/* String search routines for GNU Emacs.
- Copyright (C) 1985, 1986, 1987, 1993 Free Software Foundation, Inc.
+ Copyright (C) 1985, 1986, 1987, 1993, 1994 Free Software Foundation, Inc.
This file is part of GNU Emacs.
GNU Emacs is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
-the Free Software Foundation; either version 1, or (at your option)
+the Free Software Foundation; either version 2, or (at your option)
any later version.
GNU Emacs is distributed in the hope that it will be useful,
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, 675 Mass Ave, Cambridge, MA 02139, USA. */
+the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA. */
#include <config.h>
#include "lisp.h"
#include "syntax.h"
+#include "category.h"
#include "buffer.h"
+#include "charset.h"
+#include "region-cache.h"
#include "commands.h"
#include "blockinput.h"
#include <sys/types.h>
#include "regex.h"
-#define max(a, b) ((a) > (b) ? (a) : (b))
-#define min(a, b) ((a) < (b) ? (a) : (b))
+#define REGEXP_CACHE_SIZE 20
-/* We compile regexps into this buffer and then use it for searching. */
-
-struct re_pattern_buffer searchbuf;
+/* If the regexp is non-nil, then the buffer contains the compiled form
+ of that regexp, suitable for searching. */
+struct regexp_cache
+{
+ struct regexp_cache *next;
+ Lisp_Object regexp;
+ struct re_pattern_buffer buf;
+ char fastmap[0400];
+ /* Nonzero means regexp was compiled to do full POSIX backtracking. */
+ char posix;
+};
-char search_fastmap[0400];
+/* The instances of that struct. */
+struct regexp_cache searchbufs[REGEXP_CACHE_SIZE];
-/* Last regexp we compiled */
+/* The head of the linked list; points to the most recently used buffer. */
+struct regexp_cache *searchbuf_head;
-Lisp_Object last_regexp;
/* Every call to re_match, etc., must pass &search_regs as the regs
argument unless you can show it is unnecessary (i.e., if re_match
been allocated by checking search_regs.num_regs.
The regex code keeps track of whether it has allocated the search
- buffer using bits in searchbuf. This means that whenever you
- compile a new pattern, it completely forgets whether it has
+ buffer using bits in the re_pattern_buffer. This means that whenever
+ you compile a new pattern, it completely forgets whether it has
allocated any registers, and will allocate new registers the next
time you call a searching or matching function. Therefore, we need
to call re_set_registers after compiling a new pattern or after
Qnil if no searching has been done yet. */
static Lisp_Object last_thing_searched;
-/* error condition signalled when regexp compile_pattern fails */
+/* error condition signaled when regexp compile_pattern fails */
Lisp_Object Qinvalid_regexp;
+static void set_search_regs ();
+static void save_search_regs ();
+
+static int search_buffer ();
+
static void
matcher_overflow ()
{
#define CONST
#endif
-/* Compile a regexp and signal a Lisp error if anything goes wrong. */
+/* 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.
+ 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
+ subexpression bounds.
+ POSIX is nonzero if we want full backtracking (POSIX style)
+ 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. */
-compile_pattern (pattern, bufp, regp, translate)
+static void
+compile_pattern_1 (cp, pattern, translate, regp, posix, multibyte)
+ struct regexp_cache *cp;
Lisp_Object pattern;
- struct re_pattern_buffer *bufp;
+ Lisp_Object *translate;
struct re_registers *regp;
- char *translate;
+ int posix;
+ int multibyte;
{
CONST char *val;
- Lisp_Object dummy;
-
- if (EQ (pattern, last_regexp)
- && translate == bufp->translate)
- return;
+ reg_syntax_t old;
- last_regexp = Qnil;
- bufp->translate = translate;
+ cp->regexp = Qnil;
+ cp->buf.translate = translate;
+ cp->posix = posix;
+ cp->buf.multibyte = multibyte;
BLOCK_INPUT;
+ old = re_set_syntax (RE_SYNTAX_EMACS
+ | (posix ? 0 : RE_NO_POSIX_BACKTRACKING));
val = (CONST char *) re_compile_pattern ((char *) XSTRING (pattern)->data,
- XSTRING (pattern)->size, bufp);
+ XSTRING (pattern)->size, &cp->buf);
+ re_set_syntax (old);
UNBLOCK_INPUT;
if (val)
+ Fsignal (Qinvalid_regexp, Fcons (build_string (val), Qnil));
+
+ cp->regexp = Fcopy_sequence (pattern);
+}
+
+/* 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.
+ 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
+ subexpression bounds.
+ POSIX is nonzero if we want full backtracking (POSIX style)
+ for this pattern. 0 means backtrack only enough to get a valid match. */
+
+struct re_pattern_buffer *
+compile_pattern (pattern, regp, translate, posix)
+ Lisp_Object pattern;
+ struct re_registers *regp;
+ Lisp_Object *translate;
+ int posix;
+{
+ struct regexp_cache *cp, **cpp;
+ /* Should we check it here, or add an argument `multibyte' to this
+ function? */
+ int multibyte = !NILP (current_buffer->enable_multibyte_characters);
+
+ for (cpp = &searchbuf_head; ; cpp = &cp->next)
{
- dummy = build_string (val);
- while (1)
- Fsignal (Qinvalid_regexp, Fcons (dummy, Qnil));
+ cp = *cpp;
+ if (XSTRING (cp->regexp)->size == XSTRING (pattern)->size
+ && !NILP (Fstring_equal (cp->regexp, pattern))
+ && cp->buf.translate == translate
+ && cp->posix == posix
+ && cp->buf.multibyte == multibyte)
+ break;
+
+ /* If we're at the end of the cache, compile into the last cell. */
+ if (cp->next == 0)
+ {
+ compile_pattern_1 (cp, pattern, translate, regp, posix, multibyte);
+ break;
+ }
}
- last_regexp = pattern;
+ /* When we get here, cp (aka *cpp) contains the compiled pattern,
+ either because we found it in the cache or because we just compiled it.
+ Move it to the front of the queue to mark it as most recently used. */
+ *cpp = cp->next;
+ cp->next = searchbuf_head;
+ searchbuf_head = cp;
/* Advise the searching functions about the space we have allocated
for register data. */
- BLOCK_INPUT;
if (regp)
- re_set_registers (bufp, regp, regp->num_regs, regp->start, regp->end);
- UNBLOCK_INPUT;
+ re_set_registers (&cp->buf, regp, regp->num_regs, regp->start, regp->end);
- return;
+ return &cp->buf;
}
/* Error condition used for failing searches */
return Qnil;
}
\f
-DEFUN ("looking-at", Flooking_at, Slooking_at, 1, 1, 0,
- "Return t if text after point matches regular expression PAT.\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.")
