449362 - [FIX]Table border is rendered incorrectly when the table cell size changes due to javascript.

Status: RESOLVED FIXED

(Core :: Layout: Tables, defect)

Description

[Mark Lassau]

User-Agent:       Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.0.1) Gecko/2008072820 Firefox/3.0.1
Build Identifier: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.0.1) Gecko/2008072820 Firefox/3.0.1

The example page has javascript such that when you click on the bottom-right cell of the table, then bottom row must get higher.
when this happens, the bottom border on the left cell is left around after the new border is drawn further down.

If the page loses focus, then it is re-rendered correctly.

Reproducible: Always

Steps to Reproduce:
1.Load the example page.
2.Click on the bottom-right cell of the table.
3.The bottom row will need to expand to fit the new text.
4.Note the border in the bottom-left cell.
Actual Results:  
The bottom-left cell of the table still has a line where the old border used to be, as well as the new border.

Expected Results:  
The old border should disappear and be replaced by the new border.

This works on Firefox 2 and other browsers, the defect is seen on Firefox 3.
The example given is broken on Ubuntu 64 bit, but seems to work on Windows and Mac.
However, this example was distilled down from a more complicated set of HTML, CSS, and javascript where the problem was observed on both Linux and Windows, but not Mac.

Comment 1

[Mark Lassau]

Attachment: Example Page

Comment 2

[Mark Lassau]

Attachment: Screenshot of observed behaviour.

Uploaded a screenshot that shows the observed behaviour.
Note that the act of taking a screenshot got the page to re-render itself correctly, so I doctored the image to show what it looks like before the re-render.

Comment 3

[Ria Klaassen (not reading all bugmail)]

I was not able to reproduce the problem on Windows XP, so it seems to be Linux-only.

Comment 4

[Mark Lassau]

The behaviour was originally observed on both Ubuntu Linux and Windows XP in a web app. 
I used my Linux box to reduce the page down to the simplest possible static html that would reproduce the issue. Unfortunately this page is only broken for Linux.

So the problem is NOT Linux-only. It is just that the example page only reproduces the problem on Linux.

Comment 5

[Mark Lassau]

Attachment: This example is broken on all operating systems

This updated example displays the buggy behaviour on Ubuntu Linux, Win XP, and Mac OSX.

Comment 6

[Mark Lassau]

FYI, the reason the original example was OS-dependent seems to be that the text was changing in the second column, and so that column could change widths, depending on the font of the text.

Also the URL listed above has been updated to the OS-independant version.

Comment 7

[Ria Klaassen (not reading all bugmail)]

(In reply to comment #5)
> Created an attachment (id=332655) [details]

Ok, thanks. Now I can indeed reproduce it on Windows. 

Works: 2008020613
Fails: 2008020614

Regression window is:
http://bonsai.mozilla.org/cvsquery.cgi?module=PhoenixTinderbox&date=explicit&mindate=2008-02-06+12%3A00&maxdate=2008-02-06+15%3A00
Probably caused by Bug 414298.

Comment 8

[Boris Zbarsky [:bzbarsky]]

Attachment: This fixes things

Comment 9

[Reed Loden [:reed]]

Pushed as 17140:7e87332a44d4.

Comment 10

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 334517 [details] [diff] [review]
This fixes things

We should probably fix this on the branch too...

Comment 11

[Boris Zbarsky [:bzbarsky]]

Need to add a test once we have a way to test invalidation.

Comment 12

[Samuel Sidler (old account; do not CC)]

Boris, I'd like to let this bake a couple of days and approve it for landing (in 1.9.0.2) on Monday. Will you have time to land it then?

Comment 13

[Boris Zbarsky [:bzbarsky]]

Yeah, that should be fine.

Comment 14

[Samuel Sidler (old account; do not CC)]

Comment on attachment 334517 [details] [diff] [review]
This fixes things

Approved for 1.9.0.2. Please land in CVS. a=ss

Comment 15

[Boris Zbarsky [:bzbarsky]]

Fixed on branch.

Comment 16

[Al Billings [:abillings - ex-MoCo]]

Verified for 3.02 with Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.0.2) Gecko/2008090212 Firefox/3.0.2.

Comment 17

[Aaron Train [:aaronmt]]

Attachment: reftest invalidation

Comment 18

[Aaron Train [:aaronmt]]

Attachment: reftest invalidation - reference

Comment 19

[Boris Zbarsky [:bzbarsky]]

Pushed those tests as http://hg.mozilla.org/mozilla-central/rev/80abbea018b2

Sorry for the lag on that, and thanks for the tests!

Comment 20

[Boris Zbarsky [:bzbarsky]]

Hmm.  That test actually makes no sense, and will fail randomly (and just did, in one of my try server runs).

I'll try to fix it.

Comment 21

[Boris Zbarsky [:bzbarsky]]

Pushed http://hg.mozilla.org/mozilla-central/rev/72e869771e19 to address that