Billboarding causes GC [update]

[Irony]

After seeking the reason for frequent GCs in my code for half of last night I found it's the ships' engine glow (a semi-transparent plane with additive blending).
To be exact it's the billboarding on them. Removing setBillboarding(true) eliminates nearly all GCs.
I suppose I could do the billboarding calculation myself, but maybe it's possible for you to take a look. Thx!

[EgonOlsen]

Strange, because i'm using bill boarding a lot and never experienced this. I'll look into it...

[EgonOlsen]

If you are using billboarded child objects, an unoptimized part of the code will be used that creates up to 3 additional matrices per frame. If should be fixed in this version: http://jpct.de/download/beta/jpct_ae.jar. The code is untested...i hope that it still does the same thing just with less instance creations...

[Irony]

Sorry but no, the children are not where they're supposed to be..

[EgonOlsen]

Damn...i'll have a look again.

[EgonOlsen]

Please re-download the beta and try again. This time, i did some tests...it should work now.

[Irony]

Perfect! Thanks!

[Irony]

Hi Egon,
the billboarding works fine now, but I have a different problem in this beta.
The following error occurs when loading one specific model, however the exact same code works with previous versions of JPCT. Any idea?

06-24 09:43:40.470: I/jPCT-AE(8216): Processing object from 3DS-file: alien_spac
06-24 09:43:40.560: I/jPCT-AE(8216): Object 'alien_spac_jPCT269' created using 2556 polygons and 1465 vertices.
06-24 09:43:40.560: I/jPCT-AE(8216): Loading Texture...
06-24 09:43:40.590: D/dalvikvm(8216): GC_FOR_ALLOC freed 1608K, 42% free 31861K/54407K, paused 28ms
06-24 09:43:40.630: D/dalvikvm(8216): GC_FOR_ALLOC freed <1K, 40% free 32885K/54407K, paused 24ms
06-24 09:43:40.630: I/jPCT-AE(8216): Texture loaded...1048576 bytes/512*512 pixels!
06-24 09:43:40.865: I/size(8216): 0 471.4535
06-24 09:43:40.865: I/jPCT-AE(8216): Loading file from InputStream
06-24 09:43:40.865: I/jPCT-AE(8216): File from InputStream loaded...305 bytes
06-24 09:43:40.985: W/dalvikvm(8216): threadid=13: thread exiting with uncaught exception (group=0x40c5b1f8)
06-24 09:43:40.990: E/AndroidRuntime(8216): FATAL EXCEPTION: GLThread 517
06-24 09:43:40.990: E/AndroidRuntime(8216): java.lang.NullPointerException
06-24 09:43:40.990: E/AndroidRuntime(8216):    at com.threed.jpct.Loader.loadBinaryFile(Loader.java:1200)
06-24 09:43:40.990: E/AndroidRuntime(8216):    at com.threed.jpct.Loader.loadBinaryFile(Loader.java:1124)
06-24 09:43:40.990: E/AndroidRuntime(8216):    at com.threed.jpct.Loader.load3DS(Loader.java:1541)
06-24 09:43:40.990: E/AndroidRuntime(8216):    at com.threed.jpct.Loader.load3DS(Loader.java:165)
06-24 09:43:40.990: E/AndroidRuntime(8216):    at com.ironman_project.space.object.SpaceObject.loadModel(SpaceObject.java:156)
06-24 09:43:40.990: E/AndroidRuntime(8216):    at com.ironman_project.space.object.SpaceObject.setModelAndTexture(SpaceObject.java:98)
06-24 09:43:40.990: E/AndroidRuntime(8216):    at com.ironman_project.space.object.SpaceObject.<init>(SpaceObject.java:118)

[EgonOlsen]

Opps...yes, i've changed something in that method to reduce the garbage it creates. I worked fine in my test cases, but something is obviously wrong with it. I'll look into it later today.

[EgonOlsen]

Please re-download the beta. The nullpointer should be fixed now.

[Irony]

Works!