Android Screen Rotation Crash

[LunaRebirth]

Been digging into why my apk crashes whenever I rotate the screen. Looks like it works for a split second (I can see the screen rotate before crashing).
Was wondering if anyone else is having this issue, or if it is because of my own code.

[CuteAlien]

Hm, Irrlicht example doesn't seem to support rotation. If I remember right that was some manifest setting. Can you post your manifest xml and maybe a callstack of the crash?

[LunaRebirth]

I simply removed this line:

Code: Select all

android:screenOrientation="portrait"

from the manifest file to allow screen rotation.

The callstack looks like it is coming from my code, but if I continuously add returns everywhere, the crash location will change but still occur.
I'll still add it below anyways.

Code: Select all

--------- beginning of crash
09-07 20:12:35.652 31901 31901 F DEBUG   : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
09-07 20:12:35.652 31901 31901 F DEBUG   : Build fingerprint: 'motorola/kinzie_verizon/kinzie:7.0/NCK25.118-10.2/2:user/release-keys'
09-07 20:12:35.652 31901 31901 F DEBUG   : Revision: 'p301'
09-07 20:12:35.652 31901 31901 F DEBUG   : ABI: 'arm'
09-07 20:12:35.652 31901 31901 F DEBUG   : pid: 31526, tid: 31900, name: Thread-7  >>> com.Proj<<<
09-07 20:12:35.652 31901 31901 F DEBUG   : signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x20
09-07 20:12:35.652 31901 31901 F DEBUG   :     r0 cf64b400  r1 aaaaaaab  r2 00000000  r3 00000000
09-07 20:12:35.652 31901 31901 F DEBUG   :     r4 00000001  r5 cf9118e4  r6 cf911800  r7 00000000
09-07 20:12:35.652 31901 31901 F DEBUG   :     r8 dfee8d30  r9 dfee8ce0  sl cfd9fc6d  fp 00000058
09-07 20:12:35.652 31901 31901 F DEBUG   :     ip 00000076  sp e7c4e088  lr cfc3d3cd  pc cfc3d418  cpsr a0070030
09-07 20:12:35.656 31901 31901 F DEBUG   :
09-07 20:12:35.656 31901 31901 F DEBUG   : backtrace:
09-07 20:12:35.656 31901 31901 F DEBUG   :     #00 pc 0015a418  /data/app/com.Proj-1/lib/arm/libProj.so (_ZN10LuaElement6removeEP9lua_State+119)
09-07 20:12:35.657 31901 31901 F DEBUG   :     #01 pc 00151edb  /data/app/com.Proj-1/lib/arm/libProj.so (_ZN4LuaC5clearEb+118)
09-07 20:12:35.657 31901 31901 F DEBUG   :     #02 pc 001553db  /data/app/com.Proj-1/lib/arm/libProj.so (_Z10clearWorldSs+166)
09-07 20:12:35.657 31901 31901 F DEBUG   :     #03 pc 00155fd5  /data/app/com.Proj-1/lib/arm/libProj.so (_ZN6Client15setRequestWorldESs+48)
09-07 20:12:35.657 31901 31901 F DEBUG   :     #04 pc 0014c3e7  /data/app/com.Proj-1/lib/arm/libProj.so (main+262)
09-07 20:12:35.657 31901 31901 F DEBUG   :     #05 pc 0014c7dd  /data/app/com.Proj-1/lib/arm/libProj.so (android_main+340)
09-07 20:12:35.657 31901 31901 F DEBUG   :     #06 pc 002bcce3  /data/app/com.Proj-1/lib/arm/libProj.so
09-07 20:12:35.657 31901 31901 F DEBUG   :     #07 pc 00046eb3  /system/lib/libc.so (_ZL15__pthread_startPv+22)
09-07 20:12:35.657 31901 31901 F DEBUG   :     #08 pc 00019acd  /system/lib/libc.so (__start_thread+6)

When I rotate the screen, android_main() is called again.

[CuteAlien]

Example not crashing, but results still somewhat interesting. As it also doesn't seem to realize in Irrlicht anything changed (still drawing with old resolution and xy-coordinates of events are now wrong). Which might also get me closer to the bug with the soft-keyboard (which I had given up on already).

So - can't help you with the crash. But you might have helped me at least gettings some more ideas.

[LunaRebirth]

Okay thanks for checking that for me. Must be something on my end that's causing the crash.

[MartinVee]

Just a quick hint, but SEGV_MAPERR means that you tried to access memory that's not mapped to your program's address space. That's usually because :
- You dereference a null pointer ;
- You try to free the same memory twice ;
- A pointer was corrupted ;
- There's a buffer overflow in your program.