- (string)
+static Lisp_Object
+looking_at_1 (string, posix)
Lisp_Object string;
+ int posix;
{
Lisp_Object val;
unsigned char *p1, *p2;
int s1, s2;
register int i;
+ struct re_pattern_buffer *bufp;
+
+ if (running_asynch_code)
+ save_search_regs ();
CHECK_STRING (string, 0);
- compile_pattern (string, &searchbuf, &search_regs,
- !NILP (current_buffer->case_fold_search) ? DOWNCASE_TABLE : 0);
+ bufp = compile_pattern (string, &search_regs,
+ (!NILP (current_buffer->case_fold_search)
+ ? DOWNCASE_TABLE : 0),
+ posix);
immediate_quit = 1;
QUIT; /* Do a pending quit right away, to avoid paradoxical behavior */
s1 = ZV - BEGV;
s2 = 0;
}
+
+ re_match_object = Qnil;
- i = re_match_2 (&searchbuf, (char *) p1, s1, (char *) p2, s2,
- point - BEGV, &search_regs,
+ i = re_match_2 (bufp, (char *) p1, s1, (char *) p2, s2,
+ PT - BEGV, &search_regs,
ZV - BEGV);
if (i == -2)
matcher_overflow ();
search_regs.start[i] += BEGV;
search_regs.end[i] += BEGV;
}
- XSET (last_thing_searched, Lisp_Buffer, current_buffer);
+ XSETBUFFER (last_thing_searched, current_buffer);
immediate_quit = 0;
return val;
}
-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)
+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)
+ 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)
+ Lisp_Object regexp;
+{
+ return looking_at_1 (regexp, 1);
+}
+\f
+static Lisp_Object
+string_match_1 (regexp, string, start, posix)
Lisp_Object regexp, string, start;
+ int posix;
{
int val;
int s;
+ struct re_pattern_buffer *bufp;
+
+ if (running_asynch_code)
+ save_search_regs ();
CHECK_STRING (regexp, 0);
CHECK_STRING (string, 1);
CHECK_NUMBER (start, 2);
s = XINT (start);
if (s < 0 && -s <= len)
- s = len - s;
+ s = len + s;
else if (0 > s || s > len)
args_out_of_range (string, start);
}
- compile_pattern (regexp, &searchbuf, &search_regs,
- !NILP (current_buffer->case_fold_search) ? DOWNCASE_TABLE : 0);
+ bufp = compile_pattern (regexp, &search_regs,
+ (!NILP (current_buffer->case_fold_search)
+ ? DOWNCASE_TABLE : 0),
+ posix);
immediate_quit = 1;
- val = re_search (&searchbuf, (char *) XSTRING (string)->data,
+ re_match_object = string;
+
+ val = re_search (bufp, (char *) XSTRING (string)->data,
XSTRING (string)->size, s, XSTRING (string)->size - s,
&search_regs);
immediate_quit = 0;
return make_number (val);
}
+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)
+ 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)
+ Lisp_Object regexp, string, start;
+{
+ return string_match_1 (regexp, string, start, 1);
+}
+
/* Match REGEXP against STRING, searching all of STRING,
and return the index of the match, or negative on failure.
This does not clobber the match data. */
Lisp_Object regexp, string;
{
int val;
+ struct re_pattern_buffer *bufp;
- compile_pattern (regexp, &searchbuf, 0, 0);
+ bufp = compile_pattern (regexp, 0, 0, 0);
immediate_quit = 1;
- val = re_search (&searchbuf, (char *) XSTRING (string)->data,
+ re_match_object = string;
+
+ val = re_search (bufp, (char *) XSTRING (string)->data,
XSTRING (string)->size, 0, XSTRING (string)->size,
0);
immediate_quit = 0;
return val;
}
-\f
-/* Search for COUNT instances of the character TARGET, starting at START.
- If COUNT is negative, search backwards.
- If we find COUNT instances, set *SHORTAGE to zero, and return the
- position after the COUNTth match. Note that for reverse motion
- this is not the same as the usual convention for Emacs motion commands.
+/* Match REGEXP against STRING, searching all of STRING ignoring case,
+ and return the index of the match, or negative on failure.
+ This does not clobber the match data. */
- If we don't find COUNT instances before reaching the end of the
- buffer (or the beginning, if scanning backwards), set *SHORTAGE to
- the number of TARGETs left unfound, and return the end of the
- buffer we bumped up against. */
+extern Lisp_Object Vascii_downcase_table;
-scan_buffer (target, start, count, shortage)
- int *shortage, start;
- register int count, target;
+int
+fast_string_match_ignore_case (regexp, string)
+ Lisp_Object regexp;
+ char *string;
{
- int limit = ((count > 0) ? ZV - 1 : BEGV);
- int direction = ((count > 0) ? 1 : -1);
-
- register unsigned char *cursor;
- unsigned char *base;
-
- register int ceiling;
- register unsigned char *ceiling_addr;
-
- if (shortage != 0)
- *shortage = 0;
+ int val;
+ struct re_pattern_buffer *bufp;
+ int len = strlen (string);
+ bufp = compile_pattern (regexp, 0,
+ XCHAR_TABLE (Vascii_downcase_table)->contents, 0);
immediate_quit = 1;
-
- if (count > 0)
- while (start != limit + 1)
- {
- ceiling = BUFFER_CEILING_OF (start);
- ceiling = min (limit, ceiling);
- ceiling_addr = &FETCH_CHAR (ceiling) + 1;
- base = (cursor = &FETCH_CHAR (start));
- while (1)
- {
- while (*cursor != target && ++cursor != ceiling_addr)
- ;
- if (cursor != ceiling_addr)
- {
- if (--count == 0)
- {
- immediate_quit = 0;
- return (start + cursor - base + 1);
- }
- else
- if (++cursor == ceiling_addr)
- break;
- }
- else
- break;
- }
- start += cursor - base;
- }
- else
- {
- start--; /* first character we scan */
- while (start > limit - 1)
- { /* we WILL scan under start */
- ceiling = BUFFER_FLOOR_OF (start);
- ceiling = max (limit, ceiling);
- ceiling_addr = &FETCH_CHAR (ceiling) - 1;
- base = (cursor = &FETCH_CHAR (start));
- cursor++;
- while (1)
- {
- while (--cursor != ceiling_addr && *cursor != target)
- ;
- if (cursor != ceiling_addr)
- {
- if (++count == 0)
- {
- immediate_quit = 0;
- return (start + cursor - base + 1);
- }
- }
- else
- break;
- }
- start += cursor - base;
- }
- }
+ val = re_search (bufp, string, len, 0, len, 0);
immediate_quit = 0;
- if (shortage != 0)
- *shortage = count * direction;
- return (start + ((direction == 1 ? 0 : 1)));
-}
-
-int
-find_next_newline (from, cnt)
- register int from, cnt;
-{
- return (scan_buffer ('\n', from, cnt, (int *) 0));
+ return val;
}
\f
-Lisp_Object skip_chars ();
-
-DEFUN ("skip-chars-forward", Fskip_chars_forward, Sskip_chars_forward, 1, 2, 0,
- "Move point forward, stopping before a char not in CHARS, or at position LIM.\n\
-CHARS is like the inside of a `[...]' in a regular expression\n\
-except that `]' is never special and `\\' quotes `^', `-' or `\\'.\n\
-Thus, with arg \"a-zA-Z\", this skips letters stopping before first nonletter.\n\
-With arg \"^a-zA-Z\", skips nonletters stopping before first letter.\n\
-Returns the distance traveled, either zero or positive.")
