/* String search routines for GNU Emacs.
- Copyright (C) 1985, 86, 87, 93, 94, 97, 1998 Free Software Foundation, Inc.
+ Copyright (C) 1985, 1986, 1987, 1993, 1994, 1997, 1998, 1999, 2002, 2003,
+ 2004, 2005 Free Software Foundation, Inc.
This file is part of GNU Emacs.
You should have received a copy of the GNU General Public License
along with GNU Emacs; see the file COPYING. If not, write to
-the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
-Boston, MA 02111-1307, USA. */
+the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
+Boston, MA 02110-1301, USA. */
#include <config.h>
struct regexp_cache
{
struct regexp_cache *next;
- Lisp_Object regexp;
+ Lisp_Object regexp, whitespace_regexp;
struct re_pattern_buffer buf;
char fastmap[0400];
/* Nonzero means regexp was compiled to do full POSIX backtracking. */
Lisp_Object Qinvalid_regexp;
+Lisp_Object Vsearch_spaces_regexp;
+
static void set_search_regs ();
static void save_search_regs ();
-
+static int simple_search ();
+static int boyer_moore ();
static int search_buffer ();
static void
error ("Stack overflow in regexp matcher");
}
-#ifdef __STDC__
-#define CONST const
-#else
-#define CONST
-#endif
-
/* Compile a regexp and signal a Lisp error if anything goes wrong.
PATTERN is the pattern to compile.
CP is the place to put the result.
- TRANSLATE is a translation table for ignoring case, or NULL for none.
+ TRANSLATE is a translation table for ignoring case, or nil for none.
REGP is the structure that says where to store the "register"
values that will result from matching this pattern.
If it is 0, we should compile the pattern not to record any
for this pattern. 0 means backtrack only enough to get a valid match.
MULTIBYTE is nonzero if we want to handle multibyte characters in
PATTERN. 0 means all multibyte characters are recognized just as
- sequences of binary data. */
+ sequences of binary data.
+
+ The behavior also depends on Vsearch_spaces_regexp. */
static void
compile_pattern_1 (cp, pattern, translate, regp, posix, multibyte)
struct regexp_cache *cp;
Lisp_Object pattern;
- Lisp_Object *translate;
+ Lisp_Object translate;
struct re_registers *regp;
int posix;
int multibyte;
{
- char *raw_pattern;
+ unsigned char *raw_pattern;
int raw_pattern_size;
char *val;
reg_syntax_t old;
if (multibyte == STRING_MULTIBYTE (pattern))
{
- raw_pattern = (char *) XSTRING (pattern)->data;
- raw_pattern_size = XSTRING (pattern)->size_byte;
+ raw_pattern = (unsigned char *) SDATA (pattern);
+ raw_pattern_size = SBYTES (pattern);
}
else if (multibyte)
{
- raw_pattern_size = count_size_as_multibyte (XSTRING (pattern)->data,
- XSTRING (pattern)->size);
- raw_pattern = (char *) alloca (raw_pattern_size + 1);
- copy_text (XSTRING (pattern)->data, raw_pattern,
- XSTRING (pattern)->size, 0, 1);
+ raw_pattern_size = count_size_as_multibyte (SDATA (pattern),
+ SCHARS (pattern));
+ raw_pattern = (unsigned char *) alloca (raw_pattern_size + 1);
+ copy_text (SDATA (pattern), raw_pattern,
+ SCHARS (pattern), 0, 1);
}
else
{
by subtracting nonascii-insert-offset from each non-ASCII char,
so that only the multibyte chars which really correspond to
the chosen single-byte character set can possibly match. */
- raw_pattern_size = XSTRING (pattern)->size;
- raw_pattern = (char *) alloca (raw_pattern_size + 1);
- copy_text (XSTRING (pattern)->data, raw_pattern,
- XSTRING (pattern)->size, 1, 0);
+ raw_pattern_size = SCHARS (pattern);
+ raw_pattern = (unsigned char *) alloca (raw_pattern_size + 1);
+ copy_text (SDATA (pattern), raw_pattern,
+ SBYTES (pattern), 1, 0);
}
cp->regexp = Qnil;
- cp->buf.translate = translate;
+ cp->buf.translate = (! NILP (translate) ? translate : make_number (0));
cp->posix = posix;
cp->buf.multibyte = multibyte;
+ cp->whitespace_regexp = Vsearch_spaces_regexp;
BLOCK_INPUT;
old = re_set_syntax (RE_SYNTAX_EMACS
| (posix ? 0 : RE_NO_POSIX_BACKTRACKING));
- val = (char *) re_compile_pattern (raw_pattern, raw_pattern_size, &cp->buf);
+
+ re_set_whitespace_regexp (NILP (Vsearch_spaces_regexp) ? NULL
+ : SDATA (Vsearch_spaces_regexp));
+
+ val = (char *) re_compile_pattern ((char *)raw_pattern,
+ raw_pattern_size, &cp->buf);
+
+ re_set_whitespace_regexp (NULL);
+
re_set_syntax (old);
UNBLOCK_INPUT;
if (val)
cp->regexp = Fcopy_sequence (pattern);
}
+/* Shrink each compiled regexp buffer in the cache
+ to the size actually used right now.
+ This is called from garbage collection. */
+
+void
+shrink_regexp_cache ()
+{
+ struct regexp_cache *cp;
+
+ for (cp = searchbuf_head; cp != 0; cp = cp->next)
+ {
+ cp->buf.allocated = cp->buf.used;
+ cp->buf.buffer
+ = (unsigned char *) xrealloc (cp->buf.buffer, cp->buf.used);
+ }
+}
+
/* Compile a regexp if necessary, but first check to see if there's one in
the cache.
PATTERN is the pattern to compile.
- TRANSLATE is a translation table for ignoring case, or NULL for none.
+ TRANSLATE is a translation table for ignoring case, or nil for none.
REGP is the structure that says where to store the "register"
values that will result from matching this pattern.
If it is 0, we should compile the pattern not to record any
compile_pattern (pattern, regp, translate, posix, multibyte)
Lisp_Object pattern;
struct re_registers *regp;
- Lisp_Object *translate;
+ Lisp_Object translate;
int posix, multibyte;
{
struct regexp_cache *cp, **cpp;
for (cpp = &searchbuf_head; ; cpp = &cp->next)
{
cp = *cpp;
- if (XSTRING (cp->regexp)->size == XSTRING (pattern)->size
+ /* Entries are initialized to nil, and may be set to nil by
+ compile_pattern_1 if the pattern isn't valid. Don't apply
+ string accessors in those cases. However, compile_pattern_1
+ is only applied to the cache entry we pick here to reuse. So
+ nil should never appear before a non-nil entry. */
+ if (NILP (cp->regexp))
+ goto compile_it;
+ if (SCHARS (cp->regexp) == SCHARS (pattern)
+ && STRING_MULTIBYTE (cp->regexp) == STRING_MULTIBYTE (pattern)
&& !NILP (Fstring_equal (cp->regexp, pattern))
- && cp->buf.translate == translate
+ && EQ (cp->buf.translate, (! NILP (translate) ? translate : make_number (0)))
&& cp->posix == posix
- && cp->buf.multibyte == multibyte)
+ && cp->buf.multibyte == multibyte
+ && !NILP (Fequal (cp->whitespace_regexp, Vsearch_spaces_regexp)))
break;
- /* If we're at the end of the cache, compile into the last cell. */
+ /* If we're at the end of the cache, compile into the nil cell
+ we found, or the last (least recently used) cell with a
+ string value. */
if (cp->next == 0)
{
+ compile_it:
compile_pattern_1 (cp, pattern, translate, regp, posix, multibyte);
break;
}
if (running_asynch_code)
save_search_regs ();
- CHECK_STRING (string, 0);
+ CHECK_STRING (string);
bufp = compile_pattern (string, &search_regs,
(!NILP (current_buffer->case_fold_search)
- ? XCHAR_TABLE (DOWNCASE_TABLE)->contents : 0),
+ ? current_buffer->case_canon_table : Qnil),
posix,
!NILP (current_buffer->enable_multibyte_characters));
}
re_match_object = Qnil;
-
+
i = re_match_2 (bufp, (char *) p1, s1, (char *) p2, s2,
PT_BYTE - BEGV_BYTE, &search_regs,
ZV_BYTE - BEGV_BYTE);
+ immediate_quit = 0;
+
if (i == -2)
matcher_overflow ();
= BYTE_TO_CHAR (search_regs.end[i] + BEGV_BYTE);
}
XSETBUFFER (last_thing_searched, current_buffer);
- immediate_quit = 0;
return val;
}
DEFUN ("looking-at", Flooking_at, Slooking_at, 1, 1, 0,
- "Return t if text after point matches regular expression REGEXP.\n\
-This function modifies the match data that `match-beginning',\n\
-`match-end' and `match-data' access; save and restore the match\n\
-data if you want to preserve them.")
- (regexp)
+ doc: /* Return t if text after point matches regular expression REGEXP.
+This function modifies the match data that `match-beginning',
+`match-end' and `match-data' access; save and restore the match
+data if you want to preserve them. */)
+ (regexp)
Lisp_Object regexp;
{
return looking_at_1 (regexp, 0);
}
DEFUN ("posix-looking-at", Fposix_looking_at, Sposix_looking_at, 1, 1, 0,
- "Return t if text after point matches regular expression REGEXP.\n\
-Find the longest match, in accord with Posix regular expression rules.\n\
-This function modifies the match data that `match-beginning',\n\
-`match-end' and `match-data' access; save and restore the match\n\
-data if you want to preserve them.")
- (regexp)
+ doc: /* Return t if text after point matches regular expression REGEXP.
+Find the longest match, in accord with Posix regular expression rules.
+This function modifies the match data that `match-beginning',
+`match-end' and `match-data' access; save and restore the match
+data if you want to preserve them. */)
+ (regexp)
Lisp_Object regexp;
{
return looking_at_1 (regexp, 1);
if (running_asynch_code)
save_search_regs ();
- CHECK_STRING (regexp, 0);
- CHECK_STRING (string, 1);
+ CHECK_STRING (regexp);
+ CHECK_STRING (string);
if (NILP (start))
pos = 0, pos_byte = 0;
else
{
- int len = XSTRING (string)->size;
+ int len = SCHARS (string);
- CHECK_NUMBER (start, 2);
+ CHECK_NUMBER (start);
pos = XINT (start);
if (pos < 0 && -pos <= len)
pos = len + pos;
bufp = compile_pattern (regexp, &search_regs,
(!NILP (current_buffer->case_fold_search)
- ? XCHAR_TABLE (DOWNCASE_TABLE)->contents : 0),
+ ? current_buffer->case_canon_table : Qnil),
posix,
STRING_MULTIBYTE (string));
immediate_quit = 1;
re_match_object = string;
-
- val = re_search (bufp, (char *) XSTRING (string)->data,
- XSTRING (string)->size_byte, pos_byte,
- XSTRING (string)->size_byte - pos_byte,
+
+ val = re_search (bufp, (char *) SDATA (string),
+ SBYTES (string), pos_byte,
+ SBYTES (string) - pos_byte,
&search_regs);
immediate_quit = 0;
last_thing_searched = Qt;
}
DEFUN ("string-match", Fstring_match, Sstring_match, 2, 3, 0,
- "Return index of start of first match for REGEXP in STRING, or nil.\n\
-If third arg START is non-nil, start search at that index in STRING.\n\
-For index of first char beyond the match, do (match-end 0).\n\
-`match-end' and `match-beginning' also give indices of substrings\n\
-matched by parenthesis constructs in the pattern.")
- (regexp, string, start)
+ doc: /* Return index of start of first match for REGEXP in STRING, or nil.
+Case is ignored if `case-fold-search' is non-nil in the current buffer.
+If third arg START is non-nil, start search at that index in STRING.
+For index of first char beyond the match, do (match-end 0).
+`match-end' and `match-beginning' also give indices of substrings
+matched by parenthesis constructs in the pattern.
+
+You can use the function `match-string' to extract the substrings
+matched by the parenthesis constructions in REGEXP. */)
+ (regexp, string, start)
Lisp_Object regexp, string, start;
{
return string_match_1 (regexp, string, start, 0);
}
DEFUN ("posix-string-match", Fposix_string_match, Sposix_string_match, 2, 3, 0,
- "Return index of start of first match for REGEXP in STRING, or nil.\n\
-Find the longest match, in accord with Posix regular expression rules.\n\
-If third arg START is non-nil, start search at that index in STRING.\n\
-For index of first char beyond the match, do (match-end 0).\n\
-`match-end' and `match-beginning' also give indices of substrings\n\
-matched by parenthesis constructs in the pattern.")
- (regexp, string, start)
+ doc: /* Return index of start of first match for REGEXP in STRING, or nil.
+Find the longest match, in accord with Posix regular expression rules.
+Case is ignored if `case-fold-search' is non-nil in the current buffer.
+If third arg START is non-nil, start search at that index in STRING.
+For index of first char beyond the match, do (match-end 0).
