; substitute
- the extreme value. */
- last_start = range < 0 ? 0 : length;
- }
- else
- {
- if (BE (length < last_start, 0))
- last_start = length;
- else if (BE (last_start < 0, 0))
- last_start = 0;
- }
- }
+ if (BE (length < last_start || (0 <= range && last_start < start), 0))
+ last_start = length;
+ else if (BE (last_start < 0 || (range < 0 && start <= last_start), 0))
+ last_start = 0;
- __libc_lock_lock (dfa->lock);
+ lock_lock (dfa->lock);
- eflags |= (bufp->re_not_bol) ? REG_NOTBOL : 0;
- eflags |= (bufp->re_not_eol) ? REG_NOTEOL : 0;
+ eflags |= (bufp->not_bol) ? REG_NOTBOL : 0;
+ eflags |= (bufp->not_eol) ? REG_NOTEOL : 0;
/* Compile fastmap if we haven't yet. */
- if (start < last_start && bufp->re_fastmap != NULL
- && !bufp->re_fastmap_accurate)
+ if (start < last_start && bufp->fastmap != NULL && !bufp->fastmap_accurate)
re_compile_fastmap (bufp);
- if (BE (bufp->re_no_sub, 0))
+ if (BE (bufp->no_sub, 0))
regs = NULL;
/* We need at least 1 register. */
if (regs == NULL)
nregs = 1;
- else if (BE (bufp->re_regs_allocated == REG_FIXED
- && regs->rm_num_regs <= bufp->re_nsub, 0))
+ else if (BE (bufp->regs_allocated == REGS_FIXED
+ && regs->num_regs <= bufp->re_nsub, 0))
{
- nregs = regs->rm_num_regs;
+ nregs = regs->num_regs;
if (BE (nregs < 1, 0))
{
/* Nothing can be copied to regs. */
}
else
nregs = bufp->re_nsub + 1;
- pmatch = re_xmalloc (regmatch_t, nregs);
+ pmatch = re_malloc (regmatch_t, nregs);
if (BE (pmatch == NULL, 0))
{
rval = -2;
rval = 0;
- /* I hope we needn't fill ther regs with -1's when no match was found. */
+ /* I hope we needn't fill their regs with -1's when no match was found. */
if (result != REG_NOERROR)
- rval = -1;
+ rval = result == REG_NOMATCH ? -1 : -2;
else if (regs != NULL)
{
/* If caller wants register contents data back, copy them. */
- bufp->re_regs_allocated = re_copy_regs (regs, pmatch, nregs,
- bufp->re_regs_allocated);
- if (BE (bufp->re_regs_allocated == REG_UNALLOCATED, 0))
+ bufp->regs_allocated = re_copy_regs (regs, pmatch, nregs,
+ bufp->regs_allocated);
+ if (BE (bufp->regs_allocated == REGS_UNALLOCATED, 0))
rval = -2;
}
}
re_free (pmatch);
out:
- __libc_lock_unlock (dfa->lock);
+ lock_unlock (dfa->lock);
return rval;
}
static unsigned
-internal_function
re_copy_regs (struct re_registers *regs, regmatch_t *pmatch, Idx nregs,
int regs_allocated)
{
- int rval = REG_REALLOCATE;
+ int rval = REGS_REALLOCATE;
Idx i;
Idx need_regs = nregs + 1;
- /* We need one extra element beyond `rm_num_regs' for the `-1' marker GNU code
+ /* We need one extra element beyond 'num_regs' for the '-1' marker GNU code
uses. */
/* Have the register data arrays been allocated? */
- if (regs_allocated == REG_UNALLOCATED)
+ if (regs_allocated == REGS_UNALLOCATED)
{ /* No. So allocate them with malloc. */
- regs->rm_start = re_xmalloc (regoff_t, need_regs);
- regs->rm_end = re_malloc (regoff_t, need_regs);
- if (BE (regs->rm_start == NULL, 0) || BE (regs->rm_end == NULL, 0))
- return REG_UNALLOCATED;
- regs->rm_num_regs = need_regs;
+ regs->start = re_malloc (regoff_t, need_regs);
+ if (BE (regs->start == NULL, 0))
+ return REGS_UNALLOCATED;
+ regs->end = re_malloc (regoff_t, need_regs);
+ if (BE (regs->end == NULL, 0))
+ {
+ re_free (regs->start);
+ return REGS_UNALLOCATED;
+ }
+ regs->num_regs = need_regs;
}
- else if (regs_allocated == REG_REALLOCATE)
+ else if (regs_allocated == REGS_REALLOCATE)
{ /* Yes. If we need more elements than were already
allocated, reallocate them. If we need fewer, just
leave it alone. */
- if (BE (need_regs > regs->rm_num_regs, 0))
+ if (BE (need_regs > regs->num_regs, 0))
{
- regoff_t *new_start =
- re_xrealloc (regs->rm_start, regoff_t, need_regs);
- regoff_t *new_end = re_realloc (regs->rm_end, regoff_t, need_regs);
- if (BE (new_start == NULL, 0) || BE (new_end == NULL, 0))
- return REG_UNALLOCATED;
- regs->rm_start = new_start;
- regs->rm_end = new_end;
- regs->rm_num_regs = need_regs;
+ regoff_t *new_start = re_realloc (regs->start, regoff_t, need_regs);
+ regoff_t *new_end;
+ if (BE (new_start == NULL, 0))
+ return REGS_UNALLOCATED;
+ new_end = re_realloc (regs->end, regoff_t, need_regs);
+ if (BE (new_end == NULL, 0))
+ {
+ re_free (new_start);
+ return REGS_UNALLOCATED;
+ }
+ regs->start = new_start;
+ regs->end = new_end;
+ regs->num_regs = need_regs;
}
}
else
{
- assert (regs_allocated == REG_FIXED);
- /* This function may not be called with REG_FIXED and nregs too big. */
- assert (regs->rm_num_regs >= nregs);
- rval = REG_FIXED;
+ assert (regs_allocated == REGS_FIXED);
+ /* This function may not be called with REGS_FIXED and nregs too big. */
+ assert (regs->num_regs >= nregs);
+ rval = REGS_FIXED;
}
/* Copy the regs. */
for (i = 0; i < nregs; ++i)
{
- regs->rm_start[i] = pmatch[i].rm_so;
- regs->rm_end[i] = pmatch[i].rm_eo;
+ regs->start[i] = pmatch[i].rm_so;
+ regs->end[i] = pmatch[i].rm_eo;
}
- for ( ; i < regs->rm_num_regs; ++i)
- regs->rm_start[i] = regs->rm_end[i] = -1;
+ for ( ; i < regs->num_regs; ++i)
+ regs->start[i] = regs->end[i] = -1;
return rval;
}
freeing the old data. */
void
-re_set_registers (struct re_pattern_buffer *bufp, struct re_registers *regs,
- __re_size_t num_regs, regoff_t *starts, regoff_t *ends)
+re_set_registers (bufp, regs, num_regs, starts, ends)
+ struct re_pattern_buffer *bufp;
+ struct re_registers *regs;
+ __re_size_t num_regs;
+ regoff_t *starts, *ends;
{
if (num_regs)
{
- bufp->re_regs_allocated = REG_REALLOCATE;
- regs->rm_num_regs = num_regs;
- regs->rm_start = starts;
- regs->rm_end = ends;
+ bufp->regs_allocated = REGS_REALLOCATE;
+ regs->num_regs = num_regs;
+ regs->start = starts;
+ regs->end = ends;
}
else
{
- bufp->re_regs_allocated = REG_UNALLOCATED;
- regs->rm_num_regs = 0;
- regs->rm_start = regs->rm_end = NULL;
+ bufp->regs_allocated = REGS_UNALLOCATED;
+ regs->num_regs = 0;
+ regs->start = regs->end = NULL;
}
}
#ifdef _LIBC
# ifdef _LIBC
weak_function
# endif
-re_exec (const char *s)
+re_exec (s)
+ const char *s;
{
return 0 == regexec (&re_comp_buf, s, 0, NULL, 0);
}
(0 <= LAST_START && LAST_START <= LENGTH) */
static reg_errcode_t
-internal_function
+__attribute_warn_unused_result__
re_search_internal (const regex_t *preg,
const char *string, Idx length,