- (string, lim)
- Lisp_Object string, lim;
-{
- return skip_chars (1, 0, string, lim);
-}
+/* max and min. */
-DEFUN ("skip-chars-backward", Fskip_chars_backward, Sskip_chars_backward, 1, 2, 0,
- "Move point backward, stopping after a char not in CHARS, or at position LIM.\n\
-See `skip-chars-forward' for details.\n\
-Returns the distance traveled, either zero or negative.")
- (string, lim)
- Lisp_Object string, lim;
+static int
+max (a, b)
+ int a, b;
{
- return skip_chars (0, 0, string, lim);
+ return ((a > b) ? a : b);
}
-DEFUN ("skip-syntax-forward", Fskip_syntax_forward, Sskip_syntax_forward, 1, 2, 0,
- "Move point forward across chars in specified syntax classes.\n\
-SYNTAX is a string of syntax code characters.\n\
-Stop before a char whose syntax is not in SYNTAX, or at position LIM.\n\
-If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.\n\
-This function returns the distance traveled, either zero or positive.")
- (syntax, lim)
- Lisp_Object syntax, lim;
+static int
+min (a, b)
+ int a, b;
{
- return skip_chars (1, 1, syntax, lim);
+ return ((a < b) ? a : b);
}
-DEFUN ("skip-syntax-backward", Fskip_syntax_backward, Sskip_syntax_backward, 1, 2, 0,
- "Move point backward across chars in specified syntax classes.\n\
-SYNTAX is a string of syntax code characters.\n\
-Stop on reaching a char whose syntax is not in SYNTAX, or at position LIM.\n\
-If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.\n\
-This function returns the distance traveled, either zero or negative.")
- (syntax, lim)
- Lisp_Object syntax, lim;
+\f
+/* The newline cache: remembering which sections of text have no newlines. */
+
+/* If the user has requested newline caching, make sure it's on.
+ Otherwise, make sure it's off.
+ This is our cheezy way of associating an action with the change of
+ state of a buffer-local variable. */
+static void
+newline_cache_on_off (buf)
+ struct buffer *buf;
{
- return skip_chars (0, 1, syntax, lim);
+ if (NILP (buf->cache_long_line_scans))
+ {
+ /* It should be off. */
+ if (buf->newline_cache)
+ {
+ free_region_cache (buf->newline_cache);
+ buf->newline_cache = 0;
+ }
+ }
+ else
+ {
+ /* It should be on. */
+ if (buf->newline_cache == 0)
+ buf->newline_cache = new_region_cache ();
+ }
}
-Lisp_Object
-skip_chars (forwardp, syntaxp, string, lim)
- int forwardp, syntaxp;
- Lisp_Object string, lim;
-{
- register unsigned char *p, *pend;
- register unsigned char c;
- unsigned char fastmap[0400];
- int negate = 0;
- register int i;
+\f
+/* Search for COUNT instances of the character TARGET between START and END.
- CHECK_STRING (string, 0);
+ If COUNT is positive, search forwards; END must be >= START.
+ If COUNT is negative, search backwards for the -COUNTth instance;
+ END must be <= START.
+ If COUNT is zero, do anything you please; run rogue, for all I care.
- if (NILP (lim))
- XSET (lim, Lisp_Int, forwardp ? ZV : BEGV);
- else
- CHECK_NUMBER_COERCE_MARKER (lim, 1);
+ If END is zero, use BEGV or ZV instead, as appropriate for the
+ direction indicated by COUNT.
- /* In any case, don't allow scan outside bounds of buffer. */
+ If we find COUNT instances, set *SHORTAGE to zero, and return the
+ position after the COUNTth match. Note that for reverse motion
+ this is not the same as the usual convention for Emacs motion commands.
- /* I don't know what things this breaks, and there is no entry in the
- ChangeLog, so I reinstated the end of buffer limit check. This code
- breaks without it. (bfox) */
-/* #if 0 /* This breaks some things... jla. */
-#if 1
- if (XFASTINT (lim) > ZV)
- XFASTINT (lim) = ZV;
-#endif
-#if 0
- if (XFASTINT (lim) < BEGV)
- XFASTINT (lim) = BEGV;
-#endif
+ If we don't find COUNT instances before reaching END, set *SHORTAGE
+ to the number of TARGETs left unfound, and return END.
- p = XSTRING (string)->data;
- pend = p + XSTRING (string)->size;
- bzero (fastmap, sizeof fastmap);
+ If ALLOW_QUIT is non-zero, set immediate_quit. That's good to do
+ except when inside redisplay. */
- if (p != pend && *p == '^')
+scan_buffer (target, start, end, count, shortage, allow_quit)
+ register int target;
+ int start, end;
+ int count;
+ int *shortage;
+ int allow_quit;
+{
+ struct region_cache *newline_cache;
+ int direction;
+
+ if (count > 0)
{
- negate = 1; p++;
+ direction = 1;
+ if (! end) end = ZV;
}
-
- /* Find the characters specified and set their elements of fastmap.
- If syntaxp, each character counts as itself.