+`match-end' and `match-beginning' also give indices of substrings
+matched by parenthesis constructs in the pattern. */)
+ (regexp, string, start)
Lisp_Object regexp, string, start;
{
return string_match_1 (regexp, string, start, 1);
int val;
struct re_pattern_buffer *bufp;
- bufp = compile_pattern (regexp, 0, 0, 0, STRING_MULTIBYTE (string));
+ bufp = compile_pattern (regexp, 0, Qnil,
+ 0, STRING_MULTIBYTE (string));
immediate_quit = 1;
re_match_object = string;
-
- val = re_search (bufp, (char *) XSTRING (string)->data,
- XSTRING (string)->size, 0, XSTRING (string)->size,
- 0);
+
+ val = re_search (bufp, (char *) SDATA (string),
+ SBYTES (string), 0,
+ SBYTES (string), 0);
immediate_quit = 0;
return val;
}
int
fast_c_string_match_ignore_case (regexp, string)
Lisp_Object regexp;
- char *string;
+ const char *string;
{
int val;
struct re_pattern_buffer *bufp;
regexp = string_make_unibyte (regexp);
re_match_object = Qt;
bufp = compile_pattern (regexp, 0,
- XCHAR_TABLE (Vascii_downcase_table)->contents, 0,
+ Vascii_canon_table, 0,
0);
immediate_quit = 1;
val = re_search (bufp, string, len, 0, len, 0);
immediate_quit = 0;
return val;
}
-\f
-/* max and min. */
-static int
-max (a, b)
- int a, b;
-{
- return ((a > b) ? a : b);
-}
+/* Like fast_string_match but ignore case. */
-static int
-min (a, b)
- int a, b;
+int
+fast_string_match_ignore_case (regexp, string)
+ Lisp_Object regexp, string;
{
- return ((a < b) ? a : b);
-}
+ int val;
+ struct re_pattern_buffer *bufp;
+
+ bufp = compile_pattern (regexp, 0, Vascii_canon_table,
+ 0, STRING_MULTIBYTE (string));
+ immediate_quit = 1;
+ re_match_object = string;
+ val = re_search (bufp, (char *) SDATA (string),
+ SBYTES (string), 0,
+ SBYTES (string), 0);
+ immediate_quit = 0;
+ return val;
+}
\f
/* The newline cache: remembering which sections of text have no newlines. */
direction indicated by COUNT.
If we find COUNT instances, set *SHORTAGE to zero, and return the
- position after the COUNTth match. Note that for reverse motion
+ position past the COUNTth match. Note that for reverse motion
this is not the same as the usual convention for Emacs motion commands.
If we don't find COUNT instances before reaching END, set *SHORTAGE
If ALLOW_QUIT is non-zero, set immediate_quit. That's good to do
except when inside redisplay. */
+int
scan_buffer (target, start, end, count, shortage, allow_quit)
register int target;
int start, end;
int allow_quit;
{
struct region_cache *newline_cache;
- int direction;
+ int direction;
if (count > 0)
{
examine. */
int ceiling_byte = CHAR_TO_BYTE (end) - 1;
int start_byte = CHAR_TO_BYTE (start);
+ int tem;
/* If we're looking for a newline, consult the newline cache
to see where we can avoid some scanning. */
bytes. BUFFER_CEILING_OF returns the last character
position that is contiguous, so the ceiling is the
position after that. */
- ceiling_byte = min (BUFFER_CEILING_OF (start_byte), ceiling_byte);
+ tem = BUFFER_CEILING_OF (start_byte);
+ ceiling_byte = min (tem, ceiling_byte);
{
- /* The termination address of the dumb loop. */
+ /* The termination address of the dumb loop. */
register unsigned char *ceiling_addr
= BYTE_POS_ADDR (ceiling_byte) + 1;
register unsigned char *cursor
/* The last character to check before the next obstacle. */
int ceiling_byte = CHAR_TO_BYTE (end);
int start_byte = CHAR_TO_BYTE (start);
+ int tem;
/* Consult the newline cache, if appropriate. */
if (target == '\n' && newline_cache)
}
/* Stop scanning before the gap. */
- ceiling_byte = max (BUFFER_FLOOR_OF (start_byte - 1), ceiling_byte);
+ tem = BUFFER_FLOOR_OF (start_byte - 1);
+ ceiling_byte = max (tem, ceiling_byte);
{
/* The termination address of the dumb loop. */
int old_immediate_quit = immediate_quit;
- /* If we are not in selective display mode,
- check only for newlines. */
- int selective_display = (!NILP (current_buffer->selective_display)
- && !INTEGERP (current_buffer->selective_display));
-
/* The code that follows is like scan_buffer
but checks for either newline or carriage return. */
}
else
{
- int start_byte = CHAR_TO_BYTE (start);
while (start_byte > limit_byte)
{
ceiling = BUFFER_FLOOR_OF (start_byte - 1);
if (shortage == 0)
pos--;
-
+
return pos;
}
\f
int posix;
{
register int np;
- int lim;
+ int lim, lim_byte;
int n = direction;
if (!NILP (count))
{
- CHECK_NUMBER (count, 3);
+ CHECK_NUMBER (count);
n *= XINT (count);
}
- CHECK_STRING (string, 0);
+ CHECK_STRING (string);
if (NILP (bound))
- lim = n > 0 ? ZV : BEGV;
+ {
+ if (n > 0)
+ lim = ZV, lim_byte = ZV_BYTE;
+ else
+ lim = BEGV, lim_byte = BEGV_BYTE;
+ }
else
{
- CHECK_NUMBER_COERCE_MARKER (bound, 1);
+ CHECK_NUMBER_COERCE_MARKER (bound);
lim = XINT (bound);
if (n > 0 ? lim < PT : lim > PT)
error ("Invalid search bound (wrong side of point)");
if (lim > ZV)
- lim = ZV;
- if (lim < BEGV)
- lim = BEGV;
+ lim = ZV, lim_byte = ZV_BYTE;
+ else if (lim < BEGV)
+ lim = BEGV, lim_byte = BEGV_BYTE;
+ else
+ lim_byte = CHAR_TO_BYTE (lim);
}
- np = search_buffer (string, PT, lim, n, RE,
+ np = search_buffer (string, PT, PT_BYTE, lim, lim_byte, n, RE,
(!NILP (current_buffer->case_fold_search)
- ? XCHAR_TABLE (current_buffer->case_canon_table)->contents
- : 0),
+ ? current_buffer->case_canon_table
+ : Qnil),
(!NILP (current_buffer->case_fold_search)
- ? XCHAR_TABLE (current_buffer->case_eqv_table)->contents
- : 0),
+ ? current_buffer->case_eqv_table
+ : Qnil),
posix);
if (np <= 0)
{
{
if (lim < BEGV || lim > ZV)
abort ();
- SET_PT (lim);
+ SET_PT_BOTH (lim, lim_byte);
return Qnil;
#if 0 /* This would be clean, but maybe programs depend on
a value of nil here. */
trivial_regexp_p (regexp)
Lisp_Object regexp;
{
- int len = XSTRING (regexp)->size;
- unsigned char *s = XSTRING (regexp)->data;
- unsigned char c;
+ int len = SBYTES (regexp);
+ unsigned char *s = SDATA (regexp);
while (--len >= 0)
{
switch (*s++)
{
case '|': case '(': case ')': case '`': case '\'': case 'b':
case 'B': case '<': case '>': case 'w': case 'W': case 's':
- case 'S': case '=':
+ case 'S': case '=': case '{': case '}': case '_':
case 'c': case 'C': /* for categoryspec and notcategoryspec */
case '1': case '2': case '3': case '4': case '5':
case '6': case '7': case '8': case '9':
If N is positive, searching is forward and LIM must be greater than POS.
If N is negative, searching is backward and LIM must be less than POS.
- Returns -x if only N-x occurrences found (x > 0),
+ Returns -x if x occurrences remain to be found (x > 0),
or else the position at the beginning of the Nth occurrence
(if searching backward) or the end (if searching forward).
POSIX is nonzero if we want full backtracking (POSIX style)
for this pattern. 0 means backtrack only enough to get a valid match. */
+#define TRANSLATE(out, trt, d) \
+do \
+ { \
+ if (! NILP (trt)) \
+ { \
+ Lisp_Object temp; \
+ temp = Faref (trt, make_number (d)); \
+ if (INTEGERP (temp)) \
+ out = XINT (temp); \
+ else \
+ out = d; \
+ } \
+ else \
+ out = d; \
+ } \
+while (0)
+
static int
-search_buffer (string, pos, lim, n, RE, trt, inverse_trt, posix)
+search_buffer (string, pos, pos_byte, lim, lim_byte, n,
+ RE, trt, inverse_trt, posix)
Lisp_Object string;
int pos;
+ int pos_byte;
int lim;
+ int lim_byte;
int n;
int RE;
- Lisp_Object *trt;
- Lisp_Object *inverse_trt;
+ Lisp_Object trt;
+ Lisp_Object inverse_trt;
int posix;
{
- int len = XSTRING (string)->size;
- int len_byte = XSTRING (string)->size_byte;
- unsigned char *base_pat = XSTRING (string)->data;
- register int *BM_tab;
- int *BM_tab_base;
- register int direction = ((n > 0) ? 1 : -1);
- register int dirlen;
- int infinity, limit, k, stride_for_teases;
- register unsigned char *pat, *cursor, *p_limit;
- register int i, j;
- unsigned char *p1, *p2;
- int s1, s2;
+ int len = SCHARS (string);
+ int len_byte = SBYTES (string);
+ register int i;
if (running_asynch_code)
save_search_regs ();
+ /* Searching 0 times means don't move. */
/* Null string is found at starting position. */
- if (len == 0)
+ if (len == 0 || n == 0)
{
- set_search_regs (pos, 0);
+ set_search_regs (pos_byte, 0);
return pos;
}
- /* Searching 0 times means don't move. */
- if (n == 0)
- return pos;
-
- if (RE && !trivial_regexp_p (string))
+ if (RE && !(trivial_regexp_p (string) && NILP (Vsearch_spaces_regexp)))
{
+ unsigned char *p1, *p2;
+ int s1, s2;
struct re_pattern_buffer *bufp;
bufp = compile_pattern (string, &search_regs, trt, posix,
s2 = 0;
}
re_match_object = Qnil;
-
+
while (n < 0)
{
int val;
val = re_search_2 (bufp, (char *) p1, s1, (char *) p2, s2,
- pos - BEGV, lim - pos, &search_regs,
+ pos_byte - BEGV_BYTE, lim_byte - pos_byte,
+ &search_regs,
/* Don't allow match past current point */
- pos - BEGV);
+ pos_byte - BEGV_BYTE);
if (val == -2)
{
matcher_overflow ();
}
if (val >= 0)
{
+ pos_byte = search_regs.start[0] + BEGV_BYTE;
for (i = 0; i < search_regs.num_regs; i++)
if (search_regs.start[i] >= 0)
{
{
int val;
val = re_search_2 (bufp, (char *) p1, s1, (char *) p2, s2,
- pos - BEGV, lim - pos, &search_regs,
- lim - BEGV);
+ pos_byte - BEGV_BYTE, lim_byte - pos_byte,
+ &search_regs,
+ lim_byte - BEGV_BYTE);
if (val == -2)
{
matcher_overflow ();
}
if (val >= 0)
{
+ pos_byte = search_regs.end[0] + BEGV_BYTE;
for (i = 0; i < search_regs.num_regs; i++)
if (search_regs.start[i] >= 0)
{
}
else /* non-RE case */
{
- int pos_byte = CHAR_TO_BYTE (pos);
- int lim_byte = CHAR_TO_BYTE (lim);
-#ifdef C_ALLOCA
- int BM_tab_space[0400];
- BM_tab = &BM_tab_space[0];
-#else
- BM_tab = (int *) alloca (0400 * sizeof (int));
-#endif
- {
- unsigned char *raw_pattern;
- int raw_pattern_size;
- unsigned char *patbuf;
- int multibyte = !NILP (current_buffer->enable_multibyte_characters);
+ unsigned char *raw_pattern, *pat;
+ int raw_pattern_size;
+ int raw_pattern_size_byte;
+ unsigned char *patbuf;
+ int multibyte = !NILP (current_buffer->enable_multibyte_characters);
+ unsigned char *base_pat = SDATA (string);
+ /* Set to nozero if we find a non-ASCII char that need
+ translation. */
+ int charset_base = 0;
+ int boyer_moore_ok = 1;
+
+ /* MULTIBYTE says whether the text to be searched is multibyte.
+ We must convert PATTERN to match that, or we will not really
+ find things right. */
+
+ if (multibyte == STRING_MULTIBYTE (string))
+ {
+ raw_pattern = (unsigned char *) SDATA (string);
+ raw_pattern_size = SCHARS (string);
+ raw_pattern_size_byte = SBYTES (string);
+ }
+ else if (multibyte)
+ {
+ raw_pattern_size = SCHARS (string);
+ raw_pattern_size_byte
+ = count_size_as_multibyte (SDATA (string),
+ raw_pattern_size);
+ raw_pattern = (unsigned char *) alloca (raw_pattern_size_byte + 1);
+ copy_text (SDATA (string), raw_pattern,
+ SCHARS (string), 0, 1);
+ }
+ else
+ {
+ /* Converting multibyte to single-byte.
