/* Extended regular expression matching and search library.
- Copyright (C) 2002-2016 Free Software Foundation, Inc.
+ Copyright (C) 2002-2021 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
You should have received a copy of the GNU General Public
License along with the GNU C Library; if not, see
- <http://www.gnu.org/licenses/>. */
+ <https://www.gnu.org/licenses/>. */
static reg_errcode_t match_ctx_init (re_match_context_t *cache, int eflags,
- Idx n) internal_function;
-static void match_ctx_clean (re_match_context_t *mctx) internal_function;
-static void match_ctx_free (re_match_context_t *cache) internal_function;
+ Idx n);
+static void match_ctx_clean (re_match_context_t *mctx);
+static void match_ctx_free (re_match_context_t *cache);
static reg_errcode_t match_ctx_add_entry (re_match_context_t *cache, Idx node,
- Idx str_idx, Idx from, Idx to)
- internal_function;
-static Idx search_cur_bkref_entry (const re_match_context_t *mctx, Idx str_idx)
- internal_function;
+ Idx str_idx, Idx from, Idx to);
+static Idx search_cur_bkref_entry (const re_match_context_t *mctx, Idx str_idx);
static reg_errcode_t match_ctx_add_subtop (re_match_context_t *mctx, Idx node,
- Idx str_idx) internal_function;
+ Idx str_idx);
static re_sub_match_last_t * match_ctx_add_sublast (re_sub_match_top_t *subtop,
- Idx node, Idx str_idx)
- internal_function;
+ Idx node, Idx str_idx);
static void 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)
- internal_function;
+ Idx last_str_idx);
static reg_errcode_t re_search_internal (const regex_t *preg,
const char *string, Idx length,
Idx start, Idx last_start, Idx stop,
size_t nmatch, regmatch_t pmatch[],
- int eflags) internal_function;
+ int eflags);
static regoff_t re_search_2_stub (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, bool ret_len) internal_function;
+ Idx stop, bool ret_len);
static regoff_t re_search_stub (struct re_pattern_buffer *bufp,
const char *string, Idx length, Idx start,
regoff_t range, Idx stop,
struct re_registers *regs,
- bool ret_len) internal_function;
+ bool ret_len);
static unsigned re_copy_regs (struct re_registers *regs, regmatch_t *pmatch,
- Idx nregs, int regs_allocated) internal_function;
-static reg_errcode_t prune_impossible_nodes (re_match_context_t *mctx)
- internal_function;
+ Idx nregs, int regs_allocated);
+static reg_errcode_t prune_impossible_nodes (re_match_context_t *mctx);
static Idx check_matching (re_match_context_t *mctx, bool fl_longest_match,
- Idx *p_match_first) internal_function;
+ Idx *p_match_first);
static Idx check_halt_state_context (const re_match_context_t *mctx,
- const re_dfastate_t *state, Idx idx)
- internal_function;
+ const re_dfastate_t *state, Idx idx);
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;
+ Idx cur_idx, Idx nmatch);
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;
+ regmatch_t *regs, regmatch_t *prevregs,
+ re_node_set *eps_via_nodes);
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;
+ bool fl_backtrack);
+static reg_errcode_t free_fail_stack_return (struct re_fail_stack_t *fs);
#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);
#endif /* RE_ENABLE_I18N */
static reg_errcode_t sift_states_backward (const re_match_context_t *mctx,
- re_sift_context_t *sctx)
- internal_function;
+ re_sift_context_t *sctx);
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;
+ re_node_set *cur_dest);
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;
+ re_node_set *dest_nodes);
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);
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;
+ Idx src_idx);
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);
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;
+ Idx bkref_idx);
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;
+ Idx str_idx);
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;
+ Idx str_idx, const re_node_set *candidates);
static reg_errcode_t merge_state_array (const re_dfa_t *dfa,
re_dfastate_t **dst,
- re_dfastate_t **src, Idx num)
- internal_function;
+ re_dfastate_t **src, Idx num);
static re_dfastate_t *find_recover_state (reg_errcode_t *err,
- re_match_context_t *mctx) internal_function;
+ re_match_context_t *mctx);
static re_dfastate_t *transit_state (reg_errcode_t *err,
re_match_context_t *mctx,
- re_dfastate_t *state) internal_function;
+ re_dfastate_t *state);
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);
static reg_errcode_t check_subexp_matching_top (re_match_context_t *mctx,
re_node_set *cur_nodes,
- Idx str_idx) internal_function;
+ Idx str_idx);
#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);
#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);
#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);
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);
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);
static Idx find_subexp_node (const re_dfa_t *dfa, const re_node_set *nodes,
- Idx subexp_idx, int type) internal_function;
+ Idx subexp_idx, int type);
static reg_errcode_t 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;
+ int type);
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;
+ re_node_set *next_nodes);
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;
+ Idx ex_subexp, int type);
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;
+ int type);
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 (const re_dfa_t *dfa,
- re_dfastate_t *state) internal_function;
+ Idx subexp_num, int type);
+static bool build_trtable (const re_dfa_t *dfa, re_dfastate_t *state);
#ifdef RE_ENABLE_I18N
static int check_node_accept_bytes (const re_dfa_t *dfa, Idx node_idx,
- const re_string_t *input, Idx idx)
- internal_function;
+ const re_string_t *input, Idx idx);
# ifdef _LIBC
static unsigned int find_collation_sequence_value (const unsigned char *mbs,
- size_t name_len)
- internal_function;
+ size_t name_len);
# 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_t *states_ch) internal_function;
+ bitset_t *states_ch);
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, int min_len)
- internal_function;
+ const re_token_t *node, Idx idx);
+static reg_errcode_t extend_buffers (re_match_context_t *mctx, int min_len);
\f
/* Entry point for POSIX code. */
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. */
+ Return 0 if a match is found, REG_NOMATCH if not, REG_BADPAT if
+ EFLAGS is invalid. */
int
-regexec (const regex_t *_Restrict_ preg, const char *_Restrict_ string,
+regexec (const regex_t *__restrict preg, const char *__restrict string,
size_t nmatch, regmatch_t pmatch[], int eflags)
{
reg_errcode_t err;
}
#ifdef _LIBC
+libc_hidden_def (__regexec)
+
# include <shlib-compat.h>
versioned_symbol (libc, __regexec, regexec, GLIBC_2_3_4);
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,
strings.)
