-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am 2015-04-30 um 22:33 schrieb Henri Verbeet:
I think WINED3DFMT_FLAG_FBO_ATTACHABLE mostly does what we want as long as FBO ORM is used, perhaps it should be generalized to e.g. WINED3DFMT_FLAG_RENDERABLE. I'm inclined to keep it called WINED3DFMT_FLAG_FBO_ATTACHABLE and just not check it when we're using backbuffer ORM. We already have WINED3DFMT_FLAG_RENDERTARGET that more or less mirrors WINED3DFMT_FLAG_FBO_ATTACHABLE for color buffers. Having a FLAG_RENDERABLE in addition would be confusing IMO.
There's an argument to be made that with something that doesn't have FBO in its name we could mark depth formats supported or unsupported in the format table initialization and simplify the non-FBO depth format checks in check_device_format. But that doesn't really work because depth format compatibility depends on the onscreen depth format we end up using with onscreen ORM. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBAgAGBQJVR1sLAAoJEN0/YqbEcdMwc0IP/j1+Y8GynkGpe5vJ1oatNmip Zp2ndEoglG7C5K4dpqX0uYqUA0sTY0NcccZrwzFEKIf/1+V4532g5s/Nqf729Tcm Xnqjm2BU/SZltnR35ZlXdm4zB244SpdMbDSY39zbHDzsI/bxXd2w1RcFrKWqU50K Joe0SyX+yY3FKg5VgSEf2uXsk42ItRJetPCMzUemjxg3u8fzGyHAokOjMnov5Ml5 NPhytT4wA0/Oo6Wzm/GIk2ZXD7JN10H+kuYCXKstyTpV0nHxA3L5n82Spi/ANuls U9Mp5A+JRUOd69ysKeVmg7XStuBhYGcRAFIFfS/lWoFaMF6Ej1VtZWyYwfURZJCQ KwhBOwgo4E5m7HigB6wa8Ejg9pgx3bhs/YFnLttv+5/LsDpP6tcT5K29ddf5AOBg S3V6wvlrkedq21OS6aVzXzF37I7k0AchIw3Ha3h/WLN8wjC5PCwv3Y7a4ivkoZ6F 9R1wvgIGpbeSyC5BPohmD04BrwlesZVJlUZUr714+yjmgAZyTRHQavUNazwRmyI+ 6UIaqI2EFsYmvG7OvGDVyQ1hv2Jv77tQxGaSNrsG+d2DpKnl/FE5nwiTO5xtmmQN i+qKEfZU75vyKBdveJboBOt+l57hLcVi4Yg0r9j859M7Y0ateG8GEyOmRqwqu0qZ k4VzvBVINOdGdSkiSszS =g3D+ -----END PGP SIGNATURE-----