- Otherwise, handle backslashes and ranges specially */
-
- while (p != pend)
+ else
{
- c = *p++;
- if (syntaxp)
- fastmap[c] = 1;
- else
- {
- if (c == '\\')
- {
- if (p == pend) break;
- c = *p++;
- }
- if (p != pend && *p == '-')
- {
- p++;
- if (p == pend) break;
- while (c <= *p)
- {
- fastmap[c] = 1;
- c++;
- }
- p++;
- }
- else
- fastmap[c] = 1;
- }
+ direction = -1;
+ if (! end) end = BEGV;
}
- /* If ^ was the first character, complement the fastmap. */
+ newline_cache_on_off (current_buffer);
+ newline_cache = current_buffer->newline_cache;
- if (negate)
- for (i = 0; i < sizeof fastmap; i++)
- fastmap[i] ^= 1;
+ if (shortage != 0)
+ *shortage = 0;
- {
- int start_point = point;
+ immediate_quit = allow_quit;
- immediate_quit = 1;
- if (syntaxp)
+ if (count > 0)
+ while (start != end)
{
-
- if (forwardp)
- {
- while (point < XINT (lim)
- && fastmap[(unsigned char) syntax_code_spec[(int) SYNTAX (FETCH_CHAR (point))]])
- SET_PT (point + 1);
- }
- else
- {
- while (point > XINT (lim)
- && fastmap[(unsigned char) syntax_code_spec[(int) SYNTAX (FETCH_CHAR (point - 1))]])
- SET_PT (point - 1);
- }
+ /* Our innermost scanning loop is very simple; it doesn't know
+ about gaps, buffer ends, or the newline cache. ceiling is
+ the position of the last character before the next such
+ obstacle --- the last character the dumb search loop should
+ examine. */
+ register int ceiling = end - 1;
+
+ /* If we're looking for a newline, consult the newline cache
+ to see where we can avoid some scanning. */
+ if (target == '\n' && newline_cache)
+ {
+ int next_change;
+ immediate_quit = 0;
+ while (region_cache_forward
+ (current_buffer, newline_cache, start, &next_change))
+ start = next_change;
+ immediate_quit = allow_quit;
+
+ /* start should never be after end. */
+ if (start >= end)
+ start = end - 1;
+
+ /* Now the text after start is an unknown region, and
+ next_change is the position of the next known region. */
+ ceiling = min (next_change - 1, ceiling);
+ }
+
+ /* The dumb loop can only scan text stored in contiguous
+ bytes. BUFFER_CEILING_OF returns the last character
+ position that is contiguous, so the ceiling is the
+ position after that. */
+ ceiling = min (BUFFER_CEILING_OF (start), ceiling);
+
+ {
+ /* The termination address of the dumb loop. */
+ register unsigned char *ceiling_addr = POS_ADDR (ceiling) + 1;
+ register unsigned char *cursor = POS_ADDR (start);
+ unsigned char *base = cursor;
+
+ while (cursor < ceiling_addr)
+ {
+ unsigned char *scan_start = cursor;
+
+ /* The dumb loop. */
+ while (*cursor != target && ++cursor < ceiling_addr)
+ ;
+
+ /* If we're looking for newlines, cache the fact that
+ the region from start to cursor is free of them. */
+ if (target == '\n' && newline_cache)
+ know_region_cache (current_buffer, newline_cache,
+ start + scan_start - base,
+ start + cursor - base);
+
+ /* Did we find the target character? */
+ if (cursor < ceiling_addr)
+ {
+ if (--count == 0)
+ {
+ immediate_quit = 0;
+ return (start + cursor - base + 1);
+ }
+ cursor++;
+ }
+ }
+
+ start += cursor - base;
+ }
}
- else
+ else
+ while (start > end)
{
- if (forwardp)
- {
- while (point < XINT (lim) && fastmap[FETCH_CHAR (point)])
- SET_PT (point + 1);
- }
- else
- {
- while (point > XINT (lim) && fastmap[FETCH_CHAR (point - 1)])
- SET_PT (point - 1);
- }
+ /* The last character to check before the next obstacle. */
+ register int ceiling = end;
+
+ /* Consult the newline cache, if appropriate. */
+ if (target == '\n' && newline_cache)
+ {
+ int next_change;
+ immediate_quit = 0;
+ while (region_cache_backward
+ (current_buffer, newline_cache, start, &next_change))
+ start = next_change;
+ immediate_quit = allow_quit;
+
+ /* Start should never be at or before end. */
+ if (start <= end)
+ start = end + 1;
+
+ /* Now the text before start is an unknown region, and
+ next_change is the position of the next known region. */
+ ceiling = max (next_change, ceiling);
+ }
+
+ /* Stop scanning before the gap. */
+ ceiling = max (BUFFER_FLOOR_OF (start - 1), ceiling);
+
+ {
+ /* The termination address of the dumb loop. */
+ register unsigned char *ceiling_addr = POS_ADDR (ceiling);
+ register unsigned char *cursor = POS_ADDR (start - 1);
+ unsigned char *base = cursor;
+
+ while (cursor >= ceiling_addr)
+ {
+ unsigned char *scan_start = cursor;
+
+ while (*cursor != target && --cursor >= ceiling_addr)
+ ;
+
+ /* If we're looking for newlines, cache the fact that
+ the region from after the cursor to start is free of them. */
+ if (target == '\n' && newline_cache)
+ know_region_cache (current_buffer, newline_cache,
+ start + cursor - base,
+ start + scan_start - base);
+
+ /* Did we find the target character? */
+ if (cursor >= ceiling_addr)
+ {
+ if (++count >= 0)
+ {
+ immediate_quit = 0;
+ return (start + cursor - base);
+ }
+ cursor--;
+ }
+ }
+
+ start += cursor - base;
+ }
}
- immediate_quit = 0;
- return make_number (point - start_point);
- }
+ immediate_quit = 0;
+ if (shortage != 0)
+ *shortage = count * direction;
+ return start;
+}
+
+int
+find_next_newline_no_quit (from, cnt)
+ register int from, cnt;
+{
+ return scan_buffer ('\n', from, 0, cnt, (int *) 0, 0);
+}
+
+int
+find_next_newline (from, cnt)
+ register int from, cnt;
+{
+ return scan_buffer ('\n', from, 0, cnt, (int *) 0, 1);
+}
+
+
+/* Like find_next_newline, but returns position before the newline,
+ not after, and only search up to TO. This isn't just
+ find_next_newline (...)-1, because you might hit TO. */
+int
+find_before_next_newline (from, to, cnt)
+ int from, to, cnt;
+{
+ int shortage;
+ int pos = scan_buffer ('\n', from, to, cnt, &shortage, 1);
+
+ if (shortage == 0)
+ pos--;
+
+ return pos;
}
\f
/* Subroutines of Lisp buffer search functions. */
static Lisp_Object
-search_command (string, bound, noerror, count, direction, RE)
+search_command (string, bound, noerror, count, direction, RE, posix)
Lisp_Object string, bound, noerror, count;
int direction;
int RE;
+ int posix;
{
register int np;
int lim;
{
CHECK_NUMBER_COERCE_MARKER (bound, 1);
lim = XINT (bound);
- if (n > 0 ? lim < point : lim > point)
+ if (n > 0 ? lim < PT : lim > PT)
error ("Invalid search bound (wrong side of point)");
if (lim > ZV)
lim = ZV;
lim = BEGV;
}
- np = search_buffer (string, point, lim, n, RE,
+ np = search_buffer (string, PT, lim, n, RE,
(!NILP (current_buffer->case_fold_search)
- ? XSTRING (current_buffer->case_canon_table)->data : 0),
+ ? XCHAR_TABLE (current_buffer->case_canon_table)->contents
+ : 0),
(!NILP (current_buffer->case_fold_search)
- ? XSTRING (current_buffer->case_eqv_table)->data : 0));
+ ? XCHAR_TABLE (current_buffer->case_eqv_table)->contents
+ : 0),
+ posix);
if (np <= 0)
{
if (NILP (noerror))
return make_number (np);
}
\f
-/* search for the n'th occurrence of STRING in the current buffer,
+static int
+trivial_regexp_p (regexp)
+ Lisp_Object regexp;
+{
+ int len = XSTRING (regexp)->size;
+ unsigned char *s = XSTRING (regexp)->data;
+ unsigned char c;
+ while (--len >= 0)
+ {
+ switch (*s++)
+ {
+ case '.': case '*': case '+': case '?': case '[': case '^': case '$':
+ return 0;
+ case '\\':
+ if (--len < 0)
+ return 0;
+ switch (*s++)
+ {
+ case '|': case '(': case ')': case '`': case '\'': case 'b':
+ case 'B': case '<': case '>': case 'w': case 'W': case 's':
+ case 'S': 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':
+ return 0;
+ }
+ }
+ }
+ return 1;
+}
+
+/* Search for the n'th occurrence of STRING in the current buffer,
starting at position POS and stopping at position LIM,
- treating PAT as a literal string if RE is false or as
+ treating STRING as a literal string if RE is false or as
a regular expression if RE is true.
