[Webkit-unassigned] [Bug 40779] [Chromium] Perf regression following landing of HTML5 parser

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=40779





--- Comment #5 from Tony Gentilcore <tonyg at chromium.org>  2010-06-17 13:37:25 PST ---
(In reply to comment #4)
> (In reply to comment #3)
> > > Notice that in the reference run, all times are 3 digits. In the sample run, most times seem to line up with the occasional 5 digit time.
> > 
> > Does that correspond to individual pages that we're bad at?
> 
> I talked to chase about this. He was unclear if they should line up or not.
> 
> One strange thing is that the number of times is not evenly divisible by the number of urls.
> 
> > 
> > @tonyg, would you be willing to investigate?
> 
> I'm trying to investigate now, but haven't gotten any solid leads yet.

Finally have a solid local repro.

Our local versions of news.163.com and sport24.gr will repro consistently.

I'm still determining where/how we stall.

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