Android Question Lot of Fatal signal 11 (SIGSEGV) under Android L_PREVIEW v4.4W

[TheWind777]

I have been getting a lot of Fatal signal 11 (SIGSEGV), code 1, fault addr 0x49002e in tid 2512 (b4a.acimwow) when I switch between panels under the new L_PREVIEW version (Pre-Lollipop version for Developers) of Android.

Anybody else been having an increase of Fatal Signal 11 errors under the new Preview version of Android 5 (4.4W)?

It looks like that error tends to occur when something hasn't been initialized, yet is used... or with Null Pointer errors; yet I can't see any place where I'm not initializing things (and I always check that something is initialized before I use it).

Other places the error tends to occur, it seems, is if a timer tries to access variables which haven't yet been set up fully in a subroutine (generally a timer and flag fixes this).

Error seemed to be related to my upgrade to Android L_PREVIEW; and was wondering if anybody else is experimenting with L_PREVIEW yet?

 

[Erel]

I've previously tested the first developers release of Android L. It was working fine (after the fixes applied in B4A v3.82).

Are you using v3.82?

Can you upload an example that crashes?

 

[TheWind777]

I've previously tested the first developers release of Android L. It was working fine (after the fixes applied in B4A v3.82).

Are you using v3.82?

Can you upload an example that crashes?

Yep. Using 3.82.

Must be something that I'm initializing in the Process_Global area which I'm really supposed to be initializing somewhere else (or vice-versa).

Can't upload an example because it's huge. Just going to have to search to see that any odd library is being initialized properly and in the right place. As it is I have put all the odd Process_Globals declarations at the end of Process_Globals because I was having strange crashes. It's a problem with one of those functions (built by other people, not you). One function doesn't even give you an initialize, so they must be initializing inside their routine.

I'll find it. I just wanted to know whether the problem might have been that I just upgraded to L_PREVIEW... and that's not the problem. Thanks.

Today I'm going to verify that each of those odd libraries is being given what they want (being properly initialized) in the right places.

Also am going to verify that it isn't some variable that is being set-up which isn't being set-up fast enough, or isn't ready yet when it uses the variable. At least the error has become evident rather than being intermittent. I hate when there's an intermittent problem.

It'll be the last major bug once I find it. All others have been ironed-out. Annoying because it seems to work so seamlessly, otherwise.