rpms/grep
5
0
Fork 0
Code Issues Pull Requests Releases Activity

- Further UTF-8 processing avoided since a '\n' byte is always an

Browse Source 
end-of-line character in that encoding.
This commit is contained in:
Tim Waugh 2004-12-14 16:21:41 +00:00
parent 5b2463aa13
commit c3fd6f2bbf
2 changed files with 102 additions and 44 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

140
grep-2.5.1-egf-speedup.patch
View File

@ -1,16 +1,46 @@
--- grep-2.5.1/src/search.c.egf-speedup 2004-11-05 12:50:25.934736684 +0000
+++ grep-2.5.1/src/search.c 2004-11-05 13:52:33.819394140 +0000
@@ -70,9 +70,6 @@
--- grep-2.5.1/src/search.c 2004-12-14 15:08:58.985159277 +0000
+++ grep-2.5.1/src/search.c 2004-12-14 15:55:21.257729918 +0000
@@ -39,6 +39,9 @@
#ifdef HAVE_LIBPCRE
# include <pcre.h>
#endif
+#ifdef HAVE_LANGINFO_CODESET
+# include <langinfo.h>
+#endif
#define NCHAR (UCHAR_MAX + 1)
@@ -70,9 +73,10 @@
call the regexp matcher at all. */
static int kwset_exact_matches;
-#if defined(MBS_SUPPORT)
-static char* check_multibyte_string PARAMS ((char const *buf, size_t size));
-#endif
+/* UTF-8 encoding allows some optimizations that we can't otherwise
+ assume in a multibyte encoding. */
+static int using_utf8;
+
static void kwsinit PARAMS ((void));
static void kwsmusts PARAMS ((void));
static void Gcompile PARAMS ((char const *, size_t));
@@ -141,47 +138,6 @@
@@ -84,6 +88,15 @@
static size_t Pexecute PARAMS ((char const *, size_t, size_t *, int));
void
+check_utf8 (void)
+{
+#ifdef HAVE_LANGINFO_CODESET
+ if (strcmp (nl_langinfo (CODESET), "UTF-8") == 0)
+ using_utf8 = 1;
+#endif
+}
+
+void
dfaerror (char const *mesg)
{
error (2, 0, mesg);
@@ -141,47 +154,6 @@
}
}
@ -58,7 +88,23 @@
static void
Gcompile (char const *pattern, size_t size)
{
@@ -350,18 +306,8 @@
@@ -190,6 +162,7 @@
size_t total = size;
char const *motif = pattern;
+ check_utf8 ();
re_set_syntax (RE_SYNTAX_GREP | RE_HAT_LISTS_NOT_NEWLINE | (match_icase ? RE_ICASE : 0));
dfasyntax (RE_SYNTAX_GREP | RE_HAT_LISTS_NOT_NEWLINE, match_icase, eolbyte);
@@ -266,6 +239,7 @@
size_t total = size;
char const *motif = pattern;
+ check_utf8 ();
if (strcmp (matcher, "awk") == 0)
{
re_set_syntax (RE_SYNTAX_AWK | (match_icase ? RE_ICASE : 0));
@@ -350,18 +324,8 @@
struct kwsmatch kwsm;
size_t i, ret_val;
#ifdef MBS_SUPPORT
@ -79,7 +125,7 @@
#endif /* MBS_SUPPORT */
buflim = buf + size;
@@ -373,18 +319,48 @@
@@ -373,18 +337,48 @@
if (kwset)
{
/* Find a possible match using the KWset matcher. */
@ -90,7 +136,7 @@
if (offset == (size_t) -1)
goto failure;
+#ifdef MBS_SUPPORT
+ if (MB_CUR_MAX > 1)
+ if (MB_CUR_MAX > 1 && !using_utf8)
+ {
+ bytes_left = offset;
+ while (bytes_left)
@ -130,7 +176,7 @@
while (beg > buf && beg[-1] != eol)
--beg;
if (kwsm.index < kwset_exact_matches)
@@ -395,13 +371,47 @@
@@ -395,13 +389,47 @@
else
{
/* No good fixed strings; start with DFA. */
@ -142,7 +188,7 @@
break;
/* Narrow down to the line we've found. */
+#ifdef MBS_SUPPORT
+ if (MB_CUR_MAX > 1)
+ if (MB_CUR_MAX > 1 && !using_utf8)
+ {
+ bytes_left = offset;
+ while (bytes_left)
@ -178,7 +224,7 @@
while (beg > buf && beg[-1] != eol)
--beg;
}