Idx start, Idx last_start, Idx stop,
int eflags)
{
reg_errcode_t err;
- re_dfa_t *dfa = (re_dfa_t *) preg->re_buffer;
+ const re_dfa_t *dfa = preg->buffer;
Idx left_lim, right_lim;
int incr;
bool fl_longest_match;
int match_kind;
- Idx match_first, match_last = REG_MISSING;
+ Idx match_first;
+ Idx match_last = REG_MISSING;
Idx extra_nmatch;
bool sb;
int ch;
#else
re_match_context_t mctx;
#endif
- char *fastmap = ((preg->re_fastmap != NULL && preg->re_fastmap_accurate
- && start != last_start && !preg->re_can_be_null)
- ? preg->re_fastmap : NULL);
- unsigned REG_TRANSLATE_TYPE t =
- (unsigned REG_TRANSLATE_TYPE) preg->re_translate;
+ char *fastmap = ((preg->fastmap != NULL && preg->fastmap_accurate
+ && start != last_start && !preg->can_be_null)
+ ? preg->fastmap : NULL);
+ RE_TRANSLATE_TYPE t = preg->translate;
#if !(defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L))
memset (&mctx, '\0', sizeof (re_match_context_t));
nmatch -= extra_nmatch;
/* Check if the DFA haven't been compiled. */
- if (BE (preg->re_used == 0 || dfa->init_state == NULL
+ if (BE (preg->used == 0 || dfa->init_state == NULL
|| dfa->init_state_word == NULL || dfa->init_state_nl == NULL
|| dfa->init_state_begbuf == NULL, 0))
return REG_NOMATCH;
#endif
/* If initial states with non-begbuf contexts have no elements,
- the regex must be anchored. If preg->re_newline_anchor is set,
+ the regex must be anchored. If preg->newline_anchor is set,
we'll never use init_state_nl, so do not check it. */
if (dfa->init_state->nodes.nelem == 0
&& dfa->init_state_word->nodes.nelem == 0
&& (dfa->init_state_nl->nodes.nelem == 0
- || !preg->re_newline_anchor))
+ || !preg->newline_anchor))
{
if (start != 0 && last_start != 0)
return REG_NOMATCH;
fl_longest_match = (nmatch != 0 || dfa->nbackref);
err = re_string_allocate (&mctx.input, string, length, dfa->nodes_len + 1,
- preg->re_translate,
- preg->re_syntax & REG_IGNORE_CASE, dfa);
+ preg->translate, (preg->syntax & RE_ICASE) != 0,
+ dfa);
if (BE (err != REG_NOERROR, 0))
goto free_return;
mctx.input.stop = stop;
mctx.input.raw_stop = stop;
- mctx.input.newline_anchor = preg->re_newline_anchor;
+ mctx.input.newline_anchor = preg->newline_anchor;
err = match_ctx_init (&mctx, eflags, dfa->nbackref * 2);
if (BE (err != REG_NOERROR, 0))
multi character collating element. */
if (nmatch > 1 || dfa->has_mb_node)
{
- mctx.state_log = re_xmalloc (re_dfastate_t *, mctx.input.bufs_len + 1);
+ /* Avoid overflow. */
+ if (BE ((MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *))
+ <= mctx.input.bufs_len), 0))
+ {
+ err = REG_ESPACE;
+ goto free_return;
+ }
+
+ mctx.state_log = re_malloc (re_dfastate_t *, mctx.input.bufs_len + 1);
if (BE (mctx.state_log == NULL, 0))
{
err = REG_ESPACE;
mctx.input.tip_context = (eflags & REG_NOTBOL) ? CONTEXT_BEGBUF
: CONTEXT_NEWLINE | CONTEXT_BEGBUF;
- /* Check incrementally whether of not the input string match. */
+ /* Check incrementally whether the input string matches. */
incr = (last_start < start) ? -1 : 1;
left_lim = (last_start < start) ? last_start : start;
right_lim = (last_start < start) ? start : last_start;
sb = dfa->mb_cur_max == 1;
match_kind =
(fastmap
- ? ((sb || !(preg->re_syntax & REG_IGNORE_CASE || t) ? 4 : 0)
+ ? ((sb || !(preg->syntax & RE_ICASE || t) ? 4 : 0)
| (start <= last_start ? 2 : 0)
| (t != NULL ? 1 : 0))
: 8);
break;
match_first += incr;
if (match_first < left_lim || match_first > right_lim)
- {
- err = REG_NOMATCH;
- goto free_return;
- }
+ {
+ err = REG_NOMATCH;
+ goto free_return;
+ }
}
break;
}
else
{
mctx.match_last = match_last;
- if ((!preg->re_no_sub && nmatch > 1) || dfa->nbackref)
+ if ((!preg->no_sub && nmatch > 1) || dfa->nbackref)
{
re_dfastate_t *pstate = mctx.state_log[match_last];
mctx.last_node = check_halt_state_context (&mctx, pstate,
match_last);
}
- if ((!preg->re_no_sub && nmatch > 1 && dfa->has_plural_match)
+ if ((!preg->no_sub && nmatch > 1 && dfa->has_plural_match)
|| dfa->nbackref)
{
err = prune_impossible_nodes (&mctx);
the maximum possible regoff_t value. We need a new error
code REG_OVERFLOW. */
- if (!preg->re_no_sub && nmatch > 1)
+ if (!preg->no_sub && nmatch > 1)
{
err = set_regs (preg, &mctx, nmatch, pmatch,
dfa->has_plural_match && dfa->nbackref > 0);
goto free_return;
}
- /* At last, add the offset to the each registers, since we slided
+ /* At last, add the offset to each register, since we slid
the buffers so that we could assume that the matching starts
from 0. */
for (reg_idx = 0; reg_idx < nmatch; ++reg_idx)
}
if (dfa->subexp_map)
- 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;
+ }
}
free_return:
}
static reg_errcode_t
-internal_function
+__attribute_warn_unused_result__
prune_impossible_nodes (re_match_context_t *mctx)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
Idx halt_node, match_last;
reg_errcode_t ret;
re_dfastate_t **sifted_states;
#endif
match_last = mctx->match_last;
halt_node = mctx->last_node;
- sifted_states = re_xmalloc (re_dfastate_t *, match_last + 1);
+
+ /* Avoid overflow. */
+ if (BE (MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *)) <= match_last, 0))
+ return REG_ESPACE;
+
+ sifted_states = re_malloc (re_dfastate_t *, match_last + 1);
if (BE (sifted_states == NULL, 0))
{
ret = REG_ESPACE;
}
if (dfa->nbackref)
{
- lim_states = re_xmalloc (re_dfastate_t *, match_last + 1);
+ lim_states = re_malloc (re_dfastate_t *, match_last + 1);
if (BE (lim_states == NULL, 0))
{
ret = REG_ESPACE;
re_node_set_free (&sctx.limits);
if (BE (ret != REG_NOERROR, 0))
goto free_return;
+ if (sifted_states[0] == NULL)
+ {
+ ret = REG_NOMATCH;
+ goto free_return;
+ }
}
re_free (mctx->state_log);
mctx->state_log = sifted_states;
since initial states may have constraints like "\<", "^", etc.. */
static inline re_dfastate_t *
-__attribute ((always_inline)) internal_function
+__attribute__ ((always_inline)) internal_function
acquire_init_state_context (reg_errcode_t *err, const re_match_context_t *mctx,
Idx idx)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
if (dfa->init_state->has_constraint)
{
unsigned int context;
FL_LONGEST_MATCH means we want the POSIX longest matching.
If P_MATCH_FIRST is not NULL, and the match fails, it is set to the
next place where we may want to try matching.
