* [PATCH] Regex infrastructure for UTF-8 optimizations
@ 2003-11-12 13:44 Jakub Jelinek
0 siblings, 0 replies; only message in thread
From: Jakub Jelinek @ 2003-11-12 13:44 UTC (permalink / raw)
To: Ulrich Drepper, Roland McGrath; +Cc: Glibc hackers
Hi!
This is just easy MB_CUR_MAX -> something->mb_cur_max replacement,
needed by following patches. At the same time it reshuffles re_dfa_t a
little so that it is smaller on 64-bit arches.
2003-11-12 Jakub Jelinek <jakub@redhat.com>
* posix/regex_internal.h (struct re_string_t): Add is_utf8
and mb_cur_max fields.
(struct re_dfa_t): Likewise. Reorder fields to make structure
smaller on 64-bit arches.
(re_string_allocate, re_string_construct): Add mb_cur_max and
is_utf8 arguments.
(re_string_char_size_at, re_string_wchar_at): Use pstr->mb_cur_max
instead of MB_CUR_MAX.
* posix/regcomp.c (re_compile_fastmap_iter): Use dfa->mb_cur_max
instead of MB_CUR_MAX.
(re_compile_internal): Pass new arguments to re_string_construct.
(init_dfa): Initialize mb_cur_max and is_utf8 fields.
(peek_token, peek_token_bracket): Use input->mb_cur_max instead
of MB_CUR_MAX.
(parse_expression, parse_bracket_exp, parse_charclass_op): Use
dfa->mb_cur_max instead of MB_CUR_MAX.
* posix/regex_internal.c (re_string_construct_common): Add
mb_cur_max and is_utf8 arguments. Initialize fields with them.
(re_string_allocate, re_string_construct): Add mb_cur_max and
is_utf8 arguments, pass them to re_string_construct_common.
Use mb_cur_max instead of MB_CUR_MAX.
(re_string_realloc_buffers): Use pstr->mb_cur_max instead of
MB_CUR_MAX.
(re_string_reconstruct): Likewise.
(re_string_context_at): Use input->mb_cur_max instead of
MB_CUR_MAX.
(create_ci_newstate, create_cd_newstate): Use dfa->mb_cur_max
instead of MB_CUR_MAX.
* posix/regexec.c (re_search_internal): Likewise.
Pass new arguments to re_string_allocate.
(check_matching, transit_state_sb): Use dfa->mb_cur_max instead of
MB_CUR_MAX.
(extend_buffers): Use pstr->mb_cur_max instead of MB_CUR_MAX.
--- libc/posix/regcomp.c.jj 2003-11-12 08:25:35.000000000 +0100
+++ libc/posix/regcomp.c 2003-11-12 09:26:47.000000000 +0100
@@ -312,7 +312,7 @@ re_compile_fastmap_iter (bufp, init_stat
{
re_dfa_t *dfa = (re_dfa_t *) bufp->buffer;
int node_cnt;
- int icase = (MB_CUR_MAX == 1 && (bufp->syntax & RE_ICASE));
+ int icase = (dfa->mb_cur_max == 1 && (bufp->syntax & RE_ICASE));
for (node_cnt = 0; node_cnt < init_state->nodes.nelem; ++node_cnt)
{
int node = init_state->nodes.elems[node_cnt];
@@ -324,7 +324,7 @@ re_compile_fastmap_iter (bufp, init_stat
#ifdef RE_ENABLE_I18N
if ((bufp->syntax & RE_ICASE) && !icase)
{
- unsigned char *buf = alloca (MB_CUR_MAX), *p;
+ unsigned char *buf = alloca (dfa->mb_cur_max), *p;
wchar_t wc;
mbstate_t state;
@@ -376,7 +376,7 @@ re_compile_fastmap_iter (bufp, init_stat
re_set_fastmap (fastmap, icase, ch);
}
# else
- if (MB_CUR_MAX > 1)
+ if (dfa->mb_cur_max > 1)
