On Fri, 20 Aug 2021 at 22:26, Matteo Bruni matteo.mystral@gmail.com wrote:
On Wed, Aug 18, 2021 at 9:52 AM Nikolay Sivov nsivov@codeweavers.com wrote:
Signed-off-by: Nikolay Sivov nsivov@codeweavers.com
This change allows effects with such variables to be parsed at all, next
step to make it actually
useful would be to parse declaration string to
D3D10_SO_DECLARATION_ENTRY[] array.
Since GetShaderDesc() is not implemented, for testing purposes
declaration string could be checked via +d3d10 traces.
Sure. We could also just implement GetShaderDesc(), it doesn't seem that it will need to do much that isn't just copying the info that we already have.
I do have GetShaderDesc() fully working now, including stripping some of the information on Optimize(). That’s a small number of patches.
Also worth noting that compiler seems to do only mask validation for
declaration string, meaning that successful
compilation does not mean that effect will be successfully created.
Effect parsing then does further validation
by checking semantics.
Maybe we could add some failure tests with invalid effects. We have some for the HLSL compiler.