Bug 3509 – REGRESSION(125-412) Portion of nested table is not painted correctly on cafepress site

• Status: RESOLVED
• Resolution: FIXED
• Component: Tables

[Chris Petersen]

Expected Results: No painting issue occurs when table is rendered

What I got: A portion of the nested table is not painted correctly. This issue only happens with Version 
2.0 (412) since I can't reproduce under Safari 1.2.4 (10.3.8)

Steps to reproduce:

1) Under 10.4.1 (8B15), launch Safari Version 2.0 (412)
2) Go to http://www.cafepress.com/thestudiostore.4699899
3) After page is displayed, a portion of the table is painted incorrectly. See attached screen shot of 
page.


3/7/05 10:35 AM chris petersen:
I have created a reduced test case.
It contains two tables (one nested is the parent table) using width percentages. Also, the child table is 
nested in a form element. You  should be able to reproduce loading the sample test case. If not, please 
try several reloads of the page.

[Chris Petersen]

Created attachment 2301 [details]
Reduced test case of problem

[Chris Petersen]

Apple Bug: <rdar://problem/4039615>

[Chris Petersen]

Created attachment 2302 [details]
Page screen shot

[Chris Petersen]

Created attachment 2303 [details]
Test case screen shot

[Joost de Valk (AlthA)]

Can't confirm this one i'm afraid, works on 13 june build.

[Chris Petersen]

This issue is still occuring on TOT WebKit (06/13) for me. The best way to reproduce for me is too:

Save a copy of test case to local drive
Reset Safari and quit
Relaunch Safari and drag the test case into the browser window
Table isn't correctly rendered (same as screen shot).

[Daniel Udey]

Created attachment 4653 [details]
Test case screen shot - ToT

This bug is still present in ToT Nov 9th. It seems that it can only be
reproduced once per browser session - further reloads seemed to render
properly, likely making this a tricky bug to track.

[Eric Seidel (no email)]

I'm pretty sure this is a dupe of 4899, or rather the other way around.  As a regression, this gets a p1.

[Eric Seidel (no email)]

*** Bug 4899 has been marked as a duplicate of this bug. ***

Created attachment 5384 [details]
Further reduced testcase

The bug seems to be about "quirk margins".

In the "further reduced testcase", you can Select All after clicking test to trigger repainting of the missing 
piece.

[David Kilzer (:ddkilzer)]

Another litmus test for this bug is to hide Safari, then switch back to it to see the correct rendering.

[David Kilzer (:ddkilzer)]

*** Bug 6278 has been marked as a duplicate of this bug. ***

Created attachment 5807 [details]
Force child repaint if some of its layout was done at an intermediate position

[Darin Adler]

Comment on attachment 5807 [details]
Force child repaint if some of its layout was done at an intermediate position

Looks like a great solution to this issue. r=me

[Joost de Valk (AlthA)]

Removing keywords cause bug is closed.

[Eric Seidel (no email)]

Removing Regression keyword from bugs already fixed.