Summary: | Resizing Gmail inbox vertically results in whitespace at the bottom of the window (caused by a CSS error) | ||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Darin Fisher (:fishd, Google) <fishd> | ||||||||||||||
Component: | Layout and Rendering | Assignee: | Nobody <webkit-unassigned> | ||||||||||||||
Status: | RESOLVED FIXED | ||||||||||||||||
Severity: | Normal | CC: | bdakin, hyatt, mitz | ||||||||||||||
Priority: | P2 | Keywords: | GoogleBug, HasReduction, InRadar | ||||||||||||||
Version: | 528+ (Nightly build) | ||||||||||||||||
Hardware: | All | ||||||||||||||||
OS: | All | ||||||||||||||||
URL: | http://mail.google.com/ | ||||||||||||||||
Attachments: |
|
Description
Darin Fisher (:fishd, Google)
2008-09-08 01:02:16 PDT
Created attachment 23241 [details]
reduction
Comment on attachment 23241 [details]
reduction
oops.. let me try that again.
Created attachment 23242 [details]
reduction: frame source (click the other one)
Created attachment 23243 [details]
reduction: top frame source (click on me)
If you open attachment 23243 [details], and then resize the window in the vertical direction, increasing the size of the window, you should notice that it does not paint the background properly.
FWIW this is reproducible on Mac as well, but you have to be careful to resize vertical only. You get a white stripe at the bottom. If you resize horizontally even a little the problem fixes itself. Created attachment 23257 [details]
Reduction
Might need to do something similar to the WIP patch attached to bug 15384. Created attachment 23451 [details] Correct reduction While behavior with attachment 23257 [details] is buggy, it does not reflect what happens on Gmail. Created attachment 23455 [details] Patch for this bug and bug 15384 I did try something simpler to address this bug alone (without bug 15384), which was basically for layoutBlockChildren() to always look at grandchildren of anonymous blocks and see if any of them had percentage heights. That code got hit quite frequently, so I think this might be a better solution. Comment on attachment 23455 [details] Patch for this bug and bug 15384 r=me, assuming you are confident it's not a performance regression. Fixed in <http://trac.webkit.org/changeset/36513>. |