+
+ ??? Perhaps this conversion should be done in a special way
+ by subtracting nonascii-insert-offset from each non-ASCII char,
+ so that only the multibyte chars which really correspond to
+ the chosen single-byte character set can possibly match. */
+ raw_pattern_size = SCHARS (string);
+ raw_pattern_size_byte = SCHARS (string);
+ raw_pattern = (unsigned char *) alloca (raw_pattern_size + 1);
+ copy_text (SDATA (string), raw_pattern,
+ SBYTES (string), 1, 0);
+ }
+
+ /* Copy and optionally translate the pattern. */
+ len = raw_pattern_size;
+ len_byte = raw_pattern_size_byte;
+ patbuf = (unsigned char *) alloca (len_byte);
+ pat = patbuf;
+ base_pat = raw_pattern;
+ if (multibyte)
+ {
+ /* Fill patbuf by translated characters in STRING while
+ checking if we can use boyer-moore search. If TRT is
+ non-nil, we can use boyer-moore search only if TRT can be
+ represented by the byte array of 256 elements. For that,
+ all non-ASCII case-equivalents of all case-senstive
+ characters in STRING must belong to the same charset and
+ row. */
+
+ while (--len >= 0)
+ {
+ unsigned char str_base[MAX_MULTIBYTE_LENGTH], *str;
+ int c, translated, inverse;
+ int in_charlen, charlen;
+
+ /* If we got here and the RE flag is set, it's because we're
+ dealing with a regexp known to be trivial, so the backslash
+ just quotes the next character. */
+ if (RE && *base_pat == '\\')
+ {
+ len--;
+ raw_pattern_size--;
+ len_byte--;
+ base_pat++;
+ }
+
+ c = STRING_CHAR_AND_LENGTH (base_pat, len_byte, in_charlen);
- /* MULTIBYTE says whether the text to be searched is multibyte.
- We must convert PATTERN to match that, or we will not really
- find things right. */
+ if (NILP (trt))
+ {
+ str = base_pat;
+ charlen = in_charlen;
+ }
+ else
+ {
+ /* Translate the character. */
+ TRANSLATE (translated, trt, c);
+ charlen = CHAR_STRING (translated, str_base);
+ str = str_base;
+
+ /* Check if C has any other case-equivalents. */
+ TRANSLATE (inverse, inverse_trt, c);
+ /* If so, check if we can use boyer-moore. */
+ if (c != inverse && boyer_moore_ok)
+ {
+ /* Check if all equivalents belong to the same
+ charset & row. Note that the check of C
+ itself is done by the last iteration. Note
+ also that we don't have to check ASCII
+ characters because boyer-moore search can
+ always handle their translation. */
+ while (1)
+ {
+ if (! ASCII_BYTE_P (inverse))
+ {
+ if (SINGLE_BYTE_CHAR_P (inverse))
+ {
+ /* Boyer-moore search can't handle a
+ translation of an eight-bit
+ character. */
+ boyer_moore_ok = 0;
+ break;
+ }
+ else if (charset_base == 0)
+ charset_base = inverse & ~CHAR_FIELD3_MASK;
+ else if ((inverse & ~CHAR_FIELD3_MASK)
+ != charset_base)
+ {
+ boyer_moore_ok = 0;
+ break;
+ }
+ }
+ if (c == inverse)
+ break;
+ TRANSLATE (inverse, inverse_trt, inverse);
+ }
+ }
+ }
+
+ /* Store this character into the translated pattern. */
+ bcopy (str, pat, charlen);
+ pat += charlen;
+ base_pat += in_charlen;
+ len_byte -= in_charlen;
+ }
+ }
+ else
+ {
+ /* Unibyte buffer. */
+ charset_base = 0;
+ while (--len >= 0)
+ {
+ int c, translated;
+
+ /* If we got here and the RE flag is set, it's because we're
+ dealing with a regexp known to be trivial, so the backslash
+ just quotes the next character. */
+ if (RE && *base_pat == '\\')
+ {
+ len--;
+ raw_pattern_size--;
+ base_pat++;
+ }
+ c = *base_pat++;
+ TRANSLATE (translated, trt, c);
+ *pat++ = translated;
+ }
+ }
+
+ len_byte = pat - patbuf;
+ len = raw_pattern_size;
+ pat = base_pat = patbuf;
+
+ if (boyer_moore_ok)
+ return boyer_moore (n, pat, len, len_byte, trt, inverse_trt,
+ pos, pos_byte, lim, lim_byte,
+ charset_base);
+ else
+ return simple_search (n, pat, len, len_byte, trt,
+ pos, pos_byte, lim, lim_byte);
+ }
+}
+\f
+/* Do a simple string search N times for the string PAT,
+ whose length is LEN/LEN_BYTE,
+ from buffer position POS/POS_BYTE until LIM/LIM_BYTE.
+ TRT is the translation table.
- if (multibyte == STRING_MULTIBYTE (string))
+ Return the character position where the match is found.
+ Otherwise, if M matches remained to be found, return -M.
+
+ This kind of search works regardless of what is in PAT and
+ regardless of what is in TRT. It is used in cases where
+ boyer_moore cannot work. */
+
+static int
+simple_search (n, pat, len, len_byte, trt, pos, pos_byte, lim, lim_byte)
+ int n;
+ unsigned char *pat;
+ int len, len_byte;
+ Lisp_Object trt;
+ int pos, pos_byte;
+ int lim, lim_byte;
+{
+ int multibyte = ! NILP (current_buffer->enable_multibyte_characters);
+ int forward = n > 0;
+
+ if (lim > pos && multibyte)
+ while (n > 0)
+ {
+ while (1)
{
- raw_pattern = (char *) XSTRING (string)->data;
- raw_pattern_size = XSTRING (string)->size_byte;
+ /* Try matching at position POS. */
+ int this_pos = pos;
+ int this_pos_byte = pos_byte;
+ int this_len = len;
+ int this_len_byte = len_byte;
+ unsigned char *p = pat;
+ if (pos + len > lim)
+ goto stop;
+
+ while (this_len > 0)
+ {
+ int charlen, buf_charlen;
+ int pat_ch, buf_ch;
+
+ pat_ch = STRING_CHAR_AND_LENGTH (p, this_len_byte, charlen);
+ buf_ch = STRING_CHAR_AND_LENGTH (BYTE_POS_ADDR (this_pos_byte),
+ ZV_BYTE - this_pos_byte,
+ buf_charlen);
+ TRANSLATE (buf_ch, trt, buf_ch);
+
+ if (buf_ch != pat_ch)
+ break;
+
+ this_len_byte -= charlen;
+ this_len--;
+ p += charlen;
+
+ this_pos_byte += buf_charlen;
+ this_pos++;
+ }
+
+ if (this_len == 0)
+ {
+ pos += len;
+ pos_byte += len_byte;
+ break;
+ }
+
+ INC_BOTH (pos, pos_byte);
}
- else if (multibyte)
+
+ n--;
+ }
+ else if (lim > pos)
+ while (n > 0)
+ {
+ while (1)
{
- raw_pattern_size = count_size_as_multibyte (XSTRING (string)->data,
- XSTRING (string)->size);
- raw_pattern = (char *) alloca (raw_pattern_size + 1);
- copy_text (XSTRING (string)->data, raw_pattern,
- XSTRING (string)->size, 0, 1);
+ /* Try matching at position POS. */
+ int this_pos = pos;
+ int this_len = len;
+ unsigned char *p = pat;
+
+ if (pos + len > lim)
+ goto stop;
+
+ while (this_len > 0)
+ {
+ int pat_ch = *p++;
+ int buf_ch = FETCH_BYTE (this_pos);
+ TRANSLATE (buf_ch, trt, buf_ch);
+
+ if (buf_ch != pat_ch)
+ break;
+
+ this_len--;
+ this_pos++;
+ }
+
+ if (this_len == 0)
+ {
+ pos += len;
+ break;
+ }
+
+ pos++;
}
- else
+
+ n--;
+ }
+ /* Backwards search. */
+ else if (lim < pos && multibyte)
+ while (n < 0)
+ {
+ while (1)
{
- /* Converting multibyte to single-byte.
-
- ??? Perhaps this conversion should be done in a special way
- by subtracting nonascii-insert-offset from each non-ASCII char,
- so that only the multibyte chars which really correspond to
- the chosen single-byte character set can possibly match. */
- raw_pattern_size = XSTRING (string)->size;
- raw_pattern = (char *) alloca (raw_pattern_size + 1);
- copy_text (XSTRING (string)->data, raw_pattern,
- XSTRING (string)->size, 1, 0);
+ /* Try matching at position POS. */
+ int this_pos = pos - len;
+ int this_pos_byte = pos_byte - len_byte;
+ int this_len = len;
+ int this_len_byte = len_byte;
+ unsigned char *p = pat;
+
+ if (pos - len < lim)
+ goto stop;
+
+ while (this_len > 0)
+ {
+ int charlen, buf_charlen;
+ int pat_ch, buf_ch;
+
+ pat_ch = STRING_CHAR_AND_LENGTH (p, this_len_byte, charlen);
+ buf_ch = STRING_CHAR_AND_LENGTH (BYTE_POS_ADDR (this_pos_byte),
+ ZV_BYTE - this_pos_byte,
+ buf_charlen);
+ TRANSLATE (buf_ch, trt, buf_ch);
+
+ if (buf_ch != pat_ch)
+ break;
+
+ this_len_byte -= charlen;
+ this_len--;
+ p += charlen;
+ this_pos_byte += buf_charlen;
+ this_pos++;
+ }
+
+ if (this_len == 0)
+ {
+ pos -= len;
+ pos_byte -= len_byte;
+ break;
+ }
+
+ DEC_BOTH (pos, pos_byte);
}
- len_byte = raw_pattern_size;
- patbuf = (unsigned char *) alloca (len_byte);
- pat = patbuf;
- base_pat = raw_pattern;
- while (--len_byte >= 0)
+ n++;
+ }
+ else if (lim < pos)
+ while (n < 0)
+ {
+ while (1)
{
- /* If we got here and the RE flag is set, it's because we're
- dealing with a regexp known to be trivial, so the backslash
- just quotes the next character. */
- if (RE && *base_pat == '\\')
+ /* Try matching at position POS. */
+ int this_pos = pos - len;
+ int this_len = len;
+ unsigned char *p = pat;
+
+ if (pos - len < lim)
+ goto stop;
+
+ while (this_len > 0)
{
- len_byte--;
- base_pat++;
+ int pat_ch = *p++;
+ int buf_ch = FETCH_BYTE (this_pos);
+ TRANSLATE (buf_ch, trt, buf_ch);
+
+ if (buf_ch != pat_ch)
+ break;
+ this_len--;
+ this_pos++;
}
- *pat++ = (trt ? XINT (trt[*base_pat++]) : *base_pat++);
+
+ if (this_len == 0)
+ {
+ pos -= len;
+ break;
+ }
+
+ pos--;
}
- len_byte = pat - patbuf;
- pat = base_pat = patbuf;
+
+ n++;
}
- /* The general approach is that we are going to maintain that we know */
- /* the first (closest to the present position, in whatever direction */
- /* we're searching) character that could possibly be the last */
- /* (furthest from present position) character of a valid match. We */
- /* advance the state of our knowledge by looking at that character */
- /* and seeing whether it indeed matches the last character of the */
- /* pattern. If it does, we take a closer look. If it does not, we */
- /* move our pointer (to putative last characters) as far as is */
- /* logically possible. This amount of movement, which I call a */
- /* stride, will be the length of the pattern if the actual character */
- /* appears nowhere in the pattern, otherwise it will be the distance */
- /* from the last occurrence of that character to the end of the */
- /* pattern. */
- /* As a coding trick, an enormous stride is coded into the table for */
- /* characters that match the last character. This allows use of only */
- /* a single test, a test for having gone past the end of the */
- /* permissible match region, to test for both possible matches (when */
- /* the stride goes past the end immediately) and failure to */
- /* match (where you get nudged past the end one stride at a time). */
-
- /* Here we make a "mickey mouse" BM table. The stride of the search */
- /* is determined only by the last character of the putative match. */
- /* If that character does not match, we will stride the proper */
- /* distance to propose a match that superimposes it on the last */
- /* instance of a character that matches it (per trt), or misses */
- /* it entirely if there is none. */
-
- dirlen = len_byte * direction;
- infinity = dirlen - (lim_byte + pos_byte + len_byte + len_byte) * direction;
- if (direction < 0)
- pat = (base_pat += len_byte - 1);
- BM_tab_base = BM_tab;
- BM_tab += 0400;
- j = dirlen; /* to get it in a register */
- /* A character that does not appear in the pattern induces a */
- /* stride equal to the pattern length. */
- while (BM_tab_base != BM_tab)
+
+ stop:
+ if (n == 0)
+ {
+ if (forward)
+ set_search_regs ((multibyte ? pos_byte : pos) - len_byte, len_byte);
+ else
+ set_search_regs (multibyte ? pos_byte : pos, len_byte);
+
+ return pos;
+ }
+ else if (n > 0)
+ return -n;
+ else
+ return n;
+}
+\f
+/* Do Boyer-Moore search N times for the string BASE_PAT,
+ whose length is LEN/LEN_BYTE,
+ from buffer position POS/POS_BYTE until LIM/LIM_BYTE.
+ DIRECTION says which direction we search in.
+ TRT and INVERSE_TRT are translation tables.
+ Characters in PAT are already translated by TRT.
+
+ This kind of search works if all the characters in BASE_PAT that
+ have nontrivial translation are the same aside from the last byte.
+ This makes it possible to translate just the last byte of a
+ character, and do so after just a simple test of the context.
+ CHARSET_BASE is nonzero iff there is such a non-ASCII character.