for (i = 0; i < SBC_MAX; ++i)
if (__btowc (i) == WEOF)
re_set_fastmap (fastmap, icase, i);
@@ -744,7 +744,8 @@ re_compile_internal (preg, pattern, leng
#endif
err = re_string_construct (®exp, pattern, length, preg->translate,
- syntax & RE_ICASE);
+ syntax & RE_ICASE, dfa->mb_cur_max,
+ dfa->is_utf8);
if (BE (err != REG_NOERROR, 0))
{
re_free (dfa);
@@ -812,6 +813,13 @@ init_dfa (dfa, pat_len)
dfa->subexps = re_malloc (re_subexp_t, dfa->subexps_alloc);
dfa->word_char = NULL;
+ dfa->mb_cur_max = MB_CUR_MAX;
+#ifdef _LIBC
+ if (dfa->mb_cur_max > 1
+ && strcmp (_NL_CURRENT (LC_CTYPE, _NL_CTYPE_CODESET_NAME), "UTF-8") == 0)
+ dfa->is_utf8 = 1;
+#endif
+
if (BE (dfa->nodes == NULL || dfa->state_table == NULL
|| dfa->subexps == NULL, 0))
{
@@ -1522,7 +1530,7 @@ peek_token (token, input, syntax)
#ifdef RE_ENABLE_I18N
token->mb_partial = 0;
- if (MB_CUR_MAX > 1 &&
+ if (input->mb_cur_max > 1 &&
!re_string_first_byte (input, re_string_cur_idx (input)))
{
token->type = CHARACTER;
@@ -1738,7 +1746,7 @@ peek_token_bracket (token, input, syntax
token->opr.c = c;
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1 &&
+ if (input->mb_cur_max > 1 &&
!re_string_first_byte (input, re_string_cur_idx (input)))
{
token->type = CHARACTER;
@@ -1976,7 +1984,7 @@ parse_expression (regexp, preg, token, s
return NULL;
}
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (dfa->mb_cur_max > 1)
{
while (!re_string_eoi (regexp)
&& !re_string_first_byte (regexp, re_string_cur_idx (regexp)))
@@ -2111,7 +2119,7 @@ parse_expression (regexp, preg, token, s
*err = REG_ESPACE;
return NULL;
}
- if (MB_CUR_MAX > 1)
+ if (dfa->mb_cur_max > 1)
dfa->has_mb_node = 1;
break;
case OP_WORD:
@@ -2883,7 +2891,7 @@ parse_bracket_exp (regexp, dfa, token, s
goto parse_bracket_exp_free_return;
}
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (dfa->mb_cur_max > 1)
for (i = 0; i < SBC_MAX; ++i)
if (__btowc (i) == WEOF)
bitset_set (sbcset, i);
@@ -3042,8 +3050,8 @@ parse_bracket_exp (regexp, dfa, token, s
#ifdef RE_ENABLE_I18N
if (mbcset->nmbchars || mbcset->ncoll_syms || mbcset->nequiv_classes
- || mbcset->nranges || (MB_CUR_MAX > 1 && (mbcset->nchar_classes
- || mbcset->non_match)))
+ || mbcset->nranges || (dfa->mb_cur_max > 1 && (mbcset->nchar_classes
+ || mbcset->non_match)))
{
re_token_t alt_token;
bin_tree_t *mbc_tree;
@@ -3377,7 +3385,7 @@ build_charclass_op (dfa, trans, class_na
bitset_set(cset->sbcset, '\0');
*/
mbcset->non_match = 1;
- if (MB_CUR_MAX > 1)
+ if (dfa->mb_cur_max > 1)
for (i = 0; i < SBC_MAX; ++i)
if (__btowc (i) == WEOF)
bitset_set (sbcset, i);
@@ -3423,7 +3431,7 @@ build_charclass_op (dfa, trans, class_na
goto build_word_op_espace;
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (dfa->mb_cur_max > 1)
{
re_token_t alt_token;
bin_tree_t *mbc_tree;
--- libc/posix/regex_internal.h.jj 2003-11-11 12:19:25.000000000 +0100
