Android 4.3 Issues thread

NJDude

Expert
Licensed User
Longtime User
There's another thread HERE but I think it would be better to consolidate the issues into one.

Another problem with 4.3 is the Camera, for example, if you use B4A VideoRecordApp, it just doesn't work, I found out via one of my apps (which uses a different code but it's camera related), after upgrading I get null point errors.
 
Last edited:

NJDude

Expert
Licensed User
Longtime User
I tried the new lib and the app crashes, here is the log:
B4X:
WAIT_FOR_CONCURRENT_GC blocked 54ms
GC_FOR_ALLOC freed 677K, 9% free 9863K/10824K, paused 28ms, total 28ms
 
GC_FOR_ALLOC freed <1K, 4% free 10483K/10824K, paused 27ms, total 27ms
GC_FOR_ALLOC freed 676K, 9% free 10395K/11360K, paused 19ms, total 19ms
 
GC_FOR_ALLOC freed <1K, 4% free 11016K/11360K, paused 19ms, total 19ms
GC_FOR_ALLOC freed 676K, 9% free 10927K/11896K, paused 21ms, total 21ms
 
GC_FOR_ALLOC freed <1K, 3% free 11548K/11896K, paused 19ms, total 19ms
closeCamera
 
Closing effects
 
Effects are already closed. Nothing to do
Camera fd open as: 120
 
OMX_ERRORTYPE android::NvOmxCamera::getCameraStereoMode(NvxComponent*, NvOmxCameraUserStereoMode&): Error: invalid NVX mode 0.
 
OMX_ERRORTYPE android::NvOmxCamera::getCameraStereoModeAndCaptureInfo(NvxComponent*, NvOmxCameraUserStereoMode&, NVX_STEREOCAPTUREINFO&): getCameraStereoMode failed with 0x00000000
 
Camera fd close (MI1040)
 
Already called release()
 
Destroying camera 0
Already called release()
threadid=18: thread exiting with uncaught exception (group=0x41a09700)
session id 22 not found for pid 124
session id 23 not found for pid 124
FATAL EXCEPTION: Thread-176
java.lang.RuntimeException: startPreview failed
at com.android.camera.VideoModule.startPreview(VideoModule.java:867)
at com.android.camera.VideoModule.access$800(VideoModule.java:73)
at com.android.camera.VideoModule$4.run(VideoModule.java:411)
at java.lang.Thread.run(Thread.java:841)
Caused by: android.view.ViewRootImpl$CalledFromWrongThreadException: Only the original thread that created a view hierarchy can touch its views.
at android.view.ViewRootImpl.checkThread(ViewRootImpl.java:5908)
at android.view.ViewRootImpl.invalidateChildInParent(ViewRootImpl.java:869)
at android.view.ViewGroup.invalidateChild(ViewGroup.java:4250)
at android.view.View.invalidate(View.java:10539)
at android.view.View.setEnabled(View.java:5865)
at com.android.camera.VideoUI.enableShutter(VideoUI.java:341)
at com.android.camera.VideoModule.onPreviewStarted(VideoModule.java:884)
at com.android.camera.VideoModule.startPreview(VideoModule.java:858)
... 3 more
  Force finishing activity com.google.android.gallery3d/com.android.camera.VideoCamera
GC_FOR_ALLOC freed 1354K, 17% free 13949K/16668K, paused 53ms, total 53ms
 
closeCamera
 
already stopped.
Releasing media recorder.
sending message to waiting queue (OnActivityResult)
 
running waiting messages (1)
 
destroy
 
Thread[GAThread,5,main]: Service unavailable (code=1), will retry.
 
Unable to start service Intent { act=com.google.android.gms.analytics.service.START (has extras) } U=0: not found
Thread[GAThread,5,main]: No campaign data found.
 
Thread[Service Reconnect,5,main]: Service unavailable (code=1), using local store.
 
Unable to start service Intent { act=com.google.android.gms.analytics.service.START (has extras) } U=0: not found
 
Upvote 0

barx

Well-Known Member
Licensed User
Longtime User
anything to do with restricted accounts. Big G's docs state that errors can happen with restricted accounts...
 
Upvote 0

barx

Well-Known Member
Licensed User
Longtime User
nice, that's just how big G rolls.
 
Upvote 0

rgately

Member
Licensed User
Longtime User
I've just encountered a nasty problem with Android 4.3 that was not there before. My app uses IME to enlarge/reduce the height of a textbox, when the keyboard is opened/closed. Everything works like it did before - except, that the image of the textbox is not completely cleared when the keyboard closes, making the textbox height look wrong. The affected area is the left over portion, when the smaller size is subtracted from the larger size textbox. When I move a seekbar located in that area, portions of the screen get refreshed. I have tried numerous things and am still looking for a fix.
 
Upvote 0

rgately

Member
Licensed User
Longtime User
Erel,

I ran the IME example again and found that it works fine when the targetSdkVersion=14. Anything less, causes the same problem that I am now experiencing. In my app, I specify version 10 because I did not like the look of the new seekbars and the loss of the menu.

Please let me know if there is a problem using IME with older targetSdkVersions or if there is a problem associated with Android 4.3.

Thanks
 
Upvote 0

rgately

Member
Licensed User
Longtime User
There isn't any known issue with IME and older target values. It is most probably a bug in the new OS.
That's what I'm thinking. I will try to move on to a higher API (14 and above) but I will have to figure out what to do about the menu and other design issues. Thanks again.
 
Upvote 0

rgately

Member
Licensed User
Longtime User
I discovered a way to get rid of the textbox visual artifact, when using IME to alter the height in Android 4.3 on my Nexus 7.

The problem goes away after checking "Force GPU Rendering (Force use of GPU for 2nd drawing)". But this seems to cause problems of its own. So?
 
Last edited:
Upvote 0

canalrun

Well-Known Member
Licensed User
Longtime User
My Nexus 7 tablet just got the 4.3 upgrade. I ran some quick checks of my existing apps. One app which uses VideoRecordApp encountered the same problem that was mentioned above. I have tested this on several devices with android versions 2.2 or 2.3 through 3.X up to the Nexus when it was running 4.2.2. It worked perfectly.

The first time I ran the app under 4.3 and tried to record a video using VideoRecordApp a message popped up saying something like "Gallery needs to close" (I was using VideoRecordApp not the Gallery). I tried to record a video a second and third time; it appeared to record the video but did not save the file suggesting that it never ran the _Result routine.

I will set up the debugger and provide some Log traces, probably tomorrow. But, I can confirm the behavior encountered in the first post of this thread. Any ideas?

Barry.
 
Upvote 0
Top