+
+ If that criterion is not satisfied, do not call this function. */
+
+static int
+boyer_moore (n, base_pat, len, len_byte, trt, inverse_trt,
+ pos, pos_byte, lim, lim_byte, charset_base)
+ int n;
+ unsigned char *base_pat;
+ int len, len_byte;
+ Lisp_Object trt;
+ Lisp_Object inverse_trt;
+ int pos, pos_byte;
+ int lim, lim_byte;
+ int charset_base;
+{
+ int direction = ((n > 0) ? 1 : -1);
+ register int dirlen;
+ int infinity, limit, stride_for_teases = 0;
+ register int *BM_tab;
+ int *BM_tab_base;
+ register unsigned char *cursor, *p_limit;
+ register int i, j;
+ unsigned char *pat, *pat_end;
+ int multibyte = ! NILP (current_buffer->enable_multibyte_characters);
+
+ unsigned char simple_translate[0400];
+ /* These are set to the preceding bytes of a byte to be translated
+ if charset_base is nonzero. As the maximum byte length of a
+ multibyte character is 4, we have to check at most three previous
+ bytes. */
+ int translate_prev_byte1 = 0;
+ int translate_prev_byte2 = 0;
+ int translate_prev_byte3 = 0;
+
+#ifdef C_ALLOCA
+ int BM_tab_space[0400];
+ BM_tab = &BM_tab_space[0];
+#else
+ BM_tab = (int *) alloca (0400 * sizeof (int));
+#endif
+ /* The general approach is that we are going to maintain that we know */
+ /* the first (closest to the present position, in whatever direction */
+ /* we're searching) character that could possibly be the last */
+ /* (furthest from present position) character of a valid match. We */
+ /* advance the state of our knowledge by looking at that character */
+ /* and seeing whether it indeed matches the last character of the */
+ /* pattern. If it does, we take a closer look. If it does not, we */
+ /* move our pointer (to putative last characters) as far as is */
+ /* logically possible. This amount of movement, which I call a */
+ /* stride, will be the length of the pattern if the actual character */
+ /* appears nowhere in the pattern, otherwise it will be the distance */
+ /* from the last occurrence of that character to the end of the */
+ /* pattern. */
+ /* As a coding trick, an enormous stride is coded into the table for */
+ /* characters that match the last character. This allows use of only */
+ /* a single test, a test for having gone past the end of the */
+ /* permissible match region, to test for both possible matches (when */
+ /* the stride goes past the end immediately) and failure to */
+ /* match (where you get nudged past the end one stride at a time). */
+
+ /* Here we make a "mickey mouse" BM table. The stride of the search */
+ /* is determined only by the last character of the putative match. */
+ /* If that character does not match, we will stride the proper */
+ /* distance to propose a match that superimposes it on the last */
+ /* instance of a character that matches it (per trt), or misses */
+ /* it entirely if there is none. */
+
+ dirlen = len_byte * direction;
+ infinity = dirlen - (lim_byte + pos_byte + len_byte + len_byte) * direction;
+
+ /* Record position after the end of the pattern. */
+ pat_end = base_pat + len_byte;
+ /* BASE_PAT points to a character that we start scanning from.
+ It is the first character in a forward search,
+ the last character in a backward search. */
+ if (direction < 0)
+ base_pat = pat_end - 1;
+
+ BM_tab_base = BM_tab;
+ BM_tab += 0400;
+ j = dirlen; /* to get it in a register */
+ /* A character that does not appear in the pattern induces a */
+ /* stride equal to the pattern length. */
+ while (BM_tab_base != BM_tab)
+ {
+ *--BM_tab = j;
+ *--BM_tab = j;
+ *--BM_tab = j;
+ *--BM_tab = j;
+ }
+
+ /* We use this for translation, instead of TRT itself.
+ We fill this in to handle the characters that actually
+ occur in the pattern. Others don't matter anyway! */
+ bzero (simple_translate, sizeof simple_translate);
+ for (i = 0; i < 0400; i++)
+ simple_translate[i] = i;
+
+ if (charset_base)
+ {
+ /* Setup translate_prev_byte1/2/3 from CHARSET_BASE. Only a
+ byte following them are the target of translation. */
+ int sample_char = charset_base | 0x20;
+ unsigned char str[MAX_MULTIBYTE_LENGTH];
+ int len = CHAR_STRING (sample_char, str);
+
+ translate_prev_byte1 = str[len - 2];
+ if (len > 2)
{
- *--BM_tab = j;
- *--BM_tab = j;
- *--BM_tab = j;
- *--BM_tab = j;
+ translate_prev_byte2 = str[len - 3];
+ if (len > 3)
+ translate_prev_byte3 = str[len - 4];
}
- i = 0;
- while (i != infinity)
+ }
+
+ i = 0;
+ while (i != infinity)
+ {
+ unsigned char *ptr = base_pat + i;
+ i += direction;
+ if (i == dirlen)
+ i = infinity;
+ if (! NILP (trt))
{
- j = pat[i]; i += direction;
- if (i == dirlen) i = infinity;
- if (trt != 0)
+ /* If the byte currently looking at is the last of a
+ character to check case-equivalents, set CH to that
+ character. An ASCII character and a non-ASCII character
+ matching with CHARSET_BASE are to be checked. */
+ int ch = -1;
+
+ if (ASCII_BYTE_P (*ptr) || ! multibyte)
+ ch = *ptr;
+ else if (charset_base
+ && (pat_end - ptr) == 1 || CHAR_HEAD_P (ptr[1]))
{
- k = (j = XINT (trt[j]));
- if (i == infinity)
- stride_for_teases = BM_tab[j];
- BM_tab[j] = dirlen - i;
- /* A translation table is accompanied by its inverse -- see */
- /* comment following downcase_table for details */
- while ((j = (unsigned char) XINT (inverse_trt[j])) != k)
- BM_tab[j] = dirlen - i;
+ unsigned char *charstart = ptr - 1;
+
+ while (! (CHAR_HEAD_P (*charstart)))
+ charstart--;
+ ch = STRING_CHAR (charstart, ptr - charstart + 1);
+ if (charset_base != (ch & ~CHAR_FIELD3_MASK))
+ ch = -1;
}
+
+ if (ch > 0400)
+ j = ((unsigned char) ch) | 0200;
else
+ j = *ptr;
+
+ if (i == infinity)
+ stride_for_teases = BM_tab[j];
+
+ BM_tab[j] = dirlen - i;
+ /* A translation table is accompanied by its inverse -- see */
+ /* comment following downcase_table for details */
+ if (ch >= 0)
{
- if (i == infinity)
- stride_for_teases = BM_tab[j];
- BM_tab[j] = dirlen - i;
+ int starting_ch = ch;
+ int starting_j = j;
+
+ while (1)
+ {
+ TRANSLATE (ch, inverse_trt, ch);
+ if (ch > 0400)
+ j = ((unsigned char) ch) | 0200;
+ else
+ j = (unsigned char) ch;
+
+ /* For all the characters that map into CH,
+ set up simple_translate to map the last byte
+ into STARTING_J. */
+ simple_translate[j] = starting_j;
+ if (ch == starting_ch)
+ break;
+ BM_tab[j] = dirlen - i;
+ }
}
- /* stride_for_teases tells how much to stride if we get a */
- /* match on the far character but are subsequently */
- /* disappointed, by recording what the stride would have been */
- /* for that character if the last character had been */
- /* different. */
}
- infinity = dirlen - infinity;
- pos_byte += dirlen - ((direction > 0) ? direction : 0);
- /* loop invariant - POS_BYTE points at where last char (first
- char if reverse) of pattern would align in a possible match. */
- while (n != 0)
+ else
+ {
+ j = *ptr;
+
+ if (i == infinity)
+ stride_for_teases = BM_tab[j];
+ BM_tab[j] = dirlen - i;
+ }
+ /* stride_for_teases tells how much to stride if we get a */
+ /* match on the far character but are subsequently */
+ /* disappointed, by recording what the stride would have been */
+ /* for that character if the last character had been */
+ /* different. */
+ }
+ infinity = dirlen - infinity;
+ pos_byte += dirlen - ((direction > 0) ? direction : 0);
+ /* loop invariant - POS_BYTE points at where last char (first
+ char if reverse) of pattern would align in a possible match. */
+ while (n != 0)
+ {
+ int tail_end;
+ unsigned char *tail_end_ptr;
+
+ /* It's been reported that some (broken) compiler thinks that
+ Boolean expressions in an arithmetic context are unsigned.
+ Using an explicit ?1:0 prevents this. */
+ if ((lim_byte - pos_byte - ((direction > 0) ? 1 : 0)) * direction
+ < 0)
+ return (n * (0 - direction));
+ /* First we do the part we can by pointers (maybe nothing) */
+ QUIT;
+ pat = base_pat;
+ limit = pos_byte - dirlen + direction;
+ if (direction > 0)
{
- /* It's been reported that some (broken) compiler thinks that
- Boolean expressions in an arithmetic context are unsigned.
- Using an explicit ?1:0 prevents this. */
- if ((lim_byte - pos_byte - ((direction > 0) ? 1 : 0)) * direction
- < 0)
- return (n * (0 - direction));
- /* First we do the part we can by pointers (maybe nothing) */
- QUIT;
- pat = base_pat;
- limit = pos_byte - dirlen + direction;
- limit = ((direction > 0)
- ? BUFFER_CEILING_OF (limit)
- : BUFFER_FLOOR_OF (limit));
+ limit = BUFFER_CEILING_OF (limit);
/* LIMIT is now the last (not beyond-last!) value POS_BYTE
can take on without hitting edge of buffer or the gap. */
- limit = ((direction > 0)
- ? min (lim_byte - 1, min (limit, pos_byte + 20000))
- : max (lim_byte, max (limit, pos_byte - 20000)));
- if ((limit - pos_byte) * direction > 20)
+ limit = min (limit, pos_byte + 20000);
+ limit = min (limit, lim_byte - 1);
+ }
+ else
+ {
+ limit = BUFFER_FLOOR_OF (limit);
+ /* LIMIT is now the last (not beyond-last!) value POS_BYTE
+ can take on without hitting edge of buffer or the gap. */
+ limit = max (limit, pos_byte - 20000);
+ limit = max (limit, lim_byte);
+ }
+ tail_end = BUFFER_CEILING_OF (pos_byte) + 1;
+ tail_end_ptr = BYTE_POS_ADDR (tail_end);
+
+ if ((limit - pos_byte) * direction > 20)
+ {
+ unsigned char *p2;
+
+ p_limit = BYTE_POS_ADDR (limit);
+ p2 = (cursor = BYTE_POS_ADDR (pos_byte));
+ /* In this loop, pos + cursor - p2 is the surrogate for pos */
+ while (1) /* use one cursor setting as long as i can */
{
- p_limit = BYTE_POS_ADDR (limit);
- p2 = (cursor = BYTE_POS_ADDR (pos_byte));
- /* In this loop, pos + cursor - p2 is the surrogate for pos */
- while (1) /* use one cursor setting as long as i can */
+ if (direction > 0) /* worth duplicating */
{
- if (direction > 0) /* worth duplicating */
- {
- /* Use signed comparison if appropriate
- to make cursor+infinity sure to be > p_limit.
- Assuming that the buffer lies in a range of addresses
- that are all "positive" (as ints) or all "negative",
- either kind of comparison will work as long
- as we don't step by infinity. So pick the kind
- that works when we do step by infinity. */
- if ((EMACS_INT) (p_limit + infinity) > (EMACS_INT) p_limit)
- while ((EMACS_INT) cursor <= (EMACS_INT) p_limit)
- cursor += BM_tab[*cursor];
- else
- while ((EMACS_UINT) cursor <= (EMACS_UINT) p_limit)
- cursor += BM_tab[*cursor];
- }
+ /* Use signed comparison if appropriate
+ to make cursor+infinity sure to be > p_limit.
