02-28-2013 10:17 AM
I have an issue that i was not able to trace, and which seems to be linked with OS misbehavior as well, all in all quite strange.
How to trigger it: I have no idea. We had different hypotheses, but the issue is not reproducable.
No clicks are executed, with the exception of the back button. Peek is also possible, and a long click opens a context menu (if present). The context menu can be expanded, but triggering actions does not do anything.
I can show the application menu by swiping down, but pressing the actions does not do anything.
When i close the application the app symbol stays greyed out forever, and the target navigator shows that the application is actually running. However, it is not shown in the "active apps view".
On a long press on the main screen there is no delete icon.
The last log entry is "Application shrinks to thumbnail" which we log ourselves when the app signals thumbnail().
To resolve this situation i can either send a SIGTERM to the application process or reboot the device.
Did somebody see something similar happen?
As i don't know how to reproduce the issue i am hesistant to log something in JIRA.
02-28-2013 10:31 AM
02-28-2013 10:51 AM
Hi, not going to help you much, sorry, but whenever I've seen this happening it's usually because of some kind of memory corruption or the app getting stuck in a loop trying to free up resources on shutdown.
On going through the process of wiping out all memory leaks in order to pass B4BB this behaviour subsequently went away.
02-28-2013 11:23 AM
02-28-2013 11:39 AM
I know what you mean some of the cascades componants took me by surprise for when to delete, deleteLater, leave alone etc.
I just used the momentic qnx memory analysis tool which worked well for me, mind you I haven't passed that section on B4BB yet.
02-28-2013 11:41 AM
03-01-2013 03:18 AM
Also, consider looking into "pidin". You can type "use pidin" to learn more, or read the docs about it http://developer.blackberry.com/native/reference/b
Using that, with appropriate options (and there are many) you can probably check the state of each thread when it's experiencing this problem, as well as many other possibly useful things.
Thanks, pidin looks like a very promising tool. We haven't looked at each individual thread yet, but noticed yesterday that the number of running threads increased over time... not good.
We have fixed this already, it could have been the root of the problem.
03-01-2013 04:54 AM
03-01-2013 09:15 AM