On Mon Mar 11 21:05:41 2024 +0000, Nikolay Sivov wrote:
A better option might be to replace this with exact copy of D3DCompile2() from vkd3d, shouldn't be any reason to have them differ.
Removing spurious differences in the whole function is certainly good, although we probably want to do that over a few (separate) patches, if nothing else for future regression testing concerns.
For this specific API_VERSION change, my idea was to sort of enforce that we have a recent-enough vkd3d-shader version around, but that's probably not the way to do it.