+ Assuming that the buffer lies in a range of addresses
+ that are all "positive" (as ints) or all "negative",
+ either kind of comparison will work as long
+ as we don't step by infinity. So pick the kind
+ that works when we do step by infinity. */
+ if ((EMACS_INT) (p_limit + infinity) > (EMACS_INT) p_limit)
+ while ((EMACS_INT) cursor <= (EMACS_INT) p_limit)
+ cursor += BM_tab[*cursor];
else
- {
- if ((EMACS_INT) (p_limit + infinity) < (EMACS_INT) p_limit)
- while ((EMACS_INT) cursor >= (EMACS_INT) p_limit)
- cursor += BM_tab[*cursor];
- else
- while ((EMACS_UINT) cursor >= (EMACS_UINT) p_limit)
- cursor += BM_tab[*cursor];
- }
+ while ((EMACS_UINT) cursor <= (EMACS_UINT) p_limit)
+ cursor += BM_tab[*cursor];
+ }
+ else
+ {
+ if ((EMACS_INT) (p_limit + infinity) < (EMACS_INT) p_limit)
+ while ((EMACS_INT) cursor >= (EMACS_INT) p_limit)
+ cursor += BM_tab[*cursor];
+ else
+ while ((EMACS_UINT) cursor >= (EMACS_UINT) p_limit)
+ cursor += BM_tab[*cursor];
+ }
/* If you are here, cursor is beyond the end of the searched region. */
- /* This can happen if you match on the far character of the pattern, */
- /* because the "stride" of that character is infinity, a number able */
- /* to throw you well beyond the end of the search. It can also */
- /* happen if you fail to match within the permitted region and would */
- /* otherwise try a character beyond that region */
- if ((cursor - p_limit) * direction <= len_byte)
- break; /* a small overrun is genuine */
- cursor -= infinity; /* large overrun = hit */
- i = dirlen - direction;
- if (trt != 0)
+/* This can happen if you match on the far character of the pattern, */
+/* because the "stride" of that character is infinity, a number able */
+/* to throw you well beyond the end of the search. It can also */
+/* happen if you fail to match within the permitted region and would */
+/* otherwise try a character beyond that region */
+ if ((cursor - p_limit) * direction <= len_byte)
+ break; /* a small overrun is genuine */
+ cursor -= infinity; /* large overrun = hit */
+ i = dirlen - direction;
+ if (! NILP (trt))
+ {
+ while ((i -= direction) + direction != 0)
{
- while ((i -= direction) + direction != 0)
- if (pat[i] != XINT (trt[*(cursor -= direction)]))
- break;
+ int ch;
+ cursor -= direction;
+ /* Translate only the last byte of a character. */
+ if (! multibyte
+ || ((cursor == tail_end_ptr
+ || CHAR_HEAD_P (cursor[1]))
+ && (CHAR_HEAD_P (cursor[0])
+ /* Check if this is the last byte of
+ a translable character. */
+ || (translate_prev_byte1 == cursor[-1]
+ && (CHAR_HEAD_P (translate_prev_byte1)
+ || (translate_prev_byte2 == cursor[-2]
+ && (CHAR_HEAD_P (translate_prev_byte2)
+ || (translate_prev_byte3 == cursor[-3]))))))))
+ ch = simple_translate[*cursor];
+ else
+ ch = *cursor;
+ if (pat[i] != ch)
+ break;
}
- else
+ }
+ else
+ {
+ while ((i -= direction) + direction != 0)
{
- while ((i -= direction) + direction != 0)
- if (pat[i] != *(cursor -= direction))
- break;
+ cursor -= direction;
+ if (pat[i] != *cursor)
+ break;
}
- cursor += dirlen - i - direction; /* fix cursor */
- if (i + direction == 0)
- {
- int position;
+ }
+ cursor += dirlen - i - direction; /* fix cursor */
+ if (i + direction == 0)
+ {
+ int position;
- cursor -= direction;
+ cursor -= direction;
- position = pos_byte + cursor - p2 + ((direction > 0)
- ? 1 - len_byte : 0);
- set_search_regs (position, len_byte);
+ position = pos_byte + cursor - p2 + ((direction > 0)
+ ? 1 - len_byte : 0);
+ set_search_regs (position, len_byte);
- if ((n -= direction) != 0)
- cursor += dirlen; /* to resume search */
- else
- return ((direction > 0)
- ? search_regs.end[0] : search_regs.start[0]);
- }
+ if ((n -= direction) != 0)
+ cursor += dirlen; /* to resume search */
else
- cursor += stride_for_teases; /* <sigh> we lose - */
+ return ((direction > 0)
+ ? search_regs.end[0] : search_regs.start[0]);
}
- pos_byte += cursor - p2;
+ else
+ cursor += stride_for_teases; /* <sigh> we lose - */
}
- else
- /* Now we'll pick up a clump that has to be done the hard */
- /* way because it covers a discontinuity */
+ pos_byte += cursor - p2;
+ }
+ else
+ /* Now we'll pick up a clump that has to be done the hard */
+ /* way because it covers a discontinuity */
+ {
+ limit = ((direction > 0)
+ ? BUFFER_CEILING_OF (pos_byte - dirlen + 1)
+ : BUFFER_FLOOR_OF (pos_byte - dirlen - 1));
+ limit = ((direction > 0)
+ ? min (limit + len_byte, lim_byte - 1)
+ : max (limit - len_byte, lim_byte));
+ /* LIMIT is now the last value POS_BYTE can have
+ and still be valid for a possible match. */
+ while (1)
{
- limit = ((direction > 0)
- ? BUFFER_CEILING_OF (pos_byte - dirlen + 1)
- : BUFFER_FLOOR_OF (pos_byte - dirlen - 1));
- limit = ((direction > 0)
- ? min (limit + len_byte, lim_byte - 1)
- : max (limit - len_byte, lim_byte));
- /* LIMIT is now the last value POS_BYTE can have
- and still be valid for a possible match. */
- while (1)
+ /* This loop can be coded for space rather than */
+ /* speed because it will usually run only once. */
+ /* (the reach is at most len + 21, and typically */
+ /* does not exceed len) */
+ while ((limit - pos_byte) * direction >= 0)
+ pos_byte += BM_tab[FETCH_BYTE (pos_byte)];
+ /* now run the same tests to distinguish going off the */
+ /* end, a match or a phony match. */
+ if ((pos_byte - limit) * direction <= len_byte)
+ break; /* ran off the end */
+ /* Found what might be a match.
+ Set POS_BYTE back to last (first if reverse) pos. */
+ pos_byte -= infinity;
+ i = dirlen - direction;
+ while ((i -= direction) + direction != 0)
{
- /* This loop can be coded for space rather than */
- /* speed because it will usually run only once. */
- /* (the reach is at most len + 21, and typically */
- /* does not exceed len) */
- while ((limit - pos_byte) * direction >= 0)
- pos_byte += BM_tab[FETCH_BYTE (pos_byte)];
- /* now run the same tests to distinguish going off the */
- /* end, a match or a phony match. */
- if ((pos_byte - limit) * direction <= len_byte)
- break; /* ran off the end */
- /* Found what might be a match.
- Set POS_BYTE back to last (first if reverse) pos. */
- pos_byte -= infinity;
- i = dirlen - direction;
- while ((i -= direction) + direction != 0)
- {
- pos_byte -= direction;
- if (pat[i] != (trt != 0
- ? XINT (trt[FETCH_BYTE (pos_byte)])
- : FETCH_BYTE (pos_byte)))
- break;
- }
- /* Above loop has moved POS_BYTE part or all the way
- back to the first pos (last pos if reverse).
- Set it once again at the last (first if reverse) char. */
- pos_byte += dirlen - i- direction;
- if (i + direction == 0)
- {
- int position;
- pos_byte -= direction;
+ int ch;
+ unsigned char *ptr;
+ pos_byte -= direction;
+ ptr = BYTE_POS_ADDR (pos_byte);
+ /* Translate only the last byte of a character. */
+ if (! multibyte
+ || ((ptr == tail_end_ptr
+ || CHAR_HEAD_P (ptr[1]))
+ && (CHAR_HEAD_P (ptr[0])
+ /* Check if this is the last byte of a
+ translable character. */
+ || (translate_prev_byte1 == ptr[-1]
+ && (CHAR_HEAD_P (translate_prev_byte1)
+ || (translate_prev_byte2 == ptr[-2]
+ && (CHAR_HEAD_P (translate_prev_byte2)
+ || translate_prev_byte3 == ptr[-3])))))))
+ ch = simple_translate[*ptr];
+ else
+ ch = *ptr;
+ if (pat[i] != ch)
+ break;
+ }
+ /* Above loop has moved POS_BYTE part or all the way
+ back to the first pos (last pos if reverse).
+ Set it once again at the last (first if reverse) char. */
+ pos_byte += dirlen - i- direction;
+ if (i + direction == 0)
+ {
+ int position;
+ pos_byte -= direction;
- position = pos_byte + ((direction > 0) ? 1 - len_byte : 0);
+ position = pos_byte + ((direction > 0) ? 1 - len_byte : 0);
- set_search_regs (position, len_byte);
+ set_search_regs (position, len_byte);
- if ((n -= direction) != 0)
- pos_byte += dirlen; /* to resume search */
- else
- return ((direction > 0)
- ? search_regs.end[0] : search_regs.start[0]);
- }
+ if ((n -= direction) != 0)
+ pos_byte += dirlen; /* to resume search */
else
- pos_byte += stride_for_teases;
+ return ((direction > 0)
+ ? search_regs.end[0] : search_regs.start[0]);
}
- }
- /* We have done one clump. Can we continue? */
- if ((lim_byte - pos_byte) * direction < 0)
- return ((0 - n) * direction);
- }
- return BYTE_TO_CHAR (pos_byte);
+ else
+ pos_byte += stride_for_teases;
+ }
+ }
+ /* We have done one clump. Can we continue? */
+ if ((lim_byte - pos_byte) * direction < 0)
+ return ((0 - n) * direction);
}
+ return BYTE_TO_CHAR (pos_byte);
}
/* Record beginning BEG_BYTE and end BEG_BYTE + NBYTES
- for a match just found in the current buffer. */
+ for the overall match just found in the current buffer.
+ Also clear out the match data for registers 1 and up. */
static void
set_search_regs (beg_byte, nbytes)
int beg_byte, nbytes;
{
+ int i;
+
/* Make sure we have registers in which to store
the match position. */
if (search_regs.num_regs == 0)
search_regs.num_regs = 2;
}
+ /* Clear out the other registers. */
+ for (i = 1; i < search_regs.num_regs; i++)
+ {
+ search_regs.start[i] = -1;
+ search_regs.end[i] = -1;
+ }
+
search_regs.start[0] = BYTE_TO_CHAR (beg_byte);
search_regs.end[0] = BYTE_TO_CHAR (beg_byte + nbytes);
XSETBUFFER (last_thing_searched, current_buffer);
int prev_c = 0;
int adjust;
- CHECK_STRING (string, 0);
- p = XSTRING (string)->data;
- len = XSTRING (string)->size;
+ CHECK_STRING (string);
+ p = SDATA (string);
+ len = SCHARS (string);
for (i = 0, i_byte = 0; i < len; )
{
int c;
-
- if (STRING_MULTIBYTE (string))
- FETCH_STRING_CHAR_ADVANCE (c, string, i, i_byte);
- else
- c = XSTRING (string)->data[i++];
+
+ FETCH_STRING_CHAR_ADVANCE (c, string, i, i_byte);
if (SYNTAX (c) != Sword)
{
if (SYNTAX (prev_c) == Sword)
word_count++;
if (!word_count)
- return build_string ("");
+ return empty_string;
adjust = - punct_count + 5 * (word_count - 1) + 4;
- val = make_uninit_multibyte_string (len + adjust,
- XSTRING (string)->size_byte + adjust);
+ if (STRING_MULTIBYTE (string))
+ val = make_uninit_multibyte_string (len + adjust,
+ SBYTES (string)
+ + adjust);
+ else
+ val = make_uninit_string (len + adjust);
- o = XSTRING (val)->data;
+ o = SDATA (val);
*o++ = '\\';
*o++ = 'b';
+ prev_c = 0;
- for (i = 0; i < XSTRING (val)->size_byte; i++)
- if (SYNTAX (p[i]) == Sword)
- *o++ = p[i];
- else if (i > 0 && SYNTAX (p[i-1]) == Sword && --word_count)
- {
- *o++ = '\\';
- *o++ = 'W';
- *o++ = '\\';
- *o++ = 'W';
- *o++ = '*';
- }
+ for (i = 0, i_byte = 0; i < len; )
+ {
+ int c;
+ int i_byte_orig = i_byte;
+
+ FETCH_STRING_CHAR_ADVANCE (c, string, i, i_byte);
+
+ if (SYNTAX (c) == Sword)
+ {
+ bcopy (SDATA (string) + i_byte_orig, o,
+ i_byte - i_byte_orig);
+ o += i_byte - i_byte_orig;
+ }
+ else if (i > 0 && SYNTAX (prev_c) == Sword && --word_count)
+ {
+ *o++ = '\\';
+ *o++ = 'W';
+ *o++ = '\\';
+ *o++ = 'W';
+ *o++ = '*';
+ }
+
+ prev_c = c;
+ }
*o++ = '\\';
*o++ = 'b';
}
\f
DEFUN ("search-backward", Fsearch_backward, Ssearch_backward, 1, 4,
- "MSearch backward: ",
- "Search backward from point for STRING.\n\
-Set point to the beginning of the occurrence found, and return point.\n\
-An optional second argument bounds the search; it is a buffer position.\n\
-The match found must not extend before that position.\n\
-Optional third argument, if t, means if fail just return nil (no error).\n\
- If not nil and not t, position at limit of search and return nil.\n\
-Optional fourth argument is repeat count--search for successive occurrences.\n\
-See also the functions `match-beginning', `match-end' and `replace-match'.")
- (string, bound, noerror, count)
+ "MSearch backward: ",
+ doc: /* Search backward from point for STRING.
+Set point to the beginning of the occurrence found, and return point.
+An optional second argument bounds the search; it is a buffer position.
+The match found must not extend before that position.
+Optional third argument, if t, means if fail just return nil (no error).
+ If not nil and not t, position at limit of search and return nil.
+Optional fourth argument is repeat count--search for successive occurrences.
+
+Search case-sensitivity is determined by the value of the variable
+`case-fold-search', which see.
+
+See also the functions `match-beginning', `match-end' and `replace-match'. */)
+ (string, bound, noerror, count)
Lisp_Object string, bound, noerror, count;
{
return search_command (string, bound, noerror, count, -1, 0, 0);
}
DEFUN ("search-forward", Fsearch_forward, Ssearch_forward, 1, 4, "MSearch: ",
- "Search forward from point for STRING.\n\
-Set point to the end of the occurrence found, and return point.\n\
-An optional second argument bounds the search; it is a buffer position.\n\
-The match found must not extend after that position. nil is equivalent\n\
- to (point-max).\n\
-Optional third argument, if t, means if fail just return nil (no error).\n\
- If not nil and not t, move to limit of search and return nil.\n\
-Optional fourth argument is repeat count--search for successive occurrences.\n\
-See also the functions `match-beginning', `match-end' and `replace-match'.")
