1103165 - Clipped image layer

Status: RESOLVED DUPLICATE of bug 1098266

(Core :: Graphics, defect)

Description

[Markus Stange [:mstange]]

Attachment: screen recording with layer borders

STR:
1. Go to http://www.yelp.com/biz/bar-karaoke-lounge-toronto
2. Look at the three pictures and wait for them to disappear.
3. Scroll down while the images are gone.
4. See the first image reappear in a clip that stays fixed relative to the window.
5. After a short time the images reappear again.

Ideally, the pictures shouldn't disappear in the first place.

Comment 1

[Alice0775 White]

[Tracking Requested - why for this release]: regression

Regression window:
Good:
https://hg.mozilla.org/integration/mozilla-inbound/rev/d7680ff0e960
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:34.0) Gecko/20100101 Firefox/34.0 ID:20140807194932
Bad:
https://hg.mozilla.org/integration/mozilla-inbound/rev/23ea95cbd92f
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:34.0) Gecko/20100101 Firefox/34.0 ID:20140807204245
Pushlog:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=d7680ff0e960&tochange=23ea95cbd92f

In local Build
Last Good; 5948f9164d2d
First Bad: 23ea95cbd92f
Regressed by;
23ea95cbd92f	Timothy Nikkel — Bug 1042772. Update the reference frame and current frame offset when moving to outside a transform so that we can use the correct offset to compute the initial visible rect for wrap list display items. r=roc * * * The current reference frame is still the same as our reference frame because we set and restore it in nsFrame::BuildDisplayListForChild before this. So we need to actually compute it.

Comment 2

[Timothy Nikkel (:tnikkel)]

(In reply to Markus Stange [:mstange] from comment #0)
> 4. See the first image reappear in a clip that stays fixed relative to the
> window.
> 5. After a short time the images reappear again.
> 
> Ideally, the pictures shouldn't disappear in the first place.

I just tried quickly but I wasn't able to reproduce 4. I can reproduce the pictures disappearing completely, that part at least is fixed by the patch in bug 1098266 with the additional change that I suggested there.

Comment 3

[Lawrence Mandel [:lmandel] (use needinfo)]

Given that we're building 34 RC today, we don't have a patch, and this doesn't look like a stop ship issue, I think we're going to have to live with this in 34. I'm going to mark for tracking to consider as a ride-along in case we need to do a point release.

Comment 4

[Timothy Nikkel (:tnikkel)]

Bug 1098266 is fixed now. Do you still see any issue?

Comment 5

[Markus Stange [:mstange]]

I can no longer reproduce this in today's nightly. Thanks!

Comment 6

[Lawrence Mandel [:lmandel] (use needinfo)]

Clearing the tracking and status flags. The flags have been set on bug 1098266.