On success, re_match* functions return the length of the match, re_search*
- return the position of the start of the match. Return value -1 means no
- match was found and -2 indicates an internal error. */
+ return the position of the start of the match. They return -1 on
+ match failure, -2 on error. */
regoff_t
re_match (struct re_pattern_buffer *bufp, const char *string, Idx length,
#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, Idx start,
regoff_t range, struct re_registers *regs,
Idx len;
char *s = NULL;
- if (BE ((length1 < 0 || length2 < 0 || stop < 0
- || INT_ADD_WRAPV (length1, length2, &len)),
- 0))
+ if (__glibc_unlikely ((length1 < 0 || length2 < 0 || stop < 0
+ || INT_ADD_WRAPV (length1, length2, &len))))
return -2;
/* Concatenate the strings. */
{
s = re_malloc (char, len);
- if (BE (s == NULL, 0))
+ if (__glibc_unlikely (s == NULL))
return -2;
#ifdef _LIBC
memcpy (__mempcpy (s, string1, length1), string2, length2);
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,
bool ret_len)
Idx last_start = start + range;
/* Check for out-of-range. */
- if (BE (start < 0 || start > length, 0))
+ if (__glibc_unlikely (start < 0 || start > length))
return -1;
- if (BE (length < last_start || (0 <= range && last_start < start), 0))
+ if (__glibc_unlikely (length < last_start
+ || (0 <= range && last_start < start)))
last_start = length;
- else if (BE (last_start < 0 || (range < 0 && start <= last_start), 0))
+ else if (__glibc_unlikely (last_start < 0
+ || (range < 0 && start <= last_start)))
last_start = 0;
lock_lock (dfa->lock);
if (start < last_start && bufp->fastmap != NULL && !bufp->fastmap_accurate)
re_compile_fastmap (bufp);
- if (BE (bufp->no_sub, 0))
+ if (__glibc_unlikely (bufp->no_sub))
regs = NULL;
/* We need at least 1 register. */
if (regs == NULL)
nregs = 1;
- else if (BE (bufp->regs_allocated == REGS_FIXED
- && regs->num_regs <= bufp->re_nsub, 0))
+ else if (__glibc_unlikely (bufp->regs_allocated == REGS_FIXED
+ && regs->num_regs <= bufp->re_nsub))
{
nregs = regs->num_regs;
- if (BE (nregs < 1, 0))
+ if (__glibc_unlikely (nregs < 1))
{
/* Nothing can be copied to regs. */
regs = NULL;
else
nregs = bufp->re_nsub + 1;
pmatch = re_malloc (regmatch_t, nregs);
- if (BE (pmatch == NULL, 0))
+ if (__glibc_unlikely (pmatch == NULL))
{
rval = -2;
goto out;
/* If caller wants register contents data back, copy them. */
bufp->regs_allocated = re_copy_regs (regs, pmatch, nregs,
bufp->regs_allocated);
- if (BE (bufp->regs_allocated == REGS_UNALLOCATED, 0))
+ if (__glibc_unlikely (bufp->regs_allocated == REGS_UNALLOCATED))
rval = -2;
}
- if (BE (rval == 0, 1))
+ if (__glibc_likely (rval == 0))
{
if (ret_len)
{
- assert (pmatch[0].rm_so == start);
+ DEBUG_ASSERT (pmatch[0].rm_so == start);
rval = pmatch[0].rm_eo - start;
}
else
}
static unsigned
-internal_function
re_copy_regs (struct re_registers *regs, regmatch_t *pmatch, Idx nregs,
int regs_allocated)
{
if (regs_allocated == REGS_UNALLOCATED)
{ /* No. So allocate them with malloc. */
regs->start = re_malloc (regoff_t, need_regs);
- if (BE (regs->start == NULL, 0))
+ if (__glibc_unlikely (regs->start == NULL))
return REGS_UNALLOCATED;
regs->end = re_malloc (regoff_t, need_regs);
- if (BE (regs->end == NULL, 0))
+ if (__glibc_unlikely (regs->end == NULL))
{
re_free (regs->start);
return REGS_UNALLOCATED;
{ /* 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->num_regs, 0))
+ if (__glibc_unlikely (need_regs > regs->num_regs))
{
regoff_t *new_start = re_realloc (regs->start, regoff_t, need_regs);
regoff_t *new_end;
- if (BE (new_start == NULL, 0))
+ if (__glibc_unlikely (new_start == NULL))
return REGS_UNALLOCATED;
new_end = re_realloc (regs->end, regoff_t, need_regs);
- if (BE (new_end == NULL, 0))
+ if (__glibc_unlikely (new_end == NULL))
{
re_free (new_start);
return REGS_UNALLOCATED;
}
else
{
- assert (regs_allocated == REGS_FIXED);
+ DEBUG_ASSERT (regs_allocated == REGS_FIXED);
/* This function may not be called with REGS_FIXED and nregs too big. */
- assert (regs->num_regs >= nregs);
+ DEBUG_ASSERT (nregs <= regs->num_regs);
rval = REGS_FIXED;
}
(0 <= LAST_START && LAST_START <= LENGTH) */
static reg_errcode_t
-__attribute_warn_unused_result__ 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, size_t nmatch,
regmatch_t pmatch[], int eflags)
Idx extra_nmatch;
bool sb;
int ch;
-#if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
re_match_context_t mctx = { .dfa = dfa };
-#else
- re_match_context_t mctx;
-#endif
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));
- mctx.dfa = dfa;
-#endif
-
extra_nmatch = (nmatch > preg->re_nsub) ? nmatch - (preg->re_nsub + 1) : 0;
nmatch -= extra_nmatch;
/* Check if the DFA haven't been compiled. */
- 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))
+ if (__glibc_unlikely (preg->used == 0 || dfa->init_state == NULL
+ || dfa->init_state_word == NULL
+ || dfa->init_state_nl == NULL
+ || dfa->init_state_begbuf == NULL))
return REG_NOMATCH;
-#ifdef DEBUG
/* We assume front-end functions already check them. */
- assert (0 <= last_start && last_start <= length);
-#endif
+ DEBUG_ASSERT (0 <= last_start && last_start <= length);
/* If initial states with non-begbuf contexts have no elements,
the regex must be anchored. If preg->newline_anchor is set,
err = re_string_allocate (&mctx.input, string, length, dfa->nodes_len + 1,
preg->translate, (preg->syntax & RE_ICASE) != 0,
dfa);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
mctx.input.stop = stop;
mctx.input.raw_stop = stop;
mctx.input.newline_anchor = preg->newline_anchor;
err = match_ctx_init (&mctx, eflags, dfa->nbackref * 2);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
/* We will log all the DFA states through which the dfa pass,
if (nmatch > 1 || dfa->has_mb_node)
{
/* Avoid overflow. */
- if (BE ((MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *))
- <= mctx.input.bufs_len), 0))
+ if (__glibc_unlikely ((MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *))
+ <= mctx.input.bufs_len)))
{
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))
+ if (__glibc_unlikely (mctx.state_log == NULL))
{
err = REG_ESPACE;
goto free_return;
}
}
- else
- mctx.state_log = NULL;
match_first = start;
mctx.input.tip_context = (eflags & REG_NOTBOL) ? CONTEXT_BEGBUF
case 7:
/* Fastmap with single-byte translation, match forward. */
- while (BE (match_first < right_lim, 1)
+ while (__glibc_likely (match_first < right_lim)
&& !fastmap[t[(unsigned char) string[match_first]]])
++match_first;
goto forward_match_found_start_or_reached_end;
case 6:
/* Fastmap without translation, match forward. */
- while (BE (match_first < right_lim, 1)
+ while (__glibc_likely (match_first < right_lim)
&& !fastmap[(unsigned char) string[match_first]])
++match_first;
forward_match_found_start_or_reached_end:
- if (BE (match_first == right_lim, 0))
+ if (__glibc_unlikely (match_first == right_lim))
{
ch = match_first >= length
? 0 : (unsigned char) string[match_first];
/* If MATCH_FIRST is out of the valid range, reconstruct the
buffers. */
__re_size_t offset = match_first - mctx.input.raw_mbs_idx;
- if (BE (offset >= (__re_size_t) mctx.input.valid_raw_len, 0))
+ if (__glibc_unlikely (offset
+ >= (__re_size_t) mctx.input.valid_raw_len))
{
err = re_string_reconstruct (&mctx.input, match_first,
eflags);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
offset = match_first - mctx.input.raw_mbs_idx;
/* Reconstruct the buffers so that the matcher can assume that
the matching starts from the beginning of the buffer. */
err = re_string_reconstruct (&mctx.input, match_first, eflags);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
#ifdef RE_ENABLE_I18N
start <= last_start ? &match_first : NULL);
if (match_last != -1)
{
- if (BE (match_last == -2, 0))
+ if (__glibc_unlikely (match_last == -2))
{
err = REG_ESPACE;
goto free_return;
err = prune_impossible_nodes (&mctx);
if (err == REG_NOERROR)
break;
- if (BE (err != REG_NOMATCH, 0))
+ if (__glibc_unlikely (err != REG_NOMATCH))
goto free_return;
match_last = -1;
}
match_ctx_clean (&mctx);
}
-#ifdef DEBUG
- assert (match_last != -1);
- assert (err == REG_NOERROR);
-#endif
+ DEBUG_ASSERT (match_last != -1);
+ DEBUG_ASSERT (err == REG_NOERROR);
/* Set pmatch[] if we need. */
if (nmatch > 0)
{
err = set_regs (preg, &mctx, nmatch, pmatch,
dfa->has_plural_match && dfa->nbackref > 0);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
}
if (pmatch[reg_idx].rm_so != -1)
{
#ifdef RE_ENABLE_I18N
- if (BE (mctx.input.offsets_needed != 0, 0))
+ if (__glibc_unlikely (mctx.input.offsets_needed != 0))
{
pmatch[reg_idx].rm_so =
(pmatch[reg_idx].rm_so == mctx.input.valid_len
: mctx.input.offsets[pmatch[reg_idx].rm_eo]);
}
#else
- assert (mctx.input.offsets_needed == 0);
+ DEBUG_ASSERT (mctx.input.offsets_needed == 0);
#endif
pmatch[reg_idx].rm_so += match_first;
