Open GoogleCodeExporter opened 9 years ago
[deleted comment]
Sorry, there are typos in line 2 and 3 of my table of packed varyings. The
correct version is:
1: a a a a (a = v_TexCoord[0])
2: b b b b (b = v_TexCoord[1])
3: c c c c (c = v_TexCoord[2])
4: d d d d (d = v_FrontColor)
5: e e e e (e = v_BackColor)
6: f f f f (f = v_FrontSecondaryColor)
7: g g g g (g = v_BackSecondaryColor)
8: h m x x (h = v_ClipDistance[0], m = v_FogFactor, x = unused)
9: i x x x (i = v_ClipDistance[1], x = unused)
10: j x x x (j = v_ClipDistance[2], x = unused)
11: k x x x (k = v_ClipDistance[3], x = unused)
12: l x x x (l = v_ClipDistance[4], x = unused)
Original comment by pdae...@gmail.com
on 2 Sep 2014 at 4:02
See https://chromium-review.googlesource.com/#/c/214108/
I think the above CL will fix a lot of the varying packing issues.
Original comment by jmad...@chromium.org
on 17 Sep 2014 at 2:37
It seems that the CL above does not work in D3D9. I'm not sure at the moment if
it's a regression or simply doesn't address the packing problems in D3D9.
Original comment by jmad...@chromium.org
on 1 Oct 2014 at 8:18
My change was designed to fix the D3D11 packing issues. I didn't know that D3D9
was broken (since the tests weren't running against D3D9 at the time). It is
likely that my change simply didn't fix the D3D9 packing issues.
Original comment by aukin...@microsoft.com
on 1 Oct 2014 at 8:23
It seems Shader Model 3 does not allow TEXCOORDs higher than 16, so there's a
chance we would have broken something that might have been working previously
-- not sure exactly if such a case exists.
Original comment by jmad...@chromium.org
on 1 Oct 2014 at 9:11
Original issue reported on code.google.com by
pdae...@gmail.com
on 2 Sep 2014 at 3:57Attachments: