Bug 107179 – REGRESSION (r134166-134170?): Fixed position div unexpectedly moves after table row removed

• Component: Layout and Rendering
• Status: RESOLVED DUPLICATE

[Kevin M. Dean]

Created attachment 183282 [details]
Before

A table row is removed via javascript and the fixed div jumps down the page until you click.

This in WordPress.

I've set up a wordpress test area if you want to login and take a look. Just respond here or send me an email directly for the access information and instructions to trigger the bug.

[Kevin M. Dean]

Created attachment 183283 [details]
After

[Radar WebKit Bug Importer]

<rdar://problem/13037130>

[Alexey Proskuryakov]

The regression range is very surprising. There weren't any even remotely related changes in r139554-r139561.

> I've set up a wordpress test area if you want to login and take a look. Just respond here

Yes, please.

[Kevin M. Dean]

(In reply to comment #3)
> > I've set up a wordpress test area if you want to login and take a look. Just respond here
> 
> Yes, please.

Access info emailed to you.

[Kevin M. Dean]

(In reply to comment #3)
> The regression range is very surprising. There weren't any even remotely related changes in r139554-r139561.

Re-checking the range. I may have listed the wrong releases.

[Kevin M. Dean]

I see what happened. If I didn't scroll the page down a little first it didn't always show the bug causing me to see the wrong range. New range forthcoming.

[Alexey Proskuryakov]

This is not 100% reliably reproducible, so take this with a grain of salt, but my range is <http://trac.webkit.org/log/trunk/?rev=136455&stop_rev=136447>, with <http://trac.webkit.org/r136452> being the only suspicious revision in that range.

[Kevin M. Dean]

(In reply to comment #7)
> This is not 100% reliably reproducible, so take this with a grain of salt, but my range is <http://trac.webkit.org/log/trunk/?rev=136455&stop_rev=136447>, with <http://trac.webkit.org/r136452> being the only suspicious revision in that range.

In my testing so far, even though it gets really hard to produce, the issue is at least on or before r135805. Still checking...

[Kevin M. Dean]

Anything of interest between r135421-r135472

I've been able to reproduce with r135472, but struggling to get it to happen with r135421

[Kevin M. Dean]

(In reply to comment #9)
> Anything of interest between r135421-r135472
> 
> I've been able to reproduce with r135472, but struggling to get it to happen with r135421

...and no sooner than I said that, r135421 showed the bug.

[Kevin M. Dean]

My latest range is r134166-134170. Anything possible there?

It's definitely easy to reproduce in the most recent builds, but much harder in the old builds. Something must have changed recently that exposes the bug more.

[Alexey Proskuryakov]

Yes, there are two patches that could be related.

There is also a chance that this got broken multiple times, because it certainly at least became much easier to reproduce around my range.

[Kevin M. Dean]

(In reply to comment #12)
> Yes, there are two patches that could be related.
> 
> There is also a chance that this got broken multiple times, because it certainly at least became much easier to reproduce around my range.

Yes, I just confirmed that the range for me where it get's easier to reproduce is r136402-r136460, which your previous suspicious revision r136452 falls within.

[Robert Hogan]

Kevin: can you add instructions for reproducing the bug please? Or even better, a reduced test case?

[Kevin M. Dean]

In WordPress on the Wordfence page, you'll see a New Issues section. Click on the "I have fixed the issue" links to remove rows and trigger the bug.

Once you've removed all the rows, you'll need to click "Start a Wordfence Scan", to bring the Issues back for further testing.

[Kevin M. Dean]

Looks like the problem is no longer occurring in the recent nightly builds or it has gone back to being more difficult to reproduce like in builds prior to r136447

[Beth Dakin]

(In reply to comment #16)
> Looks like the problem is no longer occurring in the recent nightly builds or it has gone back to being more difficult to reproduce like in builds prior to r136447

Kevin, could you send me the access information as well? I will try to narrow down when it became less reproducible or was fixed.

[Kevin M. Dean]

(In reply to comment #17)
> Kevin, could you send me the access information as well? I will try to narrow down when it became less reproducible or was fixed.

Emailed.

[Beth Dakin]

My testing indicates that this was fixed (or made less reproducible) by http://trac.webkit.org/changeset/140758

But that is pretty surprising! Since I am not sure how to explain that, I think we should keep this open for now in case this bug does still exist. Kevin, please let me know if you see the bug again.

[Beth Dakin]

Okay, I'm going to assume this was fixed since we haven't seen this again lately.

*** This bug has been marked as a duplicate of bug 107882 ***