pmatch[reg_idx].rm_eo += match_first;
}
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__attribute_warn_unused_result__
prune_impossible_nodes (re_match_context_t *mctx)
{
const re_dfa_t *const dfa = mctx->dfa;
re_dfastate_t **sifted_states;
re_dfastate_t **lim_states = NULL;
re_sift_context_t sctx;
-#ifdef DEBUG
- assert (mctx->state_log != NULL);
-#endif
+ DEBUG_ASSERT (mctx->state_log != NULL);
match_last = mctx->match_last;
halt_node = mctx->last_node;
/* Avoid overflow. */
- if (BE (MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *)) <= match_last, 0))
+ if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *))
+ <= match_last))
return REG_ESPACE;
sifted_states = re_malloc (re_dfastate_t *, match_last + 1);
- if (BE (sifted_states == NULL, 0))
+ if (__glibc_unlikely (sifted_states == NULL))
{
ret = REG_ESPACE;
goto free_return;
if (dfa->nbackref)
{
lim_states = re_malloc (re_dfastate_t *, match_last + 1);
- if (BE (lim_states == NULL, 0))
+ if (__glibc_unlikely (lim_states == NULL))
{
ret = REG_ESPACE;
goto free_return;
match_last);
ret = sift_states_backward (mctx, &sctx);
re_node_set_free (&sctx.limits);
- if (BE (ret != REG_NOERROR, 0))
+ if (__glibc_unlikely (ret != REG_NOERROR))
goto free_return;
if (sifted_states[0] != NULL || lim_states[0] != NULL)
break;
match_last + 1);
re_free (lim_states);
lim_states = NULL;
- if (BE (ret != REG_NOERROR, 0))
+ if (__glibc_unlikely (ret != REG_NOERROR))
goto free_return;
}
else
sift_ctx_init (&sctx, sifted_states, lim_states, halt_node, match_last);
ret = sift_states_backward (mctx, &sctx);
re_node_set_free (&sctx.limits);
- if (BE (ret != REG_NOERROR, 0))
+ if (__glibc_unlikely (ret != REG_NOERROR))
goto free_return;
if (sifted_states[0] == NULL)
{
since initial states may have constraints like "\<", "^", etc.. */
static inline re_dfastate_t *
-__attribute__ ((always_inline)) internal_function
+__attribute__ ((always_inline))
acquire_init_state_context (reg_errcode_t *err, const re_match_context_t *mctx,
Idx idx)
{
index of the buffer. */
static Idx
-internal_function __attribute_warn_unused_result__
+__attribute_warn_unused_result__
check_matching (re_match_context_t *mctx, bool fl_longest_match,
Idx *p_match_first)
{
err = REG_NOERROR;
cur_state = acquire_init_state_context (&err, mctx, cur_str_idx);
/* An initial state must not be NULL (invalid). */
- if (BE (cur_state == NULL, 0))
+ if (__glibc_unlikely (cur_state == NULL))
{
- assert (err == REG_ESPACE);
+ DEBUG_ASSERT (err == REG_ESPACE);
return -2;
}
/* Check OP_OPEN_SUBEXP in the initial state in case that we use them
later. E.g. Processing back references. */
- if (BE (dfa->nbackref, 0))
+ if (__glibc_unlikely (dfa->nbackref))
{
at_init_state = false;
err = check_subexp_matching_top (mctx, &cur_state->nodes, 0);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
if (cur_state->has_backref)
{
err = transit_state_bkref (mctx, &cur_state->nodes);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
}
}
/* If the RE accepts NULL string. */
- if (BE (cur_state->halt, 0))
+ if (__glibc_unlikely (cur_state->halt))
{
if (!cur_state->has_constraint
|| check_halt_state_context (mctx, cur_state, cur_str_idx))
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)
+ if ((__glibc_unlikely (next_char_idx >= mctx->input.bufs_len)
&& mctx->input.bufs_len < mctx->input.len)
- || (BE (next_char_idx >= mctx->input.valid_len, 0)
+ || (__glibc_unlikely (next_char_idx >= mctx->input.valid_len)
&& mctx->input.valid_len < mctx->input.len))
{
err = extend_buffers (mctx, next_char_idx + 1);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
- assert (err == REG_ESPACE);
+ DEBUG_ASSERT (err == REG_ESPACE);
return -2;
}
}
/* Reached the invalid state or an error. Try to recover a valid
state using the state log, if available and if we have not
already found a valid (even if not the longest) match. */
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return -2;
if (mctx->state_log == NULL
break;
}
- if (BE (at_init_state, 0))
+ if (__glibc_unlikely (at_init_state))
{
if (old_state == cur_state)
next_start_idx = next_char_idx;
/* Check NODE match the current context. */
static bool
-internal_function
check_halt_node_context (const re_dfa_t *dfa, Idx node, unsigned int context)
{
re_token_type_t type = dfa->nodes[node].type;
match the context, return the node. */
static Idx
-internal_function
check_halt_state_context (const re_match_context_t *mctx,
const re_dfastate_t *state, Idx idx)
{
Idx i;
unsigned int context;
-#ifdef DEBUG
- assert (state->halt);
-#endif
+ DEBUG_ASSERT (state->halt);
context = re_string_context_at (&mctx->input, idx, mctx->eflags);
for (i = 0; i < state->nodes.nelem; ++i)
if (check_halt_node_context (mctx->dfa, state->nodes.elems[i], context))
/* Compute the next node to which "NFA" transit from NODE("NFA" is a NFA
corresponding to the DFA).
Return the destination node, and update EPS_VIA_NODES;
- return -1 in case of errors. */
+ return -1 on match failure, -2 on error. */
static Idx
-internal_function
proceed_next_node (const re_match_context_t *mctx, Idx nregs, regmatch_t *regs,
+ regmatch_t *prevregs,
Idx *pidx, Idx node, re_node_set *eps_via_nodes,
struct re_fail_stack_t *fs)
{
const re_dfa_t *const dfa = mctx->dfa;
- Idx i;
- bool ok;
if (IS_EPSILON_NODE (dfa->nodes[node].type))
{
re_node_set *cur_nodes = &mctx->state_log[*pidx]->nodes;
re_node_set *edests = &dfa->edests[node];
- Idx dest_node;
- ok = re_node_set_insert (eps_via_nodes, node);
- if (BE (! ok, 0))
+ bool ok = re_node_set_insert (eps_via_nodes, node);
+ if (__glibc_unlikely (! ok))
return -2;
- /* Pick up a valid destination, or return -1 if none
- is found. */
- for (dest_node = -1, i = 0; i < edests->nelem; ++i)
+
+ /* Pick a valid destination, or return -1 if none is found. */
+ Idx dest_node = -1;
+ for (Idx i = 0; i < edests->nelem; i++)
{
Idx candidate = edests->elems[i];
if (!re_node_set_contains (cur_nodes, 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))
+ prevregs, eps_via_nodes))
return -2;
/* We know we are going to exit. */
if (type == OP_BACK_REF)
{
Idx subexp_idx = dfa->nodes[node].opr.idx + 1;
- naccepted = regs[subexp_idx].rm_eo - regs[subexp_idx].rm_so;
+ if (subexp_idx < nregs)
+ naccepted = regs[subexp_idx].rm_eo - regs[subexp_idx].rm_so;
if (fs != NULL)
{
- if (regs[subexp_idx].rm_so == -1 || regs[subexp_idx].rm_eo == -1)
+ if (subexp_idx >= nregs
+ || regs[subexp_idx].rm_so == -1
+ || regs[subexp_idx].rm_eo == -1)
return -1;
else if (naccepted)
{
char *buf = (char *) re_string_get_buffer (&mctx->input);
- if (memcmp (buf + regs[subexp_idx].rm_so, buf + *pidx,
- naccepted) != 0)
+ if (mctx->input.valid_len - *pidx < naccepted
+ || (memcmp (buf + regs[subexp_idx].rm_so, buf + *pidx,
+ naccepted)
+ != 0))
return -1;
}
}
if (naccepted == 0)
{
Idx dest_node;
- ok = re_node_set_insert (eps_via_nodes, node);
- if (BE (! ok, 0))
+ bool ok = re_node_set_insert (eps_via_nodes, node);
+ if (__glibc_unlikely (! ok))
return -2;
dest_node = dfa->edests[node].elems[0];
if (re_node_set_contains (&mctx->state_log[*pidx]->nodes,
}
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__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 nregs, regmatch_t *regs, regmatch_t *prevregs,
+ re_node_set *eps_via_nodes)
{
reg_errcode_t err;
Idx num = fs->num++;
if (fs->num == 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));
+ new_array = re_realloc (fs->stack, struct re_fail_stack_ent_t,
+ fs->alloc * 2);
if (new_array == NULL)
return REG_ESPACE;
fs->alloc *= 2;
}
fs->stack[num].idx = str_idx;
fs->stack[num].node = dest_node;
- fs->stack[num].regs = re_malloc (regmatch_t, nregs);
+ fs->stack[num].regs = re_malloc (regmatch_t, 2 * nregs);
if (fs->stack[num].regs == NULL)
return REG_ESPACE;
memcpy (fs->stack[num].regs, regs, sizeof (regmatch_t) * nregs);
+ memcpy (fs->stack[num].regs + nregs, prevregs, sizeof (regmatch_t) * nregs);
err = re_node_set_init_copy (&fs->stack[num].eps_via_nodes, eps_via_nodes);
return err;
}
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)
+ regmatch_t *regs, regmatch_t *prevregs,
+ re_node_set *eps_via_nodes)
{
+ if (fs == NULL || fs->num == 0)
+ return -1;
Idx num = --fs->num;
- assert (num >= 0);
*pidx = fs->stack[num].idx;
memcpy (regs, fs->stack[num].regs, sizeof (regmatch_t) * nregs);
+ memcpy (prevregs, fs->stack[num].regs + nregs, sizeof (regmatch_t) * nregs);
re_node_set_free (eps_via_nodes);
re_free (fs->stack[num].regs);
*eps_via_nodes = fs->stack[num].eps_via_nodes;
+ DEBUG_ASSERT (0 <= fs->stack[num].node);
return fs->stack[num].node;
}
+
+#define DYNARRAY_STRUCT regmatch_list
+#define DYNARRAY_ELEMENT regmatch_t
+#define DYNARRAY_PREFIX regmatch_list_
+#include <malloc/dynarray-skeleton.c>
+
/* Set the positions where the subexpressions are starts/ends to registers
PMATCH.
