Re: [PATCH 3/5] wined3d: Remove ffp_proj_control flag from the GLSL shader backend.
On 19 June 2015 at 00:07, Matteo Bruni <mbruni(a)codeweavers.com> wrote:
Using the GLSL shader backend implies GLSL vp and fp replacements, An argument could perhaps be made that it should, but currently that's not true. You can still use the the GLSL shader backend with e.g. fixed function fragment processing with some level of success. The only reason it doesn't happen much in practice is the way select_vertex_implementation() and select_fragment_implementation() are implemented, but that's external to the shader backend.
2015-06-19 11:06 GMT+02:00 Henri Verbeet <hverbeet(a)gmail.com>:
On 19 June 2015 at 00:07, Matteo Bruni <mbruni(a)codeweavers.com> wrote:
Using the GLSL shader backend implies GLSL vp and fp replacements, An argument could perhaps be made that it should, but currently that's not true. You can still use the the GLSL shader backend with e.g. fixed function fragment processing with some level of success. The only reason it doesn't happen much in practice is the way select_vertex_implementation() and select_fragment_implementation() are implemented, but that's external to the shader backend.
That's true. It's likely I need to enforce it at some point though, at least for core profile. Actually using non-GLSL vp with the GLSL shader backend is already broken by some of my latest changes in set_glsl_shader_program(). I can fix that part but I'm not sure there aren't other blockers after it. I guess I'll look into adding some mechanism to the shader backends for accepting or rejecting vp and fp implementations, or something along those lines.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am 2015-06-19 um 11:06 schrieb Henri Verbeet:
You can still use the the GLSL shader backend with e.g. fixed function fragment processing with some level of success. The level success with that is quite high actually. I use this to test the nvts / nvrc backends when I touch something that might have an impact there. I prefer to keep unexpected mixes of shader / vertex / fragment pipelines working.
-----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBAgAGBQJVh+twAAoJEN0/YqbEcdMwKOcP/iE9kijd2k4iYZrAKIFKTTCZ j6HOVmSLpRpmcx/m2//qihqGVEL1+7kQfX0yY5QGcuJEw4+2MX9ohitxii0nQEvi 4TwcHuvsJDkcakfkRCcF9ZRPsCcHa22FUUevsiMhdlobjFLg6FI0Fu1iyxEXTF0B pUxwm9xDd243je1XubUu7xhy3BTkfypwuIEQxOCt+LnAx4HpCSqFLqT0gBUcu+64 tkuYTwetLefPxCyRVWyoGIbgMDFuNPFOlVRoHR/5EI0S7WKwLK0SOjtoA3g2Bdzu XS7Hz8JTwKGIyUGjIYOOl9ILkpyHvclKggs7c/prZp8hXYx97kv/06UQpXfAIEGH y1jJUThxJU3Soq050YPHnTduNT6OY61B7lc2ePlhf793PnCg2QsBmzXlt0Fg9txr ++6IGUp6YFzPfaYvpV6v8ZK3tf7231rphyjskOStvH7yanaQUrtrc50FHI9o28gg 8uski5eGxbFcvujCLLabaFsnRyBel0mP04Iu/UwCMHmvcuRl4YXxNSvbuGx3mqCi RCOp/TP1bLEtGklJCWjdjOHKlEoJcejzjAlaC7hCRfuSB/gE+HeaTWoUL3GK7NIP jvG5HMaJDRhjitJE5wucYLqe7mg/xAwcDUUHtwdnmd7hJy1rQwQCwx1HlbBPHLmZ rUfkzEKc1XnrmCrSg4dh =d+vf -----END PGP SIGNATURE-----
participants (3)
-
Henri Verbeet -
Matteo Bruni -
Stefan Dösinger