+++ libc/posix/regex_internal.h 2003-11-12 09:47:13.000000000 +0100
@@ -334,6 +334,8 @@ struct re_string_t
RE_TRANSLATE_TYPE trans;
/* 1 if REG_ICASE. */
unsigned int icase : 1;
+ unsigned int is_utf8 : 1;
+ int mb_cur_max;
};
typedef struct re_string_t re_string_t;
/* In case of REG_ICASE, we allocate the buffer dynamically for mbs. */
@@ -345,10 +347,12 @@ typedef struct re_string_t re_string_t;
static reg_errcode_t re_string_allocate (re_string_t *pstr, const char *str,
int len, int init_len,
- RE_TRANSLATE_TYPE trans, int icase);
+ RE_TRANSLATE_TYPE trans, int icase,
+ int mb_cur_max, int is_utf8);
static reg_errcode_t re_string_construct (re_string_t *pstr, const char *str,
int len, RE_TRANSLATE_TYPE trans,
- int icase);
+ int icase, int mb_cur_max,
+ int is_utf8);
static reg_errcode_t re_string_reconstruct (re_string_t *pstr, int idx,
int eflags, int newline);
static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
@@ -571,11 +575,7 @@ struct re_fail_stack_t
struct re_dfa_t
{
re_bitset_ptr_t word_char;
-
- /* number of subexpressions `re_nsub' is in regex_t. */
- int subexps_alloc;
re_subexp_t *subexps;
-
re_token_t *nodes;
int nodes_alloc;
int nodes_len;
@@ -586,11 +586,14 @@ struct re_dfa_t
re_node_set *eclosures;
re_node_set *inveclosures;
struct re_state_table_entry *state_table;
- unsigned int state_hash_mask;
re_dfastate_t *init_state;
re_dfastate_t *init_state_word;
re_dfastate_t *init_state_nl;
re_dfastate_t *init_state_begbuf;
+
+ /* number of subexpressions `re_nsub' is in regex_t. */
+ int subexps_alloc;
+ unsigned int state_hash_mask;
int states_alloc;
int init_node;
int nbackref; /* The number of backreference in this dfa. */
@@ -604,6 +607,8 @@ struct re_dfa_t
a node which can accept multibyte character or multi character
collating element. */
unsigned int has_mb_node : 1;
+ unsigned int is_utf8 : 1;
+ int mb_cur_max;
};
typedef struct re_dfa_t re_dfa_t;
@@ -700,7 +705,7 @@ re_string_char_size_at (pstr, idx)
int idx;
{
int byte_idx;
- if (MB_CUR_MAX == 1)
+ if (pstr->mb_cur_max == 1)
return 1;
for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
if (pstr->wcs[idx + byte_idx] != WEOF)
@@ -713,7 +718,7 @@ re_string_wchar_at (pstr, idx)
const re_string_t *pstr;
int idx;
{
- if (MB_CUR_MAX == 1)
+ if (pstr->mb_cur_max == 1)
return (wint_t) pstr->mbs[idx];
return (wint_t) pstr->wcs[idx];
}
--- libc/posix/regex_internal.c.jj 2003-11-12 08:26:45.000000000 +0100
+++ libc/posix/regex_internal.c 2003-11-12 09:48:11.000000000 +0100
@@ -20,7 +20,8 @@
static void re_string_construct_common (const char *str, int len,
re_string_t *pstr,
- RE_TRANSLATE_TYPE trans, int icase);
+ RE_TRANSLATE_TYPE trans, int icase,
+ int mb_cur_max, int is_utf8);
#ifdef RE_ENABLE_I18N
static int re_string_skip_chars (re_string_t *pstr, int new_raw_idx,
wint_t *last_wc);
@@ -46,15 +47,17 @@ static unsigned int inline calc_state_ha
re_string_reconstruct before using the object. */