Note: We assume that pmatch[0] is already set, and
pmatch[i].rm_so == pmatch[i].rm_eo == -1 for 0 < i < nmatch. */
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__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_node_set eps_via_nodes;
struct re_fail_stack_t *fs;
struct re_fail_stack_t fs_body = { 0, 2, NULL };
- regmatch_t *prev_idx_match;
- bool prev_idx_match_malloced = false;
+ struct regmatch_list prev_match;
+ regmatch_list_init (&prev_match);
-#ifdef DEBUG
- assert (nmatch > 1);
- assert (mctx->state_log != NULL);
-#endif
+ DEBUG_ASSERT (nmatch > 1);
+ DEBUG_ASSERT (mctx->state_log != NULL);
if (fl_backtrack)
{
fs = &fs_body;
cur_node = dfa->init_node;
re_node_set_init_empty (&eps_via_nodes);
- if (__libc_use_alloca (nmatch * sizeof (regmatch_t)))
- prev_idx_match = (regmatch_t *) alloca (nmatch * sizeof (regmatch_t));
- else
+ if (!regmatch_list_resize (&prev_match, nmatch))
{
- prev_idx_match = re_malloc (regmatch_t, nmatch);
- if (prev_idx_match == NULL)
- {
- free_fail_stack_return (fs);
- return REG_ESPACE;
- }
- prev_idx_match_malloced = true;
+ regmatch_list_free (&prev_match);
+ free_fail_stack_return (fs);
+ return REG_ESPACE;
}
+ regmatch_t *prev_idx_match = regmatch_list_begin (&prev_match);
memcpy (prev_idx_match, pmatch, sizeof (regmatch_t) * nmatch);
for (idx = pmatch[0].rm_so; idx <= pmatch[0].rm_eo ;)
{
update_regs (dfa, pmatch, prev_idx_match, cur_node, idx, nmatch);
- if (idx == pmatch[0].rm_eo && cur_node == mctx->last_node)
+ if ((idx == pmatch[0].rm_eo && cur_node == mctx->last_node)
+ || re_node_set_contains (&eps_via_nodes, cur_node))
{
Idx reg_idx;
+ cur_node = -1;
if (fs)
{
for (reg_idx = 0; reg_idx < nmatch; ++reg_idx)
if (pmatch[reg_idx].rm_so > -1 && pmatch[reg_idx].rm_eo == -1)
- break;
- if (reg_idx == nmatch)
- {
- re_node_set_free (&eps_via_nodes);
- if (prev_idx_match_malloced)
- re_free (prev_idx_match);
- return free_fail_stack_return (fs);
- }
- cur_node = pop_fail_stack (fs, &idx, nmatch, pmatch,
- &eps_via_nodes);
+ {
+ cur_node = pop_fail_stack (fs, &idx, nmatch, pmatch,
+ prev_idx_match, &eps_via_nodes);
+ break;
+ }
}
- else
+ if (cur_node < 0)
{
re_node_set_free (&eps_via_nodes);
- if (prev_idx_match_malloced)
- re_free (prev_idx_match);
- return REG_NOERROR;
+ regmatch_list_free (&prev_match);
+ return free_fail_stack_return (fs);
}
}
/* Proceed to next node. */
- cur_node = proceed_next_node (mctx, nmatch, pmatch, &idx, cur_node,
+ cur_node = proceed_next_node (mctx, nmatch, pmatch, prev_idx_match,
+ &idx, cur_node,
&eps_via_nodes, fs);
- if (BE (cur_node < 0, 0))
+ if (__glibc_unlikely (cur_node < 0))
{
- if (BE (cur_node == -2, 0))
+ if (__glibc_unlikely (cur_node == -2))
{
re_node_set_free (&eps_via_nodes);
- if (prev_idx_match_malloced)
- re_free (prev_idx_match);
+ regmatch_list_free (&prev_match);
free_fail_stack_return (fs);
return REG_ESPACE;
}
- if (fs)
- cur_node = pop_fail_stack (fs, &idx, nmatch, pmatch,
- &eps_via_nodes);
- else
+ cur_node = pop_fail_stack (fs, &idx, nmatch, pmatch,
+ prev_idx_match, &eps_via_nodes);
+ if (cur_node < 0)
{
re_node_set_free (&eps_via_nodes);
- if (prev_idx_match_malloced)
- re_free (prev_idx_match);
+ regmatch_list_free (&prev_match);
+ free_fail_stack_return (fs);
return REG_NOMATCH;
}
}
}
re_node_set_free (&eps_via_nodes);
- if (prev_idx_match_malloced)
- re_free (prev_idx_match);
+ regmatch_list_free (&prev_match);
return free_fail_stack_return (fs);
}
static reg_errcode_t
-internal_function
free_fail_stack_return (struct re_fail_stack_t *fs)
{
if (fs)
}
static void
-internal_function
update_regs (const re_dfa_t *dfa, regmatch_t *pmatch,
regmatch_t *prev_idx_match, Idx cur_node, Idx cur_idx, Idx nmatch)
{
}
else if (type == OP_CLOSE_SUBEXP)
{
+ /* We are at the last node of this sub expression. */
Idx reg_num = dfa->nodes[cur_node].opr.idx + 1;
if (reg_num < nmatch)
{
- /* We are at the last node of this sub expression. */
if (pmatch[reg_num].rm_so < cur_idx)
{
pmatch[reg_num].rm_eo = cur_idx;
((state) != NULL && re_node_set_contains (&(state)->nodes, node))
static reg_errcode_t
-internal_function
sift_states_backward (const re_match_context_t *mctx, re_sift_context_t *sctx)
{
reg_errcode_t err;
Idx str_idx = sctx->last_str_idx;
re_node_set cur_dest;
-#ifdef DEBUG
- assert (mctx->state_log != NULL && mctx->state_log[str_idx] != NULL);
-#endif
+ DEBUG_ASSERT (mctx->state_log != NULL && mctx->state_log[str_idx] != NULL);
/* Build sifted state_log[str_idx]. It has the nodes which can epsilon
transit to the last_node and the last_node itself. */
err = re_node_set_init_1 (&cur_dest, sctx->last_node);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
err = update_cur_sifted_state (mctx, sctx, str_idx, &cur_dest);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
/* Then check each states in the state_log. */
if (mctx->state_log[str_idx])
{
err = build_sifted_states (mctx, sctx, str_idx, &cur_dest);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
}
- It is in CUR_SRC.