If N is positive, searching is forward and LIM must be greater than POS.
Returns -x if only N-x occurrences found (x > 0),
or else the position at the beginning of the Nth occurrence
- (if searching backward) or the end (if searching forward). */
+ (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. */
-search_buffer (string, pos, lim, n, RE, trt, inverse_trt)
+static int
+search_buffer (string, pos, lim, n, RE, trt, inverse_trt, posix)
Lisp_Object string;
int pos;
int lim;
int n;
int RE;
- register unsigned char *trt;
- register unsigned char *inverse_trt;
+ Lisp_Object *trt;
+ Lisp_Object *inverse_trt;
+ int posix;
{
int len = XSTRING (string)->size;
unsigned char *base_pat = XSTRING (string)->data;
unsigned char *p1, *p2;
int s1, s2;
+ if (running_asynch_code)
+ save_search_regs ();
+
/* Null string is found at starting position. */
if (len == 0)
- return pos;
+ {
+ set_search_regs (pos, 0);
+ return pos;
+ }
/* Searching 0 times means don't move. */
if (n == 0)
return pos;
- if (RE)
- compile_pattern (string, &searchbuf, &search_regs, (char *) trt);
-
- if (RE /* Here we detect whether the */
- /* generality of an RE search is */
- /* really needed. */
- /* first item is "exact match" */
- && *(searchbuf.buffer) == (char) RE_EXACTN_VALUE
- && searchbuf.buffer[1] + 2 == searchbuf.used) /*first is ONLY item */
+ if (RE && !trivial_regexp_p (string))
{
- RE = 0; /* can do straight (non RE) search */
- pat = (base_pat = (unsigned char *) searchbuf.buffer + 2);
- /* trt already applied */
- len = searchbuf.used - 2;
- }
- else if (!RE)
- {
- pat = (unsigned char *) alloca (len);
+ struct re_pattern_buffer *bufp;
- for (i = len; i--;) /* Copy the pattern; apply trt */
- *pat++ = (((int) trt) ? trt [*base_pat++] : *base_pat++);
- pat -= len; base_pat = pat;
- }
+ bufp = compile_pattern (string, &search_regs, trt, posix);
- if (RE)
- {
immediate_quit = 1; /* Quit immediately if user types ^G,
because letting this function finish
can take too long. */
s1 = ZV - BEGV;
s2 = 0;
}
+ re_match_object = Qnil;
+
while (n < 0)
{
int val;
- val = re_search_2 (&searchbuf, (char *) p1, s1, (char *) p2, s2,
+ val = re_search_2 (bufp, (char *) p1, s1, (char *) p2, s2,
pos - BEGV, lim - pos, &search_regs,
/* Don't allow match past current point */
pos - BEGV);
if (val == -2)
- matcher_overflow ();
+ {
+ matcher_overflow ();
+ }
if (val >= 0)
{
j = BEGV;
search_regs.start[i] += j;
search_regs.end[i] += j;
}
- XSET (last_thing_searched, Lisp_Buffer, current_buffer);
+ XSETBUFFER (last_thing_searched, current_buffer);
/* Set pos to the new position. */
pos = search_regs.start[0];
}
while (n > 0)
{
int val;
- val = re_search_2 (&searchbuf, (char *) p1, s1, (char *) p2, s2,
+ val = re_search_2 (bufp, (char *) p1, s1, (char *) p2, s2,
pos - BEGV, lim - pos, &search_regs,
lim - BEGV);
if (val == -2)
- matcher_overflow ();
+ {
+ matcher_overflow ();
+ }
if (val >= 0)
{
j = BEGV;
search_regs.start[i] += j;
search_regs.end[i] += j;
}
- XSET (last_thing_searched, Lisp_Buffer, current_buffer);
+ XSETBUFFER (last_thing_searched, current_buffer);
pos = search_regs.end[0];
}
else
#else
BM_tab = (int *) alloca (0400 * sizeof (int));
#endif
+ {
+ unsigned char *patbuf = (unsigned char *) alloca (len);
+ pat = patbuf;
+ while (--len >= 0)
+ {
+ /* 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--;
+ base_pat++;
+ }
+ *pat++ = (trt ? trt[*base_pat++] : *base_pat++);
+ }
+ len = pat - patbuf;
+ pat = base_pat = patbuf;
+ }
/* 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 */
{
j = pat[i]; i += direction;
if (i == dirlen) i = infinity;
- if ((int) trt)
+ if (trt != 0)
{
k = (j = trt[j]);
if (i == infinity)
BM_tab[j] = dirlen - i;
/* A translation table is accompanied by its inverse -- see */
/* comment following downcase_table for details */
- while ((j = inverse_trt[j]) != k)
+ while ((j = (unsigned char) inverse_trt[j]) != k)
BM_tab[j] = dirlen - i;
}
else
of pattern would align in a possible match. */
while (n != 0)
{
- if ((lim - pos - (direction > 0)) * 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 - pos - ((direction > 0) ? 1 : 0)) * direction < 0)
return (n * (0 - direction));
/* First we do the part we can by pointers (maybe nothing) */
QUIT;
: max (lim, max (limit, pos - 20000)));
if ((limit - pos) * direction > 20)
{
- p_limit = &FETCH_CHAR (limit);
- p2 = (cursor = &FETCH_CHAR (pos));
+ p_limit = POS_ADDR (limit);
+ p2 = (cursor = POS_ADDR (pos));
/* In this loop, pos + cursor - p2 is the surrogate for pos */
while (1) /* use one cursor setting as long as i can */
{
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 ((int) (p_limit + infinity) > (int) p_limit)
- while ((int) cursor <= (int) p_limit)
+ if ((EMACS_INT) (p_limit + infinity) > (EMACS_INT) p_limit)
+ while ((EMACS_INT) cursor <= (EMACS_INT) p_limit)
cursor += BM_tab[*cursor];
else
- while ((unsigned int) cursor <= (unsigned int) p_limit)
+ while ((EMACS_UINT) cursor <= (EMACS_UINT) p_limit)
cursor += BM_tab[*cursor];
}
else
{
- if ((int) (p_limit + infinity) < (int) p_limit)
- while ((int) cursor >= (int) p_limit)
+ if ((EMACS_INT) (p_limit + infinity) < (EMACS_INT) p_limit)
+ while ((EMACS_INT) cursor >= (EMACS_INT) p_limit)
cursor += BM_tab[*cursor];
else
- while ((unsigned int) cursor >= (unsigned int) p_limit)
+ 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. */