- Note that the matcher assume that the maching starts from the current
+ Note that the matcher assumes that the matching starts from the current
index of the buffer. */
static Idx
-internal_function
+internal_function __attribute_warn_unused_result__
check_matching (re_match_context_t *mctx, bool fl_longest_match,
Idx *p_match_first)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
reg_errcode_t err;
Idx match = 0;
Idx match_last = REG_MISSING;
{
err = transit_state_bkref (mctx, &cur_state->nodes);
if (BE (err != REG_NOERROR, 0))
- return err;
+ return err;
}
}
}
re_dfastate_t *old_state = cur_state;
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))
+ if ((BE (next_char_idx >= mctx->input.bufs_len, 0)
+ && mctx->input.bufs_len < mctx->input.len)
+ || (BE (next_char_idx >= mctx->input.valid_len, 0)
+ && mctx->input.valid_len < mctx->input.len))
+ {
+ err = extend_buffers (mctx, next_char_idx + 1);
+ if (BE (err != REG_NOERROR, 0))
{
assert (err == REG_ESPACE);
return REG_ERROR;
}
- }
+ }
cur_state = transit_state (&err, mctx, cur_state);
if (mctx->state_log != NULL)
static Idx
internal_function
-proceed_next_node (const re_match_context_t *mctx,
- Idx nregs, regmatch_t *regs, Idx *pidx, Idx node,
- re_node_set *eps_via_nodes, struct re_fail_stack_t *fs)
+proceed_next_node (const re_match_context_t *mctx, Idx nregs, regmatch_t *regs,
+ Idx *pidx, Idx node, re_node_set *eps_via_nodes,
+ struct re_fail_stack_t *fs)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
Idx i;
bool ok;
if (IS_EPSILON_NODE (dfa->nodes[node].type))
if (dest_node == REG_MISSING)
dest_node = candidate;
- else
+ else
{
/* In order to avoid infinite loop like "(a*)*", return the second
- epsilon-transition if the first was already considered. */
+ epsilon-transition if the first was already considered. */
if (re_node_set_contains (eps_via_nodes, dest_node))
- return candidate;
+ return candidate;
/* Otherwise, push the second epsilon-transition on the fail stack. */
else if (fs != NULL
&& push_fail_stack (fs, *pidx, candidate, nregs, regs,
- eps_via_nodes))
+ eps_via_nodes))
return REG_ERROR;
/* We know we are going to exit. */
}
static reg_errcode_t
-internal_function
+internal_function __attribute_warn_unused_result__
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)
{
Idx num = fs->num++;
if (fs->num == fs->alloc)
{
- struct re_fail_stack_ent_t *new_array =
- re_x2realloc (fs->stack, struct re_fail_stack_ent_t, &fs->alloc);
+ struct re_fail_stack_ent_t *new_array;
+ new_array = realloc (fs->stack, (sizeof (struct re_fail_stack_ent_t)
+ * fs->alloc * 2));
if (new_array == NULL)
return REG_ESPACE;
+ fs->alloc *= 2;
fs->stack = new_array;
}
fs->stack[num].idx = str_idx;
fs->stack[num].node = dest_node;
- fs->stack[num].regs = re_xmalloc (regmatch_t, nregs);
+ fs->stack[num].regs = re_malloc (regmatch_t, nregs);
if (fs->stack[num].regs == NULL)
return REG_ESPACE;
memcpy (fs->stack[num].regs, regs, sizeof (regmatch_t) * nregs);
static Idx
internal_function
-pop_fail_stack (struct re_fail_stack_t *fs, Idx *pidx,
- Idx nregs, regmatch_t *regs, re_node_set *eps_via_nodes)
+pop_fail_stack (struct re_fail_stack_t *fs, Idx *pidx, Idx nregs,
+ regmatch_t *regs, re_node_set *eps_via_nodes)
{
Idx num = --fs->num;
assert (REG_VALID_INDEX (num));
pmatch[i].rm_so == pmatch[i].rm_eo == -1 for 0 < i < nmatch. */
static reg_errcode_t
-internal_function
-set_regs (const regex_t *preg, const re_match_context_t *mctx,
- size_t nmatch, regmatch_t *pmatch, bool fl_backtrack)
+internal_function __attribute_warn_unused_result__
+set_regs (const regex_t *preg, const re_match_context_t *mctx, size_t nmatch,
+ regmatch_t *pmatch, bool fl_backtrack)
{
- re_dfa_t *dfa = (re_dfa_t *) preg->re_buffer;
+ const re_dfa_t *dfa = preg->buffer;
Idx idx, cur_node;
re_node_set eps_via_nodes;
struct re_fail_stack_t *fs;
if (fl_backtrack)
{
fs = &fs_body;
- fs->stack = re_xmalloc (struct re_fail_stack_ent_t, fs->alloc);
+ fs->stack = re_malloc (struct re_fail_stack_ent_t, fs->alloc);
if (fs->stack == NULL)
return REG_ESPACE;
}
cur_node = dfa->init_node;
re_node_set_init_empty (&eps_via_nodes);
- if (re_alloc_oversized (nmatch, sizeof (regmatch_t)))
- {
- free_fail_stack_return (fs);
- return REG_ESPACE;
- }
if (__libc_use_alloca (nmatch * sizeof (regmatch_t)))
prev_idx_match = (regmatch_t *) alloca (nmatch * sizeof (regmatch_t));
else
static void
internal_function
-update_regs (re_dfa_t *dfa, regmatch_t *pmatch, regmatch_t *prev_idx_match,
- Idx cur_node, Idx cur_idx, Idx nmatch)
+update_regs (const re_dfa_t *dfa, regmatch_t *pmatch,
+ regmatch_t *prev_idx_match, Idx cur_node, Idx cur_idx, Idx nmatch)
{
int type = dfa->nodes[cur_node].type;
if (type == OP_OPEN_SUBEXP)
and sift the nodes in each states according to the following rules.
Updated state_log will be wrote to STATE_LOG.
- Rules: We throw away the Node `a' in the STATE_LOG[STR_IDX] if...
+ Rules: We throw away the Node 'a' in the STATE_LOG[STR_IDX] if...
1. When STR_IDX == MATCH_LAST(the last index in the state_log):
- If `a' isn't the LAST_NODE and `a' can't epsilon transit to
- the LAST_NODE, we throw away the node `a'.
- 2. When 0 <= STR_IDX < MATCH_LAST and `a' accepts
- string `s' and transit to `b':
+ If 'a' isn't the LAST_NODE and 'a' can't epsilon transit to
+ the LAST_NODE, we throw away the node 'a'.
+ 2. When 0 <= STR_IDX < MATCH_LAST and 'a' accepts
+ string 's' and transit to 'b':
i. If 'b' isn't in the STATE_LOG[STR_IDX+strlen('s')], we throw
- away the node `a'.
+ away the node 'a'.
ii. If 'b' is in the STATE_LOG[STR_IDX+strlen('s')] but 'b' is
- thrown away, we throw away the node `a'.
+ thrown away, we throw away the node 'a'.
3. When 0 <= STR_IDX < MATCH_LAST and 'a' epsilon transit to 'b':
i. If 'b' isn't in the STATE_LOG[STR_IDX], we throw away the
- node `a'.
+ node 'a'.
ii. If 'b' is in the STATE_LOG[STR_IDX] but 'b' is thrown away,
- we throw away the node `a'. */
+ we throw away the node 'a'. */
#define STATE_NODE_CONTAINS(state,node) \
((state) != NULL && re_node_set_contains (&(state)->nodes, node))
static reg_errcode_t
internal_function
-sift_states_backward (re_match_context_t *mctx, re_sift_context_t *sctx)
+sift_states_backward (const re_match_context_t *mctx, re_sift_context_t *sctx)
{
reg_errcode_t err;
int null_cnt = 0;
if (mctx->state_log[str_idx])
{
err = build_sifted_states (mctx, sctx, str_idx, &cur_dest);
- if (BE (err != REG_NOERROR, 0))
+ if (BE (err != REG_NOERROR, 0))
goto free_return;
}
}
static reg_errcode_t
-internal_function
-build_sifted_states (re_match_context_t *mctx, re_sift_context_t *sctx,
+internal_function __attribute_warn_unused_result__
+build_sifted_states (const re_match_context_t *mctx, re_sift_context_t *sctx,
Idx str_idx, re_node_set *cur_dest)
{
- re_dfa_t *const dfa = mctx->dfa;
- re_node_set *cur_src = &mctx->state_log[str_idx]->non_eps_nodes;
+ const re_dfa_t *const dfa = mctx->dfa;
+ const re_node_set *cur_src = &mctx->state_log[str_idx]->non_eps_nodes;
Idx i;
/* Then build the next sifted state.
- We build the next sifted state on `cur_dest', and update
- `sifted_states[str_idx]' with `cur_dest'.
+ We build the next sifted state on 'cur_dest', and update
+ 'sifted_states[str_idx]' with 'cur_dest'.
Note:
- `cur_dest' is the sifted state from `state_log[str_idx + 1]'.
- `cur_src' points the node_set of the old `state_log[str_idx]'
+ 'cur_dest' is the sifted state from 'state_log[str_idx + 1]'.
