[webkit-reviews] review requested: [Bug 30261] Optimization: Reduce calls to StringImpl's upper / lower / isLower : [Attachment 41136] patch 3

[bugzilla-daemon at webkit.org]

Jens Alfke <snej at chromium.org> has asked  for review:
Bug 30261: Optimization: Reduce calls to StringImpl's upper / lower / isLower
https://bugs.webkit.org/show_bug.cgi?id=30261

Attachment 41136: patch 3
https://bugs.webkit.org/attachment.cgi?id=41136&action=review

------- Additional Comments from Jens Alfke <snej at chromium.org>
OK, this fixes those issues. I'm usually reluctant to shadow a member name with
a local, but if that's the convention I'll change it.

I wrote a micro-benchmark that just runs that one method and the
pointer-chasing was faster. Most of the added speed comes from the change to
the way noUpper is set, though; you'd think the two forms would optimize the
same, but they don't.