[Webkit-unassigned] [Bug 42079] SVG Large curve path segment OOM crash

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=42079





--- Comment #9 from Dirk Schulze <krit at webkit.org>  2010-08-11 23:44:39 PST ---
(In reply to comment #8)
> (In reply to comment #6)
> > Created an attachment (id=64147)
 --> (https://bugs.webkit.org/attachment.cgi?id=64147) [details] [details]
> > Recursive version of curveLength in PathTraversalState
> > 
> > Recursive version of curveLength in PathTraversalState. It avoids the stack and returns earlier, if length is reached in getPointAtLength.
> > 
> > The patch is for testing and not for review. Can someone test it on Skia please? I'd like to know, if the test works now, and how long it takes to get the totalLength of a path.
> 
> Can you explain why you want to test this on different platforms? This is cross-platform code, AFAIK?

Just on the first few.
The problem is, that we use the data of the platform path on pointAtLength as well as on getTotalLength. E.g. Cairo reduces the curve, so that the curve fits into the internal bounds of the canvas. That means we'll get different results across platforms (at least on very big shapes, like the curve in the example).
You may noticed, that for the repro test, opera and FF give different results for the length. This is the reason why.
On WebKitGtk I get the same results like FireFox, because both are using cairo as graphic engine.

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