[webkit-reviews] review granted: [Bug 5677] speed up KWQ timers by using CFRunLoopTimer instead of NSTimer : [Attachment 4640] version that seems to work well

bugzilla-request-daemon at opendarwin.org bugzilla-request-daemon at opendarwin.org
Sat Dec 3 14:00:13 PST 2005


Eric Seidel <macdome at opendarwin.org> has granted Darin Adler
<darin at apple.com>'s request for review:
Bug 5677: speed up KWQ timers by using CFRunLoopTimer instead of NSTimer
http://bugzilla.opendarwin.org/show_bug.cgi?id=5677

Attachment 4640: version that seems to work well
http://bugzilla.opendarwin.org/attachment.cgi?id=4640&action=edit

------- Additional Comments from Eric Seidel <macdome at opendarwin.org>
Shouldn't PausedTimeouts create and destroy the actual array?  Or maybe it
should take them, but construction should be the last step in the fuction. 
Would be more obvious if we had PassRefPtr, but we don't yet.

Also, I'm curious as to why you might wish to use a raw c array and length
pointer (for the paused timeouts array), instead of something like a QMemArray
which would hold the two together for you?

Wow.  That was a big patch.  And I don't feel like my brain managed to soak in
every line of code change, but in general it looke sane, besides the manual
memory management noted above.



More information about the webkit-reviews mailing list