static reg_errcode_t
-re_string_allocate (pstr, str, len, init_len, trans, icase)
+re_string_allocate (pstr, str, len, init_len, trans, icase,
+ mb_cur_max, is_utf8)
re_string_t *pstr;
const char *str;
- int len, init_len, icase;
+ int len, init_len, icase, mb_cur_max, is_utf8;
RE_TRANSLATE_TYPE trans;
{
reg_errcode_t ret;
int init_buf_len = (len + 1 < init_len) ? len + 1: init_len;
- re_string_construct_common (str, len, pstr, trans, icase);
+ re_string_construct_common (str, len, pstr, trans, icase,
+ mb_cur_max, is_utf8);
pstr->stop = pstr->len;
ret = re_string_realloc_buffers (pstr, init_buf_len);
@@ -65,21 +68,22 @@ re_string_allocate (pstr, str, len, init
: (unsigned char *) str);
pstr->mbs = MBS_ALLOCATED (pstr) ? pstr->mbs : pstr->mbs_case;
pstr->valid_len = (MBS_CASE_ALLOCATED (pstr) || MBS_ALLOCATED (pstr)
- || MB_CUR_MAX > 1) ? pstr->valid_len : len;
+ || mb_cur_max > 1) ? pstr->valid_len : len;
return REG_NOERROR;
}
/* This function allocate the buffers, and initialize them. */
static reg_errcode_t
-re_string_construct (pstr, str, len, trans, icase)
+re_string_construct (pstr, str, len, trans, icase, mb_cur_max, is_utf8)
re_string_t *pstr;
const char *str;
- int len, icase;
+ int len, icase, mb_cur_max, is_utf8;
RE_TRANSLATE_TYPE trans;
{
reg_errcode_t ret;
- re_string_construct_common (str, len, pstr, trans, icase);
+ re_string_construct_common (str, len, pstr, trans, icase,
+ mb_cur_max, is_utf8);
pstr->stop = pstr->len;
/* Set 0 so that this function can initialize whole buffers. */
pstr->valid_len = 0;
@@ -97,7 +101,7 @@ re_string_construct (pstr, str, len, tra
if (icase)
{
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (mb_cur_max > 1)
build_wcs_upper_buffer (pstr);
else
#endif /* RE_ENABLE_I18N */
@@ -106,7 +110,7 @@ re_string_construct (pstr, str, len, tra
else
{
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (mb_cur_max > 1)
build_wcs_buffer (pstr);
else
#endif /* RE_ENABLE_I18N */
@@ -131,7 +135,7 @@ re_string_realloc_buffers (pstr, new_buf
int new_buf_len;
{
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (pstr->mb_cur_max > 1)
{
wint_t *new_array = re_realloc (pstr->wcs, wint_t, new_buf_len);
if (BE (new_array == NULL, 0))
@@ -163,18 +167,20 @@ re_string_realloc_buffers (pstr, new_buf
static void
-re_string_construct_common (str, len, pstr, trans, icase)
+re_string_construct_common (str, len, pstr, trans, icase, mb_cur_max, is_utf8)
const char *str;
int len;
re_string_t *pstr;
RE_TRANSLATE_TYPE trans;
- int icase;
+ int icase, mb_cur_max, is_utf8;
{
memset (pstr, '\0', sizeof (re_string_t));
pstr->raw_mbs = (const unsigned char *) str;
pstr->len = len;
pstr->trans = trans;
pstr->icase = icase ? 1 : 0;
+ pstr->mb_cur_max = mb_cur_max;
+ pstr->is_utf8 = is_utf8;
}
#ifdef RE_ENABLE_I18N
@@ -372,7 +378,7 @@ re_string_translate_buffer (pstr)
}
/* This function re-construct the buffers.