break; /* a small overrun is genuine */
cursor -= infinity; /* large overrun = hit */
i = dirlen - direction;
- if ((int) trt)
+ if (trt != 0)
{
while ((i -= direction) + direction != 0)
if (pat[i] != trt[*(cursor -= direction)])
{
cursor -= direction;
- /* Make sure we have registers in which to store
- the match position. */
- if (search_regs.num_regs == 0)
- {
- regoff_t *starts, *ends;
-
- starts =
- (regoff_t *) xmalloc (2 * sizeof (regoff_t));
- ends =
- (regoff_t *) xmalloc (2 * sizeof (regoff_t));
- BLOCK_INPUT;
- re_set_registers (&searchbuf,
- &search_regs,
- 2, starts, ends);
- UNBLOCK_INPUT;
- }
+ set_search_regs (pos + cursor - p2 + ((direction > 0)
+ ? 1 - len : 0),
+ len);
- search_regs.start[0]
- = pos + cursor - p2 + ((direction > 0)
- ? 1 - len : 0);
- search_regs.end[0] = len + search_regs.start[0];
- XSET (last_thing_searched, Lisp_Buffer, current_buffer);
if ((n -= direction) != 0)
cursor += dirlen; /* to resume search */
else
/* (the reach is at most len + 21, and typically */
/* does not exceed len) */
while ((limit - pos) * direction >= 0)
- pos += BM_tab[FETCH_CHAR(pos)];
+ pos += BM_tab[FETCH_BYTE (pos)];
/* now run the same tests to distinguish going off the */
/* end, a match or a phony match. */
if ((pos - limit) * direction <= len)
while ((i -= direction) + direction != 0)
{
pos -= direction;
- if (pat[i] != (((int) trt)
- ? trt[FETCH_CHAR(pos)]
- : FETCH_CHAR (pos)))
+ if (pat[i] != (trt != 0
+ ? trt[FETCH_BYTE (pos)]
+ : FETCH_BYTE (pos)))
break;
}
/* Above loop has moved POS part or all the way
{
pos -= direction;
- /* Make sure we have registers in which to store
- the match position. */
- if (search_regs.num_regs == 0)
- {
- regoff_t *starts, *ends;
-
- starts =
- (regoff_t *) xmalloc (2 * sizeof (regoff_t));
- ends =
- (regoff_t *) xmalloc (2 * sizeof (regoff_t));
- BLOCK_INPUT;
- re_set_registers (&searchbuf,
- &search_regs,
- 2, starts, ends);
- UNBLOCK_INPUT;
- }
+ set_search_regs (pos + ((direction > 0) ? 1 - len : 0),
+ len);
- search_regs.start[0]
- = pos + ((direction > 0) ? 1 - len : 0);
- search_regs.end[0] = len + search_regs.start[0];
- XSET (last_thing_searched, Lisp_Buffer, current_buffer);
if ((n -= direction) != 0)
pos += dirlen; /* to resume search */
else
return pos;
}
}
+
+/* Record beginning BEG and end BEG + LEN
+ for a match just found in the current buffer. */
+
+static void
+set_search_regs (beg, len)
+ int beg, len;
+{
+ /* Make sure we have registers in which to store
+ the match position. */
+ if (search_regs.num_regs == 0)
+ {
+ search_regs.start = (regoff_t *) xmalloc (2 * sizeof (regoff_t));
+ search_regs.end = (regoff_t *) xmalloc (2 * sizeof (regoff_t));
+ search_regs.num_regs = 2;
+ }
+
+ search_regs.start[0] = beg;
+ search_regs.end[0] = beg + len;
+ XSETBUFFER (last_thing_searched, current_buffer);
+}
\f
/* Given a string of words separated by word delimiters,
compute a regexp that matches those exact words
(string, bound, noerror, count)
Lisp_Object string, bound, noerror, count;
{
- return search_command (string, bound, noerror, count, -1, 0);
+ return search_command (string, bound, noerror, count, -1, 0, 0);
}
DEFUN ("search-forward", Fsearch_forward, Ssearch_forward, 1, 4, "sSearch: ",
(string, bound, noerror, count)
Lisp_Object string, bound, noerror, count;
{
- return search_command (string, bound, noerror, count, 1, 0);
+ return search_command (string, bound, noerror, count, 1, 0, 0);
}
DEFUN ("word-search-backward", Fword_search_backward, Sword_search_backward, 1, 4,
(string, bound, noerror, count)
Lisp_Object string, bound, noerror, count;
{
- return search_command (wordify (string), bound, noerror, count, -1, 1);
+ return search_command (wordify (string), bound, noerror, count, -1, 1, 0);
}
DEFUN ("word-search-forward", Fword_search_forward, Sword_search_forward, 1, 4,
(string, bound, noerror, count)
Lisp_Object string, bound, noerror, count;
{
- return search_command (wordify (string), bound, noerror, count, 1, 1);
+ return search_command (wordify (string), bound, noerror, count, 1, 1, 0);
}
DEFUN ("re-search-backward", Fre_search_backward, Sre_search_backward, 1, 4,
"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 place the origin of the search.\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'.")
- (string, bound, noerror, count)
- Lisp_Object string, bound, noerror, count;
+ (regexp, bound, noerror, count)
+ Lisp_Object regexp, bound, noerror, count;
{
- return search_command (string, bound, noerror, count, -1, 1);
+ return search_command (regexp, bound, noerror, count, -1, 1, 0);
}
DEFUN ("re-search-forward", Fre_search_forward, Sre_search_forward, 1, 4,
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)
- Lisp_Object string, bound, noerror, count;
+ (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)
+ 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)
+ Lisp_Object regexp, bound, noerror, count;
{
- return search_command (string, bound, noerror, count, 1, 1);
+ return search_command (regexp, bound, noerror, count, 1, 1, 1);
}
\f
-DEFUN ("replace-match", Freplace_match, Sreplace_match, 1, 3, 0,
+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 convert to all caps or cap initials, like replaced 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\
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.")
