[Webkit-unassigned] [Bug 37292] http://trac.webkit.org/changeset/57215 caused perf regressions

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=37292





--- Comment #23 from Ojan Vafai <ojan at chromium.org>  2010-04-28 12:05:24 PST ---
(In reply to comment #22)
> (From update of attachment 54597 [details])
> I'd really like someone to take a crack at fixing the underlying problem.  Dan
> has agreed to look into this, but others are welcome to try as well.  Would it
> be possible to leave the #define set to 0 until someone tries and fails to fix
> this?  If it really seems unlikely we'll be able to fix the underlying problem,
> then we can change the default value of the #define.

It's certainly possible, but I don't think we should leave performance
regressions in trunk by default. We wouldn't leave failing tests in the tree,
would we? Performance and memory regressions are even worse since there is only
one scale to measure them on. That means that other changes that affect
performance will mask this one and it becomes nearly impossible to tell if the
fix actually fixes the entirety of the regression.

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