And update state_log. */
err = update_cur_sifted_state (mctx, sctx, str_idx, &cur_dest);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
}
err = REG_NOERROR;
}
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__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)
{
Idx prev_node = cur_src->elems[i];
int naccepted = 0;
bool ok;
+ DEBUG_ASSERT (!IS_EPSILON_NODE (dfa->nodes[prev_node].type));
-#ifdef DEBUG
- re_token_type_t type = dfa->nodes[prev_node].type;
- assert (!IS_EPSILON_NODE (type));
-#endif
#ifdef RE_ENABLE_I18N
/* If the node may accept "multi byte". */
if (dfa->nodes[prev_node].accept_mb)
continue;
}
ok = re_node_set_insert (cur_dest, prev_node);
- if (BE (! ok, 0))
+ if (__glibc_unlikely (! ok))
return REG_ESPACE;
}
/* Helper functions. */
static reg_errcode_t
-internal_function
clean_state_log_if_needed (re_match_context_t *mctx, Idx next_state_log_idx)
{
Idx top = mctx->state_log_top;
{
reg_errcode_t err;
err = extend_buffers (mctx, next_state_log_idx + 1);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
}
static reg_errcode_t
-internal_function
merge_state_array (const re_dfa_t *dfa, re_dfastate_t **dst,
re_dfastate_t **src, Idx num)
{
re_node_set merged_set;
err = re_node_set_init_union (&merged_set, &dst[st_idx]->nodes,
&src[st_idx]->nodes);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
dst[st_idx] = re_acquire_state (&err, dfa, &merged_set);
re_node_set_free (&merged_set);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
}
}
static reg_errcode_t
-internal_function
update_cur_sifted_state (const re_match_context_t *mctx,
re_sift_context_t *sctx, Idx str_idx,
re_node_set *dest_nodes)
/* At first, add the nodes which can epsilon transit to a node in
DEST_NODE. */
err = add_epsilon_src_nodes (dfa, dest_nodes, candidates);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
/* Then, check the limitations in the current sift_context. */
{
err = check_subexp_limits (dfa, dest_nodes, candidates, &sctx->limits,
mctx->bkref_ents, str_idx);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
}
sctx->sifted_states[str_idx] = re_acquire_state (&err, dfa, dest_nodes);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
if (candidates && mctx->state_log[str_idx]->has_backref)
{
err = sift_states_bkref (mctx, sctx, str_idx, candidates);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
return REG_NOERROR;
}
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__attribute_warn_unused_result__
add_epsilon_src_nodes (const re_dfa_t *dfa, re_node_set *dest_nodes,
const re_node_set *candidates)
{
Idx i;
re_dfastate_t *state = re_acquire_state (&err, dfa, dest_nodes);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
if (!state->inveclosure.alloc)
{
err = re_node_set_alloc (&state->inveclosure, dest_nodes->nelem);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return REG_ESPACE;
for (i = 0; i < dest_nodes->nelem; i++)
{
err = re_node_set_merge (&state->inveclosure,
dfa->inveclosures + dest_nodes->elems[i]);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return REG_ESPACE;
}
}
}
static reg_errcode_t
-internal_function
sub_epsilon_src_nodes (const re_dfa_t *dfa, Idx node, re_node_set *dest_nodes,
const re_node_set *candidates)
{
{
err = re_node_set_add_intersect (&except_nodes, candidates,
dfa->inveclosures + cur_node);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&except_nodes);
return err;
}
static bool
-internal_function
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)
{
}
static int
-internal_function
check_dst_limits_calc_pos_1 (const re_match_context_t *mctx, int boundaries,
Idx subexp_idx, Idx from_node, Idx bkref_idx)
{
}
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)
which are against limitations from DEST_NODES. */
static reg_errcode_t
-internal_function
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)
{
err = sub_epsilon_src_nodes (dfa, ops_node, dest_nodes,
candidates);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
Remove it form the current sifted state. */
err = sub_epsilon_src_nodes (dfa, node, dest_nodes,
candidates);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
--node_idx;
}
Remove it form the current sifted state. */
err = sub_epsilon_src_nodes (dfa, node, dest_nodes,
candidates);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
}
}
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__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)
{
{
local_sctx = *sctx;
err = re_node_set_init_copy (&local_sctx.limits, &sctx->limits);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
}
local_sctx.last_node = node;
local_sctx.last_str_idx = str_idx;
ok = re_node_set_insert (&local_sctx.limits, enabled_idx);
- if (BE (! ok, 0))
+ if (__glibc_unlikely (! ok))
{
err = REG_ESPACE;
goto free_return;
}
cur_state = local_sctx.sifted_states[str_idx];
err = sift_states_backward (mctx, &local_sctx);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
if (sctx->limited_states != NULL)
{
err = merge_state_array (dfa, sctx->limited_states,
local_sctx.sifted_states,
str_idx + 1);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
}
local_sctx.sifted_states[str_idx] = cur_state;
#ifdef RE_ENABLE_I18N
static int
-internal_function
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)
{
int naccepted;
/* 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]))
+ 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
destination was already thrown away, then the node
- could't accept the current input "multi byte". */
+ couldn't accept the current input "multi byte". */
naccepted = 0;
/* Otherwise, it is sure that the node could accept
'naccepted' bytes input. */
/* Return the next state to which the current state STATE will transit by
accepting the current input byte, and update STATE_LOG if necessary.
+ Return NULL on failure.
If STATE can accept a multibyte char/collating element/back reference
update the destination of STATE_LOG. */
static re_dfastate_t *
-internal_function __attribute_warn_unused_result__
+__attribute_warn_unused_result__
transit_state (reg_errcode_t *err, re_match_context_t *mctx,
re_dfastate_t *state)
{
#ifdef RE_ENABLE_I18N
/* If the current state can accept multibyte. */
- if (BE (state->accept_mb, 0))
+ if (__glibc_unlikely (state->accept_mb))
{
*err = transit_state_mb (mctx, state);
- if (BE (*err != REG_NOERROR, 0))
+ if (__glibc_unlikely (*err != REG_NOERROR))
return NULL;
}
#endif /* RE_ENABLE_I18N */
for (;;)
{
trtable = state->trtable;
- if (BE (trtable != NULL, 1))
+ if (__glibc_likely (trtable != NULL))
return trtable[ch];
trtable = state->word_trtable;
- if (BE (trtable != NULL, 1))
+ if (__glibc_likely (trtable != NULL))
{
unsigned int context;
context
/* Update the state_log if we need */
static re_dfastate_t *
-internal_function
merge_state_with_log (reg_errcode_t *err, re_match_context_t *mctx,
re_dfastate_t *next_state)
{
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 (__glibc_unlikely (*err != REG_NOERROR))
return NULL;
}
else
re_node_set_free (&next_nodes);
}
- if (BE (dfa->nbackref, 0) && next_state != NULL)
+ if (__glibc_unlikely (dfa->nbackref) && next_state != NULL)
{
/* Check OP_OPEN_SUBEXP in the current state in case that we use them
later. We must check them here, since the back references in the
next state might use them. */
*err = check_subexp_matching_top (mctx, &next_state->nodes,
cur_idx);
- if (BE (*err != REG_NOERROR, 0))
+ if (__glibc_unlikely (*err != REG_NOERROR))
return NULL;
/* If the next state has back references. */
if (next_state->has_backref)
{
*err = transit_state_bkref (mctx, &next_state->nodes);
- if (BE (*err != REG_NOERROR, 0))
+ if (__glibc_unlikely (*err != REG_NOERROR))
return NULL;
next_state = mctx->state_log[cur_idx];
}
multi-byte match, then look in the log for a state
from which to restart matching. */
static re_dfastate_t *
-internal_function
find_recover_state (reg_errcode_t *err, re_match_context_t *mctx)
{
re_dfastate_t *cur_state;
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)
{
& ((bitset_word_t) 1 << dfa->nodes[node].opr.idx)))
{
err = match_ctx_add_subtop (mctx, node, str_idx);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
}
#if 0
/* Return the next state to which the current state STATE will transit by
- accepting the current input byte. */
+ accepting the current input byte. Return NULL on failure. */
static re_dfastate_t *
transit_state_sb (reg_errcode_t *err, re_match_context_t *mctx,
unsigned int context;
*err = re_node_set_alloc (&next_nodes, state->nodes.nelem + 1);
- if (BE (*err != REG_NOERROR, 0))
+ if (__glibc_unlikely (*err != REG_NOERROR))
return NULL;
for (node_cnt = 0; node_cnt < state->nodes.nelem; ++node_cnt)
{
{
*err = re_node_set_merge (&next_nodes,
dfa->eclosures + dfa->nexts[cur_node]);
- if (BE (*err != REG_NOERROR, 0))
+ if (__glibc_unlikely (*err != REG_NOERROR))
{
re_node_set_free (&next_nodes);
return NULL;
#ifdef RE_ENABLE_I18N
static reg_errcode_t
-internal_function
transit_state_mb (re_match_context_t *mctx, re_dfastate_t *pstate)
{
const re_dfa_t *const dfa = mctx->dfa;
mctx->max_mb_elem_len = ((mctx->max_mb_elem_len < naccepted) ? naccepted
: mctx->max_mb_elem_len);
err = clean_state_log_if_needed (mctx, dest_idx);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
-#ifdef DEBUG
- assert (dfa->nexts[cur_node_idx] != -1);
-#endif
+ DEBUG_ASSERT (dfa->nexts[cur_node_idx] != -1);
new_nodes = dfa->eclosures + dfa->nexts[cur_node_idx];
dest_state = mctx->state_log[dest_idx];
{
err = re_node_set_init_union (&dest_nodes,
dest_state->entrance_nodes, new_nodes);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
context = re_string_context_at (&mctx->input, dest_idx - 1,
= re_acquire_state_context (&err, dfa, &dest_nodes, context);
if (dest_state != NULL)
re_node_set_free (&dest_nodes);
- if (BE (mctx->state_log[dest_idx] == NULL && err != REG_NOERROR, 0))
+ if (__glibc_unlikely (mctx->state_log[dest_idx] == NULL
+ && err != REG_NOERROR))
return err;
}
return REG_NOERROR;
#endif /* RE_ENABLE_I18N */
static reg_errcode_t
-internal_function
transit_state_bkref (re_match_context_t *mctx, const re_node_set *nodes)
{
const re_dfa_t *const dfa = mctx->dfa;
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))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
/* And add the epsilon closures (which is 'new_dest_nodes') of
the backreference to appropriate state_log. */
-#ifdef DEBUG
- assert (dfa->nexts[node_idx] != -1);
-#endif
+ DEBUG_ASSERT (dfa->nexts[node_idx] != -1);
for (; bkc_idx < mctx->nbkref_ents; ++bkc_idx)
{
Idx subexp_len;