@@ -469,15 +479,6 @@
@@ -469,15 +497,6 @@
} /* for (beg = end ..) */
failure:
@ -194,7 +240,7 @@
return (size_t) -1;
success_in_beg_and_end:
@@ -486,15 +487,6 @@
@@ -486,15 +505,6 @@
/* FALLTHROUGH */
success_in_start_and_len:
@ -210,7 +256,15 @@
*match_size = len;
return start;
}
@@ -531,17 +523,8 @@
@@ -504,6 +514,7 @@
{
char const *beg, *lim, *err;
+ check_utf8 ();
kwsinit ();
beg = pattern;
do
@@ -531,17 +542,8 @@
struct kwsmatch kwsmatch;
size_t ret_val;
#ifdef MBS_SUPPORT
@ -230,13 +284,13 @@
#endif /* MBS_SUPPORT */
for (beg = buf; beg <= buf + size; ++beg)
@@ -550,8 +533,33 @@
@@ -550,8 +552,33 @@
if (offset == (size_t) -1)
goto failure;
#ifdef MBS_SUPPORT
- if (MB_CUR_MAX > 1 && mb_properties[offset+beg-buf] == 0)
- continue; /* It is a part of multibyte character. */
+ if (MB_CUR_MAX > 1)
+ if (MB_CUR_MAX > 1 && !using_utf8)
+ {
+ size_t bytes_left = offset;
+ while (bytes_left)
@ -266,12 +320,12 @@
#endif /* MBS_SUPPORT */
beg += offset;
len = kwsmatch.size[0];
@@ -587,6 +595,36 @@
@@ -587,6 +614,36 @@
if (offset == -1) {
break; /* Try a different anchor. */
}
+#ifdef MBS_SUPPORT
+ if (MB_CUR_MAX > 1)
+ if (MB_CUR_MAX > 1 && !using_utf8)
+ {
+ size_t bytes_left = offset;
+ while (bytes_left)
@ -303,47 +357,47 @@
beg += offset;
len = kwsmatch.size[0];
}
@@ -597,20 +635,30 @@
@@ -597,19 +654,31 @@
}
failure:
-#ifdef MBS_SUPPORT
+ return -1;
+
+ success:
#ifdef MBS_SUPPORT
- if (MB_CUR_MAX > 1)
- {
+ if (MB_CUR_MAX > 1 && !using_utf8)
{
- if (match_icase)
- free((char *) buf);
- if (mb_properties)
- free(mb_properties);
- }
-#endif /* MBS_SUPPORT */
return -1;
success:
+#ifdef MBS_SUPPORT
+ end = beg + len;
+ while (end < buf + size)
+ {
+ size_t len = mbrlen (end, buf + size - end, &mbs);
+ if (len == (size_t) -1 || len == (size_t) -2 || len == 0)
+ end = beg + len;
+ while (end < buf + size)
+ {
+ memset (&mbs, '\0', sizeof (mbstate_t));
+ len = 1;
+ }
+ if (len == 1 && *end == eol)
+ break;
+ size_t len = mbrlen (end, buf + size - end, &mbs);
+ if (len == (size_t) -1 || len == (size_t) -2 || len == 0)
+ {
+ memset (&mbs, '\0', sizeof (mbstate_t));
+ len = 1;
+ }
+ if (len == 1 && *end == eol)
+ break;
+
+ end += len;
+ }
+ end++;
+#else
+ end += len;
+ }
}
+ else
#endif /* MBS_SUPPORT */
- return -1;
-
- success:
end = memchr (beg + len, eol, (buf + size) - (beg + len));
+
end++;
+#endif /* MBS_SUPPORT */
+ /* Hmm, is this correct for multibyte? */
while (buf < beg && beg[-1] != eol)
--beg;
len = end - beg;
@@ -618,15 +666,6 @@
@@ -618,15 +687,6 @@
success_in_beg_and_len:
*match_size = len;

6
grep.spec
View File

@ -1,7 +1,7 @@
Summary: The GNU versions of grep pattern matching utilities.
Name: grep
Version: 2.5.1
Release: 41
Release: 42
License: GPL
Group: Applications/Text
Source: ftp://ftp.gnu.org/pub/gnu/grep/grep-%{version}.tar.bz2
@ -85,6 +85,10 @@ fi
%{_mandir}/*/*
%changelog
* Tue Dec 14 2004 Tim Waugh <twaugh@redhat.com> 2.5.1-42
- Further UTF-8 processing avoided since a '\n' byte is always an
end-of-line character in that encoding.
* Fri Dec 3 2004 Tim Waugh <twaugh@redhat.com> 2.5.1-41
- Fixed a busy loop in the egf-speedup patch (bug #140781).