[Webkit-unassigned] [Bug 45751] Extend YARR Interpreter with beginning character look-up optimization

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=45751





--- Comment #8 from Peter Varga <pvarga at inf.u-szeged.hu>  2010-10-01 05:33:28 PST ---
(In reply to comment #7)

Thanks, Michael for the advices.

Avoiding this optimization in case of fixed strings was a really good idea
because it significantly reduces the number of cases where the optimization is
searching for beginning characters in case of the v8-regexp test.
The new version of patch has been uploaded at 
https://bugs.webkit.org/show_bug.cgi?id=45748 but I haven't noticed a significant
performance improvement compared to the previous results.

Your other idea to try to avoid the normal matching of prefixes seems to be too 
complex and probably it doesn't yield significant performance improvements on the
v8-regexp test, because the beginning character look-up optimization
isn't able to make this test faster. The cause for this is that there are a lot
of special patterns (eg. which contain assertions like BOL) in v8-regexp.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.