993462 - Call screen is not visible when making a call from the phone app

Status: RESOLVED DUPLICATE of bug 993732

(Core :: DOM: Core & HTML, defect)

Description

[Florin Strugariu [:Bebe]]

Attachment: logcat of the issue

Description:
When making a call the call screen is not visibile

Environment:
Gaia      1958454595b1fa0e061f0652ae965629993f5708
Gecko     https://hg.mozilla.org/mozilla-central/rev/8883360b1edb
BuildID   20140408040204
Version   31.0a1
ro.build.version.incremental=324
ro.build.date=Thu Dec 19 14:04:55 CST 2013

Reproducible manually: YES

STR:
1. Open the Phone app
2. Dial number
3. Tap call

Expected:
The phone app switches to the call screen and we can see the call screen information 

Actual
The phone app closes and we can not stop the call
The call is made and we can talk if the other person answers 
There is no Phone app in recent apps

If we reopen the phone app the call screen is opend

Comment 2

[[:AndreiH]]

This is reproducible also for Messages app.
Sending a SMS to a number also closes the Messages app!

Comment 3

[Florin Strugariu [:Bebe]]

It might be the same issue I see in the stack trace similar errors:

Lot of:
D/HwcUtils( 5515): Skip layer

and:
E/GeckoConsole( 5515): [JavaScript Error: "IndexedDB UnknownErr: IDBTransaction.cpp:864"]

and 

/Gecko   ( 5821): ###!!! [Child][MessageChannel] Error: Channel closing: too late to send/recv, messages will be lost

@JSmith these to tests (Make call and Send SMS) look like to have the same root from the logcat

Comment 4

[:gerard-majax]

6b82f43a5fd93e7e70fc252088f24e61789abdcc is fine, d0849cfd3c632a0bb60df28567d48fc01dae387c is not. Anyone willing to bisect?

Comment 5

[:gerard-majax]

This is very blocking.

Comment 6

[Jason Smith [:jsmith]]

Two things:

1. We need a window
2. We need to check if this happens on 1.4

Comment 8

[:gerard-majax]

(In reply to Jason Smith [:jsmith] from comment #6)
> Two things:
> 
> 1. We need a window
> 2. We need to check if this happens on 1.4

We already have a set of git commit ids that reproduces/does not reproduce.

Comment 9

[Anthony Ricaud (:rik)]

From which (In reply to Alexandre LISSY :gerard-majax from comment #8)
> We already have a set of git commit ids that reproduces/does not reproduce.
From which tree are those commit ids? Gecko/gaia? Mercurial/git.m.o/github?

Comment 10

[Harvey]

Issue does not reproduce on the latest 1.4

Mozilla Inbound Regression Window:

Last Working:
Device: Buri 1.5 MOZ
BuildID: 20140407034726
Gaia: f1a98bfaa3ab2480945bd7018831fd56c61cdc24
Gecko: d459cc1b5a94
Version: 31.0a1
Firmware Version: v1.2-device.cfg

First Broken:
Device: Buri 1.5 MOZ
BuildID: 20140407040527
Gaia: f1a98bfaa3ab2480945bd7018831fd56c61cdc24
Gecko: f94df8c36ab3
Version: 31.0a1
Firmware Verision: V1.2-device.cfg

Note: Gaia is the same on both builds so this seems to be a Gecko issue.

Gecko Pushlog: http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=d459cc1b5a94&tochange=f94df8c36ab3

Comment 11

[Jason Smith [:jsmith]]

The two possible bugs here that could have caused this are both DOM-related, so sending this over to the DOM component.

Comment 12

[Jason Smith [:jsmith]]

No-Jun tested a local backout of bug 906164 and was unable to reproduce the bug.

Comment 13

[:gerard-majax]

(In reply to Jason Smith [:jsmith] from comment #12)
> No-Jun tested a local backout of bug 906164 and was unable to reproduce the
> bug.

I do confirm this, we need to backout this.

Comment 14

[Jason Smith [:jsmith]]

Resolved as part of bug 993732.