+ 'cur_src' points the node_set of the old 'state_log[str_idx]'
(with the epsilon nodes pre-filtered out). */
for (i = 0; i < cur_src->nelem; i++)
{
assert (!IS_EPSILON_NODE (type));
#endif
#ifdef RE_ENABLE_I18N
- /* If the node may accept `multi byte'. */
+ /* If the node may accept "multi byte". */
if (dfa->nodes[prev_node].accept_mb)
naccepted = sift_states_iter_mb (mctx, sctx, prev_node,
str_idx, sctx->last_str_idx);
{
Idx top = mctx->state_log_top;
- if (next_state_log_idx >= mctx->input.bufs_len
+ if ((next_state_log_idx >= mctx->input.bufs_len
+ && mctx->input.bufs_len < mctx->input.len)
|| (next_state_log_idx >= mctx->input.valid_len
&& mctx->input.valid_len < mctx->input.len))
{
reg_errcode_t err;
- err = extend_buffers (mctx);
+ err = extend_buffers (mctx, next_state_log_idx + 1);
if (BE (err != REG_NOERROR, 0))
return err;
}
static reg_errcode_t
internal_function
-merge_state_array (re_dfa_t *dfa, re_dfastate_t **dst, re_dfastate_t **src,
- Idx num)
+merge_state_array (const re_dfa_t *dfa, re_dfastate_t **dst,
+ re_dfastate_t **src, Idx num)
{
Idx st_idx;
reg_errcode_t err;
static reg_errcode_t
internal_function
-update_cur_sifted_state (re_match_context_t *mctx, re_sift_context_t *sctx,
- Idx str_idx, re_node_set *dest_nodes)
+update_cur_sifted_state (const re_match_context_t *mctx,
+ re_sift_context_t *sctx, Idx str_idx,
+ re_node_set *dest_nodes)
{
- re_dfa_t *const dfa = mctx->dfa;
- reg_errcode_t err;
+ const re_dfa_t *const dfa = mctx->dfa;
+ reg_errcode_t err = REG_NOERROR;
const re_node_set *candidates;
candidates = ((mctx->state_log[str_idx] == NULL) ? NULL
: &mctx->state_log[str_idx]->nodes);
}
static reg_errcode_t
-internal_function
-add_epsilon_src_nodes (re_dfa_t *dfa, re_node_set *dest_nodes,
+internal_function __attribute_warn_unused_result__
+add_epsilon_src_nodes (const re_dfa_t *dfa, re_node_set *dest_nodes,
const re_node_set *candidates)
{
reg_errcode_t err = REG_NOERROR;
{
err = re_node_set_alloc (&state->inveclosure, dest_nodes->nelem);
if (BE (err != REG_NOERROR, 0))
- return REG_ESPACE;
+ return REG_ESPACE;
for (i = 0; i < dest_nodes->nelem; i++)
- 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;
+ }
}
return re_node_set_add_intersect (dest_nodes, candidates,
&state->inveclosure);
static reg_errcode_t
internal_function
-sub_epsilon_src_nodes (re_dfa_t *dfa, Idx node, re_node_set *dest_nodes,
+sub_epsilon_src_nodes (const re_dfa_t *dfa, Idx node, re_node_set *dest_nodes,
const re_node_set *candidates)
{
Idx ecl_idx;
check_dst_limits (const re_match_context_t *mctx, const re_node_set *limits,
Idx dst_node, Idx dst_idx, Idx src_node, Idx src_idx)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
Idx lim_idx, src_pos, dst_pos;
Idx dst_bkref_idx = search_cur_bkref_entry (mctx, dst_idx);
check_dst_limits_calc_pos_1 (const re_match_context_t *mctx, int boundaries,
Idx subexp_idx, Idx from_node, Idx bkref_idx)
{
- re_dfa_t *const dfa = mctx->dfa;
- re_node_set *eclosures = dfa->eclosures + from_node;
+ const re_dfa_t *const dfa = mctx->dfa;
+ const re_node_set *eclosures = dfa->eclosures + from_node;
Idx node_idx;
/* Else, we are on the boundary: examine the nodes on the epsilon
{
struct re_backref_cache_entry *ent = mctx->bkref_ents + bkref_idx;
do
- {
+ {
Idx dst;
int cpos;
if (ent->node != node)
continue;
- if (subexp_idx
- < CHAR_BIT * sizeof ent->eps_reachable_subexps_map
- && !(ent->eps_reachable_subexps_map & (1u << subexp_idx)))
+ if (subexp_idx < BITSET_WORD_BITS
+ && !(ent->eps_reachable_subexps_map
+ & ((bitset_word_t) 1 << subexp_idx)))
continue;
/* Recurse trying to reach the OP_OPEN_SUBEXP and
if (dst == from_node)
{
if (boundaries & 1)
- return -1;
+ return -1;
else /* if (boundaries & 2) */
- return 0;
+ return 0;
}
cpos =
if (cpos == 0 && (boundaries & 2))
return 0;
- if (subexp_idx
- < CHAR_BIT * sizeof ent->eps_reachable_subexps_map)
- ent->eps_reachable_subexps_map &= ~(1u << subexp_idx);
- }
+ if (subexp_idx < BITSET_WORD_BITS)
+ ent->eps_reachable_subexps_map
+ &= ~((bitset_word_t) 1 << subexp_idx);
+ }
while (ent++->more);
}
break;
static int
internal_function
-check_dst_limits_calc_pos (const re_match_context_t *mctx,
- Idx limit, Idx subexp_idx,
- Idx from_node, Idx str_idx, Idx bkref_idx)
+check_dst_limits_calc_pos (const re_match_context_t *mctx, Idx limit,
+ Idx subexp_idx, Idx from_node, Idx str_idx,
+ Idx bkref_idx)
{
struct re_backref_cache_entry *lim = mctx->bkref_ents + limit;
int boundaries;
static reg_errcode_t
internal_function
-check_subexp_limits (re_dfa_t *dfa, re_node_set *dest_nodes,
+check_subexp_limits (const re_dfa_t *dfa, re_node_set *dest_nodes,
const re_node_set *candidates, re_node_set *limits,
struct re_backref_cache_entry *bkref_ents, Idx str_idx)
{
}
static reg_errcode_t
-internal_function
-sift_states_bkref (re_match_context_t *mctx, re_sift_context_t *sctx,
+internal_function __attribute_warn_unused_result__
+sift_states_bkref (const re_match_context_t *mctx, re_sift_context_t *sctx,
Idx str_idx, const re_node_set *candidates)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
reg_errcode_t err;
Idx node_idx, node;
re_sift_context_t local_sctx;
enabled_idx = first_idx;
do
{
+ Idx subexp_len;
+ Idx to_idx;
+ Idx dst_node;
bool ok;
- Idx subexp_len, to_idx, dst_node;
re_dfastate_t *cur_state;
if (entry->node != node)
re_node_set_remove (&local_sctx.limits, enabled_idx);
/* mctx->bkref_ents may have changed, reload the pointer. */
- entry = mctx->bkref_ents + enabled_idx;
+ entry = mctx->bkref_ents + enabled_idx;
}
while (enabled_idx++, entry++->more);
}
sift_states_iter_mb (const re_match_context_t *mctx, re_sift_context_t *sctx,
Idx node_idx, Idx str_idx, Idx max_str_idx)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
int naccepted;
- /* Check the node can accept `multi byte'. */
+ /* Check the node can accept "multi byte". */
naccepted = check_node_accept_bytes (dfa, node_idx, &mctx->input, str_idx);
if (naccepted > 0 && str_idx + naccepted <= max_str_idx &&
!STATE_NODE_CONTAINS (sctx->sifted_states[str_idx + naccepted],
dfa->nexts[node_idx]))
- /* The node can't accept the `multi byte', or the
+ /* The node can't accept the "multi byte", or the
destination was already thrown away, then the node
- could't accept the current input `multi byte'. */
+ could't accept the current input "multi byte". */
naccepted = 0;
/* Otherwise, it is sure that the node could accept
- `naccepted' bytes input. */
+ 'naccepted' bytes input. */
return naccepted;
}
#endif /* RE_ENABLE_I18N */
update the destination of STATE_LOG. */
static re_dfastate_t *
-internal_function
+internal_function __attribute_warn_unused_result__
transit_state (reg_errcode_t *err, re_match_context_t *mctx,
re_dfastate_t *state)
{
trtable = state->word_trtable;
if (BE (trtable != NULL, 1))
- {
+ {
unsigned int context;
context
= re_string_context_at (&mctx->input,
}
/* Update the state_log if we need */
-re_dfastate_t *
+static re_dfastate_t *
internal_function
merge_state_with_log (reg_errcode_t *err, re_match_context_t *mctx,
re_dfastate_t *next_state)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
Idx cur_idx = re_string_cur_idx (&mctx->input);
if (cur_idx > mctx->state_log_top)
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. */
pstate = mctx->state_log[cur_idx];
log_nodes = pstate->entrance_nodes;
if (next_state != NULL)
- {
- 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,
log_nodes);
- if (BE (*err != REG_NOERROR, 0))
+ if (BE (*err != REG_NOERROR, 0))
return NULL;
- }
+ }
else
- next_nodes = *log_nodes;
+ next_nodes = *log_nodes;
/* 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_acquire_state_context (err, dfa, &next_nodes, context);
+ = re_acquire_state_context (err, dfa, &next_nodes, context);
/* We don't need to check errors here, since the return value of
- this function is next_state and ERR is already set. */
+ this function is next_state and ERR is already set. */
if (table_nodes != NULL)
- re_node_set_free (&next_nodes);
+ re_node_set_free (&next_nodes);
}
if (BE (dfa->nbackref, 0) && next_state != NULL)
internal_function
find_recover_state (reg_errcode_t *err, re_match_context_t *mctx)
{
- re_dfastate_t *cur_state = NULL;
+ re_dfastate_t *cur_state;
do
{
Idx max = mctx->state_log_top;
do
{
- 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);
}
while (mctx->state_log[cur_str_idx] == NULL);
/* From the node set CUR_NODES, pick up the nodes whose types are
OP_OPEN_SUBEXP and which have corresponding back references in the regular
expression. And register them to use them later for evaluating the
- correspoding back references. */
+ corresponding back references. */
static reg_errcode_t
internal_function
check_subexp_matching_top (re_match_context_t *mctx, re_node_set *cur_nodes,
Idx str_idx)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
Idx node_idx;
reg_errcode_t err;
{
Idx node = cur_nodes->elems[node_idx];
if (dfa->nodes[node].type == OP_OPEN_SUBEXP
- && dfa->nodes[node].opr.idx < CHAR_BIT * sizeof dfa->used_bkref_map
- && dfa->used_bkref_map & (1u << dfa->nodes[node].opr.idx))
+ && dfa->nodes[node].opr.idx < BITSET_WORD_BITS
+ && (dfa->used_bkref_map
+ & ((bitset_word_t) 1 << dfa->nodes[node].opr.idx)))
{
err = match_ctx_add_subtop (mctx, node, str_idx);
if (BE (err != REG_NOERROR, 0))
transit_state_sb (reg_errcode_t *err, re_match_context_t *mctx,
re_dfastate_t *state)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
re_node_set next_nodes;
re_dfastate_t *next_state;
Idx node_cnt, cur_str_idx = re_string_cur_idx (&mctx->input);
internal_function
transit_state_mb (re_match_context_t *mctx, re_dfastate_t *pstate)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
reg_errcode_t err;
Idx i;
re_dfastate_t *dest_state;
if (!dfa->nodes[cur_node_idx].accept_mb)
- continue;
+ continue;
if (dfa->nodes[cur_node_idx].constraint)
{
if (naccepted == 0)
continue;
- /* The node can accepts `naccepted' bytes. */
+ /* The node can accepts 'naccepted' bytes. */
dest_idx = re_string_cur_idx (&mctx->input) + naccepted;
mctx->max_mb_elem_len = ((mctx->max_mb_elem_len < naccepted) ? naccepted
: mctx->max_mb_elem_len);
if (BE (err != REG_NOERROR, 0))
return err;
}
- context = re_string_context_at (&mctx->input, dest_idx - 1, mctx->eflags);
+ context = re_string_context_at (&mctx->input, dest_idx - 1,
+ mctx->eflags);
mctx->state_log[dest_idx]
= re_acquire_state_context (&err, dfa, &dest_nodes, context);
if (dest_state != NULL)
internal_function
transit_state_bkref (re_match_context_t *mctx, const re_node_set *nodes)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
reg_errcode_t err;
Idx i;
Idx cur_str_idx = re_string_cur_idx (&mctx->input);
const re_token_t *node = dfa->nodes + node_idx;
re_node_set *new_dest_nodes;
- /* Check whether `node' is a backreference or not. */
+ /* Check whether 'node' is a backreference or not. */
if (node->type != OP_BACK_REF)
continue;
continue;
}
- /* `node' is a backreference.