- (string, bound, noerror, count)
+ doc: /* Search forward from point for STRING.
+Set point to the end of the occurrence found, and return point.
+An optional second argument bounds the search; it is a buffer position.
+The match found must not extend after that position. nil is equivalent
+ to (point-max).
+Optional third argument, if t, means if fail just return nil (no error).
+ If not nil and not t, move to limit of search and return nil.
+Optional fourth argument is repeat count--search for successive occurrences.
+
+Search case-sensitivity is determined by the value of the variable
+`case-fold-search', which see.
+
+See also the functions `match-beginning', `match-end' and `replace-match'. */)
+ (string, bound, noerror, count)
Lisp_Object string, bound, noerror, count;
{
return search_command (string, bound, noerror, count, 1, 0, 0);
}
DEFUN ("word-search-backward", Fword_search_backward, Sword_search_backward, 1, 4,
- "sWord search backward: ",
- "Search backward from point for STRING, ignoring differences in punctuation.\n\
-Set point to the beginning of the occurrence found, and return point.\n\
-An optional second argument bounds the search; it is a buffer position.\n\
-The match found must not extend before that position.\n\
-Optional third argument, if t, means if fail just return nil (no error).\n\
- If not nil and not t, move to limit of search and return nil.\n\
-Optional fourth argument is repeat count--search for successive occurrences.")
- (string, bound, noerror, count)
+ "sWord search backward: ",
+ doc: /* Search backward from point for STRING, ignoring differences in punctuation.
+Set point to the beginning of the occurrence found, and return point.
+An optional second argument bounds the search; it is a buffer position.
+The match found must not extend before that position.
+Optional third argument, if t, means if fail just return nil (no error).
+ If not nil and not t, move to limit of search and return nil.
+Optional fourth argument is repeat count--search for successive occurrences. */)
+ (string, bound, noerror, count)
Lisp_Object string, bound, noerror, count;
{
return search_command (wordify (string), bound, noerror, count, -1, 1, 0);
}
DEFUN ("word-search-forward", Fword_search_forward, Sword_search_forward, 1, 4,
- "sWord search: ",
- "Search forward from point for STRING, ignoring differences in punctuation.\n\
-Set point to the end of the occurrence found, and return point.\n\
-An optional second argument bounds the search; it is a buffer position.\n\
-The match found must not extend after that position.\n\
-Optional third argument, if t, means if fail just return nil (no error).\n\
- If not nil and not t, move to limit of search and return nil.\n\
-Optional fourth argument is repeat count--search for successive occurrences.")
- (string, bound, noerror, count)
+ "sWord search: ",
+ doc: /* Search forward from point for STRING, ignoring differences in punctuation.
+Set point to the end of the occurrence found, and return point.
+An optional second argument bounds the search; it is a buffer position.
+The match found must not extend after that position.
+Optional third argument, if t, means if fail just return nil (no error).
+ If not nil and not t, move to limit of search and return nil.
+Optional fourth argument is repeat count--search for successive occurrences. */)
+ (string, bound, noerror, count)
Lisp_Object string, bound, noerror, count;
{
return search_command (wordify (string), bound, noerror, count, 1, 1, 0);
}
DEFUN ("re-search-backward", Fre_search_backward, Sre_search_backward, 1, 4,
- "sRE search backward: ",
- "Search backward from point for match for regular expression REGEXP.\n\
-Set point to the beginning of the match, and return point.\n\
-The match found is the one starting last in the buffer\n\
-and yet ending before the origin of the search.\n\
-An optional second argument bounds the search; it is a buffer position.\n\
-The match found must start at or after that position.\n\
-Optional third argument, if t, means if fail just return nil (no error).\n\
- If not nil and not t, move to limit of search and return nil.\n\
-Optional fourth argument is repeat count--search for successive occurrences.\n\
-See also the functions `match-beginning', `match-end' and `replace-match'.")
- (regexp, bound, noerror, count)
+ "sRE search backward: ",
+ doc: /* Search backward from point for match for regular expression REGEXP.
+Set point to the beginning of the match, and return point.
+The match found is the one starting last in the buffer
+and yet ending before the origin of the search.
+An optional second argument bounds the search; it is a buffer position.
+The match found must start at or after that position.
+Optional third argument, if t, means if fail just return nil (no error).
+ If not nil and not t, move to limit of search and return nil.
+Optional fourth argument is repeat count--search for successive occurrences.
+See also the functions `match-beginning', `match-end', `match-string',
+and `replace-match'. */)
+ (regexp, bound, noerror, count)
Lisp_Object regexp, bound, noerror, count;
{
return search_command (regexp, bound, noerror, count, -1, 1, 0);
}
DEFUN ("re-search-forward", Fre_search_forward, Sre_search_forward, 1, 4,
- "sRE search: ",
- "Search forward from point for regular expression REGEXP.\n\
-Set point to the end of the occurrence found, and return point.\n\
-An optional second argument bounds the search; it is a buffer position.\n\
-The match found must not extend after that position.\n\
-Optional third argument, if t, means if fail just return nil (no error).\n\
- If not nil and not t, move to limit of search and return nil.\n\
-Optional fourth argument is repeat count--search for successive occurrences.\n\
-See also the functions `match-beginning', `match-end' and `replace-match'.")
- (regexp, bound, noerror, count)
+ "sRE search: ",
+ doc: /* Search forward from point for regular expression REGEXP.
+Set point to the end of the occurrence found, and return point.
+An optional second argument bounds the search; it is a buffer position.
+The match found must not extend after that position.
+Optional third argument, if t, means if fail just return nil (no error).
+ If not nil and not t, move to limit of search and return nil.
+Optional fourth argument is repeat count--search for successive occurrences.
+See also the functions `match-beginning', `match-end', `match-string',
+and `replace-match'. */)
+ (regexp, bound, noerror, count)
Lisp_Object regexp, bound, noerror, count;
{
return search_command (regexp, bound, noerror, count, 1, 1, 0);
}
DEFUN ("posix-search-backward", Fposix_search_backward, Sposix_search_backward, 1, 4,
- "sPosix search backward: ",
- "Search backward from point for match for regular expression REGEXP.\n\
-Find the longest match in accord with Posix regular expression rules.\n\
-Set point to the beginning of the match, and return point.\n\
-The match found is the one starting last in the buffer\n\
-and yet ending before the origin of the search.\n\
-An optional second argument bounds the search; it is a buffer position.\n\
-The match found must start at or after that position.\n\
-Optional third argument, if t, means if fail just return nil (no error).\n\
- If not nil and not t, move to limit of search and return nil.\n\
-Optional fourth argument is repeat count--search for successive occurrences.\n\
-See also the functions `match-beginning', `match-end' and `replace-match'.")
- (regexp, bound, noerror, count)
+ "sPosix search backward: ",
+ doc: /* Search backward from point for match for regular expression REGEXP.
+Find the longest match in accord with Posix regular expression rules.
+Set point to the beginning of the match, and return point.
+The match found is the one starting last in the buffer
+and yet ending before the origin of the search.
+An optional second argument bounds the search; it is a buffer position.
+The match found must start at or after that position.
+Optional third argument, if t, means if fail just return nil (no error).
+ If not nil and not t, move to limit of search and return nil.
+Optional fourth argument is repeat count--search for successive occurrences.
+See also the functions `match-beginning', `match-end', `match-string',
+and `replace-match'. */)
+ (regexp, bound, noerror, count)
Lisp_Object regexp, bound, noerror, count;
{
return search_command (regexp, bound, noerror, count, -1, 1, 1);
}
DEFUN ("posix-search-forward", Fposix_search_forward, Sposix_search_forward, 1, 4,
- "sPosix search: ",
- "Search forward from point for regular expression REGEXP.\n\
-Find the longest match in accord with Posix regular expression rules.\n\
-Set point to the end of the occurrence found, and return point.\n\
-An optional second argument bounds the search; it is a buffer position.\n\
-The match found must not extend after that position.\n\
-Optional third argument, if t, means if fail just return nil (no error).\n\
- If not nil and not t, move to limit of search and return nil.\n\
-Optional fourth argument is repeat count--search for successive occurrences.\n\
-See also the functions `match-beginning', `match-end' and `replace-match'.")
- (regexp, bound, noerror, count)
+ "sPosix search: ",
+ doc: /* Search forward from point for regular expression REGEXP.
+Find the longest match in accord with Posix regular expression rules.
+Set point to the end of the occurrence found, and return point.
+An optional second argument bounds the search; it is a buffer position.
+The match found must not extend after that position.
+Optional third argument, if t, means if fail just return nil (no error).
+ If not nil and not t, move to limit of search and return nil.
+Optional fourth argument is repeat count--search for successive occurrences.
+See also the functions `match-beginning', `match-end', `match-string',
+and `replace-match'. */)
+ (regexp, bound, noerror, count)
Lisp_Object regexp, bound, noerror, count;
{
return search_command (regexp, bound, noerror, count, 1, 1, 1);
}
\f
DEFUN ("replace-match", Freplace_match, Sreplace_match, 1, 5, 0,
- "Replace text matched by last search with NEWTEXT.\n\
-If second arg FIXEDCASE is non-nil, do not alter case of replacement text.\n\
-Otherwise maybe capitalize the whole text, or maybe just word initials,\n\
-based on the replaced text.\n\
-If the replaced text has only capital letters\n\
-and has at least one multiletter word, convert NEWTEXT to all caps.\n\
-If the replaced text has at least one word starting with a capital letter,\n\
-then capitalize each word in NEWTEXT.\n\n\
-If third arg LITERAL is non-nil, insert NEWTEXT literally.\n\
-Otherwise treat `\\' as special:\n\
- `\\&' in NEWTEXT means substitute original matched text.\n\
- `\\N' means substitute what matched the Nth `\\(...\\)'.\n\
- If Nth parens didn't match, substitute nothing.\n\
- `\\\\' means insert one `\\'.\n\
-FIXEDCASE and LITERAL are optional arguments.\n\
-Leaves point at end of replacement text.\n\
-\n\
-The optional fourth argument STRING can be a string to modify.\n\
-In that case, this function creates and returns a new string\n\
-which is made by replacing the part of STRING that was matched.\n\
-\n\
-The optional fifth argument SUBEXP specifies a subexpression of the match.\n\
-It says to replace just that subexpression instead of the whole match.\n\
-This is useful only after a regular expression search or match\n\
-since only regular expressions have distinguished subexpressions.")
- (newtext, fixedcase, literal, string, subexp)
+ doc: /* Replace text matched by last search with NEWTEXT.
+Leave point at the end of the replacement text.
+
+If second arg FIXEDCASE is non-nil, do not alter case of replacement text.
+Otherwise maybe capitalize the whole text, or maybe just word initials,
+based on the replaced text.
+If the replaced text has only capital letters
+and has at least one multiletter word, convert NEWTEXT to all caps.
+Otherwise if all words are capitalized in the replaced text,
+capitalize each word in NEWTEXT.
+
+If third arg LITERAL is non-nil, insert NEWTEXT literally.
+Otherwise treat `\\' as special:
+ `\\&' in NEWTEXT means substitute original matched text.
+ `\\N' means substitute what matched the Nth `\\(...\\)'.
+ If Nth parens didn't match, substitute nothing.
+ `\\\\' means insert one `\\'.
+Case conversion does not apply to these substitutions.
+
+FIXEDCASE and LITERAL are optional arguments.
+
+The optional fourth argument STRING can be a string to modify.
+This is meaningful when the previous match was done against STRING,
+using `string-match'. When used this way, `replace-match'
+creates and returns a new string made by copying STRING and replacing
+the part of STRING that was matched.
+
+The optional fifth argument SUBEXP specifies a subexpression;
+it says to replace just that subexpression with NEWTEXT,
+rather than replacing the entire matched text.
+This is, in a vague sense, the inverse of using `\\N' in NEWTEXT;
+`\\N' copies subexp N into NEWTEXT, but using N as SUBEXP puts
+NEWTEXT in place of subexp N.
