Fast Restart with Headless Player Service Causes Crash:
This post is by request of Amazon engineers after a workshop session: I am unable to provide further details here; please get to me through our Amazon contact if you like actually really really need something here cuz they say they have what they need. I do have a video showing this occur so please DM me and I can send to you directly.
If you use the BackHandler.exitApp() [I know of no other proper way to ‘exit’ the app at the moment] and then quickly re-open the app you will get a crash in a release build or a strange bug report when in a debug build: ‘Cannot initialize hmrClient twice’.
Hi @MatthewThies,
Thank you for reporting this. We will check with the concerned team on this. We will let you know if more information is needed.
Regards,
Siva
We also experienced a similar issue.
When we exit the app and then relaunch it, the app doesn’t behave as expected (for us it was just showing a loading spinner and not loading the view).
What we noticed was that when the app is exited, the ‘main’ interactive process instance is killed and the ‘service’ instance still remains. (vda shell vlcm list).
1 Like
I believe there is known issue in the destroy() command of the player client for the headless service; we were initially instructed NOT to use this command but if you ‘vlcm list’ on your device after existing you may see a hanging service (your headless player service). We noticed this as our only recourse for app exit is BackHandler.exitApp() and we always came back in a strange app state when attempting to relaunch from the Kepler UI; there were a lot of rabbit holes here but regardless you might ask your Amazon contact about this destroy() command and work-arounds until they fix the signal 11 crash it currently causes….I just realized now you reported using vlcm list to see your hanging service. Apologies; Me tired!
Closing this topic as issue is fixed/resolved.
