On Sat Dec 2 17:59:48 2023 +0000, Petrichor Park wrote:
> Doing this would require changes to other test harness machinery;
> specifically it would stop printing any context if the shader fails to
> compile, or if there's some other problem.
> It would require a lot of changes to everything the harness could handle
> to provide better feedback in each case; I think that's kind of out of
> scope of this MR and probably out of scope for the harness in general.
(the change is to remove the `vkd3d_test_push_context` call on L1439, and then I'd need to remove the matching `vkd3d_test_pop_context` call that's somewhere)
--
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/499#note_54756
On Sat Dec 2 17:59:48 2023 +0000, Giovanni Mascellani wrote:
> Right now this prints both the line where the section starts and the
> line there the `probe` directive here. This feels a bit too much.
Doing this would require changes to other test harness machinery; specifically it would stop printing any context if the shader fails to compile, or if there's some other problem.
It would require a lot of changes to everything the harness could handle to provide better feedback in each case; I think that's kind of out of scope of this MR and probably out of scope for the harness in general.
--
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/499#note_54755