-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Interesting. I can't remember if I tested my test on my ATI card, but I'll give it a try. It looks like ATI happily accepts the size mismatch. If it does we could just make the test accept the result and keep our implementation the same or extend the test to see how the depth test behaves outside the DS surface and see if we can implement it.
Am 28.03.2011 um 14:26 schrieb Henri Verbeet:
This fails for me on native (Windows 7, Radeon HD 6950):
visual.c:11813: Test failed: DS size test: Pixel (2, 2) after clear is 0xff0000, expected 0x000000FF visual.c:11815: Test failed: DS size test: Pixel (31, 31) after clear is 0xff0000, expected 0x000000FF visual.c:11817: Test failed: DS size test: Pixel (32, 32) after clear is 0xff0000, expected 0x000000FF visual.c:11819: Test failed: DS size test: Pixel (63, 63) after clear is 0xff0000, expected 0x000000FF visual.c:11825: Test failed: IDirect3DDevice9_ValidateDevice returned 0, expected D3DERR_CONFLICTINGRENDERSTATE. visual.c:11832: Test failed: IDirect3DDevice9_ValidateDevice returned 0, expected D3DERR_CONFLICTINGRENDERSTATE.