Forums | developer.brewmp.com Forums | developer.brewmp.com

Developer

Forums

We are testing our application on LGVX4400 handset with Grinder tool and we getting crashes (power cycles) if event rate is too high (100mb delay between events).
However, if we configure grinder to use 1000ms delay - everything goes fine for quite a long time (whole night).
I have a suspicion, that hight event rate floods event queue and triggers watchdog.
Is this possible?
Somehow from posts I get an information that NSTL tests with 250ms delay, but their documents say 4000 event for 2-3 hours, and this makes about 1 event in 2 seconds. Which figure is right?

Sure, if your app isn't processing the keystrokes fast enough, the event queue is going to get full. In my testing, the VX4400 usually has the least problem with this, so if it's happening on the 4400, you'll probably have worse problems on other handsets. Certain handsets seem to need a small breather (idle cpu) every once in a while.
I suppose you saw this thread: http://brewforums.qualcomm.com/showthread.php?threadid=821. I was drawing and overdrawing... had to make a dirty rect system.
Since 250ms is more stringent, I'd go with that. Actually, I test at about 150ms now.
--t

Sure, if your app isn't processing the keystrokes fast enough, the event queue is going to get full. In my testing, the VX4400 usually has the least problem with this, so if it's happening on the 4400, you'll probably have worse problems on other handsets. Certain handsets seem to need a small breather (idle cpu) every once in a while.
I suppose you saw this thread: http://brewforums.qualcomm.com/showthread.php?threadid=821. I was drawing and overdrawing... had to make a dirty rect system.
Since 250ms is more stringent, I'd go with that. Actually, I test at about 150ms now.
--t