05:01 adavy: q4a: I think so
05:02 adavy: Now with it should show different results windows vs others
05:03 adavy: now ideally you'd want to show impact of programmable pixel shader that replaces depth
05:03 adavy: as well as simple programmable vs
05:06 q4a: Ok. I would like to boot in Linux and test Wine+OpenGL, Wine+Nine (with and without patch https://gitlab.freedesktop.org/q4a/mesa/-/commit/27645c05f83a7e762e17476b9dfb6b0ef5bd15aa )
05:06 q4a: Or it's to early?
05:08 adavy: when you feel your test values won't change, you can check windows and write the correct expected values
05:22 adavy: If needed I have a very old laptop with win xp and nvidia card not able of full dx9
05:22 adavy: but I think when you submit a wine test they complete for other hw they have
14:12 q4a: I added Windows build for Nine tests and add my test `test_table_shader_fog` with 48 cases: https://github.com/q4a/nine-tests/tree/build-win... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/wlfYbQPrJJzyTfUbQPkkqIzO>)
14:27 q4a: Mesa tests ran on Radeon HD 8400 and I'm not sure: should I run same tests in Wine+OpenGL or on hw with HD Graphics (Ivy Bridge GT1)?
19:10 adavy: what is the whole set of failures on nine with wfog patch ?
19:22 adavy: q4a: what does mean //other ?
19:25 adavy: The values of the test seem to indicate what was intuited from the doc and wasn't done correctly before by all implementations: that the projection matrix being orthogonal impacts even when you use a programmable vs
19:26 q4a: //other - meant that it has different color compared to version without ps/vs
19:28 adavy: I suspect your shader is invalid to use with xyzrhw
19:28 adavy: dcl_position v0 -> probably positiont
19:28 adavy: that might explain why xyzrhw is so close to ff vs
19:29 adavy: maybe it reverts to ff vs
19:29 adavy: that is good to know
19:29 q4a: what is the whole set of failures on nine with wfog patch ? - 6 cases: 20-22, 44-46
19:30 adavy: likely the wfog patch doesn't do correctly programmable ps
19:31 adavy: I think basically what remains for the test to be complete is to have a ps variant that overwrites depth
19:31 adavy: to see if - and when - that depth is used for fog
19:31 adavy: you could set the depth to a constant
19:31 adavy: in that ps
22:15 q4a: https://github.com/q4a/nine-tests/commit/a22c55ec2b83ee4685e077888c76a9140e54bdaa - added pixel_shader_code2 with constant color... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/zWkeboyyamvBtJhOWLKQBRab>)
22:16 q4a: s/pixel_shader_code2/pixel\_shader\_code2/, s/264/24/
22:17 q4a: s/pixel_shader_code2/pixel\_shader\_code2/, s/264/26/