907900 - Page Elements overlap right scrollbar

Status: RESOLVED DUPLICATE of bug 906199

(Core :: Web Painting, defect)

Description

[Clint Milner]

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:23.0) Gecko/20100101 Firefox/23.0 (Beta/Release)
Build ID: 20130814063812

Steps to reproduce:

I visited this site: http://tridiv.com/

I re-sized the browser and some of the 3D shapes at the top of the page overlap the vertical scroll bar.  


Actual results:

Scroll bar is still functional, but the scroll bar isn't being rendered on top.


Expected results:

Perhaps a this is a 'feature.' I would expect that nothing can leave the bounds of the page and/or the scroll bars are rendered on top of everything.

Comment 1

[Tiziana Sellitto [:tiziana]]

Thanks for reporting this!
Clint could you attack a screenshot to better understand the problem?

Comment 2

[Markus Dermietzel]

We encountered the same issue in our project. We had a slideshow with css-based-3-d items in it. Firefox 23.0 (MacOSX 10.6.8) ignores the "overflow: hidden" of the surrounding div and renders the items out to right border of the browser on top of the scrollbar.

I suppose, this must be a rendering issue, caused by experimental css-implementation especially of "transform-style: preserve-3d;". When commenting out this line in css, everything works fine (except 3D ;) ). 

You can see the issue here: https://paperc.com/store/overview (online without fix until this evening approx. Aug. 28th 18:00 Central European Time).

Comment 3

[Markus Dermietzel]

Attachment: Screenshot of css-based-3d item overlapping right scrollbar, overflow: hidden of surrounding div is ignored.

Comment 4

[Tiziana Sellitto [:tiziana]]

Markus can you test it on Firefox beta version and/or latest Nightly with a clean profile (https://support.mozilla.org/en-US/kb/profile-manager-create-and-remove-firefox-profiles)?

http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/24.0b6-candidates/
http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/latest-mozilla-central/

It seems to me that this problem has been fixed. I can reproduce it on Firefox 23 but not on the other Firefox channels. 

In Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:24.0) Gecko/20100101 Firefox/24.0 ID:20130826142034 

In Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:26.0) Gecko/20100101 Firefox/26.0
buildID 20130828030202

Comment 5

[Markus Dermietzel]

I tested it with firefox 24.0b6 and the problem did not occur.

So the issue seems to be fixed in future versions.
Thank you!

Comment 6

[Tiziana Sellitto [:tiziana]]

changing this to core product even if I don't know if this should be fixed against the current release or closed.

Comment 7

[Lukas Blakk [:lsblakk] use ?needinfo]

Stephen - does this look like something left over from pulling out the Lion scrollbars from 23 which have been restored in 24?

Comment 8

[Liz Henry (:lizzard) (relman/hg->git project)]

This is a bit odd. I verified this issue happens in FF23.0.1.  It is working correctly in Aurora, and was broken in yesterday's Nightly, fixed in today's Nightly, 26.0a1 (2013-08-28).  I will try narrowing down what's going on here!

Comment 9

[Liz Henry (:lizzard) (relman/hg->git project)]

bisecting a little further, actually it was broken last week and fixed on the 24th. Hope that helps!  

Last good nightly: 2013-08-24
First bad nightly: 2013-08-23

Comment 10

[Stephen A Pohl [:spohl]]

Taking this bug for now. It does seem like it could have been caused by the scrollbar changes, but mozregression is telling me a different story right now. Looking into it.

Comment 11

[Stephen A Pohl [:spohl]]

> OS: Mac OS X → All

Is this true? Does this affect platforms other than Mac OSX?

Comment 12

[Stephen A Pohl [:spohl]]

I bisected this and the patches for bug 841192 seem to have caused this problem:
Last good nightly: 2013-04-05
First bad nightly: 2013-04-06

Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=55f9e3e3dae7&tochange=768af8d8fad4

Due to skipped revisions, the first bad revision could be any of:
changeset:   127797:698d23ec564b
user:        Robert O'Callahan <[e-mail]>
date:        Mon Mar 04 22:55:59 2013 +1300
summary:     Bug 841192. Part 2: Move FrameLayerBuilder::Clip to DisplayItemClip. r=mattwoodrow

changeset:   127798:fb7633e8733e
user:        Robert O'Callahan <[e-mail]>
date:        Mon Mar 04 22:56:00 2013 +1300
summary:     Bug 841192. Part 3: Make DisplayItemClip members private and encapsulate them in a real API. r=mattwoodrow

changeset:   127799:ec3f4cfe3866
user:        Robert O'Callahan <[e-mail]>
date:        Mon Mar 04 22:56:00 2013 +1300
summary:     Bug 841192. Part 4: Create DisplayListClipState. r=mattwoodrow

changeset:   127800:d6f4ddf03ffb
user:        Robert O'Callahan <[e-mail]>
date:        Mon Mar 04 22:56:01 2013 +1300
summary:     Bug 841192. Part 5: Add "current DisplayListClipState" to nsDisplayListBuilder. r=mattwoodrow

changeset:   127801:dc2f8190c7dd
user:        Robert O'Callahan <[e-mail]>
date:        Mon Mar 04 22:56:01 2013 +1300
summary:     Bug 841192. Part 6: Save and restore DisplayListClipState when we start building display lists for a frame. r=mattwoodrow

