Summary: | ASSERTION FAILED: d->m_view && !d->m_view->needsLayout() in Frame::Paint | ||||||
---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Matt Lilek <dev+webkit> | ||||
Component: | Layout and Rendering | Assignee: | Nobody <webkit-unassigned> | ||||
Status: | RESOLVED FIXED | ||||||
Severity: | Normal | CC: | mitz | ||||
Priority: | P1 | Keywords: | HasReduction, Regression | ||||
Version: | 523.x (Safari 3) | ||||||
Hardware: | Mac | ||||||
OS: | OS X 10.4 | ||||||
URL: | http://broadband.tsn.ca | ||||||
Attachments: |
|
Description
Matt Lilek
2007-10-06 15:41:27 PDT
Like bug 14899, !d->m_view->needsLayout() is causing this to fail. This bug belongs to the general "letting arbitrary stuff happen under attach()" category. In this case, the plugin that was just created puts up a sheet which changes the window's key status and triggers a redisplay. Reduction: <object type="application/x-oleobject"></object> Created attachment 16825 [details]
Defer updateWidget() until after attach()
Comment on attachment 16825 [details]
Defer updateWidget() until after attach()
Looks like gold to me.
Comment on attachment 16825 [details]
Defer updateWidget() until after attach()
This patch caused test regressions. I think waiting until after layout was a mistake. The updating should happen sooner.
Bug 15804 has a patch that should fix this instance of the bug, but not the general problem. |