Closed Bug 832563 Opened 12 years ago Closed 12 years ago

ScriptTimeoutException when attempting to kill the video app

Categories

(Remote Protocol :: Marionette, defect)

ARM
Android
defect
Not set
major

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: davehunt, Assigned: mdas)

References

Details

Attachments

(1 file)

We're having issues killing the video app, which is hanging and causing any following tests from running.
Assignee: nobody → mdas
Severity: normal → major
So far, I can see the 'appterminated' event is fired and the video app appears to close, but it is still listed when we do getRunningApps (perhaps it's just being minimized?). I'm currently investigating further.
When I experienced this issue I found that I could not even terminate the video player on the phone manually. It would appear in the cards view, and killing it from there (either via upwards flick or "x" button) would appear to kill it, but bringing up the cards view again would show it as still running. The only way I could get rid of it was by rebooting the phone.
(In reply to Bob Silverberg from comment #2) > When I experienced this issue I found that I could not even terminate the > video player on the phone manually. It would appear in the cards view, and > killing it from there (either via upwards flick or "x" button) would appear > to kill it, but bringing up the cards view again would show it as still > running. The only way I could get rid of it was by rebooting the phone. How did the phone get in this state? After running a test, or by using it manually?
Vivien might also be seeing this, as per IRC: 15:48 mdas: vingtetun: do you know where it gets removed from the runningApps list? I just see it getting a new 'killed' attribute, but not being removed from the runningApps list. 15:49 vingtetun: mdas: i would say the removeFrame method should do it. looking.... 15:49 mdas: thanks 15:49 vingtetun: delete runningApps[origin] 15:50 vingtetun: it should be removed 15:50 vingtetun: or maybe something goes wrong before this line is called and the frame is never removed... 15:51 vingtetun: that would explain a strange behavior i have seen with the Cards View 15:53 vingtetun: i have seen an issue where an application that is killed is still in the cards view
(In reply to Jonathan Griffin (:jgriffin) from comment #3) > (In reply to Bob Silverberg from comment #2) > > When I experienced this issue I found that I could not even terminate the > > video player on the phone manually. It would appear in the cards view, and > > killing it from there (either via upwards flick or "x" button) would appear > > to kill it, but bringing up the cards view again would show it as still > > running. The only way I could get rid of it was by rebooting the phone. > > How did the phone get in this state? After running a test, or by using it > manually? That happened after running a test a few times, so perhaps it was the killAll that put it into that state.
I feel like the issue lives in https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/window_manager.js#L1787 This code call the removeFrame function (the one that is responsible of removing the app from the runningApps array) only once the transition from the application to the homescreen is finished. But if any other transition happens at the same time then the callback is cleared and never called apparently. Someone need to dig more but that seems a good place to investigate.
(In reply to Vivien Nicolas (:vingtetun) (:21) from comment #6) > I feel like the issue lives in > https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/ > window_manager.js#L1787 > > This code call the removeFrame function (the one that is responsible of > removing the app from the runningApps array) only once the transition from > the application to the homescreen is finished. But if any other transition > happens at the same time then the callback is cleared and never called > apparently. Someone need to dig more but that seems a good place to > investigate. So the transistion to homescreen completes, and I can confirm that removeFrame() and the `delete runningApps(origin)` line is called, but for some reason, when we call getRunningApps, video is still listed. I've raised Bug 834293 about this problem.
I believe this has been fixed for a while.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
It's not been causing us an issue because we perform a restart between tests, but I can attempt to replicate this with a latest build and no restart to see if it's still relevant.
I can confirm this is no longer an issue.
Thanks Dave.
Product: Testing → Remote Protocol
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: