"Difficulties with one device" I have no way of knowing how many makes/models of device this problem is on. I know that it happens all the time on my Samsung Galaxy (I9000). Others report exactly the same thing as being a feature of a number of HTC devices - these two make up a major part of the market I believe. I think I may well have two faults - one is the "standard" eglSwapBuffers hang, which none of the suggested fixes have had any effect on -square textures, different GLSurfaceView, wait, flush, finish etc. The second one, which I think <might> be related, is my intermittent framebuffer corruption, I am willing to accept this may be my fault in some way I have been unable to pinpoint. http://www.fordartworks.co.uk/dodgyS.png - the ractilinear dodgy area just above the left and right arrow buttons.
"smacks of throwing the baby out..." At the moment, I am just hugely frustrated by the long list of problems I have faced with Android - halts on writing to SD card, gradients that don't work at one particular angle and so on. I have been a software designer and developer for over 25 years - parallel architecture research, trading room systems, real time etc (assembler, Fortran, Prolog, Occam, C, C++, Java) and I have never been so frustrated by any system - there has always been a different way of doing things, getting round problems. With this /these? OpenGL problem(s) it appears to be a black box. Without a genuine understanding of what is wrong, any "fix" is pure conjecture. "should just be ignored" I am extremely reluctant to put out an App that looks flakey or, much worse, shuts down people's phones. "Ignoring" a problem that I can't get round is not a viable solution I feel. "plugging and unplugging the USB cable" The USB cable stuff was a problem I had with very jerky OpenGL performance when the rendering load was heavy - this was caused by the CPU clock reducing when the GPU was active, I believe. "seeing the behavior elsewhere" I haven't experienced this problem with other programs on this phone. (I have not detected the problem on my Asus Transformer either, but I have tested that far less. It will also have very different performance characteristics.) I don't play games, 2D or 3D. Many of the openGL demonstrators, testers or benchmarkers don't stress the system sufficiently - i can get 50 or 60 fps. The problem is very timing dependent- i have one picture definition for my App that fails vaguely regularly - i have 2 or 3 others that do very occasionally, and I have 100+ that i haven't seen go wrong yet. I simply haven't the time to sit and fiddle about with a 3rd party app on the off-chance of catching a glitch. If I did - that would prove relatively little. I could sit there till doomsday with my own app under the right conditions and never see anything untoward. I'll maybe just take a break from this (go and paint a few pictures :-) ) - i've been on this problem 7 days a week for nearly 3 weeks now. I might even go get a real job again, lol. "It's insane" Thank you for your concerns over my mental health :-) On Oct 11, 7:28 pm, Christopher Van Kirk <christopher.vank...@gmail.com> wrote: > It's insane to throw away months of work. > > Are you seeing this behavior in other 2 or 3d games running on that > device? If so, it's probably a device fault that you simply can't get > around, and should just be ignored. There are other devices in the world > that, in theory, it should work on. In your market publish, just mention > that you had problems with that device and that it's not recommended to > be run there, then move on. > > If you aren't seeing the behavior elsewhere, then what you have is an > application problem. Have you considered that plugging and unplugging > the USB cable is introducing thread locking that is absent in your code? > When you connect the USB cable, if you have development enabled, you > instantly start reading out logging info, which may have locking in > there that's fixing some problem you have. > > Anyway, giving up smacks of throwing the baby out with the bath water > because of difficulties with one device. I wouldn't do it. > > On 10/12/2011 12:02 AM, calderwa wrote: > > > > > Nope, 'fraid not. That code seems if anything to makes things a little > > worse for me = intermittent screen glitches, with rectangular areas > > garbled, and the less frequesnt hangup for a second or so, with the > > even less frequent reboot. I am using Samsung Galaxy S, 2.3.3, so > > clearly not restricted to HTC. All hangs seem to be locking problems > > in SharedBufferStack. > > glFlush calls have no effect on my problem. Switching from > > glDrawArrays to VBOs and glDrawElements had no effect. > > > This bug is SO bad and i have spent SO long on it now that I am pretty > > much at the stage of junking many months of work and not publishing > > the openGL version of my app. > > > This comes on top of unacknowledged problems in java 2d drawing > > routines, the ridiculous DirectBuffer problem in 3.0+ where it > > allocates 4* as much memory as you ask for. The list is nearly > > endless. > > > Apple, here I come? > -- You received this message because you are subscribed to the Google Groups "Android Developers" group. To post to this group, send email to android-developers@googlegroups.com To unsubscribe from this group, send email to android-developers+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/android-developers?hl=en