+This is useful only after a regular expression search or match,
+since only regular expressions have distinguished subexpressions. */)
+ (newtext, fixedcase, literal, string, subexp)
Lisp_Object newtext, fixedcase, literal, string, subexp;
{
enum { nochange, all_caps, cap_initial } case_action;
- register int pos, last;
+ register int pos, pos_byte;
int some_multiletter_word;
int some_lowercase;
int some_uppercase;
int some_nonuppercase_initial;
register int c, prevc;
- int inslen;
int sub;
int opoint, newpoint;
- CHECK_STRING (newtext, 0);
+ CHECK_STRING (newtext);
if (! NILP (string))
- CHECK_STRING (string, 4);
+ CHECK_STRING (string);
case_action = nochange; /* We tried an initialization */
/* but some C compilers blew it */
if (search_regs.num_regs <= 0)
- error ("replace-match called before any match found");
+ error ("`replace-match' called before any match found");
if (NILP (subexp))
sub = 0;
else
{
- CHECK_NUMBER (subexp, 3);
+ CHECK_NUMBER (subexp);
sub = XINT (subexp);
if (sub < 0 || sub >= search_regs.num_regs)
args_out_of_range (subexp, make_number (search_regs.num_regs));
{
if (search_regs.start[sub] < 0
|| search_regs.start[sub] > search_regs.end[sub]
- || search_regs.end[sub] > XSTRING (string)->size)
+ || search_regs.end[sub] > SCHARS (string))
args_out_of_range (make_number (search_regs.start[sub]),
make_number (search_regs.end[sub]));
}
if (NILP (fixedcase))
{
- int beg;
/* Decide how to casify by examining the matched text. */
+ int last;
- if (NILP (string))
- last = CHAR_TO_BYTE (search_regs.end[sub]);
- else
- last = search_regs.end[sub];
+ pos = search_regs.start[sub];
+ last = search_regs.end[sub];
if (NILP (string))
- beg = CHAR_TO_BYTE (search_regs.start[sub]);
+ pos_byte = CHAR_TO_BYTE (pos);
else
- beg = search_regs.start[sub];
+ pos_byte = string_char_to_byte (string, pos);
prevc = '\n';
case_action = all_caps;
some_nonuppercase_initial = 0;
some_uppercase = 0;
- for (pos = beg; pos < last; pos++)
+ while (pos < last)
{
if (NILP (string))
- c = FETCH_BYTE (pos);
+ {
+ c = FETCH_CHAR (pos_byte);
+ INC_BOTH (pos, pos_byte);
+ }
else
- c = XSTRING (string)->data[pos];
+ FETCH_STRING_CHAR_ADVANCE (c, string, pos, pos_byte);
if (LOWERCASEP (c))
{
if desired. */
if (NILP (literal))
{
- int lastpos = -1;
- int lastpos_byte = -1;
+ int lastpos = 0;
+ int lastpos_byte = 0;
/* We build up the substituted string in ACCUM. */
Lisp_Object accum;
Lisp_Object middle;
- int pos_byte;
+ int length = SBYTES (newtext);
accum = Qnil;
- for (pos_byte = 0, pos = 0; pos_byte < XSTRING (newtext)->size_byte;)
+ for (pos_byte = 0, pos = 0; pos_byte < length;)
{
int substart = -1;
- int subend;
+ int subend = 0;
int delbackslash = 0;
FETCH_STRING_CHAR_ADVANCE (c, newtext, pos, pos_byte);
if (c == '\\')
{
FETCH_STRING_CHAR_ADVANCE (c, newtext, pos, pos_byte);
+
if (c == '&')
{
substart = search_regs.start[sub];
subend = search_regs.end[sub];
}
- else if (c >= '1' && c <= '9' && c <= search_regs.num_regs + '0')
+ else if (c >= '1' && c <= '9')
{
- if (search_regs.start[c - '0'] >= 0)
+ if (search_regs.start[c - '0'] >= 0
+ && c <= search_regs.num_regs + '0')
{
substart = search_regs.start[c - '0'];
subend = search_regs.end[c - '0'];
}
+ else
+ {
+ /* If that subexp did not match,
+ replace \\N with nothing. */
+ substart = 0;
+ subend = 0;
+ }
}
else if (c == '\\')
delbackslash = 1;
}
if (substart >= 0)
{
- if (pos - 1 != lastpos + 1)
- middle = substring_both (newtext, lastpos + 1,
- lastpos_byte + 1,
- pos - 1, pos_byte - 1);
+ if (pos - 2 != lastpos)
+ middle = substring_both (newtext, lastpos,
+ lastpos_byte,
+ pos - 2, pos_byte - 2);
else
middle = Qnil;
accum = concat3 (accum, middle,
}
else if (delbackslash)
{
- middle = substring_both (newtext, lastpos + 1,
- lastpos_byte + 1,
- pos, pos_byte);
+ middle = substring_both (newtext, lastpos,
+ lastpos_byte,
+ pos - 1, pos_byte - 1);
accum = concat2 (accum, middle);
lastpos = pos;
}
}
- if (pos != lastpos + 1)
- middle = substring_both (newtext, lastpos + 1,
- lastpos_byte + 1,
+ if (pos != lastpos)
+ middle = substring_both (newtext, lastpos,
+ lastpos_byte,
pos, pos_byte);
else
middle = Qnil;
return concat3 (before, newtext, after);
}
- /* Record point, the move (quietly) to the start of the match. */
- if (PT > search_regs.start[sub])
+ /* Record point, then move (quietly) to the start of the match. */
+ if (PT >= search_regs.end[sub])
opoint = PT - ZV;
+ else if (PT > search_regs.start[sub])
+ opoint = search_regs.end[sub] - ZV;
else
opoint = PT;
- TEMP_SET_PT (search_regs.start[sub]);
-
- /* We insert the replacement text before the old text, and then
- delete the original text. This means that markers at the
- beginning or end of the original will float to the corresponding
- position in the replacement. */
- if (!NILP (literal))
- Finsert_and_inherit (1, &newtext);
- else
+ /* If we want non-literal replacement,
+ perform substitution on the replacement string. */
+ if (NILP (literal))
{
- struct gcpro gcpro1;
- GCPRO1 (newtext);
-
- for (pos = 0; pos < XSTRING (newtext)->size; pos++)
+ int length = SBYTES (newtext);
+ unsigned char *substed;
+ int substed_alloc_size, substed_len;
+ int buf_multibyte = !NILP (current_buffer->enable_multibyte_characters);
+ int str_multibyte = STRING_MULTIBYTE (newtext);
+ Lisp_Object rev_tbl;
+ int really_changed = 0;
+
+ rev_tbl= (!buf_multibyte && CHAR_TABLE_P (Vnonascii_translation_table)
+ ? Fchar_table_extra_slot (Vnonascii_translation_table,
+ make_number (0))
+ : Qnil);
+
+ substed_alloc_size = length * 2 + 100;
+ substed = (unsigned char *) xmalloc (substed_alloc_size + 1);
+ substed_len = 0;
+
+ /* Go thru NEWTEXT, producing the actual text to insert in
+ SUBSTED while adjusting multibyteness to that of the current
+ buffer. */
+
+ for (pos_byte = 0, pos = 0; pos_byte < length;)
{
- int offset = PT - search_regs.start[sub];
+ unsigned char str[MAX_MULTIBYTE_LENGTH];
+ unsigned char *add_stuff = NULL;
+ int add_len = 0;
+ int idx = -1;
+
+ if (str_multibyte)
+ {
+ FETCH_STRING_CHAR_ADVANCE_NO_CHECK (c, newtext, pos, pos_byte);
+ if (!buf_multibyte)
+ c = multibyte_char_to_unibyte (c, rev_tbl);
+ }
+ else
+ {
+ /* Note that we don't have to increment POS. */
+ c = SREF (newtext, pos_byte++);
+ if (buf_multibyte)
+ c = unibyte_char_to_multibyte (c);
+ }
+
+ /* Either set ADD_STUFF and ADD_LEN to the text to put in SUBSTED,
+ or set IDX to a match index, which means put that part
+ of the buffer text into SUBSTED. */
- c = XSTRING (newtext)->data[pos];
if (c == '\\')
{
- c = XSTRING (newtext)->data[++pos];
+ really_changed = 1;
+
+ if (str_multibyte)
+ {
+ FETCH_STRING_CHAR_ADVANCE_NO_CHECK (c, newtext,
+ pos, pos_byte);
+ if (!buf_multibyte && !SINGLE_BYTE_CHAR_P (c))
+ c = multibyte_char_to_unibyte (c, rev_tbl);
+ }
+ else
+ {
+ c = SREF (newtext, pos_byte++);
+ if (buf_multibyte)
+ c = unibyte_char_to_multibyte (c);
+ }
+
if (c == '&')
- Finsert_buffer_substring
- (Fcurrent_buffer (),
- make_number (search_regs.start[sub] + offset),
- make_number (search_regs.end[sub] + offset));
+ idx = sub;
else if (c >= '1' && c <= '9' && c <= search_regs.num_regs + '0')
{
if (search_regs.start[c - '0'] >= 1)
- Finsert_buffer_substring
- (Fcurrent_buffer (),
- make_number (search_regs.start[c - '0'] + offset),
- make_number (search_regs.end[c - '0'] + offset));
+ idx = c - '0';
}
else if (c == '\\')
- insert_char (c);
+ add_len = 1, add_stuff = "\\";
else
- error ("Invalid use of `\\' in replacement text");
+ {
+ xfree (substed);
+ error ("Invalid use of `\\' in replacement text");
+ }
+ }
+ else
+ {
+ add_len = CHAR_STRING (c, str);
+ add_stuff = str;
+ }
+
+ /* If we want to copy part of a previous match,
+ set up ADD_STUFF and ADD_LEN to point to it. */
+ if (idx >= 0)
+ {
+ int begbyte = CHAR_TO_BYTE (search_regs.start[idx]);
+ add_len = CHAR_TO_BYTE (search_regs.end[idx]) - begbyte;
+ if (search_regs.start[idx] < GPT && GPT < search_regs.end[idx])
+ move_gap (search_regs.start[idx]);
+ add_stuff = BYTE_POS_ADDR (begbyte);
+ }
+
+ /* Now the stuff we want to add to SUBSTED
+ is invariably ADD_LEN bytes starting at ADD_STUFF. */
+
+ /* Make sure SUBSTED is big enough. */
+ if (substed_len + add_len >= substed_alloc_size)
+ {
+ substed_alloc_size = substed_len + add_len + 500;
+ substed = (unsigned char *) xrealloc (substed,
+ substed_alloc_size + 1);
+ }
+
+ /* Now add to the end of SUBSTED. */
+ if (add_stuff)
+ {
+ bcopy (add_stuff, substed + substed_len, add_len);
+ substed_len += add_len;
+ }
+ }
+
+ if (really_changed)
+ {
+ if (buf_multibyte)
+ {
+ int nchars = multibyte_chars_in_text (substed, substed_len);
+
+ newtext = make_multibyte_string (substed, nchars, substed_len);
}
else
- insert_char (c);
+ newtext = make_unibyte_string (substed, substed_len);
}
- UNGCPRO;
+ xfree (substed);
}
- inslen = PT - (search_regs.start[sub]);
- del_range (search_regs.start[sub] + inslen, search_regs.end[sub] + inslen);
+ /* Replace the old text with the new in the cleanest possible way. */
+ replace_range (search_regs.start[sub], search_regs.end[sub],
+ newtext, 1, 0, 1);
+ newpoint = search_regs.start[sub] + SCHARS (newtext);
if (case_action == all_caps)
- Fupcase_region (make_number (PT - inslen), make_number (PT));
+ Fupcase_region (make_number (search_regs.start[sub]),
+ make_number (newpoint));
else if (case_action == cap_initial)
- Fupcase_initials_region (make_number (PT - inslen), make_number (PT));
+ Fupcase_initials_region (make_number (search_regs.start[sub]),
+ make_number (newpoint));
- newpoint = PT;
+ /* Adjust search data for this change. */
+ {
+ int oldend = search_regs.end[sub];
+ int oldstart = search_regs.start[sub];
+ int change = newpoint - search_regs.end[sub];
+ int i;
+
+ for (i = 0; i < search_regs.num_regs; i++)
+ {
+ if (search_regs.start[i] >= oldend)
+ search_regs.start[i] += change;
+ else if (search_regs.start[i] > oldstart)
+ search_regs.start[i] = oldstart;
+ if (search_regs.end[i] >= oldend)
+ search_regs.end[i] += change;
+ else if (search_regs.end[i] > oldstart)
+ search_regs.end[i] = oldstart;
+ }
+ }
/* Put point back where it was in the text. */
if (opoint <= 0)
/* Now move point "officially" to the start of the inserted replacement. */
move_if_not_intangible (newpoint);
-
+
return Qnil;
}
\f
{
register int n;
- CHECK_NUMBER (num, 0);
+ CHECK_NUMBER (num);
n = XINT (num);
- if (n < 0 || n >= search_regs.num_regs)
- args_out_of_range (num, make_number (search_regs.num_regs));
- if (search_regs.num_regs <= 0
+ if (n < 0)
+ args_out_of_range (num, make_number (0));
+ if (search_regs.num_regs <= 0)
+ error ("No match data, because no search succeeded");
+ if (n >= search_regs.num_regs
|| search_regs.start[n] < 0)
return Qnil;
return (make_number ((beginningp) ? search_regs.start[n]
}
DEFUN ("match-beginning", Fmatch_beginning, Smatch_beginning, 1, 1, 0,
- "Return position of start of text matched by last search.\n\
-SUBEXP, a number, specifies which parenthesized expression in the last\n\
- regexp.\n\
-Value is nil if SUBEXPth pair didn't match, or there were less than\n\
- SUBEXP pairs.\n\
-Zero means the entire text matched by the whole regexp or whole string.")
- (subexp)
+ doc: /* Return position of start of text matched by last search.
+SUBEXP, a number, specifies which parenthesized expression in the last
+ regexp.
+Value is nil if SUBEXPth pair didn't match, or there were less than
+ SUBEXP pairs.
+Zero means the entire text matched by the whole regexp or whole string. */)
+ (subexp)
Lisp_Object subexp;
{
return match_limit (subexp, 1);
}
DEFUN ("match-end", Fmatch_end, Smatch_end, 1, 1, 0,
- "Return position of end of text matched by last search.\n\
-SUBEXP, a number, specifies which parenthesized expression in the last\n\
- regexp.\n\
-Value is nil if SUBEXPth pair didn't match, or there were less than\n\
- SUBEXP pairs.\n\
-Zero means the entire text matched by the whole regexp or whole string.")
- (subexp)
+ doc: /* Return position of end of text matched by last search.
+SUBEXP, a number, specifies which parenthesized expression in the last
+ regexp.
+Value is nil if SUBEXPth pair didn't match, or there were less than
+ SUBEXP pairs.