mctx->state_log[dest_str_idx]
= re_acquire_state_context (&err, dfa, new_dest_nodes,
context);
- if (BE (mctx->state_log[dest_str_idx] == NULL
- && err != REG_NOERROR, 0))
+ if (__glibc_unlikely (mctx->state_log[dest_str_idx] == NULL
+ && err != REG_NOERROR))
goto free_return;
}
else
err = re_node_set_init_union (&dest_nodes,
dest_state->entrance_nodes,
new_dest_nodes);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&dest_nodes);
goto free_return;
mctx->state_log[dest_str_idx]
= re_acquire_state_context (&err, dfa, &dest_nodes, context);
re_node_set_free (&dest_nodes);
- if (BE (mctx->state_log[dest_str_idx] == NULL
- && err != REG_NOERROR, 0))
+ if (__glibc_unlikely (mctx->state_log[dest_str_idx] == NULL
+ && err != REG_NOERROR))
goto free_return;
}
/* We need to check recursively if the backreference can epsilon
{
err = check_subexp_matching_top (mctx, new_dest_nodes,
cur_str_idx);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
err = transit_state_bkref (mctx, new_dest_nodes);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto free_return;
}
}
delay these checking for prune_impossible_nodes(). */
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__attribute_warn_unused_result__
get_subexp (re_match_context_t *mctx, Idx bkref_node, Idx bkref_str_idx)
{
const re_dfa_t *const dfa = mctx->dfa;
at the back reference? */
if (sl_str_diff > 0)
{
- if (BE (bkref_str_off + sl_str_diff > mctx->input.valid_len, 0))
+ if (__glibc_unlikely (bkref_str_off + sl_str_diff
+ > mctx->input.valid_len))
{
/* Not enough chars for a successful match. */
if (bkref_str_off + sl_str_diff > mctx->input.len)
err = clean_state_log_if_needed (mctx,
bkref_str_off
+ sl_str_diff);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
buf = (const char *) re_string_get_buffer (&mctx->input);
}
if (err == REG_NOMATCH)
continue;
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
at the back reference? */
if (sl_str_off > 0)
{
- if (BE (bkref_str_off >= mctx->input.valid_len, 0))
+ if (__glibc_unlikely (bkref_str_off >= mctx->input.valid_len))
{
/* If we are at the end of the input, we cannot match. */
if (bkref_str_off >= mctx->input.len)
break;
err = extend_buffers (mctx, bkref_str_off + 1);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
buf = (const char *) re_string_get_buffer (&mctx->input);
OP_CLOSE_SUBEXP);
if (err == REG_NOMATCH)
continue;
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
sub_last = match_ctx_add_sublast (sub_top, cls_node, sl_str);
- if (BE (sub_last == NULL, 0))
+ if (__glibc_unlikely (sub_last == NULL))
return REG_ESPACE;
err = get_subexp_sub (mctx, sub_top, sub_last, bkref_node,
bkref_str_idx);
+ buf = (const char *) re_string_get_buffer (&mctx->input);
if (err == REG_NOMATCH)
continue;
+ if (__glibc_unlikely (err != REG_NOERROR))
+ return err;
}
}
return REG_NOERROR;
and SUB_LAST. */
static reg_errcode_t
-internal_function
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)
{
return err;
err = match_ctx_add_entry (mctx, bkref_node, bkref_str, sub_top->str_idx,
sub_last->str_idx);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
to_idx = bkref_str + sub_last->str_idx - sub_top->str_idx;
return clean_state_log_if_needed (mctx, to_idx);
E.g. RE: (a){2} */
static Idx
-internal_function
find_subexp_node (const re_dfa_t *dfa, const re_node_set *nodes,
Idx subexp_idx, int type)
{
/* Check whether the node TOP_NODE at TOP_STR can arrive to the node
LAST_NODE at LAST_STR. We record the path onto PATH since it will be
heavily reused.
- Return REG_NOERROR if it can arrive, or REG_NOMATCH otherwise. */
+ Return REG_NOERROR if it can arrive, REG_NOMATCH if it cannot,
+ REG_ESPACE if memory is exhausted. */
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__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)
{
subexp_num = dfa->nodes[top_node].opr.idx;
/* Extend the buffer if we need. */
- if (BE (path->alloc < last_str + mctx->max_mb_elem_len + 1, 0))
+ if (__glibc_unlikely (path->alloc < last_str + mctx->max_mb_elem_len + 1))
{
re_dfastate_t **new_array;
Idx old_alloc = path->alloc;
Idx incr_alloc = last_str + mctx->max_mb_elem_len + 1;
Idx new_alloc;
- if (BE (IDX_MAX - old_alloc < incr_alloc, 0))
+ if (__glibc_unlikely (IDX_MAX - old_alloc < incr_alloc))
return REG_ESPACE;
new_alloc = old_alloc + incr_alloc;
- if (BE (SIZE_MAX / sizeof (re_dfastate_t *) < new_alloc, 0))
+ if (__glibc_unlikely (SIZE_MAX / sizeof (re_dfastate_t *) < new_alloc))
return REG_ESPACE;
new_array = re_realloc (path->array, re_dfastate_t *, new_alloc);
- if (BE (new_array == NULL, 0))
+ if (__glibc_unlikely (new_array == NULL))
return REG_ESPACE;
path->array = new_array;
path->alloc = new_alloc;
if (str_idx == top_str)
{
err = re_node_set_init_1 (&next_nodes, top_node);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
err = check_arrival_expand_ecl (dfa, &next_nodes, subexp_num, type);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&next_nodes);
return err;
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 (__glibc_unlikely (err != REG_NOERROR))
return err;
}
else
{
err = expand_bkref_cache (mctx, &next_nodes, str_idx,
subexp_num, type);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&next_nodes);
return err;
}
}
cur_state = re_acquire_state_context (&err, dfa, &next_nodes, context);
- if (BE (cur_state == NULL && err != REG_NOERROR, 0))
+ if (__glibc_unlikely (cur_state == NULL && err != REG_NOERROR))
{
re_node_set_free (&next_nodes);
return err;
{
err = re_node_set_merge (&next_nodes,
&mctx->state_log[str_idx + 1]->nodes);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&next_nodes);
return err;
err = check_arrival_add_next_nodes (mctx, str_idx,
&cur_state->non_eps_nodes,
&next_nodes);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&next_nodes);
return err;
if (next_nodes.nelem)
{
err = check_arrival_expand_ecl (dfa, &next_nodes, subexp_num, type);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&next_nodes);
return err;
}
err = expand_bkref_cache (mctx, &next_nodes, str_idx,
subexp_num, type);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&next_nodes);
return err;
}
context = re_string_context_at (&mctx->input, str_idx - 1, mctx->eflags);
cur_state = re_acquire_state_context (&err, dfa, &next_nodes, context);
- if (BE (cur_state == NULL && err != REG_NOERROR, 0))
+ if (__glibc_unlikely (cur_state == NULL && err != REG_NOERROR))
{
re_node_set_free (&next_nodes);
return err;
Can't we unify them? */
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__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)
{
{
int naccepted = 0;
Idx cur_node = cur_nodes->elems[cur_idx];
-#ifdef DEBUG
- re_token_type_t type = dfa->nodes[cur_node].type;
- assert (!IS_EPSILON_NODE (type));
-#endif
+ DEBUG_ASSERT (!IS_EPSILON_NODE (dfa->nodes[cur_node].type));
+
#ifdef RE_ENABLE_I18N
/* If the node may accept "multi byte". */
if (dfa->nodes[cur_node].accept_mb)
if (dest_state)
{
err = re_node_set_merge (&union_set, &dest_state->nodes);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&union_set);
return err;
}
}
ok = re_node_set_insert (&union_set, next_node);
- if (BE (! ok, 0))
+ if (__glibc_unlikely (! ok))
{
re_node_set_free (&union_set);
return REG_ESPACE;
}
mctx->state_log[next_idx] = re_acquire_state (&err, dfa,
&union_set);
- if (BE (mctx->state_log[next_idx] == NULL
- && err != REG_NOERROR, 0))
+ if (__glibc_unlikely (mctx->state_log[next_idx] == NULL
+ && err != REG_NOERROR))
{
re_node_set_free (&union_set);
return err;
|| check_node_accept (mctx, dfa->nodes + cur_node, str_idx))
{
ok = re_node_set_insert (next_nodes, dfa->nexts[cur_node]);
- if (BE (! ok, 0))
+ if (__glibc_unlikely (! ok))
{
re_node_set_free (&union_set);
return REG_ESPACE;
*/
static reg_errcode_t
-internal_function
check_arrival_expand_ecl (const re_dfa_t *dfa, re_node_set *cur_nodes,
Idx ex_subexp, int type)
{
reg_errcode_t err;
Idx idx, outside_node;
re_node_set new_nodes;
-#ifdef DEBUG
- assert (cur_nodes->nelem);
-#endif
+ DEBUG_ASSERT (cur_nodes->nelem);
err = re_node_set_alloc (&new_nodes, cur_nodes->nelem);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
/* Create a new node set NEW_NODES with the nodes which are epsilon
closures of the node in CUR_NODES. */
{
/* There are no problematic nodes, just merge them. */
err = re_node_set_merge (&new_nodes, eclosure);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&new_nodes);
return err;
/* There are problematic nodes, re-calculate incrementally. */
err = check_arrival_expand_ecl_sub (dfa, &new_nodes, cur_node,
ex_subexp, type);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
{
re_node_set_free (&new_nodes);
return err;
problematic append it to DST_NODES. */
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__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)
{
if (type == OP_CLOSE_SUBEXP)
{
ok = re_node_set_insert (dst_nodes, cur_node);
- if (BE (! ok, 0))
+ if (__glibc_unlikely (! ok))
return REG_ESPACE;
}
break;
}
ok = re_node_set_insert (dst_nodes, cur_node);
- if (BE (! ok, 0))
+ if (__glibc_unlikely (! ok))
return REG_ESPACE;
if (dfa->edests[cur_node].nelem == 0)
break;
err = check_arrival_expand_ecl_sub (dfa, dst_nodes,
dfa->edests[cur_node].elems[1],
ex_subexp, type);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
cur_node = dfa->edests[cur_node].elems[0];
in MCTX->BKREF_ENTS. */
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__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)
{
err2 = check_arrival_expand_ecl (dfa, &new_dests, subexp_num, type);
err3 = re_node_set_merge (cur_nodes, &new_dests);
re_node_set_free (&new_dests);
- if (BE (err != REG_NOERROR || err2 != REG_NOERROR
- || err3 != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR || err2 != REG_NOERROR
+ || err3 != REG_NOERROR))
{
err = (err != REG_NOERROR ? err
: (err2 != REG_NOERROR ? err2 : err3));
err = re_node_set_init_copy (&union_set,
&mctx->state_log[to_idx]->nodes);
ok = re_node_set_insert (&union_set, next_node);
- if (BE (err != REG_NOERROR || ! ok, 0))
+ if (__glibc_unlikely (err != REG_NOERROR || ! ok))
{
re_node_set_free (&union_set);
err = err != REG_NOERROR ? err : REG_ESPACE;
else
{
err = re_node_set_init_1 (&union_set, next_node);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
return err;
}
mctx->state_log[to_idx] = re_acquire_state (&err, dfa, &union_set);
re_node_set_free (&union_set);
- if (BE (mctx->state_log[to_idx] == NULL
- && err != REG_NOERROR, 0))
+ if (__glibc_unlikely (mctx->state_log[to_idx] == NULL
+ && err != REG_NOERROR))
return err;
}
}
/* Build transition table for the state.