+ /* 'node' is a backreference.
Check the substring which the substring matched. */
bkc_idx = mctx->nbkref_ents;
err = get_subexp (mctx, node_idx, cur_str_idx);
if (BE (err != REG_NOERROR, 0))
goto free_return;
- /* And add the epsilon closures (which is `new_dest_nodes') of
+ /* And add the epsilon closures (which is 'new_dest_nodes') of
the backreference to appropriate state_log. */
#ifdef DEBUG
assert (dfa->nexts[node_idx] != REG_MISSING);
dest_state = mctx->state_log[dest_str_idx];
prev_nelem = ((mctx->state_log[cur_str_idx] == NULL) ? 0
: mctx->state_log[cur_str_idx]->nodes.nelem);
- /* Add `new_dest_node' to state_log. */
+ /* Add 'new_dest_node' to state_log. */
if (dest_state == NULL)
{
mctx->state_log[dest_str_idx]
delay these checking for prune_impossible_nodes(). */
static reg_errcode_t
-internal_function
+internal_function __attribute_warn_unused_result__
get_subexp (re_match_context_t *mctx, Idx bkref_node, Idx bkref_str_idx)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
Idx subexp_num, sub_top_idx;
const char *buf = (const char *) re_string_get_buffer (&mctx->input);
/* Return if we have already checked BKREF_NODE at BKREF_STR_IDX. */
Idx cache_idx = search_cur_bkref_entry (mctx, bkref_str_idx);
if (cache_idx != REG_MISSING)
{
- const struct re_backref_cache_entry *entry = mctx->bkref_ents + cache_idx;
+ const struct re_backref_cache_entry *entry
+ = mctx->bkref_ents + cache_idx;
do
- if (entry->node == bkref_node)
+ if (entry->node == bkref_node)
return REG_NOERROR; /* We already checked it. */
while (entry++->more);
}
buf = (const char *) re_string_get_buffer (&mctx->input);
}
if (memcmp (buf + bkref_str_off, buf + sl_str, sl_str_diff) != 0)
- break; /* We don't need to search this sub expression any more. */
+ /* We don't need to search this sub expression any more. */
+ break;
}
bkref_str_off += sl_str_diff;
sl_str += sl_str_diff;
if (bkref_str_off >= mctx->input.len)
break;
- err = extend_buffers (mctx);
+ err = extend_buffers (mctx, bkref_str_off + 1);
if (BE (err != REG_NOERROR, 0))
return err;
continue;
/* Does this state have a ')' of the sub expression? */
nodes = &mctx->state_log[sl_str]->nodes;
- cls_node = find_subexp_node (dfa, nodes, subexp_num, OP_CLOSE_SUBEXP);
+ cls_node = find_subexp_node (dfa, nodes, subexp_num,
+ OP_CLOSE_SUBEXP);
if (cls_node == REG_MISSING)
continue; /* No. */
if (sub_top->path == NULL)
{
- sub_top->path = re_calloc (state_array_t,
- sl_str - sub_top->str_idx + 1);
+ sub_top->path = calloc (sizeof (state_array_t),
+ sl_str - sub_top->str_idx + 1);
if (sub_top->path == NULL)
return REG_ESPACE;
}
/* Can the OP_OPEN_SUBEXP node arrive the OP_CLOSE_SUBEXP node
in the current context? */
err = check_arrival (mctx, sub_top->path, sub_top->node,
- sub_top->str_idx, cls_node, sl_str, OP_CLOSE_SUBEXP);
+ sub_top->str_idx, cls_node, sl_str,
+ OP_CLOSE_SUBEXP);
if (err == REG_NOMATCH)
continue;
if (BE (err != REG_NOERROR, 0))
Idx to_idx;
/* Can the subexpression arrive the back reference? */
err = check_arrival (mctx, &sub_last->path, sub_last->node,
- sub_last->str_idx, bkref_node, bkref_str, OP_OPEN_SUBEXP);
+ sub_last->str_idx, bkref_node, bkref_str,
+ OP_OPEN_SUBEXP);
if (err != REG_NOERROR)
return err;
err = match_ctx_add_entry (mctx, bkref_node, bkref_str, sub_top->str_idx,
Return REG_NOERROR if it can arrive, or REG_NOMATCH otherwise. */
static reg_errcode_t
-internal_function
-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)
+internal_function __attribute_warn_unused_result__
+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)
{
- re_dfa_t *const dfa = mctx->dfa;
- reg_errcode_t err;
+ const re_dfa_t *const dfa = mctx->dfa;
+ reg_errcode_t err = REG_NOERROR;
Idx subexp_num, backup_cur_idx, str_idx, null_cnt;
re_dfastate_t *cur_state = NULL;
re_node_set *cur_nodes, next_nodes;
{
re_dfastate_t **new_array;
Idx old_alloc = path->alloc;
- Idx new_alloc = old_alloc + last_str + mctx->max_mb_elem_len + 1;
- if (BE (new_alloc < old_alloc, 0))
+ Idx incr_alloc = last_str + mctx->max_mb_elem_len + 1;
+ Idx new_alloc;
+ if (BE (IDX_MAX - old_alloc < incr_alloc, 0))
return REG_ESPACE;
- new_array = re_xrealloc (path->array, re_dfastate_t *, new_alloc);
+ new_alloc = old_alloc + incr_alloc;
+ if (BE (SIZE_MAX / sizeof (re_dfastate_t *) < new_alloc, 0))
+ return REG_ESPACE;
+ new_array = re_realloc (path->array, re_dfastate_t *, new_alloc);
if (BE (new_array == NULL, 0))
return REG_ESPACE;
path->array = new_array;
path->alloc = new_alloc;
memset (new_array + old_alloc, '\0',
- sizeof (re_dfastate_t *) * (new_alloc - old_alloc));
+ sizeof (re_dfastate_t *) * (path->alloc - old_alloc));
}
- str_idx = path->next_idx == 0 ? top_str : path->next_idx;
+ str_idx = path->next_idx ? path->next_idx : top_str;
/* Temporary modify MCTX. */
backup_state_log = mctx->state_log;
if (cur_state && cur_state->has_backref)
{
err = re_node_set_init_copy (&next_nodes, &cur_state->nodes);
- if (BE ( err != REG_NOERROR, 0))
+ if (BE (err != REG_NOERROR, 0))
return err;
}
else
{
err = expand_bkref_cache (mctx, &next_nodes, str_idx,
subexp_num, type);
- if (BE ( err != REG_NOERROR, 0))
+ if (BE (err != REG_NOERROR, 0))
{
re_node_set_free (&next_nodes);
return err;
if (cur_state)
{
err = check_arrival_add_next_nodes (mctx, str_idx,
- &cur_state->non_eps_nodes, &next_nodes);
+ &cur_state->non_eps_nodes,
+ &next_nodes);
if (BE (err != REG_NOERROR, 0))
{
re_node_set_free (&next_nodes);
}
err = expand_bkref_cache (mctx, &next_nodes, str_idx,
subexp_num, type);
- if (BE ( err != REG_NOERROR, 0))
+ if (BE (err != REG_NOERROR, 0))
{
re_node_set_free (&next_nodes);
return err;
Can't we unify them? */
static reg_errcode_t
-internal_function
+internal_function __attribute_warn_unused_result__
check_arrival_add_next_nodes (re_match_context_t *mctx, Idx str_idx,
- re_node_set *cur_nodes,
- re_node_set *next_nodes)
+ re_node_set *cur_nodes, re_node_set *next_nodes)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
bool ok;
Idx cur_idx;
- reg_errcode_t err;
+#ifdef RE_ENABLE_I18N
