grep: use REG_STARTEND (if available) to speed up regexec
authorBenjamin Kramer <benny.kra@googlemail.com>
Tue, 26 Jan 2010 17:48:36 +0000 (18:48 +0100)
committerJunio C Hamano <gitster@pobox.com>
Tue, 26 Jan 2010 18:44:10 +0000 (10:44 -0800)
BSD and glibc have an extension to regexec which takes a buffer + length pair
instead of a NUL-terminated string. Since we already have the length computed
this can save us a strlen call inside regexec.

Signed-off-by: Benjamin Kramer <benny.kra@googlemail.com>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
grep.c

diff --git a/grep.c b/grep.c
index 8e1f7de7717da95e8954491f5205547f2091aa35..452c2cbae4c4b706c867574c437490cdf7ac9bd5 100644 (file)
--- a/grep.c
+++ b/grep.c
@@ -640,8 +640,15 @@ static int look_ahead(struct grep_opt *opt,
 
                if (p->fixed)
                        hit = !fixmatch(p->pattern, bol, p->ignore_case, &m);
-               else
+               else {
+#ifdef REG_STARTEND
+                       m.rm_so = 0;
+                       m.rm_eo = *left_p;
+                       hit = !regexec(&p->regexp, bol, 1, &m, REG_STARTEND);
+#else
                        hit = !regexec(&p->regexp, bol, 1, &m, 0);
+#endif
+               }
                if (!hit || m.rm_so < 0 || m.rm_eo < 0)
                        continue;
                if (earliest < 0 || m.rm_so < earliest)