Code

utf8.c: speculatively assume utf-8 in strbuf_add_wrapped_text()
authorRené Scharfe <rene.scharfe@lsrfire.ath.cx>
Fri, 19 Feb 2010 22:20:44 +0000 (23:20 +0100)
committerJunio C Hamano <gitster@pobox.com>
Sat, 20 Feb 2010 17:22:44 +0000 (09:22 -0800)
commit462749b728f72079a67202d4d0d1ef19ef993f61
treed88473ea817cf1c86f8cd3f51c906ef3cccd34f5
parent68ad5e1e9c10e8a640703aadbdf8b8366014373b
utf8.c: speculatively assume utf-8 in strbuf_add_wrapped_text()

is_utf8() works by calling utf8_width() for each character at the
supplied location.  In strbuf_add_wrapped_text(), we do that anyway
while wrapping the lines.  So instead of checking the encoding
beforehand, optimistically assume that it's utf-8 and wrap along
until an invalid character is hit, and when that happens start over.

This pays off if the text consists only of valid utf-8 characters.
The following command was run against the Linux kernel repo with
git 1.7.0:

$ time git log --format='%b' v2.6.32 >/dev/null

real 0m2.679s
user 0m2.580s
sys 0m0.100s

$ time git log --format='%w(60,4,8)%b' >/dev/null

real 0m4.342s
user 0m4.230s
sys 0m0.110s

And with this patch series:

$ time git log --format='%w(60,4,8)%b' >/dev/null

real 0m3.741s
user 0m3.630s
sys 0m0.110s

So the cost of wrapping is reduced to 70% in this case.

Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
utf8.c