+ reg_errcode_t err = REG_NOERROR;
+#endif
re_node_set union_set;
re_node_set_init_empty (&union_set);
for (cur_idx = 0; cur_idx < cur_nodes->nelem; ++cur_idx)
assert (!IS_EPSILON_NODE (type));
#endif
#ifdef RE_ENABLE_I18N
- /* If the node may accept `multi byte'. */
+ /* If the node may accept "multi byte". */
if (dfa->nodes[cur_node].accept_mb)
{
naccepted = check_node_accept_bytes (dfa, cur_node, &mctx->input,
static reg_errcode_t
internal_function
-check_arrival_expand_ecl (re_dfa_t *dfa, re_node_set *cur_nodes,
+check_arrival_expand_ecl (const re_dfa_t *dfa, re_node_set *cur_nodes,
Idx ex_subexp, int type)
{
reg_errcode_t err;
for (idx = 0; idx < cur_nodes->nelem; ++idx)
{
Idx cur_node = cur_nodes->elems[idx];
- re_node_set *eclosure = dfa->eclosures + cur_node;
+ const re_node_set *eclosure = dfa->eclosures + cur_node;
outside_node = find_subexp_node (dfa, eclosure, ex_subexp, type);
if (outside_node == REG_MISSING)
{
problematic append it to DST_NODES. */
static reg_errcode_t
-internal_function
-check_arrival_expand_ecl_sub (re_dfa_t *dfa, re_node_set *dst_nodes,
+internal_function __attribute_warn_unused_result__
+check_arrival_expand_ecl_sub (const re_dfa_t *dfa, re_node_set *dst_nodes,
Idx target, Idx ex_subexp, int type)
{
Idx cur_node;
break;
if (dfa->edests[cur_node].nelem == 2)
{
- reg_errcode_t ret =
- check_arrival_expand_ecl_sub (dfa, dst_nodes,
- dfa->edests[cur_node].elems[1],
- ex_subexp, type);
- if (BE (ret != REG_NOERROR, 0))
- return ret;
+ reg_errcode_t err;
+ err = check_arrival_expand_ecl_sub (dfa, dst_nodes,
+ dfa->edests[cur_node].elems[1],
+ ex_subexp, type);
+ if (BE (err != REG_NOERROR, 0))
+ return err;
}
cur_node = dfa->edests[cur_node].elems[0];
}
in MCTX->BKREF_ENTS. */
static reg_errcode_t
-internal_function
+internal_function __attribute_warn_unused_result__
expand_bkref_cache (re_match_context_t *mctx, re_node_set *cur_nodes,
Idx cur_str, Idx subexp_num, int type)
{
- re_dfa_t *const dfa = mctx->dfa;
+ const re_dfa_t *const dfa = mctx->dfa;
reg_errcode_t err;
Idx cache_idx_start = search_cur_bkref_entry (mctx, cur_str);
struct re_backref_cache_entry *ent;
static bool
internal_function
-build_trtable (re_dfa_t *dfa, re_dfastate_t *state)
+build_trtable (const re_dfa_t *dfa, re_dfastate_t *state)
{
reg_errcode_t err;
Idx i, j;
int ch;
bool need_word_trtable = false;
- unsigned int elem, mask;
- bool dests_node_malloced = false, dest_states_malloced = false;
- Idx ndests; /* Number of the destination states from `state'. */
+ bitset_word_t elem, mask;
+ bool dests_node_malloced = false;
+ bool dest_states_malloced = false;
+ Idx ndests; /* Number of the destination states from 'state'. */
re_dfastate_t **trtable;
re_dfastate_t **dest_states = NULL, **dest_states_word, **dest_states_nl;
re_node_set follows, *dests_node;
- bitset *dests_ch;
- bitset acceptable;
+ bitset_t *dests_ch;
+ bitset_t acceptable;
+
+ struct dests_alloc
+ {
+ re_node_set dests_node[SBC_MAX];
+ bitset_t dests_ch[SBC_MAX];
+ } *dests_alloc;
/* We build DFA states which corresponds to the destination nodes
- from `state'. `dests_node[i]' represents the nodes which i-th
- destination state contains, and `dests_ch[i]' represents the
+ from 'state'. 'dests_node[i]' represents the nodes which i-th
+ destination state contains, and 'dests_ch[i]' represents the
characters which i-th destination state accepts. */
- if (__libc_use_alloca ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX))
- dests_node = (re_node_set *)
- alloca ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX);
+ if (__libc_use_alloca (sizeof (struct dests_alloc)))
+ dests_alloc = (struct dests_alloc *) alloca (sizeof (struct dests_alloc));
else
{
- dests_node = (re_node_set *)
- malloc ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX);
- if (BE (dests_node == NULL, 0))
+ dests_alloc = re_malloc (struct dests_alloc, 1);
+ if (BE (dests_alloc == NULL, 0))
return false;
dests_node_malloced = true;
}
- dests_ch = (bitset *) (dests_node + SBC_MAX);
+ dests_node = dests_alloc->dests_node;
+ dests_ch = dests_alloc->dests_ch;
- /* Initialize transiton table. */
+ /* Initialize transition table. */
state->word_trtable = state->trtable = NULL;
- /* At first, group all nodes belonging to `state' into several
+ /* At first, group all nodes belonging to 'state' into several
destinations. */
ndests = group_nodes_into_DFAstates (dfa, state, dests_node, dests_ch);
if (BE (! REG_VALID_NONZERO_INDEX (ndests), 0))
{
if (dests_node_malloced)
- free (dests_node);
+ free (dests_alloc);
+ /* Return false in case of an error, true otherwise. */
if (ndests == 0)
{
- state->trtable = re_calloc (re_dfastate_t *, SBC_MAX);
+ state->trtable = (re_dfastate_t **)
+ calloc (sizeof (re_dfastate_t *), SBC_MAX);
+ if (BE (state->trtable == NULL, 0))
+ return false;
return true;
}
return false;
goto out_free;
/* Avoid arithmetic overflow in size calculation. */
- if (BE (((SIZE_MAX - (sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX)
- / (3 * sizeof (re_dfastate_t *)))
- < ndests, 0))
+ if (BE ((((SIZE_MAX - (sizeof (re_node_set) + sizeof (bitset_t)) * SBC_MAX)
+ / (3 * sizeof (re_dfastate_t *)))
+ < ndests),
+ 0))
goto out_free;
- if (__libc_use_alloca ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX
+ if (__libc_use_alloca ((sizeof (re_node_set) + sizeof (bitset_t)) * SBC_MAX
+ ndests * 3 * sizeof (re_dfastate_t *)))
dest_states = (re_dfastate_t **)
alloca (ndests * 3 * sizeof (re_dfastate_t *));
for (i = 0; i < ndests; ++i)
re_node_set_free (dests_node + i);
if (dests_node_malloced)
- free (dests_node);
+ free (dests_alloc);
return false;
}
dest_states_malloced = true;
CONTEXT_NEWLINE);
if (BE (dest_states_nl[i] == NULL && err != REG_NOERROR, 0))
goto out_free;
- }
+ }
else
{
dest_states_word[i] = dest_states[i];
character, or we are in a single-byte character set so we can
discern by looking at the character code: allocate a
256-entry transition table. */
- trtable = state->trtable = re_calloc (re_dfastate_t *, SBC_MAX);
+ trtable = state->trtable =