- (newtext, fixedcase, literal)
- Lisp_Object newtext, fixedcase, literal;
+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)
+ Lisp_Object newtext, fixedcase, literal, string, subexp;
{
enum { nochange, all_caps, cap_initial } case_action;
register int pos, last;
int some_multiletter_word;
int some_lowercase;
- int some_uppercase_initial;
+ int some_uppercase;
+ int some_nonuppercase_initial;
register int c, prevc;
int inslen;
+ int sub;
CHECK_STRING (newtext, 0);
+ if (! NILP (string))
+ CHECK_STRING (string, 4);
+
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");
- if (search_regs.start[0] < BEGV
- || search_regs.start[0] > search_regs.end[0]
- || search_regs.end[0] > ZV)
- args_out_of_range (make_number (search_regs.start[0]),
- make_number (search_regs.end[0]));
+ if (NILP (subexp))
+ sub = 0;
+ else
+ {
+ CHECK_NUMBER (subexp, 3);
+ sub = XINT (subexp);
+ if (sub < 0 || sub >= search_regs.num_regs)
+ args_out_of_range (subexp, make_number (search_regs.num_regs));
+ }
+
+ if (NILP (string))
+ {
+ if (search_regs.start[sub] < BEGV
+ || search_regs.start[sub] > search_regs.end[sub]
+ || search_regs.end[sub] > ZV)
+ args_out_of_range (make_number (search_regs.start[sub]),
+ make_number (search_regs.end[sub]));
+ }
+ else
+ {
+ if (search_regs.start[sub] < 0
+ || search_regs.start[sub] > search_regs.end[sub]
+ || search_regs.end[sub] > XSTRING (string)->size)
+ args_out_of_range (make_number (search_regs.start[sub]),
+ make_number (search_regs.end[sub]));
+ }
if (NILP (fixedcase))
{
/* Decide how to casify by examining the matched text. */
- last = search_regs.end[0];
+ last = search_regs.end[sub];
prevc = '\n';
case_action = all_caps;
is more than one letter long. */
some_multiletter_word = 0;
some_lowercase = 0;
- some_uppercase_initial = 0;
+ some_nonuppercase_initial = 0;
+ some_uppercase = 0;
- for (pos = search_regs.start[0]; pos < last; pos++)
+ for (pos = search_regs.start[sub]; pos < last; pos++)
{
- c = FETCH_CHAR (pos);
+ if (NILP (string))
+ c = FETCH_BYTE (pos);
+ else
+ c = XSTRING (string)->data[pos];
+
if (LOWERCASEP (c))
{
/* Cannot be all caps if any original char is lower case */
some_lowercase = 1;
if (SYNTAX (prevc) != Sword)
- ;
+ some_nonuppercase_initial = 1;
else
some_multiletter_word = 1;
}
else if (!NOCASEP (c))
{
+ some_uppercase = 1;
if (SYNTAX (prevc) != Sword)
- some_uppercase_initial = 1;
+ ;
else
some_multiletter_word = 1;
}
+ else
+ {
+ /* If the initial is a caseless word constituent,
+ treat that like a lowercase initial. */
+ if (SYNTAX (prevc) != Sword)
+ some_nonuppercase_initial = 1;
+ }
prevc = c;
}
and has at least one multiletter word. */
if (! some_lowercase && some_multiletter_word)
case_action = all_caps;
- /* Capitalize each word, if the old text has a capitalized word. */
- else if (some_uppercase_initial)
+ /* Capitalize each word, if the old text has all capitalized words. */
+ else if (!some_nonuppercase_initial && some_multiletter_word)
case_action = cap_initial;
+ else if (!some_nonuppercase_initial && some_uppercase)
+ /* Should x -> yz, operating on X, give Yz or YZ?
+ We'll assume the latter. */
+ case_action = all_caps;
else
case_action = nochange;
}
+ /* Do replacement in a string. */
+ if (!NILP (string))
+ {
+ Lisp_Object before, after;
+
+ before = Fsubstring (string, make_number (0),
+ make_number (search_regs.start[sub]));
+ after = Fsubstring (string, make_number (search_regs.end[sub]), Qnil);
+
+ /* Substitute parts of the match into NEWTEXT
+ if desired. */
+ if (NILP (literal))
+ {
+ int lastpos = -1;
+ /* We build up the substituted string in ACCUM. */
+ Lisp_Object accum;
+ Lisp_Object middle;
+
+ accum = Qnil;
+
+ for (pos = 0; pos < XSTRING (newtext)->size; pos++)
+ {
+ int substart = -1;
+ int subend;
+ int delbackslash = 0;
+
+ c = XSTRING (newtext)->data[pos];
+ if (c == '\\')
+ {
+ c = XSTRING (newtext)->data[++pos];
+ if (c == '&')
+ {
+ substart = search_regs.start[sub];
+ subend = search_regs.end[sub];
+ }
+ else if (c >= '1' && c <= '9' && c <= search_regs.num_regs + '0')
+ {
+ if (search_regs.start[c - '0'] >= 0)
+ {
+ substart = search_regs.start[c - '0'];
+ subend = search_regs.end[c - '0'];
+ }
+ }
+ else if (c == '\\')
+ delbackslash = 1;
+ else
+ error ("Invalid use of `\\' in replacement text");
+ }
+ if (substart >= 0)
+ {
+ if (pos - 1 != lastpos + 1)
+ middle = Fsubstring (newtext,
+ make_number (lastpos + 1),
+ make_number (pos - 1));
+ else
+ middle = Qnil;
+ accum = concat3 (accum, middle,
+ Fsubstring (string, make_number (substart),
+ make_number (subend)));
+ lastpos = pos;
+ }
+ else if (delbackslash)
+ {
+ middle = Fsubstring (newtext, make_number (lastpos + 1),
+ make_number (pos));
+ accum = concat2 (accum, middle);
+ lastpos = pos;
+ }
+ }
+
+ if (pos != lastpos + 1)
+ middle = Fsubstring (newtext, make_number (lastpos + 1),
+ make_number (pos));
+ else
+ middle = Qnil;
+
+ newtext = concat2 (accum, middle);
+ }
+
+ /* Do case substitution in NEWTEXT if desired. */
+ if (case_action == all_caps)
+ newtext = Fupcase (newtext);
+ else if (case_action == cap_initial)
+ newtext = Fupcase_initials (newtext);
+
+ return concat3 (before, newtext, after);
+ }
+
/* 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. */
- SET_PT (search_regs.start[0]);
+ SET_PT (search_regs.start[sub]);
if (!NILP (literal))
Finsert_and_inherit (1, &newtext);
else
for (pos = 0; pos < XSTRING (newtext)->size; pos++)
{
- int offset = point - search_regs.start[0];
+ int offset = PT - search_regs.start[sub];
c = XSTRING (newtext)->data[pos];
if (c == '\\')
if (c == '&')
Finsert_buffer_substring
(Fcurrent_buffer (),
- make_number (search_regs.start[0] + offset),
- make_number (search_regs.end[0] + offset));
- else if (c >= '1' && c <= search_regs.num_regs + '0')
+ make_number (search_regs.start[sub] + offset),
+ make_number (search_regs.end[sub] + offset));
+ else if (c >= '1' && c <= '9' && c <= search_regs.num_regs + '0')
{
if (search_regs.start[c - '0'] >= 1)
Finsert_buffer_substring
make_number (search_regs.start[c - '0'] + offset),
make_number (search_regs.end[c - '0'] + offset));
}
- else
+ else if (c == '\\')
insert_char (c);
+ else
+ error ("Invalid use of `\\' in replacement text");
}
else
insert_char (c);
UNGCPRO;
}
- inslen = point - (search_regs.start[0]);
- del_range (search_regs.start[0] + inslen, search_regs.end[0] + inslen);
+ inslen = PT - (search_regs.start[sub]);
+ del_range (search_regs.start[sub] + inslen, search_regs.end[sub] + inslen);
if (case_action == all_caps)
- Fupcase_region (make_number (point - inslen), make_number (point));
+ Fupcase_region (make_number (PT - inslen), make_number (PT));
else if (case_action == cap_initial)
- upcase_initials_region (make_number (point - inslen), make_number (point));
+ Fupcase_initials_region (make_number (PT - inslen), make_number (PT));
return Qnil;
}
\f
DEFUN ("match-beginning", Fmatch_beginning, Smatch_beginning, 1, 1, 0,
"Return position of start of text matched by last search.\n\
-NUM specifies which parenthesized expression in the last regexp.\n\
- Value is nil if NUMth pair didn't match, or there were less than NUM pairs.\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.")