Return true if successful. */
-static bool
-internal_function
+static bool __attribute_noinline__
build_trtable (const re_dfa_t *dfa, re_dfastate_t *state)
{
reg_errcode_t err;
int ch;
bool need_word_trtable = false;
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_t *dests_ch;
+ re_dfastate_t *dest_states[SBC_MAX];
+ re_dfastate_t *dest_states_word[SBC_MAX];
+ re_dfastate_t *dest_states_nl[SBC_MAX];
+ re_node_set follows;
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
characters which i-th destination state accepts. */
- if (__libc_use_alloca (sizeof (struct dests_alloc)))
- dests_alloc = (struct dests_alloc *) alloca (sizeof (struct dests_alloc));
- else
- {
- dests_alloc = re_malloc (struct dests_alloc, 1);
- if (BE (dests_alloc == NULL, 0))
- return false;
- dests_node_malloced = true;
- }
- dests_node = dests_alloc->dests_node;
- dests_ch = dests_alloc->dests_ch;
+ re_node_set dests_node[SBC_MAX];
+ bitset_t dests_ch[SBC_MAX];
/* Initialize transition table. */
state->word_trtable = state->trtable = NULL;
/* At first, group all nodes belonging to 'state' into several
destinations. */
ndests = group_nodes_into_DFAstates (dfa, state, dests_node, dests_ch);
- if (BE (ndests <= 0, 0))
+ if (__glibc_unlikely (ndests <= 0))
{
- if (dests_node_malloced)
- free (dests_alloc);
/* Return false in case of an error, true otherwise. */
if (ndests == 0)
{
state->trtable = (re_dfastate_t **)
calloc (sizeof (re_dfastate_t *), SBC_MAX);
- if (BE (state->trtable == NULL, 0))
+ if (__glibc_unlikely (state->trtable == NULL))
return false;
return true;
}
}
err = re_node_set_alloc (&follows, ndests + 1);
- if (BE (err != REG_NOERROR, 0))
- goto out_free;
-
- /* Avoid arithmetic overflow in size calculation. */
- 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_t)) * SBC_MAX
- + ndests * 3 * sizeof (re_dfastate_t *)))
- dest_states = (re_dfastate_t **)
- alloca (ndests * 3 * sizeof (re_dfastate_t *));
- else
+ if (__glibc_unlikely (err != REG_NOERROR))
{
- dest_states = (re_dfastate_t **)
- malloc (ndests * 3 * sizeof (re_dfastate_t *));
- if (BE (dest_states == NULL, 0))
- {
-out_free:
- if (dest_states_malloced)
- free (dest_states);
- re_node_set_free (&follows);
- for (i = 0; i < ndests; ++i)
- re_node_set_free (dests_node + i);
- if (dests_node_malloced)
- free (dests_alloc);
- return false;
- }
- dest_states_malloced = true;
+ out_free:
+ re_node_set_free (&follows);
+ for (i = 0; i < ndests; ++i)
+ re_node_set_free (dests_node + i);
+ return false;
}
- dest_states_word = dest_states + ndests;
- dest_states_nl = dest_states_word + ndests;
+
bitset_empty (acceptable);
/* Then build the states for all destinations. */
if (next_node != -1)
{
err = re_node_set_merge (&follows, dfa->eclosures + next_node);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto out_free;
}
}
dest_states[i] = re_acquire_state_context (&err, dfa, &follows, 0);
- if (BE (dest_states[i] == NULL && err != REG_NOERROR, 0))
+ if (__glibc_unlikely (dest_states[i] == NULL && err != REG_NOERROR))
goto out_free;
/* If the new state has context constraint,
build appropriate states for these contexts. */
{
dest_states_word[i] = re_acquire_state_context (&err, dfa, &follows,
CONTEXT_WORD);
- if (BE (dest_states_word[i] == NULL && err != REG_NOERROR, 0))
+ if (__glibc_unlikely (dest_states_word[i] == NULL
+ && err != REG_NOERROR))
goto out_free;
if (dest_states[i] != dest_states_word[i] && dfa->mb_cur_max > 1)
dest_states_nl[i] = re_acquire_state_context (&err, dfa, &follows,
CONTEXT_NEWLINE);
- if (BE (dest_states_nl[i] == NULL && err != REG_NOERROR, 0))
+ if (__glibc_unlikely (dest_states_nl[i] == NULL && err != REG_NOERROR))
goto out_free;
}
else
bitset_merge (acceptable, dests_ch[i]);
}
- if (!BE (need_word_trtable, 0))
+ if (!__glibc_unlikely (need_word_trtable))
{
/* We don't care about whether the following character is a word
character, or we are in a single-byte character set so we can
256-entry transition table. */
trtable = state->trtable =
(re_dfastate_t **) calloc (sizeof (re_dfastate_t *), SBC_MAX);
- if (BE (trtable == NULL, 0))
+ if (__glibc_unlikely (trtable == NULL))
goto out_free;
/* For all characters ch...: */
for (ch = i * BITSET_WORD_BITS, elem = acceptable[i], mask = 1;
elem;
mask <<= 1, elem >>= 1, ++ch)
- if (BE (elem & 1, 0))
+ if (__glibc_unlikely (elem & 1))
{
/* There must be exactly one destination which accepts
character ch. See group_nodes_into_DFAstates. */
starting at trtable[SBC_MAX]. */
trtable = state->word_trtable =
(re_dfastate_t **) calloc (sizeof (re_dfastate_t *), 2 * SBC_MAX);
- if (BE (trtable == NULL, 0))
+ if (__glibc_unlikely (trtable == NULL))
goto out_free;
/* For all characters ch...: */
for (ch = i * BITSET_WORD_BITS, elem = acceptable[i], mask = 1;
elem;
mask <<= 1, elem >>= 1, ++ch)
- if (BE (elem & 1, 0))
+ if (__glibc_unlikely (elem & 1))
{
/* There must be exactly one destination which accepts
character ch. See group_nodes_into_DFAstates. */
}
}
- if (dest_states_malloced)
- free (dest_states);
-
re_node_set_free (&follows);
for (i = 0; i < ndests; ++i)
re_node_set_free (dests_node + i);
-
- if (dests_node_malloced)
- free (dests_alloc);
-
return true;
}
/* Group all nodes belonging to STATE into several destinations.