+ (re_dfastate_t **) calloc (sizeof (re_dfastate_t *), SBC_MAX);
if (BE (trtable == NULL, 0))
goto out_free;
/* For all characters ch...: */
- for (i = 0; i < BITSET_UINTS; ++i)
- for (ch = i * UINT_BITS, elem = acceptable[i], mask = 1;
+ for (i = 0; i < BITSET_WORDS; ++i)
+ for (ch = i * BITSET_WORD_BITS, elem = acceptable[i], mask = 1;
elem;
mask <<= 1, elem >>= 1, ++ch)
if (BE (elem & 1, 0))
by looking at the character code: build two 256-entry
transition tables, one starting at trtable[0] and one
starting at trtable[SBC_MAX]. */
- trtable = state->word_trtable = re_calloc (re_dfastate_t *, 2 * SBC_MAX);
+ trtable = state->word_trtable =
+ (re_dfastate_t **) calloc (sizeof (re_dfastate_t *), 2 * SBC_MAX);
if (BE (trtable == NULL, 0))
goto out_free;
/* For all characters ch...: */
- for (i = 0; i < BITSET_UINTS; ++i)
- for (ch = i * UINT_BITS, elem = acceptable[i], mask = 1;
+ for (i = 0; i < BITSET_WORDS; ++i)
+ for (ch = i * BITSET_WORD_BITS, elem = acceptable[i], mask = 1;
elem;
mask <<= 1, elem >>= 1, ++ch)
if (BE (elem & 1, 0))
re_node_set_free (dests_node + i);
if (dests_node_malloced)
- free (dests_node);
+ free (dests_alloc);
return true;
}
static Idx
internal_function
group_nodes_into_DFAstates (const re_dfa_t *dfa, const re_dfastate_t *state,
- re_node_set *dests_node, bitset *dests_ch)
+ re_node_set *dests_node, bitset_t *dests_ch)
{
reg_errcode_t err;
bool ok;
Idx i, j, k;
- Idx ndests; /* Number of the destinations from `state'. */
- bitset accepts; /* Characters a node can accept. */
+ Idx ndests; /* Number of the destinations from 'state'. */
+ bitset_t accepts; /* Characters a node can accept. */
const re_node_set *cur_nodes = &state->nodes;
bitset_empty (accepts);
ndests = 0;
- /* For all the nodes belonging to `state', */
+ /* For all the nodes belonging to 'state', */
for (i = 0; i < cur_nodes->nelem; ++i)
{
re_token_t *node = &dfa->nodes[cur_nodes->elems[i]];
else
#endif
bitset_set_all (accepts);
- if (!(dfa->syntax & REG_DOT_NEWLINE))
+ if (!(dfa->syntax & RE_DOT_NEWLINE))
bitset_clear (accepts, '\n');
- if (dfa->syntax & REG_DOT_NOT_NULL)
+ if (dfa->syntax & RE_DOT_NOT_NULL)
bitset_clear (accepts, '\0');
}
#ifdef RE_ENABLE_I18N
else if (type == OP_UTF8_PERIOD)
- {
- memset (accepts, 255, sizeof (unsigned int) * BITSET_UINTS / 2);
- if (!(dfa->syntax & REG_DOT_NEWLINE))
+ {
+ if (ASCII_CHARS % BITSET_WORD_BITS == 0)
+ memset (accepts, -1, ASCII_CHARS / CHAR_BIT);
+ else
+ bitset_merge (accepts, utf8_sb_map);
+ if (!(dfa->syntax & RE_DOT_NEWLINE))
bitset_clear (accepts, '\n');
- if (dfa->syntax & REG_DOT_NOT_NULL)
+ if (dfa->syntax & RE_DOT_NOT_NULL)
bitset_clear (accepts, '\0');
- }
+ }
#endif
else
continue;
- /* Check the `accepts' and sift the characters which are not
+ /* Check the 'accepts' and sift the characters which are not
match it the context. */
if (constraint)
{
if (constraint & NEXT_NEWLINE_CONSTRAINT)
{
- unsigned int accepts_newline =
- bitset_contain (accepts, NEWLINE_CHAR);
+ bool accepts_newline = bitset_contain (accepts, NEWLINE_CHAR);
bitset_empty (accepts);
if (accepts_newline)
bitset_set (accepts, NEWLINE_CHAR);
if (constraint & NEXT_WORD_CONSTRAINT)
{
- unsigned int any_set = 0;
+ bitset_word_t any_set = 0;
if (type == CHARACTER && !node->word_char)
{
bitset_empty (accepts);
}
#ifdef RE_ENABLE_I18N
if (dfa->mb_cur_max > 1)
- for (j = 0; j < BITSET_UINTS; ++j)
+ for (j = 0; j < BITSET_WORDS; ++j)
any_set |= (accepts[j] &= (dfa->word_char[j] | ~dfa->sb_char[j]));
else
#endif
- for (j = 0; j < BITSET_UINTS; ++j)
+ for (j = 0; j < BITSET_WORDS; ++j)
any_set |= (accepts[j] &= dfa->word_char[j]);
if (!any_set)
continue;
}
if (constraint & NEXT_NOTWORD_CONSTRAINT)
{
- unsigned int any_set = 0;
+ bitset_word_t any_set = 0;
if (type == CHARACTER && node->word_char)
{
bitset_empty (accepts);
}
#ifdef RE_ENABLE_I18N
if (dfa->mb_cur_max > 1)
- for (j = 0; j < BITSET_UINTS; ++j)
+ for (j = 0; j < BITSET_WORDS; ++j)
any_set |= (accepts[j] &= ~(dfa->word_char[j] & dfa->sb_char[j]));
else
#endif
- for (j = 0; j < BITSET_UINTS; ++j)
+ for (j = 0; j < BITSET_WORDS; ++j)
any_set |= (accepts[j] &= ~dfa->word_char[j]);
if (!any_set)
continue;
}
}
- /* Then divide `accepts' into DFA states, or create a new
+ /* Then divide 'accepts' into DFA states, or create a new
state. Above, we make sure that accepts is not empty. */
for (j = 0; j < ndests; ++j)
{
- bitset intersec; /* Intersection sets, see below. */
- bitset remains;
+ bitset_t intersec; /* Intersection sets, see below. */
+ bitset_t remains;
/* Flags, see below. */
- unsigned int has_intersec, not_subset, not_consumed;
+ bitset_word_t has_intersec, not_subset, not_consumed;
/* Optimization, skip if this state doesn't accept the character. */
if (type == CHARACTER && !bitset_contain (dests_ch[j], node->opr.c))
continue;
- /* Enumerate the intersection set of this state and `accepts'. */
+ /* Enumerate the intersection set of this state and 'accepts'. */
has_intersec = 0;
- for (k = 0; k < BITSET_UINTS; ++k)
+ for (k = 0; k < BITSET_WORDS; ++k)
has_intersec |= intersec[k] = accepts[k] & dests_ch[j][k];
/* And skip if the intersection set is empty. */
if (!has_intersec)
continue;
- /* Then check if this state is a subset of `accepts'. */
+ /* Then check if this state is a subset of 'accepts'. */
not_subset = not_consumed = 0;
- for (k = 0; k < BITSET_UINTS; ++k)
+ for (k = 0; k < BITSET_WORDS; ++k)
{
not_subset |= remains[k] = ~accepts[k] & dests_ch[j][k];
not_consumed |= accepts[k] = accepts[k] & ~dests_ch[j][k];
}
- /* If this state isn't a subset of `accepts', create a
- new group state, which has the `remains'. */
+ /* If this state isn't a subset of 'accepts', create a
+ new group state, which has the 'remains'. */
if (not_subset)
{
bitset_copy (dests_ch[ndests], remains);
}
#ifdef RE_ENABLE_I18N
-/* Check how many bytes the node `dfa->nodes[node_idx]' accepts.
+/* Check how many bytes the node 'dfa->nodes[node_idx]' accepts.
Return the number of the bytes the node accepts.
STR_IDX is the current index of the input string.
