The Mesa version in Debian bookworm (22.3.6) we use for testing is now a bit outdated. While it's valuable to test on older drivers because they might be more representative of what users are running, it's also useful to test on recent drivers, which gives a less noisy signal to developers.
For the moment we enable it only for llvmpipe, because RADV is already passing all the tests anyway. And we don't bother with 32 bit here.
On llvmpipe using Mesa 24 fixes three shader runner tests.
--
v4: ci: Run tests on llvmpipe from Mesa 24.0.3.
ci: Make llvmpipe from Mesa 24.0.3 available in the CI image.
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/743
The Mesa version in Debian bookworm (22.3.6) we use for testing is now a bit outdated. While it's valuable to test on older drivers because they might be more representative of what users are running, it's also useful to test on recent drivers, which gives a less noisy signal to developers.
For the moment we enable it only for llvmpipe, because RADV is already passing all the tests anyway. And we don't bother with 32 bit here.
On llvmpipe using Mesa 24 fixes three shader runner tests.
--
v3: ci: Make llvmpipe from Mesa 24.0.3 available in the CI image.
ci: Run tests on llvmpipe from Mesa 24.0.3.
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/743
> > Would it make sense to start namespacing options that only apply to some shader types or profiles? For example using `--fx-include-empty-buffers` or even `--fx4-include-empty-buffers`. Not necessarily this time, but maybe start thinking about it. Also for HLSL options, for example: `--hlsl-profile` and `--hlsl-semantic-compat-map` instead of what we have now.
>
> Broadly yes, although I don't think we should just remove the existing options; at best we could deprecate them and remove them from --help.
Sure, it wouldn't be appropriate to break command lines that now work, so the current options could be added as aliases of the prefixed ones.
--
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/742#note_66784