1016782 - Can't escape from camera app when started from lockscreen

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: Gaia::System::Lockscreen, defect)

Description

[Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧]

STR: Open camera app, then try to go to the homescreen or app switcher.  This is on m-c/master, with a full flash (either with the boot.img work in progress from bug 1004195 or with the original boot.img).

What happens: The phone vibrates, but stays in the camera app.  It's possible to access the Settings app via the button in the notifications tray, because that still works, but no other apps (including the homescreen) can be accessed.  Even killing the content process for the camera app doesn't restore control of the phone; it's necessary to stop/start B2G or reboot.

Comment 1

[Mike Habicher [:mikeh] (high bugzilla latency)]

In case it's related, I've run into Lockscreen lockdown in bug 1016616.

Diego, I'm pretty sure this isn't caused by the Camera app.

Comment 2

[Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧]

Addendum to STR: it seems that the Camera app has to be opened from the lockscreen, and that has to be the first thing done after boot — if I go to the homescreen first and start the Camera app from there, or go to the homescreen then re-lock and *then* swipe for the Camera app, things work as expected.  There isn't a passcode.  Also maybe relevant: this is an eng build with NOFTU=1.

Comment 3

[Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧]

This also reproduces on keon.

Comment 4

[Jed Davis [:jld] ⟨⏰|UTC-7⟩ ⟦he/him⟧]

…and on emulator-x86-kk.  

It seems as if the lockscreen is starting the camera app but not marking the system as unlocked somehow.

Interestingly re comment #1, this doesn't reproduce if I set a passcode — that goes through a different path in ls_activateCamera.

I've added some logging and found that if I unlock-to-camera immediately (and the bug occurs), we enter ls_unlock with window.AppWindowManager == null, but if I unlock-to-camera later (and the bug doesn't occur), window.AppWindowManager != null.  So maybe it's something about that case.

Comment 5

[BMO Automation]

Firefox OS is not being worked on