- (num)
- Lisp_Object num;
+ (subexp)
+ Lisp_Object subexp;
{
- return match_limit (num, 1);
+ 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\
-ARG, a number, specifies which parenthesized expression in the last regexp.\n\
- Value is nil if ARGth pair didn't match, or there were less than ARG pairs.\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.")
- (num)
- Lisp_Object num;
+ (subexp)
+ Lisp_Object subexp;
{
- return match_limit (num, 0);
+ return match_limit (subexp, 0);
}
-DEFUN ("match-data", Fmatch_data, Smatch_data, 0, 0, 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.")
- ()
+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;
{
+ Lisp_Object tail, prev;
Lisp_Object *data;
int i, len;
if (NILP (last_thing_searched))
- error ("match-data called before any match found");
+ return Qnil;
data = (Lisp_Object *) alloca ((2 * search_regs.num_regs)
* sizeof (Lisp_Object));
int start = search_regs.start[i];
if (start >= 0)
{
- if (EQ (last_thing_searched, Qt))
+ if (EQ (last_thing_searched, Qt)
+ || ! NILP (integers))
{
- XFASTINT (data[2 * i]) = start;
- XFASTINT (data[2 * i + 1]) = search_regs.end[i];
+ XSETFASTINT (data[2 * i], start);
+ XSETFASTINT (data[2 * i + 1], search_regs.end[i]);
}
- else if (XTYPE (last_thing_searched) == Lisp_Buffer)
+ else if (BUFFERP (last_thing_searched))
{
data[2 * i] = Fmake_marker ();
Fset_marker (data[2 * i],
else
data[2 * i] = data [2 * i + 1] = Qnil;
}
- return Flist (2 * len + 2, data);
+
+ /* If REUSE is not usable, cons up the values and return them. */
+ if (! CONSP (reuse))
+ return Flist (2 * len + 2, 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)
+ {
+ if (i < 2 * len + 2)
+ XCONS (tail)->car = data[i];
+ else
+ XCONS (tail)->car = 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);
+
+ return reuse;
}
register int i;
register Lisp_Object marker;
+ if (running_asynch_code)
+ save_search_regs ();
+
if (!CONSP (list) && !NILP (list))
list = wrong_type_argument (Qconsp, list);
length * sizeof (regoff_t));
}
- BLOCK_INPUT;
- re_set_registers (&searchbuf, &search_regs, length,
- search_regs.start, search_regs.end);
- UNBLOCK_INPUT;
+ search_regs.num_regs = length;
}
}
}
else
{
- if (XTYPE (marker) == Lisp_Marker)
+ if (MARKERP (marker))
{
if (XMARKER (marker)->buffer == 0)
- XFASTINT (marker) = 0;
+ XSETFASTINT (marker, 0);
else
- XSET (last_thing_searched, Lisp_Buffer,
- XMARKER (marker)->buffer);
+ XSETBUFFER (last_thing_searched, XMARKER (marker)->buffer);
}
CHECK_NUMBER_COERCE_MARKER (marker, 0);
list = Fcdr (list);
marker = Fcar (list);
- if (XTYPE (marker) == Lisp_Marker
- && XMARKER (marker)->buffer == 0)
- XFASTINT (marker) = 0;
+ if (MARKERP (marker) && XMARKER (marker)->buffer == 0)
+ XSETFASTINT (marker, 0);
CHECK_NUMBER_COERCE_MARKER (marker, 0);
search_regs.end[i] = XINT (marker);
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;
+
+/* Called from Flooking_at, Fstring_match, search_buffer, Fstore_match_data
+ if asynchronous code (filter or sentinel) is running. */
+static void
+save_search_regs ()
+{
+ if (!search_regs_saved)
+ {
+ saved_search_regs.num_regs = search_regs.num_regs;
+ saved_search_regs.start = search_regs.start;
+ saved_search_regs.end = search_regs.end;
+ search_regs.num_regs = 0;
+ search_regs.start = 0;
+ search_regs.end = 0;
+
+ search_regs_saved = 1;
+ }
+}
+
+/* Called upon exit from filters and sentinels. */
+void
+restore_match_data ()
+{
+ if (search_regs_saved)
+ {
+ if (search_regs.num_regs > 0)
+ {
+ xfree (search_regs.start);
+ xfree (search_regs.end);
+ }
+ search_regs.num_regs = saved_search_regs.num_regs;
+ search_regs.start = saved_search_regs.start;
+ search_regs.end = saved_search_regs.end;
+
+ search_regs_saved = 0;
+ }
+}
+
/* 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.")
- (str)
- Lisp_Object str;
+ (string)
+ Lisp_Object string;
{
register unsigned char *in, *out, *end;
register unsigned char *temp;
- CHECK_STRING (str, 0);
+ CHECK_STRING (string, 0);
- temp = (unsigned char *) alloca (XSTRING (str)->size * 2);
+ temp = (unsigned char *) alloca (XSTRING (string)->size * 2);
/* Now copy the data into the new string, inserting escapes. */
- in = XSTRING (str)->data;
- end = in + XSTRING (str)->size;
+ in = XSTRING (string)->data;
+ end = in + XSTRING (string)->size;
out = temp;
for (; in != end; in++)
{
register int i;
- searchbuf.allocated = 100;
- searchbuf.buffer = (unsigned char *) malloc (searchbuf.allocated);
- searchbuf.fastmap = search_fastmap;
+ for (i = 0; i < REGEXP_CACHE_SIZE; ++i)
+ {
+ searchbufs[i].buf.allocated = 100;
+ searchbufs[i].buf.buffer = (unsigned char *) malloc (100);
+ searchbufs[i].buf.fastmap = searchbufs[i].fastmap;
+ searchbufs[i].regexp = Qnil;
+ staticpro (&searchbufs[i].regexp);
+ searchbufs[i].next = (i == REGEXP_CACHE_SIZE-1 ? 0 : &searchbufs[i+1]);
+ }
+ searchbuf_head = &searchbufs[0];
Qsearch_failed = intern ("search-failed");
staticpro (&Qsearch_failed);
Fput (Qinvalid_regexp, Qerror_message,
build_string ("Invalid regexp"));
- last_regexp = Qnil;
- staticpro (&last_regexp);
-
last_thing_searched = Qnil;
staticpro (&last_thing_searched);
- defsubr (&Sstring_match);
defsubr (&Slooking_at);
- defsubr (&Sskip_chars_forward);
- defsubr (&Sskip_chars_backward);
- defsubr (&Sskip_syntax_forward);
- defsubr (&Sskip_syntax_backward);
+ defsubr (&Sposix_looking_at);
+ defsubr (&Sstring_match);
+ defsubr (&Sposix_string_match);
defsubr (&Ssearch_forward);
defsubr (&Ssearch_backward);
defsubr (&Sword_search_forward);
defsubr (&Sword_search_backward);
defsubr (&Sre_search_forward);
defsubr (&Sre_search_backward);
+ defsubr (&Sposix_search_forward);
+ defsubr (&Sposix_search_backward);
defsubr (&Sreplace_match);
defsubr (&Smatch_beginning);
defsubr (&Smatch_end);
HCoop / bpt / emacs.git / blobdiff