Thank you for the tutorial. I'll leave this open for feedback for a while and eventually add it to the wiki. Please remind me in the future if I forget.

I have split the forum thread: viewtopic.php?f=29&p=1053627
Please don't feel afraid to open a new forum thread once the discussion diverges too much from the original topic.

Are you working with the PVR API? shadow mapping can't be used effectively on the PVR graphics chip, because it doesn't have shader support. What you need to do in shadow mapping is render from a light's viewpoint to a depth texture (which the DC already can't do), then use this texture from your ca...

If you write some good guides I'm sure you'll get wiki access quickly. If it's one thing we're missing it's documentation

I wrote a project from the ground up without Kos and newlib that creates executables in the kilobyte range.
Even that became quite fat (1mb?) when I used GCC 6. So disabling certain features, standard library, debugging, runtime checks should go a long way.

My library​ is not finished. Most importantly I did not inplement clipping to the screen plane.

Actually fragmentation could be reduced by creating buckets of textures by size or at least something them by size and then putting smaller textures on one end while putting bigger ones on the other. So deallocating a big texture makes room for exactly another new big texture without fragmentation. ...

I only know that GCC 6 worked for me for C, didn't test C++.
Just a shot in the dark, but perhaps you can disable some newer C++ library features with command line flags? I'd read the changelog

Yes, it should remove the data from VRAM. But it's also possible that the PNG loading code doesn't free its memory. Once the texture data has been passed to glTexImage2D it should be freed from RAM. And please don't send a huge zip. Just create a minimal test case. Should be simple enough by comment...

Author of the tutorial here, sorry I only saw the thread just now.
So everything's fine with the tutorial code?
What emulator did you use which failed to show things correctly?

I assume it was a fluke then.
The glOrtho implementation is correct.
";;" won't cause any problems ever, compilers just warn about it because there's no use case for it (except to make intricate Japanese smileys maybe).

Can you please post your test case (whole file, but as few code as necessary to show the problem)? I'll look into it then.

libgl uses the code from pvr.h and related files:

http//gamedev.allusion.net/docs/kos-2.0.0/pvr_8h.html#aa3845a080ca8ad85eabd8905ad2431cf

So search the libgl code for these pvr structs for setting up the depth test and texturing.

That's messed up! To render with the PVR a command list is used. It consists of a header followed by one or multiple triangle strips. It seems like this header is not set up properly. Perhaps you could look into that? In the Kos documentation it's referred to as polygon header and context (same thin...

havent find anything until now, but found somthing other strange file gl-cap.c https://github.com/ljsebald/kos-libgl/blob/master/gl-cap.c line 111-141 GLboolean APIENTRY glIsEnabled(GLenum cap) { if(cap >= GL_LIGHT0 && cap <= GL_LIGHT15) return _glKosIsLightEnabled(cap & 0xFF); switch(c...

I assume you didn't find anything else because you ran "make install clean" for the port or you used the build all script which does that. That will download the source code, compile and install it, then delete the code. I think the command you need is make fetch or something.. please look...

Thanks for informing us about what happened, Josh. I don't think anybody here had any doubts about your reputation. People just thought you were depressed about the KS failure because you worked so hard, some even wondered whether you got sick or died. But nobody entertained the thought that you wer...

What I always wished for for offline play was that all map layouts of all section IDs can be used. E.g. I'm Skyly, get Ruins 1 layout A which is possible with Skyly, then I get Ruins 2 layout B which is only possible with Redria. Is it possible to patch that? I never understood why Sonic Team limite...

In that case you seem to have found a bug in the new OpenGL implementation..
Sadly I don't have the time to test and fix this myself right now.
If you want to fix this, you can search for GL_DEPTH in kos-ports/libGL. You probably need to change some default setting or make the depth test opposite.