static int
internal_function
-check_node_accept_bytes (re_dfa_t *dfa, Idx node_idx,
+check_node_accept_bytes (const re_dfa_t *dfa, Idx node_idx,
const re_string_t *input, Idx str_idx)
{
const re_token_t *node = dfa->nodes + node_idx;
if (node->type == OP_PERIOD)
{
if (char_len <= 1)
- return 0;
+ return 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. */
- if ((!(dfa->syntax & REG_DOT_NEWLINE) &&
+ if ((!(dfa->syntax & RE_DOT_NEWLINE) &&
re_string_byte_at (input, str_idx) == '\n') ||
- ((dfa->syntax & REG_DOT_NOT_NULL) &&
+ ((dfa->syntax & RE_DOT_NOT_NULL) &&
re_string_byte_at (input, str_idx) == '\0'))
return 0;
return char_len;
const int32_t *table, *indirect;
const unsigned char *weights, *extra;
const char *collseqwc;
- int32_t idx;
/* This #include defines a local function! */
# include <locale/weight.h>
in_collseq = find_collation_sequence_value (pin, elem_len);
}
/* match with range expression? */
+ /* FIXME: Implement rational ranges here, too. */
for (i = 0; i < cset->nranges; ++i)
if (cset->range_starts[i] <= in_collseq
&& in_collseq <= cset->range_ends[i])
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
indirect = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
- idx = findidx (&cp);
+ int32_t idx = findidx (&cp, elem_len);
if (idx > 0)
for (i = 0; i < cset->nequiv_classes; ++i)
{
int32_t equiv_class_idx = cset->equiv_classes[i];
- 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))
{
Idx cnt = 0;
+
+ idx &= 0xffffff;
+ equiv_class_idx &= 0xffffff;
+
while (cnt <= weight_len
&& (weights[equiv_class_idx + 1 + cnt]
== weights[idx + 1 + cnt]))
# endif /* _LIBC */
{
/* match with range expression? */
-#if __GNUC__ >= 2
- wchar_t cmp_buf[] = {L'\0', L'\0', wc, L'\0', L'\0', L'\0'};
-#else
- wchar_t cmp_buf[] = {L'\0', L'\0', L'\0', L'\0', L'\0', L'\0'};
- cmp_buf[2] = wc;
-#endif
for (i = 0; i < cset->nranges; ++i)
{
- cmp_buf[0] = cset->range_starts[i];
- cmp_buf[4] = cset->range_ends[i];
- if (wcscoll (cmp_buf, cmp_buf + 2) <= 0
- && wcscoll (cmp_buf + 2, cmp_buf + 4) <= 0)
+ if (cset->range_starts[i] <= wc && wc <= cset->range_ends[i])
{
match_len = char_len;
goto check_node_accept_bytes_match;
# ifdef _LIBC
static unsigned int
+internal_function
find_collation_sequence_value (const unsigned char *mbs, size_t mbs_len)
{
uint32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
/* Skip the collation sequence value. */
idx += sizeof (uint32_t);
/* Skip the wide char sequence of the collating element. */
- idx = idx + sizeof (uint32_t) * (extra[idx] + 1);
+ idx = idx + sizeof (uint32_t) * (*(int32_t *) (extra + idx) + 1);
/* If we found the entry, return the sequence value. */
if (found)
return *(uint32_t *) (extra + idx);
#ifdef RE_ENABLE_I18N
case OP_UTF8_PERIOD:
- if (ch >= 0x80)
+ if (ch >= ASCII_CHARS)
return false;
/* FALLTHROUGH */
#endif
case OP_PERIOD:
- if ((ch == '\n' && !(mctx->dfa->syntax & REG_DOT_NEWLINE))
- || (ch == '\0' && (mctx->dfa->syntax & REG_DOT_NOT_NULL)))
+ if ((ch == '\n' && !(mctx->dfa->syntax & RE_DOT_NEWLINE))
+ || (ch == '\0' && (mctx->dfa->syntax & RE_DOT_NOT_NULL)))
return false;
break;
/* Extend the buffers, if the buffers have run out. */
static reg_errcode_t
-internal_function
-extend_buffers (re_match_context_t *mctx)
+internal_function __attribute_warn_unused_result__
+extend_buffers (re_match_context_t *mctx, int min_len)
{
reg_errcode_t ret;
re_string_t *pstr = &mctx->input;
- /* Double the lengthes of the buffers. */
- ret = re_string_realloc_buffers (pstr, pstr->bufs_len * 2);
+ /* Avoid overflow. */
+ if (BE (MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *)) / 2
+ <= pstr->bufs_len, 0))
+ return REG_ESPACE;
+
+ /* Double the lengths of the buffers, but allocate at least MIN_LEN. */
+ ret = re_string_realloc_buffers (pstr,
+ MAX (min_len,
+ MIN (pstr->len, pstr->bufs_len * 2)));
if (BE (ret != REG_NOERROR, 0))
return ret;
/* XXX We have no indication of the size of this buffer. If this
allocation fail we have no indication that the state_log array
does not have the right size. */
- re_dfastate_t **new_array = re_xrealloc (mctx->state_log, re_dfastate_t *,
- pstr->bufs_len + 1);
+ re_dfastate_t **new_array = re_realloc (mctx->state_log, re_dfastate_t *,
+ pstr->bufs_len + 1);
if (BE (new_array == NULL, 0))
return REG_ESPACE;
mctx->state_log = new_array;
/* Initialize MCTX. */
static reg_errcode_t
-internal_function
+internal_function __attribute_warn_unused_result__
match_ctx_init (re_match_context_t *mctx, int eflags, Idx n)
{
mctx->eflags = eflags;
mctx->match_last = REG_MISSING;
if (n > 0)
{
- mctx->bkref_ents = re_xmalloc (struct re_backref_cache_entry, n);
- mctx->sub_tops = re_xmalloc (re_sub_match_top_t *, n);
+ /* Avoid overflow. */
+ size_t max_object_size =
+ MAX (sizeof (struct re_backref_cache_entry),
+ sizeof (re_sub_match_top_t *));
+ if (BE (MIN (IDX_MAX, SIZE_MAX / max_object_size) < n, 0))
+ return REG_ESPACE;
+
+ mctx->bkref_ents = re_malloc (struct re_backref_cache_entry, n);
+ mctx->sub_tops = re_malloc (re_sub_match_top_t *, n);
if (BE (mctx->bkref_ents == NULL || mctx->sub_tops == NULL, 0))
return REG_ESPACE;
}
*/
static reg_errcode_t
-internal_function
-match_ctx_add_entry (re_match_context_t *mctx, Idx node, Idx str_idx,
- Idx from, Idx to)
+internal_function __attribute_warn_unused_result__
+match_ctx_add_entry (re_match_context_t *mctx, Idx node, Idx str_idx, Idx from,
+ Idx to)
{
if (mctx->nbkref_ents >= mctx->abkref_ents)
{
struct re_backref_cache_entry* new_entry;
- new_entry = re_x2realloc (mctx->bkref_ents, struct re_backref_cache_entry,
- &mctx->abkref_ents);
+ new_entry = re_realloc (mctx->bkref_ents, struct re_backref_cache_entry,
+ mctx->abkref_ents * 2);
if (BE (new_entry == NULL, 0))
{
re_free (mctx->bkref_ents);
}
mctx->bkref_ents = new_entry;
memset (mctx->bkref_ents + mctx->nbkref_ents, '\0',
- (sizeof (struct re_backref_cache_entry)
- * (mctx->abkref_ents - mctx->nbkref_ents)));
+ sizeof (struct re_backref_cache_entry) * mctx->abkref_ents);
+ mctx->abkref_ents *= 2;
}
if (mctx->nbkref_ents > 0
&& mctx->bkref_ents[mctx->nbkref_ents - 1].str_idx == str_idx)
at STR_IDX. */
static reg_errcode_t
-internal_function
+internal_function __attribute_warn_unused_result__
match_ctx_add_subtop (re_match_context_t *mctx, Idx node, Idx str_idx)
{
#ifdef DEBUG
#endif
if (BE (mctx->nsub_tops == mctx->asub_tops, 0))
{
- Idx new_asub_tops = mctx->asub_tops;
- re_sub_match_top_t **new_array = re_x2realloc (mctx->sub_tops,
- re_sub_match_top_t *,
- &new_asub_tops);
+ Idx new_asub_tops = mctx->asub_tops * 2;
+ re_sub_match_top_t **new_array = re_realloc (mctx->sub_tops,
+ re_sub_match_top_t *,
+ new_asub_tops);
if (BE (new_array == NULL, 0))
return REG_ESPACE;
mctx->sub_tops = new_array;
mctx->asub_tops = new_asub_tops;
}
- mctx->sub_tops[mctx->nsub_tops] = re_calloc (re_sub_match_top_t, 1);
+ mctx->sub_tops[mctx->nsub_tops] = calloc (1, sizeof (re_sub_match_top_t));
if (BE (mctx->sub_tops[mctx->nsub_tops] == NULL, 0))
return REG_ESPACE;
mctx->sub_tops[mctx->nsub_tops]->node = node;
re_sub_match_last_t *new_entry;
if (BE (subtop->nlasts == subtop->alasts, 0))
{
- Idx new_alasts = subtop->alasts;
- re_sub_match_last_t **new_array = re_x2realloc (subtop->lasts,
- re_sub_match_last_t *,
- &new_alasts);
+ Idx new_alasts = 2 * subtop->alasts + 1;
+ re_sub_match_last_t **new_array = re_realloc (subtop->lasts,
+ re_sub_match_last_t *,
+ new_alasts);
if (BE (new_array == NULL, 0))
return NULL;
subtop->lasts = new_array;
subtop->alasts = new_alasts;
}
- new_entry = re_calloc (re_sub_match_last_t, 1);
+ new_entry = calloc (1, sizeof (re_sub_match_last_t));
if (BE (new_entry != NULL, 1))
{
subtop->lasts[subtop->nlasts] = new_entry;
static void
internal_function
-sift_ctx_init (re_sift_context_t *sctx,
- re_dfastate_t **sifted_sts,
- re_dfastate_t **limited_sts,
- Idx last_node, Idx last_str_idx)
+sift_ctx_init (re_sift_context_t *sctx, re_dfastate_t **sifted_sts,
+ re_dfastate_t **limited_sts, Idx last_node, Idx last_str_idx)
{
sctx->sifted_states = sifted_sts;
sctx->limited_states = limited_sts;
projects / gnulib.git / blobdiff