Summary: | ASSERT when clicking and dragging on SVG clock | ||||||
---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Eric Seidel (no email) <eric> | ||||
Component: | SVG | Assignee: | Nobody <webkit-unassigned> | ||||
Status: | RESOLVED FIXED | ||||||
Severity: | Normal | ||||||
Priority: | P2 | ||||||
Version: | 523.x (Safari 3) | ||||||
Hardware: | Mac | ||||||
OS: | OS X 10.4 | ||||||
URL: | http://oskamp.dyndns.org/SiemensClock/SVG/ | ||||||
Bug Depends on: | 15281 | ||||||
Bug Blocks: | |||||||
Attachments: |
|
Description
Eric Seidel (no email)
2007-09-28 09:25:12 PDT
This bug in fact appears identical to the one I fixed last week. There must still be something wrong in how DOMSVGImageElement is behaving. This is the same bug as bug 15281, but it appears it's not quite fixed (at least not for all examples). Well, this is the problem: 2007-09-28 21:06:41.521 Safari[3227] Creating wrapper for: svg image http://www.w3.org/2000/svg, using class: DOMSVGElement Now the question is just *why* is it creating the wrong DOM wrapper. Ok, the problem is: typedef HashMap<const QualifiedName::QualifiedNameImpl*, Class> ObjCClassMap; static ObjCClassMap* elementClassMap; It's doing a default pointer-compare between QualifiedNameImpls, which is wrong. We don't care what the .prefix() is, only that the localName and namespaceURI are identical. I expect we have this bug other places in WebKit! Created attachment 16443 [details]
the fix
Comment on attachment 16443 [details]
the fix
r=me for the Feature branch with the test that Eric showed me.
|