Kenneth Graunke <kenn...@whitecape.org> writes: > On 08/27/2012 01:38 PM, Eric Anholt wrote: >> Kenneth Graunke <kenn...@whitecape.org> writes: >> >>> Our previous assumption, SWIZZLE_XYZW, was completely bogus for depth >>> textures. There are no Y, Z, or W components. >> >> Seems like this doesn't even need to be shadow-sampler-dependent, just >> whether the baseformat is depth/depthstencil. > > Except for one problem: the baseformat depends on the currently bound > textures. I doubt people have the correct textures bound at > glLinkShader() time, so we'd be guessing...randomly?
Of course. You're right, this is a good way to guess. > That said, it would be great to do better: I'm still seeing some > recompiles due to us guessing XYZW instead of XXX1, and I think that's > due to people sampling raw depth values by binding a depth texture to a > regular (non-shadow) sampler. Yeah, sampling depth directly is used for some soft shadow techniques. `
pgpNXlgcqJjml.pgp
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
