Summary: | REGRESSION (r18574): User profile pages display incorrectly at ebay.ca | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | crastinator | ||||||
Component: | Tables | Assignee: | Nobody <webkit-unassigned> | ||||||
Status: | RESOLVED FIXED | ||||||||
Severity: | Normal | CC: | mitz | ||||||
Priority: | P1 | Keywords: | NeedsReduction, Regression | ||||||
Version: | 420+ | ||||||||
Hardware: | Mac | ||||||||
OS: | OS X 10.4 | ||||||||
URL: | http://ebay.ca | ||||||||
Attachments: |
|
Description
crastinator
2007-02-14 23:50:44 PST
Created attachment 13183 [details] Partial reduction The regression happened in r18574 (fix for bug 5164). Changing the percentScaleFactor in Length.h to 1 fixes the bug and this reduction, making me think that the bug has to do with one of the FIXMEs in AutoTableLayout.cpp. Created attachment 13184 [details] Fix table layout bug The actual bug predates r18574 and fixing it gives better results in two existing tests. The reason for the regression is that the higher accuracy made it happen even in this bug's particular case (you can change the yellow cell's width in the reduction from 4 to 20 to get a case that fails in pre-r18574 builds). (In reply to comment #2) > Created an attachment (id=13184) [edit] > Fix table layout bug The ChangeLog says that tables/mozilla_expected_failures/core/col_span2.html is "fixed", so should it move out of the mozilla_expected_failures directory? (In reply to comment #3) > The ChangeLog says that tables/mozilla_expected_failures/core/col_span2.html is > "fixed", so should it move out of the mozilla_expected_failures directory? > That is a long test with many individual cases. In some of them WebKit's layout still don't match Firefox's, so I didn't want to move it. Perhaps a more appropriate phrasing for the change log would be "Covered by existing layout tests:". Comment on attachment 13184 [details]
Fix table layout bug
r=me
|