Bug 82074

Summary: REGRESSION: Tables in the Light-Tech phpBB theme (and others) are broken
Product: WebKit Reporter: Eric Kopf <ekopf>
Component: CSSAssignee: Nobody <webkit-unassigned>
Status: NEW ---    
Severity: Normal CC: jchaffraix, robert, webkit-bug-importer
Priority: P1 Keywords: InRadar, Regression
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   
URL: http://demo.phpbb3styles.net/Light-Tech
Attachments:
Description Flags
"Good" layout from previous webkit build.
none
screenshot of "bad" example none

Description Eric Kopf 2012-03-23 11:38:21 PDT
In the Light-Tech phpBB theme, the tables used to display forums are broken so that column headers are not in alignment with column data.

* STEPS TO REPRODUCE
1. Visit http://demo.phpbb3styles.net/Light-Tech

* RESULTS
Columns are not displayed correctly. (see attached screenshots)

Expected: 
Column headers and column data aligned properly.

* REGRESSION
Works: opensource r111699  Fails: opensource r111702
http://trac.webkit.org/log/trunk/?rev=111702&stop_rev=111700

* NOTES
A number of phpBB themes on that site that are broken in the same way, though not all are affected and I did not regress each theme to see if they broke at the same time as the Light-Tech theme.

<rdar://11106275>
Comment 1 Eric Kopf 2012-03-23 11:39:32 PDT
Created attachment 133523 [details]
"Good" layout from previous webkit build.
Comment 2 Julien Chaffraix 2012-03-23 15:34:10 PDT
I don't think I get the issue here. The column headers and the data seems to be aligned to me (even in the screenshot). I tried on an older version of Safari & Chromium and did not see anything that caught my eyes. Could you point it out for me?

The regression range doesn't look good btw. If I had to make a blind bet, I would look around r111742 as it can impact a table layout.
Comment 3 Eric Kopf 2012-03-26 15:55:20 PDT
Created attachment 133914 [details]
screenshot of "bad" example

Not sure why this didn't attach before, but here's an example of the bad layout I was seeing.
Comment 4 Eric Kopf 2012-03-26 15:57:13 PDT
Can't repro this today on any machine, currently testing w/r112115.
Comment 5 Robert Hogan 2012-03-27 11:36:43 PDT
Maybe the site had a webkit-specific hack to handle the behaviour fixed in bug 78412. They reacted mighty fast if that's the case, so I tend to doubt it.

This is a similar report to bug 82193.