Then for all destinations, set the nodes belonging to the destination
to DESTS_NODE[i] and set the characters accepted by the destination
- to DEST_CH[i]. This function return the number of destinations. */
+ to DEST_CH[i]. Return the number of destinations if successful,
+ -1 on internal error. */
static Idx
-internal_function
group_nodes_into_DFAstates (const re_dfa_t *dfa, const re_dfastate_t *state,
re_node_set *dests_node, bitset_t *dests_ch)
{
bitset_copy (dests_ch[ndests], remains);
bitset_copy (dests_ch[j], intersec);
err = re_node_set_init_copy (dests_node + ndests, &dests_node[j]);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto error_return;
++ndests;
}
/* Put the position in the current group. */
ok = re_node_set_insert (&dests_node[j], cur_nodes->elems[i]);
- if (BE (! ok, 0))
+ if (__glibc_unlikely (! ok))
goto error_return;
/* If all characters are consumed, go to next node. */
{
bitset_copy (dests_ch[ndests], accepts);
err = re_node_set_init_1 (dests_node + ndests, cur_nodes->elems[i]);
- if (BE (err != REG_NOERROR, 0))
+ if (__glibc_unlikely (err != REG_NOERROR))
goto error_return;
++ndests;
bitset_empty (accepts);
}
}
+ assume (ndests <= SBC_MAX);
return ndests;
error_return:
for (j = 0; j < ndests; ++j)
# endif
static int
-internal_function
check_node_accept_bytes (const re_dfa_t *dfa, Idx node_idx,
const re_string_t *input, Idx str_idx)
{
int char_len, elem_len;
Idx i;
- if (BE (node->type == OP_UTF8_PERIOD, 0))
+ if (__glibc_unlikely (node->type == OP_UTF8_PERIOD))
{
unsigned char c = re_string_byte_at (input, str_idx), d;
- if (BE (c < 0xc2, 1))
+ if (__glibc_likely (c < 0xc2))
return 0;
if (str_idx + 2 > input->len)
/* 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 & RE_DOT_NEWLINE) &&
- re_string_byte_at (input, str_idx) == '\n') ||
- ((dfa->syntax & RE_DOT_NOT_NULL) &&
- re_string_byte_at (input, str_idx) == '\0'))
+ if ((!(dfa->syntax & RE_DOT_NEWLINE)
+ && re_string_byte_at (input, str_idx) == '\n')
+ || ((dfa->syntax & RE_DOT_NOT_NULL)
+ && re_string_byte_at (input, str_idx) == '\0'))
return 0;
return char_len;
}
indirect = (const int32_t *)
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
int32_t idx = findidx (table, indirect, extra, &cp, elem_len);
+ int32_t rule = idx >> 24;
+ idx &= 0xffffff;
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 & 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]))
- ++cnt;
- if (cnt > weight_len)
- {
- match_len = elem_len;
- goto check_node_accept_bytes_match;
- }
- }
- }
+ {
+ size_t weight_len = weights[idx];
+ for (i = 0; i < cset->nequiv_classes; ++i)
+ {
+ int32_t equiv_class_idx = cset->equiv_classes[i];
+ int32_t equiv_class_rule = equiv_class_idx >> 24;
+ equiv_class_idx &= 0xffffff;
+ if (weights[equiv_class_idx] == weight_len
+ && equiv_class_rule == rule
+ && memcmp (weights + idx + 1,
+ weights + equiv_class_idx + 1,
+ weight_len) == 0)
+ {
+ match_len = elem_len;
+ goto check_node_accept_bytes_match;
+ }
+ }
+ }
}
}
else
# 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);
byte of the INPUT. */
static bool
-internal_function
check_node_accept (const re_match_context_t *mctx, const re_token_t *node,
Idx idx)
{
case OP_UTF8_PERIOD:
if (ch >= ASCII_CHARS)
return false;
- /* FALLTHROUGH */
+ FALLTHROUGH;
#endif
case OP_PERIOD:
if ((ch == '\n' && !(mctx->dfa->syntax & RE_DOT_NEWLINE))
/* Extend the buffers, if the buffers have run out. */
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__attribute_warn_unused_result__
extend_buffers (re_match_context_t *mctx, int min_len)
{
reg_errcode_t ret;
re_string_t *pstr = &mctx->input;
/* Avoid overflow. */
- if (BE (MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *)) / 2
- <= pstr->bufs_len, 0))
+ if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / sizeof (re_dfastate_t *)) / 2
+ <= pstr->bufs_len))
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))
+ if (__glibc_unlikely (ret != REG_NOERROR))
return ret;
if (mctx->state_log != NULL)
does not have the right size. */
re_dfastate_t **new_array = re_realloc (mctx->state_log, re_dfastate_t *,
pstr->bufs_len + 1);
- if (BE (new_array == NULL, 0))
+ if (__glibc_unlikely (new_array == NULL))
return REG_ESPACE;
mctx->state_log = new_array;
}
if (pstr->mb_cur_max > 1)
{
ret = build_wcs_upper_buffer (pstr);
- if (BE (ret != REG_NOERROR, 0))
+ if (__glibc_unlikely (ret != REG_NOERROR))
return ret;
}
else
/* Initialize MCTX. */
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__attribute_warn_unused_result__
match_ctx_init (re_match_context_t *mctx, int eflags, Idx n)
{
mctx->eflags = eflags;
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))
+ if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / max_object_size) < n))
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))
+ if (__glibc_unlikely (mctx->bkref_ents == NULL || mctx->sub_tops == NULL))
return REG_ESPACE;
}
/* Already zero-ed by the caller.
of the input, or changes the input string. */
static void
-internal_function
match_ctx_clean (re_match_context_t *mctx)
{
Idx st_idx;
re_free (top->path->array);
re_free (top->path);
}
- free (top);
+ re_free (top);
}
mctx->nsub_tops = 0;
/* Free all the memory associated with MCTX. */
static void
-internal_function
match_ctx_free (re_match_context_t *mctx)
{
/* First, free all the memory associated with MCTX->SUB_TOPS. */
*/
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__attribute_warn_unused_result__
match_ctx_add_entry (re_match_context_t *mctx, Idx node, Idx str_idx, Idx from,
Idx to)
{
struct re_backref_cache_entry* new_entry;
new_entry = re_realloc (mctx->bkref_ents, struct re_backref_cache_entry,
mctx->abkref_ents * 2);
- if (BE (new_entry == NULL, 0))
+ if (__glibc_unlikely (new_entry == NULL))
{
re_free (mctx->bkref_ents);
return REG_ESPACE;
found. Note that MCTX->BKREF_ENTS is already sorted by MCTX->STR_IDX. */
static Idx
-internal_function
search_cur_bkref_entry (const re_match_context_t *mctx, Idx str_idx)
{
Idx left, right, mid, last;
at STR_IDX. */
static reg_errcode_t
-internal_function __attribute_warn_unused_result__
+__attribute_warn_unused_result__
match_ctx_add_subtop (re_match_context_t *mctx, Idx node, Idx str_idx)
{
-#ifdef DEBUG
- assert (mctx->sub_tops != NULL);
- assert (mctx->asub_tops > 0);
-#endif
- if (BE (mctx->nsub_tops == mctx->asub_tops, 0))
+ DEBUG_ASSERT (mctx->sub_tops != NULL);
+ DEBUG_ASSERT (mctx->asub_tops > 0);
+ if (__glibc_unlikely (mctx->nsub_tops == mctx->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))
+ if (__glibc_unlikely (new_array == NULL))
return REG_ESPACE;
mctx->sub_tops = new_array;
mctx->asub_tops = new_asub_tops;
}
mctx->sub_tops[mctx->nsub_tops] = calloc (1, sizeof (re_sub_match_top_t));
- if (BE (mctx->sub_tops[mctx->nsub_tops] == NULL, 0))
+ if (__glibc_unlikely (mctx->sub_tops[mctx->nsub_tops] == NULL))
return REG_ESPACE;
mctx->sub_tops[mctx->nsub_tops]->node = node;
mctx->sub_tops[mctx->nsub_tops++]->str_idx = str_idx;
}
/* Register the node NODE, whose type is OP_CLOSE_SUBEXP, and which matches
- at STR_IDX, whose corresponding OP_OPEN_SUBEXP is SUB_TOP. */
+ at STR_IDX, whose corresponding OP_OPEN_SUBEXP is SUB_TOP.
+ Return the new entry if successful, NULL if memory is exhausted. */
static re_sub_match_last_t *
-internal_function
match_ctx_add_sublast (re_sub_match_top_t *subtop, Idx node, Idx str_idx)
{
re_sub_match_last_t *new_entry;
- if (BE (subtop->nlasts == subtop->alasts, 0))
+ if (__glibc_unlikely (subtop->nlasts == subtop->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))
+ if (__glibc_unlikely (new_array == NULL))
return NULL;
subtop->lasts = new_array;
subtop->alasts = new_alasts;
}
new_entry = calloc (1, sizeof (re_sub_match_last_t));
- if (BE (new_entry != NULL, 1))
+ if (__glibc_likely (new_entry != NULL))
{
subtop->lasts[subtop->nlasts] = new_entry;
new_entry->node = node;
}
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)
{
projects / platform / upstream / m4.git / blobdiff