From 9fa0ec14fd4a1209bb7a30911e3b44bc14000fb7 Mon Sep 17 00:00:00 2001 From: Tim Waugh Date: Nov 04 2004 15:17:42 +0000 Subject: - More improvements to egf-speedup patch (bug #138076). --- diff --git a/grep-2.5.1-egf-speedup.patch b/grep-2.5.1-egf-speedup.patch index 76d97a0..a646df5 100644 --- a/grep-2.5.1-egf-speedup.patch +++ b/grep-2.5.1-egf-speedup.patch @@ -1,5 +1,5 @@ ---- grep-2.5.1/src/search.c.egf-speedup 2004-11-04 10:42:45.000000000 +0000 -+++ grep-2.5.1/src/search.c 2004-11-04 11:12:44.688292744 +0000 +--- grep-2.5.1/src/search.c.egf-speedup 2004-11-04 14:34:20.883140606 +0000 ++++ grep-2.5.1/src/search.c 2004-11-04 14:53:29.991266613 +0000 @@ -70,9 +70,6 @@ call the regexp matcher at all. */ static int kwset_exact_matches; @@ -79,30 +79,34 @@ #endif /* MBS_SUPPORT */ buflim = buf + size; -@@ -376,15 +322,34 @@ +@@ -373,18 +319,42 @@ + if (kwset) + { + /* Find a possible match using the KWset matcher. */ ++#ifdef MBS_SUPPORT ++ size_t bytes_left = 0; ++#endif /* MBS_SUPPORT */ size_t offset = kwsexec (kwset, beg, buflim - beg, &kwsm); if (offset == (size_t) -1) goto failure; +#ifdef MBS_SUPPORT + if (MB_CUR_MAX > 1) + { -+ size_t n = offset; -+ while (n) ++ bytes_left = offset; ++ while (bytes_left) + { -+ size_t len = mbrlen (beg, n, &mbs); ++ size_t len = mbrlen (beg, bytes_left, &mbs); + if (len == (size_t) -1 || len == (size_t) -2 || len == 0) + { + /* Incomplete character. */ + memset (&mbs, '\0', sizeof (mbstate_t)); -+ beg += n; ++ beg += bytes_left; + break; + } + + beg += len; -+ n -= len; ++ bytes_left -= len; + } -+ if (n) -+ continue; + } + else +#endif /* MBS_SUPPORT */ @@ -111,44 +115,58 @@ run it through DFA. */ end = memchr(beg, eol, buflim - beg); end++; --#ifdef MBS_SUPPORT + #ifdef MBS_SUPPORT - if (MB_CUR_MAX > 1 && mb_properties[beg - buf] == 0) -- continue; ++ if (MB_CUR_MAX > 1 && bytes_left) + continue; -#endif ++#endif /* MBS_SUPPORT */ while (beg > buf && beg[-1] != eol) --beg; if (kwsm.index < kwset_exact_matches) -@@ -399,6 +364,29 @@ +@@ -395,13 +365,41 @@ + else + { + /* No good fixed strings; start with DFA. */ ++#ifdef MBS_SUPPORT ++ size_t bytes_left = 0; ++#endif /* MBS_SUPPORT */ + size_t offset = dfaexec (&dfa, beg, buflim - beg, &backref); if (offset == (size_t) -1) break; /* Narrow down to the line we've found. */ +#ifdef MBS_SUPPORT + if (MB_CUR_MAX > 1) + { -+ size_t n = offset; -+ while (n) ++ bytes_left = offset; ++ while (bytes_left) + { -+ size_t len = mbrlen (beg, n, &mbs); ++ size_t len = mbrlen (beg, bytes_left, &mbs); + if (len == (size_t) -1 || len == (size_t) -2 || len == 0) + { + /* Incomplete character. */ + memset (&mbs, '\0', sizeof (mbstate_t)); -+ beg += n; ++ beg += bytes_left; + break; + } + + beg += len; -+ n -= len; ++ bytes_left -= len; + } -+ if (n) -+ continue; + } + else +#endif /* MBS_SUPPORT */ beg += offset; end = memchr (beg, eol, buflim - beg); end++; -@@ -469,15 +457,6 @@ ++#ifdef MBS_SUPPORT ++ if (MB_CUR_MAX > 1 && bytes_left) ++ continue; ++#endif /* MBS_SUPPORT */ + while (beg > buf && beg[-1] != eol) + --beg; + } +@@ -469,15 +467,6 @@ } /* for (beg = end ..) */ failure: @@ -164,7 +182,7 @@ return (size_t) -1; success_in_beg_and_end: -@@ -486,15 +465,6 @@ +@@ -486,15 +475,6 @@ /* FALLTHROUGH */ success_in_start_and_len: @@ -180,7 +198,7 @@ *match_size = len; return start; } -@@ -531,17 +501,8 @@ +@@ -531,17 +511,8 @@ struct kwsmatch kwsmatch; size_t ret_val; #ifdef MBS_SUPPORT @@ -200,7 +218,7 @@ #endif /* MBS_SUPPORT */ for (beg = buf; beg <= buf + size; ++beg) -@@ -550,8 +511,28 @@ +@@ -550,8 +521,28 @@ if (offset == (size_t) -1) goto failure; #ifdef MBS_SUPPORT @@ -208,54 +226,54 @@ - continue; /* It is a part of multibyte character. */ + if (MB_CUR_MAX > 1) + { -+ size_t n = offset; -+ while (n) ++ size_t bytes_left = offset; ++ while (bytes_left) + { -+ size_t len = mbrlen (beg, n, &mbs); ++ size_t len = mbrlen (beg, bytes_left, &mbs); + if (len == (size_t) -1 || len == (size_t) -2 || len == 0) + { + /* Incomplete character. */ + memset (&mbs, '\0', sizeof (mbstate_t)); -+ beg += n; ++ beg += bytes_left; + break; + } + + beg += len; -+ n -= len; ++ bytes_left -= len; + } + -+ if (n) ++ if (bytes_left) + continue; + } + else #endif /* MBS_SUPPORT */ beg += offset; len = kwsmatch.size[0]; -@@ -587,6 +568,31 @@ +@@ -587,6 +578,31 @@ if (offset == -1) { break; /* Try a different anchor. */ } +#ifdef MBS_SUPPORT + if (MB_CUR_MAX > 1) + { -+ size_t n = offset; -+ while (n) ++ size_t bytes_left = offset; ++ while (bytes_left) + { -+ size_t len = mbrlen (beg, n, &mbs); ++ size_t len = mbrlen (beg, bytes_left, &mbs); + if (len == (size_t) -1 || len == (size_t) -2 || + len == 0) + { + /* Incomplete character. */ + memset (&mbs, '\0', sizeof (mbstate_t)); -+ beg += n; ++ beg += bytes_left; + break; + } + + beg += len; -+ n -= len; ++ bytes_left -= len; + } + -+ if (n) ++ if (bytes_left) + break; /* Try a different anchor. */ + } + else @@ -263,7 +281,7 @@ beg += offset; len = kwsmatch.size[0]; } -@@ -597,20 +603,30 @@ +@@ -597,20 +613,30 @@ } failure: @@ -303,7 +321,7 @@ while (buf < beg && beg[-1] != eol) --beg; len = end - beg; -@@ -618,15 +634,6 @@ +@@ -618,15 +644,6 @@ success_in_beg_and_len: *match_size = len; diff --git a/grep.spec b/grep.spec index 54c0646..a4474b7 100644 --- a/grep.spec +++ b/grep.spec @@ -1,7 +1,7 @@ Summary: The GNU versions of grep pattern matching utilities. Name: grep Version: 2.5.1 -Release: 33 +Release: 34 License: GPL Group: Applications/Text Source: ftp://ftp.gnu.org/pub/gnu/grep/grep-%{version}.tar.bz2 @@ -81,6 +81,9 @@ fi %{_mandir}/*/* %changelog +* Thu Nov 4 2004 Tim Waugh 2.5.1-34 +- More improvements to egf-speedup patch (bug #138076). + * Thu Nov 4 2004 Tim Waugh 2.5.1-33 - Small improvements to egf-speedup patch.