896099 - [Helix] Extra White or Black background showing up behind the app while minimizing it.

Status: RESOLVED WORKSFORME

(Core :: Graphics, defect)

Description

[Swaroopa Vantipalli [:swaroopa]]

Attachment: Video

Gecko  http://hg.mozilla.org/releases/mozilla-b2g18_v1_1_0_hd/rev/574fd9fa525e
Gaia   2a0b3acd3d10ca285771740c3a907b88a4c053aa
BuildID 20130719070223
Version 18.0

STR:

1.Goto the home screen and from the bottom navigation open Contacts app and click on Home button.
2. Do the same for the messages and Phone apps. Click on the app to open it and then click the home button

Expected behavior:
The change from app view to the homescreen should be neat without any extra distortions.

Actual Behavior:
Observe that as the app screen disappears, either white or black background is shown behind the app.

Attached a video..let me know if this not clear.

Comment 1

[Tim Guan-tin Chien [:timdream] (please needinfo)]

Should be a async animation issue, need to loop gfx ppl. Vivien can help.

Comment 2

[David Zbarsky (:dzbarsky)]

I'm not sure what device this is, but I can't reproduce it on my Unagi.  This might be related to all the async animation warnings we're seeing in the UI.  It may be worth bisecting gaia on a debug b2g build to find the change that regressed this.  Nick, you have any ideas?

Comment 3

[Tim Guan-tin Chien [:timdream] (please needinfo)]

(In reply to David Zbarsky (:dzbarsky) from comment #2)
> I'm not sure what device this is, but I can't reproduce it on my Unagi. 
> This might be related to all the async animation warnings we're seeing in
> the UI.  It may be worth bisecting gaia on a debug b2g build to find the
> change that regressed this.  Nick, you have any ideas?

David, I guess the difference matters on this particular is having |layout.css.devPixelsPerPix| set to 1.5. If you have a SGS2 around you should be able to reproduce it.

Comment 4

[Nick Cameron [:nrc]]

Sounds like bug 849399, could it have reappeared?

Can you repro this on desktop by pref'ing on OMTC and OMTA and zooming in on a page with animation? If so, then it would be good to have a simplified test case so we know if it is a transition on opacity or size, or whether it is due to page loading.

(Another thought is that the flash could be due to loading the page, but the page loading should start when the animation starts, so this seems unlikely).

Comment 5

[Tim Guan-tin Chien [:timdream] (please needinfo)]

(In reply to Nick Cameron [:nrc] from comment #4)
> Sounds like bug 849399, could it have reappeared?
> 
> Can you repro this on desktop by pref'ing on OMTC and OMTA and zooming in on
> a page with animation? If so, then it would be good to have a simplified
> test case so we know if it is a transition on opacity or size, or whether it
> is due to page loading.
> 
> (Another thought is that the flash could be due to loading the page, but the
> page loading should start when the animation starts, so this seems unlikely).

Ouch, I didn't realized this bug has been sit here for 3 weeks....

Anyhow, I turned on 
layers.offmainthreadcomposition.enabled
layers.offmainthreadcomposition.async-animations
and set 
layout.css.devPixelsPerPix
to 1.5 and test on Nightly, and I didn't saw any glitch. This did not rule out anything if I didn't test this correctly or if this only happens on b2g18.

Comment 6

[Nick Cameron [:nrc]]

(In reply to Tim Guan-tin Chien [:timdream] (MoCo-TPE) from comment #5)
> (In reply to Nick Cameron [:nrc] from comment #4)
> > Sounds like bug 849399, could it have reappeared?
> > 
> > Can you repro this on desktop by pref'ing on OMTC and OMTA and zooming in on
> > a page with animation? If so, then it would be good to have a simplified
> > test case so we know if it is a transition on opacity or size, or whether it
> > is due to page loading.
> > 
> > (Another thought is that the flash could be due to loading the page, but the
> > page loading should start when the animation starts, so this seems unlikely).
> 
> Ouch, I didn't realized this bug has been sit here for 3 weeks....
> 
> Anyhow, I turned on 
> layers.offmainthreadcomposition.enabled
> layers.offmainthreadcomposition.async-animations
> and set 
> layout.css.devPixelsPerPix
> to 1.5 and test on Nightly, and I didn't saw any glitch. This did not rule
> out anything if I didn't test this correctly or if this only happens on
> b2g18.

Not sure I can suggest anything else.

Can you repro on the phone but using m-c instead of b2g18? That would implicate 849399, which doesn't look like it landed on b2g18, only inbound. In which case we should try and rebase that and see if it fixes things.

Comment 7

[Tim Guan-tin Chien [:timdream] (please needinfo)]

(In reply to Nick Cameron [:nrc] from comment #6)
> (In reply to Tim Guan-tin Chien [:timdream] (MoCo-TPE) from comment #5)
> > Ouch, I didn't realized this bug has been sit here for 3 weeks....
> > 
> > Anyhow, I turned on 
> > layers.offmainthreadcomposition.enabled
> > layers.offmainthreadcomposition.async-animations
> > and set 
> > layout.css.devPixelsPerPix
> > to 1.5 and test on Nightly, and I didn't saw any glitch. This did not rule
> > out anything if I didn't test this correctly or if this only happens on
> > b2g18.
> 
> Not sure I can suggest anything else.
> 
> Can you repro on the phone but using m-c instead of b2g18? That would
> implicate 849399, which doesn't look like it landed on b2g18, only inbound.
> In which case we should try and rebase that and see if it fixes things.

I cannot reproduce this bug on m-c on the phone.

Comment 8

[Nick Cameron [:nrc]]

(In reply to Tim Guan-tin Chien [:timdream] (MoCo-TPE) from comment #7)
> (In reply to Nick Cameron [:nrc] from comment #6)
> > (In reply to Tim Guan-tin Chien [:timdream] (MoCo-TPE) from comment #5)
> > > Ouch, I didn't realized this bug has been sit here for 3 weeks....
> > > 
> > > Anyhow, I turned on 
> > > layers.offmainthreadcomposition.enabled
> > > layers.offmainthreadcomposition.async-animations
> > > and set 
> > > layout.css.devPixelsPerPix
> > > to 1.5 and test on Nightly, and I didn't saw any glitch. This did not rule
> > > out anything if I didn't test this correctly or if this only happens on
> > > b2g18.
> > 
> > Not sure I can suggest anything else.
> > 
> > Can you repro on the phone but using m-c instead of b2g18? That would
> > implicate 849399, which doesn't look like it landed on b2g18, only inbound.
> > In which case we should try and rebase that and see if it fixes things.
> 
> I cannot reproduce this bug on m-c on the phone.

Great, thanks for experimenting! So, that suggests we need to rebase bug 849399 to b2g18 (or persuade someone that we should use m-c for Helix :-) ). Do you want to give it a go, or do you want to leave it to me? (It shouldn't take too long, but I'm unlikely to get to it until next week some time).

Comment 9

[Wayne Chang [:wchang]]

This seems to have been fixed somehow, I cannot reproduce this on 20130814042202. Closing as WFM, if anyone reproduces this please provide build info and STR details.

Comment 10

[Wayne Chang [:wchang]]

Setting flag for correctness sake.