Erel,
I am still working on sorting things in this thread:
https://www.b4x.com/android/forum/t...ty-containing-a-google-map.68023/#post-431953
I believe I am getting closer to solving it but I am not yet sure.
To be sure I am running long term stress tests - programmatically randomly running through all the activities with constrained available memory.
This takes a lot of waiting.
In the meantime I am looking at workarounds - what to do if I can not solve it - or maybe just added insurance.
If the app were published with the OOM problem unresolved then obviously every so often the user is likely to encounter it.
If the app instance blows up with OOM and then stays in memory then attempts at restarting the app will bring up the current (memory allocation clogged) instance - to get out of this problem the user would have to know how to manually destroy the app instance - hardly user friendly.
Putting ExitApplication in Starter.Application_Error seems to solve this issue pretty well.
Also, users will encounter it more frequently if the app instance is sitting in memory between uses with its memory allocation semi-clogged.
If I destroy the app instance after every use by using ExitApplication in Main.Activity_Pause then the likelihood of the user encountering the problem at all is diminished.
None of this is really to the point of the question in post #1 - which basically is why does ExitApplication behave differently in different settings.
Could you also give some explanation as to why ExitApplication should be avoided (aside from protocol/accepted practice etc.).
Thanks as always for your efforts...