changeset:   127802:a9214f38c73e
user:        Robert O'Callahan <[e-mail]>
date:        Mon Mar 04 22:56:01 2013 +1300
summary:     Bug 841192. Part 7: Move child->MarkAbsoluteFramesForDisplayList call down. r=mattwoodrow

changeset:   127803:b16876942c8d
user:        Robert O'Callahan <[e-mail]>
date:        Mon Mar 04 22:56:02 2013 +1300
summary:     Bug 841192. Part 8: On encountering out-of-flow frames during display list construction, save their clip rect as well as their dirty rect for later use when we traverse the placeholder. r=mattwoodrow

changeset:   127804:30928f33d63f
user:        Robert O'Callahan <[e-mail]>
date:        Thu Mar 07 00:08:11 2013 +1300
summary:     Bug 841192. Part 9: Add extra APIs to DisplayItemClip to set the current state, intersect with another DisplayItemClip, test intersection with a rect, translate, and provide easy access to a DisplayItemClip object which doesn't clip anything. r=mattwoodrow

changeset:   127805:072417fd4d8e
user:        Robert O'Callahan <[e-mail]>
date:        Thu Mar 07 00:08:13 2013 +1300
summary:     Bug 841192. Part 10: Add extra APIs to DisplayListClipState to constrain the current clip, and an Auto class for safe saving and restoring. r=mattwoodrow

changeset:   127806:c9583f064a2d
user:        Robert O'Callahan <[e-mail]>
date:        Thu Mar 07 00:08:14 2013 +1300
summary:     Bug 841192. Part 11: Make nsDisplayOptionEventGrabber be a non-anonymous display item list wrapper. r=mattwoodrow

changeset:   127807:54952cb8f869
user:        Robert O'Callahan <[e-mail]>
date:        Thu Mar 07 00:08:15 2013 +1300
summary:     Bug 841192. Part 12: Move RoundedRectIntersectsRect from nsDisplayList.cpp to nsLayoutUtils. r=mattwoodrow

changeset:   127808:00fdcb8176a7
user:        Robert O'Callahan <[e-mail]>
date:        Thu Mar 07 00:08:16 2013 +1300
summary:     Bug 841192. Part 13: Rename nsFrame::ApplyOverflowClipping to ShouldApplyOverflowClipping. r=mattwoodrow

changeset:   127809:5f7f4e26d4a9
user:        Robert O'Callahan <[e-mail]>
date:        Mon Mar 04 22:56:02 2013 +1300
summary:     Bug 841192. Part 14: Convert all usage of nsDisplayClip(RoundedRect) to use DisplayListClipState/DisplayItemClip. r=mattwoodrow

changeset:   127810:3396e302a1e4
user:        Robert O'Callahan <[e-mail]>
date:        Fri Apr 05 00:36:45 2013 +1300
summary:     Bug 841192. Part 15: Move DisplayListClipState clipping methods to AutoSaveRestore and AutoClipMultiple helper classes for safer usage. r=mattwoodrow



The issue got fixed between 2013-07-26 and 2013-07-27, most likely by the patches for bug 896443:
Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=46d73e889cb4&tochange=fb48c7d58b8b


I'm not sure what our usual process is when later branches have a fix. Rob, is it sufficient for the fix to be in Fx24? Do you see anything else we need to do here?

Comment 13

[Alice0775 White]

Sorry, OS change. 
Same thing happens in Windows7 too, however regression and fixing range was different.

FYI, In windows7,
Regression window(m-i)
Good:
http://hg.mozilla.org/integration/mozilla-inbound/rev/905671d954c5
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:23.0) Gecko/20130405 Firefox/23.0 ID:20130405053752
Bad:
http://hg.mozilla.org/integration/mozilla-inbound/rev/1e958f7224a7
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:23.0) Gecko/20130405 Firefox/23.0 ID:20130405054152
Pushlog:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=905671d954c5&tochange=1e958f7224a7

Regressed by: Bug 841192

Fixed window(m-i)
Bad:
http://hg.mozilla.org/integration/mozilla-inbound/rev/b1e97a8eea0f
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:25.0) Gecko/20130726 Firefox/25.0 ID:20130726083824
Good:
http://hg.mozilla.org/integration/mozilla-inbound/rev/2e14be2e3cad
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:25.0) Gecko/20130726 Firefox/25.0 ID:20130726084622
Pushlog:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=b1e97a8eea0f&tochange=2e14be2e3cad
Fixed by:
2e14be2e3cad	Stephen Pohl — Bug 896443: Followup for Windows: Fix the z-ordering of overlay scrollbars to make them appear on top of content with z-index > 0. r=roc

Comment 14

[Stephen A Pohl [:spohl]]

Ah, interesting. This actually matches what I found, just that I bisected mozilla-central. Let's set this back to OS: All. Thanks!

Comment 15

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I'm not sure why bug 896443 fixed anything here, but this is definitely bug 906199 (fixed in next Firefox release).

Comment 16

[Liz Henry (:lizzard) (relman/hg->git project)]

Stephen and Robert, my concern was also it keeps breaking in Nightly. Cheers!

Comment 17

[Stephen A Pohl [:spohl]]

(In reply to Liz Henry :lizzard from comment #16)
> Stephen and Robert, my concern was also it keeps breaking in Nightly. Cheers!

Your comment 9 mentions that this is no longer an issue in 8/24. This would make sense, since bug 906199 landed on 8/23. Are you saying that something is still broken after 8/24?