1336929 - Fennec sometimes get stuck in a state where there is a perma-bar of white on the top of the screen

Status: VERIFIED FIXED

(Firefox for Android Graveyard :: Toolbar, defect, P3)

Description

[Jan Henning [:JanH]]

Companion to bug 1305969 - occasionally Fennec gets into a state where there is a permanent white space the size of the toolbar immediately *below* the toolbar.

> 4 Feb 2017
> 10:11	gcp Nightly is busted?
> 10:11	gcp I have a bar the size of the awesomebar above all content'

I've seen this myself recently, although I can't say for sure whether only on Nightly/local builds or on Release as well.

Comment 1

[Kartikaya Gupta (email:[e-mail])]

Likely we won't investigate this until bug 1335895 is done.

Comment 2

[Jan Henning [:JanH]]

Found some hopefully easy and reliable STR:
1. Cold start Firefox.
2. As soon as the UI becomes visible, press the Home button to background it again.
3. Wait a bit to give Gecko time to start up.
4. Bring Firefox into the foreground again.

There should now be a permanent bit of white space below the toolbar. After the next background/foreground cycle, this issue will fix itself.

If it helps narrowing things down, I can't reproduce this in 50.

Comment 3

[Kartikaya Gupta (email:[e-mail])]

(In reply to Jan Henning [:JanH] from comment #2)
> If it helps narrowing things down, I can't reproduce this in 50.

If this is a regression then we should probably track down a regression range to avoid shipping it.

Comment 4

[Jan Henning [:JanH]]

Regressing push: https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&revision=8aa58e6577516d8bde4ff9d4555e59d6a511fc4f&filter-searchStr=android%20api%2015%20opt

Comment 5

[Kevin Brosnan [Ex-Mozilla]]

[Tracking Requested - why for this release]: Regression in Fx 53

Comment 6

[Randall Barker [:rbarker]]

My guess is that a re-paint is failing after a resize. Just rotating device and rotating back fixes the issue. I can reproduce the issue. One change that I made in Bug 1328752 is always requesting the repaint from the UI thread instead of requesting it from both UI and main thread which IPC/IPDL doesn't like. I'll look there first.

Comment 7

[Randall Barker [:rbarker]]

The SyncResumeResizeCompositor is failing because the UiCompositorControllerChild hasn't established its connection with the compositor thread yet so its a race condition. Looking into possible solutions now.

Comment 8

[Randall Barker [:rbarker]]

Attachment: 0001-Bug-1336929-Have-UiCompositorControllerChild-cache-surface-resize-when-not-yet-initialized-17020714-57d6624e.patch

Comment 9

[Randall Barker [:rbarker]]

Attachment: 0001-Bug-1336929-Have-UiCompositorControllerChild-cache-surface-resize-when-not-yet-initialized-17020715-cc182a0.patch

Comment 10

[Randall Barker [:rbarker]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=c82ae11a5dbc61368ee073aa37da27e6da2ce853

Comment 11

[alex_mayorga]

Attachment: photo5539583883226490793.jpg

¡Hola Randall!

Is the bug pictured in the attachment this bug?

In my case, the height of the white space is a bit larger than the awesome bar.

Also taps on the screen seem to be miss interpreted by about the same amount of space.

¡Gracias!
Alex

Comment 12

[Randall Barker [:rbarker]]

(In reply to alex_mayorga from comment #11)
> Created attachment 8835580 [details]
> photo5539583883226490793.jpg
> 
> ¡Hola Randall!
> 
> Is the bug pictured in the attachment this bug?
> 
> In my case, the height of the white space is a bit larger than the awesome
> bar.
> 
> Also taps on the screen seem to be miss interpreted by about the same amount
> of space.
> 
> ¡Gracias!
> Alex

This is probably the same bug. The size of the white space has nothing to do with size of the awesome bar but instead what ever size the compositor was initialized with.

Comment 13

[Marcia Knous [:marcia]]

Tracking 53/54+ for this visible regression.

Comment 14

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/6f43298f5271
Have UiCompositorControllerChild cache surface resize when not yet initialized. r=jchen,dvander

Comment 15

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/6f43298f5271

Comment 16

[Ryan VanderMeulen [:RyanVM]]

Please request Aurora approval on this when you get a chance.

Comment 17

[Jan Henning [:JanH]]

I can no longer reproduce this on a local build based on the current mozilla-central as well as today's Nightly.

Comment 19

[Randall Barker [:rbarker]]

Comment on attachment 8834675 [details] [diff] [review]
0001-Bug-1336929-Have-UiCompositorControllerChild-cache-surface-resize-when-not-yet-initialized-17020715-cc182a0.patch

Approval Request Comment
[Feature/Bug causing the regression]:Bug 1328752 
[User impact if declined]: The page content will sometimes be shifted down causing a white bar across the top of the content page and touch events will not align with content. 
[Is this code covered by automated tests?]: No.
[Has the fix been verified in Nightly?]: Yes
[Needs manual test from QE? If yes, steps to reproduce]: No.
[List of other uplifts needed for the feature/fix]: None.
[Is the change risky?]: No
[Why is the change risky/not risky?]: It just adds a simple cache for when the class is not initialized yet.
[String changes made/needed]: None

Comment 20

[Gerry Chang [:gchang]]

Comment on attachment 8834675 [details] [diff] [review]
0001-Bug-1336929-Have-UiCompositorControllerChild-cache-surface-resize-when-not-yet-initialized-17020715-cc182a0.patch

Fix a UI issue that a white bar below the toolbar. Aurora53+.

Comment 21

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/16153c4e477e