+Zero means the entire text matched by the whole regexp or whole string. */)
+ (subexp)
Lisp_Object subexp;
{
return match_limit (subexp, 0);
-}
-
-DEFUN ("match-data", Fmatch_data, Smatch_data, 0, 2, 0,
- "Return a list containing all info on what the last search matched.\n\
-Element 2N is `(match-beginning N)'; element 2N + 1 is `(match-end N)'.\n\
-All the elements are markers or nil (nil if the Nth pair didn't match)\n\
-if the last match was on a buffer; integers or nil if a string was matched.\n\
-Use `store-match-data' to reinstate the data in this list.\n\
-\n\
-If INTEGERS (the optional first argument) is non-nil, always use integers\n\
-\(rather than markers) to represent buffer positions.\n\
-If REUSE is a list, reuse it as part of the value. If REUSE is long enough\n\
-to hold all the values, and if INTEGERS is non-nil, no consing is done.")
- (integers, reuse)
- Lisp_Object integers, reuse;
+}
+
+DEFUN ("match-data", Fmatch_data, Smatch_data, 0, 3, 0,
+ doc: /* Return a list containing all info on what the last search matched.
+Element 2N is `(match-beginning N)'; element 2N + 1 is `(match-end N)'.
+All the elements are markers or nil (nil if the Nth pair didn't match)
+if the last match was on a buffer; integers or nil if a string was matched.
+Use `store-match-data' to reinstate the data in this list.
+
+If INTEGERS (the optional first argument) is non-nil, always use
+integers \(rather than markers) to represent buffer positions. In
+this case, and if the last match was in a buffer, the buffer will get
+stored as one additional element at the end of the list.
+
+If REUSE is a list, reuse it as part of the value. If REUSE is long
+enough to hold all the values, and if INTEGERS is non-nil, no consing
+is done.
+
+If optional third arg RESEAT is non-nil, any previous markers on the
+REUSE list will be modified to point to nowhere.
+
+Return value is undefined if the last search failed. */)
+ (integers, reuse, reseat)
+ Lisp_Object integers, reuse, reseat;
{
Lisp_Object tail, prev;
Lisp_Object *data;
int i, len;
+ if (!NILP (reseat))
+ for (tail = reuse; CONSP (tail); tail = XCDR (tail))
+ if (MARKERP (XCAR (tail)))
+ {
+ unchain_marker (XMARKER (XCAR (tail)));
+ XSETCAR (tail, Qnil);
+ }
+
if (NILP (last_thing_searched))
return Qnil;
- data = (Lisp_Object *) alloca ((2 * search_regs.num_regs)
+ prev = Qnil;
+
+ data = (Lisp_Object *) alloca ((2 * search_regs.num_regs + 1)
* sizeof (Lisp_Object));
- len = -1;
+ len = 0;
for (i = 0; i < search_regs.num_regs; i++)
{
int start = search_regs.start[i];
last_thing_searched);
data[2 * i + 1] = Fmake_marker ();
Fset_marker (data[2 * i + 1],
- make_number (search_regs.end[i]),
+ make_number (search_regs.end[i]),
last_thing_searched);
}
else
/* last_thing_searched must always be Qt, a buffer, or Qnil. */
abort ();
- len = i;
+ len = 2 * i + 2;
}
else
- data[2 * i] = data [2 * i + 1] = Qnil;
+ data[2 * i] = data[2 * i + 1] = Qnil;
+ }
+
+ if (BUFFERP (last_thing_searched) && !NILP (integers))
+ {
+ data[len] = last_thing_searched;
+ len++;
}
/* If REUSE is not usable, cons up the values and return them. */
if (! CONSP (reuse))
- return Flist (2 * len + 2, data);
+ return Flist (len, data);
/* If REUSE is a list, store as many value elements as will fit
into the elements of REUSE. */
for (i = 0, tail = reuse; CONSP (tail);
- i++, tail = XCONS (tail)->cdr)
+ i++, tail = XCDR (tail))
{
- if (i < 2 * len + 2)
- XCONS (tail)->car = data[i];
+ if (i < len)
+ XSETCAR (tail, data[i]);
else
- XCONS (tail)->car = Qnil;
+ XSETCAR (tail, Qnil);
prev = tail;
}
/* If we couldn't fit all value elements into REUSE,
cons up the rest of them and add them to the end of REUSE. */
- if (i < 2 * len + 2)
- XCONS (prev)->cdr = Flist (2 * len + 2 - i, data + i);
+ if (i < len)
+ XSETCDR (prev, Flist (len - i, data + i));
return reuse;
}
+/* Internal usage only:
+ If RESEAT is `evaporate', put the markers back on the free list
+ immediately. No other references to the markers must exist in this case,
+ so it is used only internally on the unwind stack and save-match-data from
+ Lisp. */
-DEFUN ("store-match-data", Fstore_match_data, Sstore_match_data, 1, 1, 0,
- "Set internal data on last search match from elements of LIST.\n\
-LIST should have been created by calling `match-data' previously.")
- (list)
- register Lisp_Object list;
+DEFUN ("set-match-data", Fset_match_data, Sset_match_data, 1, 2, 0,
+ doc: /* Set internal data on last search match from elements of LIST.
+LIST should have been created by calling `match-data' previously.
+
+If optional arg RESEAT is non-nil, make markers on LIST point nowhere. */)
+ (list, reseat)
+ register Lisp_Object list, reseat;
{
register int i;
register Lisp_Object marker;
if (!CONSP (list) && !NILP (list))
list = wrong_type_argument (Qconsp, list);
- /* Unless we find a marker with a buffer in LIST, assume that this
- match data came from a string. */
+ /* Unless we find a marker with a buffer or an explicit buffer
+ in LIST, assume that this match data came from a string. */
last_thing_searched = Qt;
/* Allocate registers if they don't already exist. */
length * sizeof (regoff_t));
}
+ for (i = search_regs.num_regs; i < length; i++)
+ search_regs.start[i] = -1;
+
search_regs.num_regs = length;
}
- }
- for (i = 0; i < search_regs.num_regs; i++)
- {
- marker = Fcar (list);
- if (NILP (marker))
- {
- search_regs.start[i] = -1;
- list = Fcdr (list);
- }
- else
- {
- if (MARKERP (marker))
- {
- if (XMARKER (marker)->buffer == 0)
- XSETFASTINT (marker, 0);
- else
- XSETBUFFER (last_thing_searched, XMARKER (marker)->buffer);
- }
+ for (i = 0; CONSP (list); i++)
+ {
+ marker = XCAR (list);
+ if (BUFFERP (marker))
+ {
+ last_thing_searched = marker;
+ break;
+ }
+ if (i >= length)
+ break;
+ if (NILP (marker))
+ {
+ search_regs.start[i] = -1;
+ list = XCDR (list);
+ }
+ else
+ {
+ int from;
+ Lisp_Object m;
- CHECK_NUMBER_COERCE_MARKER (marker, 0);
- search_regs.start[i] = XINT (marker);
- list = Fcdr (list);
+ m = marker;
+ if (MARKERP (marker))
+ {
+ if (XMARKER (marker)->buffer == 0)
+ XSETFASTINT (marker, 0);
+ else
+ XSETBUFFER (last_thing_searched, XMARKER (marker)->buffer);
+ }
- marker = Fcar (list);
- if (MARKERP (marker) && XMARKER (marker)->buffer == 0)
- XSETFASTINT (marker, 0);
+ CHECK_NUMBER_COERCE_MARKER (marker);
+ from = XINT (marker);
- CHECK_NUMBER_COERCE_MARKER (marker, 0);
- search_regs.end[i] = XINT (marker);
- }
- list = Fcdr (list);
- }
+ if (!NILP (reseat) && MARKERP (m))
+ {
+ if (EQ (reseat, Qevaporate))
+ free_marker (m);
+ else
+ unchain_marker (XMARKER (m));
+ XSETCAR (list, Qnil);
+ }
+
+ if ((list = XCDR (list), !CONSP (list)))
+ break;
- return Qnil;
+ m = marker = XCAR (list);
+
+ if (MARKERP (marker) && XMARKER (marker)->buffer == 0)
+ XSETFASTINT (marker, 0);
+
+ CHECK_NUMBER_COERCE_MARKER (marker);
+ search_regs.start[i] = from;
+ search_regs.end[i] = XINT (marker);
+
+ if (!NILP (reseat) && MARKERP (m))
+ {
+ if (EQ (reseat, Qevaporate))
+ free_marker (m);
+ else
+ unchain_marker (XMARKER (m));
+ XSETCAR (list, Qnil);
+ }
+ }
+ list = XCDR (list);
+ }
+
+ for (; i < search_regs.num_regs; i++)
+ search_regs.start[i] = -1;
+ }
+
+ return Qnil;
}
/* If non-zero the match data have been saved in saved_search_regs
during the execution of a sentinel or filter. */
static int search_regs_saved;
static struct re_registers saved_search_regs;
+static Lisp_Object saved_last_thing_searched;
/* Called from Flooking_at, Fstring_match, search_buffer, Fstore_match_data
if asynchronous code (filter or sentinel) is running. */
saved_search_regs.num_regs = search_regs.num_regs;
saved_search_regs.start = search_regs.start;
saved_search_regs.end = search_regs.end;
+ saved_last_thing_searched = last_thing_searched;
+ last_thing_searched = Qnil;
search_regs.num_regs = 0;
search_regs.start = 0;
search_regs.end = 0;
/* Called upon exit from filters and sentinels. */
void
-restore_match_data ()
+restore_search_regs ()
{
if (search_regs_saved)
{
search_regs.num_regs = saved_search_regs.num_regs;
search_regs.start = saved_search_regs.start;
search_regs.end = saved_search_regs.end;
-
+ last_thing_searched = saved_last_thing_searched;
+ saved_last_thing_searched = Qnil;
search_regs_saved = 0;
}
}
+static Lisp_Object
+unwind_set_match_data (list)
+ Lisp_Object list;
+{
+ /* It is safe to free (evaporate) the markers immediately. */
+ return Fset_match_data (list, Qevaporate);
+}
+
+/* Called to unwind protect the match data. */
+void
+record_unwind_save_match_data ()
+{
+ record_unwind_protect (unwind_set_match_data,
+ Fmatch_data (Qnil, Qnil, Qnil));
+}
+
/* Quote a string to inactivate reg-expr chars */
DEFUN ("regexp-quote", Fregexp_quote, Sregexp_quote, 1, 1, 0,
- "Return a regexp string which matches exactly STRING and nothing else.")
- (string)
+ doc: /* Return a regexp string which matches exactly STRING and nothing else. */)
+ (string)
Lisp_Object string;
{
register unsigned char *in, *out, *end;
register unsigned char *temp;
int backslashes_added = 0;
- CHECK_STRING (string, 0);
+ CHECK_STRING (string);
- temp = (unsigned char *) alloca (XSTRING (string)->size_byte * 2);
+ temp = (unsigned char *) alloca (SBYTES (string) * 2);
/* Now copy the data into the new string, inserting escapes. */
- in = XSTRING (string)->data;
- end = in + XSTRING (string)->size_byte;
- out = temp;
+ in = SDATA (string);
+ end = in + SBYTES (string);
+ out = temp;
for (; in != end; in++)
{
*out++ = *in;
}
- return make_multibyte_string (temp,
- XSTRING (string)->size + backslashes_added,
- out - temp);
+ return make_specified_string (temp,
+ SCHARS (string) + backslashes_added,
+ out - temp,
+ STRING_MULTIBYTE (string));
}
-\f
+\f
+void
syms_of_search ()
{
register int i;
for (i = 0; i < REGEXP_CACHE_SIZE; ++i)
{
searchbufs[i].buf.allocated = 100;
- searchbufs[i].buf.buffer = (unsigned char *) malloc (100);
+ searchbufs[i].buf.buffer = (unsigned char *) xmalloc (100);
searchbufs[i].buf.fastmap = searchbufs[i].fastmap;
searchbufs[i].regexp = Qnil;
+ searchbufs[i].whitespace_regexp = Qnil;
staticpro (&searchbufs[i].regexp);
+ staticpro (&searchbufs[i].whitespace_regexp);
searchbufs[i].next = (i == REGEXP_CACHE_SIZE-1 ? 0 : &searchbufs[i+1]);
}
searchbuf_head = &searchbufs[0];
last_thing_searched = Qnil;
staticpro (&last_thing_searched);
+ saved_last_thing_searched = Qnil;
+ staticpro (&saved_last_thing_searched);
+
+ DEFVAR_LISP ("search-spaces-regexp", &Vsearch_spaces_regexp,
+ doc: /* Regexp to substitute for bunches of spaces in regexp search.
+Some commands use this for user-specified regexps.
+Spaces that occur inside character classes or repetition operators
+or other such regexp constructs are not replaced with this.
+A value of nil (which is the normal value) means treat spaces literally. */);
+ Vsearch_spaces_regexp = Qnil;
+
defsubr (&Slooking_at);
defsubr (&Sposix_looking_at);
defsubr (&Sstring_match);
defsubr (&Smatch_beginning);
defsubr (&Smatch_end);
defsubr (&Smatch_data);
- defsubr (&Sstore_match_data);
+ defsubr (&Sset_match_data);
defsubr (&Sregexp_quote);
}
+
+/* arch-tag: a6059d79-0552-4f14-a2cb-d379a4e3c78f
+ (do not change this comment) */
HCoop / bpt / emacs.git / blobdiff