- Concretely, convert to wide character in case of MB_CUR_MAX > 1,
+ Concretely, convert to wide character in case of pstr->mb_cur_max > 1,
convert to upper case in case of REG_ICASE, apply translation. */
static reg_errcode_t
@@ -385,7 +391,7 @@ re_string_reconstruct (pstr, idx, eflags
{
/* Reset buffer. */
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (pstr->mb_cur_max > 1)
memset (&pstr->cur_state, '\0', sizeof (mbstate_t));
#endif /* RE_ENABLE_I18N */
pstr->len += pstr->raw_mbs_idx;
@@ -409,7 +415,7 @@ re_string_reconstruct (pstr, idx, eflags
pstr->tip_context = re_string_context_at (pstr, offset - 1, eflags,
newline);
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (pstr->mb_cur_max > 1)
memmove (pstr->wcs, pstr->wcs + offset,
(pstr->valid_len - offset) * sizeof (wint_t));
#endif /* RE_ENABLE_I18N */
@@ -429,7 +435,7 @@ re_string_reconstruct (pstr, idx, eflags
/* No, skip all characters until IDX. */
pstr->valid_len = 0;
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (pstr->mb_cur_max > 1)
{
int wcs_idx;
wint_t wc;
@@ -467,7 +473,7 @@ re_string_reconstruct (pstr, idx, eflags
/* Then build the buffers. */
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (pstr->mb_cur_max > 1)
{
if (pstr->icase)
build_wcs_upper_buffer (pstr);
@@ -519,7 +525,7 @@ re_string_context_at (input, idx, eflags
: CONTEXT_NEWLINE | CONTEXT_ENDBUF);
}
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (input->mb_cur_max > 1)
{
wint_t wc;
int wc_idx = idx;
@@ -1157,7 +1163,7 @@ create_ci_newstate (dfa, nodes, hash)
newstate->halt = 1;
#ifdef RE_ENABLE_I18N
else if (type == COMPLEX_BRACKET
- || (type == OP_PERIOD && MB_CUR_MAX > 1))
+ || (type == OP_PERIOD && dfa->mb_cur_max > 1))
newstate->accept_mb = 1;
#endif /* RE_ENABLE_I18N */
else if (type == OP_BACK_REF)
@@ -1208,7 +1214,7 @@ create_cd_newstate (dfa, nodes, context,
newstate->halt = 1;
#ifdef RE_ENABLE_I18N
else if (type == COMPLEX_BRACKET
- || (type == OP_PERIOD && MB_CUR_MAX > 1))
+ || (type == OP_PERIOD && dfa->mb_cur_max > 1))
newstate->accept_mb = 1;
#endif /* RE_ENABLE_I18N */
else if (type == OP_BACK_REF)
--- libc/posix/regexec.c.jj 2003-10-07 11:55:50.000000000 +0200
+++ libc/posix/regexec.c 2003-11-12 09:22:45.000000000 +0100
@@ -605,7 +605,8 @@ re_search_internal (preg, string, length
fl_longest_match = (nmatch != 0 || dfa->nbackref);
err = re_string_allocate (&input, string, length, dfa->nodes_len + 1,
- preg->translate, preg->syntax & RE_ICASE);
+ preg->translate, preg->syntax & RE_ICASE,
+ dfa->mb_cur_max, dfa->is_utf8);
if (BE (err != REG_NOERROR, 0))
goto free_return;
input.stop = stop;
@@ -643,7 +644,7 @@ re_search_internal (preg, string, length
incr = (range < 0) ? -1 : 1;
left_lim = (range < 0) ? start + range : start;
right_lim = (range < 0) ? start : start + range;
- sb = MB_CUR_MAX == 1;
+ sb = dfa->mb_cur_max == 1;
fast_translate = sb || !(preg->syntax & RE_ICASE || preg->translate);
for (;;)
@@ -1018,7 +1019,7 @@ check_matching (preg, mctx, fl_search, f
/* Restart from initial state, since we are searching
the point from where matching start. */
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX == 1
+ if (dfa->mb_cur_max == 1
|| re_string_first_byte (mctx->input, cur_str_idx))
#endif /* RE_ENABLE_I18N */
cur_state = acquire_init_state_context (&err, preg, mctx,
@@ -2280,7 +2281,7 @@ transit_state_sb (err, preg, state, fl_s
{
#ifdef RE_ENABLE_I18N
int not_initial = 0;
- if (MB_CUR_MAX > 1)
+ if (dfa->mb_cur_max > 1)
for (node_cnt = 0; node_cnt < next_nodes.nelem; ++node_cnt)
if (dfa->nodes[next_nodes.elems[node_cnt]].type == CHARACTER)
{
@@ -3772,7 +3773,7 @@ extend_buffers (mctx)
if (pstr->icase)
{
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (pstr->mb_cur_max > 1)
build_wcs_upper_buffer (pstr);
else
#endif /* RE_ENABLE_I18N */
@@ -3781,7 +3782,7 @@ extend_buffers (mctx)
else
{
#ifdef RE_ENABLE_I18N
- if (MB_CUR_MAX > 1)
+ if (pstr->mb_cur_max > 1)
build_wcs_buffer (pstr);
else
#endif /* RE_ENABLE_I18N */
Jakub
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2003-11-12 13:44 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-11-12 13:44 [PATCH] Regex infrastructure for UTF-8 optimizations Jakub Jelinek
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).