Summary: | REGRESSION(r26696): GtkLauncher segfaults on WebCore::WidthIterator::advance | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Jan Alonzo <jmalonzo> | ||||||||
Component: | Text | Assignee: | Mark Rowe (bdash) <mrowe> | ||||||||
Status: | RESOLVED FIXED | ||||||||||
Severity: | Normal | CC: | darin, mrowe | ||||||||
Priority: | P2 | Keywords: | Gtk | ||||||||
Version: | 523.x (Safari 3) | ||||||||||
Hardware: | PC | ||||||||||
OS: | Linux | ||||||||||
Attachments: |
|
Description
Jan Alonzo
2007-10-20 18:58:39 PDT
Created attachment 16751 [details]
segfault backtrace
backtrace
Created attachment 16752 [details]
Reduction
I don't see the crash on launch as described, but I can reproduce this when searching Google for "bdash". I've attached a reduction of the page that demonstrates the crash. It appears to be a single Unicode character that is causing the problem.
This was introduced in http://trac.webkit.org/projects/webkit/changeset/26696. Created attachment 16755 [details]
Patch
Comment on attachment 16755 [details]
Patch
Good fix. But we really should structure this so we don't call glyphDataForCharacter twice; it can be an expensive operation. If references make this too tricky, you can use a const GlyphData*.
Comment on attachment 16755 [details]
Patch
r=me, although darin should maybe look at this when he gets a chance.
Darin, GlyphPage::glyphDataForCharacter is always an array lookup so I wouldn't consider it an expensive operation. I landed it after Dave reviewed it on IRC, but I can go ahead and make the change you suggested if you would like. (In reply to comment #8) > Darin, GlyphPage::glyphDataForCharacter is always an array lookup so I wouldn't > consider it an expensive operation. I landed it after Dave reviewed it on IRC, > but I can go ahead and make the change you suggested if you would like. OK. I can live with this the way it is, I guess. |