John Kessenich
John Kessenich
It turns out the first step (finalize stuff) cannot be done until after proper pool tracking is done (4th in the list), much as in #916. I have a couple...
#1147 is addressing the 1st and 4th items in the list.
It compiles clean for me, as a compute shader, using glslang both as a validator and for generating SPIR-V. How are you executing it? What version are you using (-v)?
BTW, buffers are fully implemented, and I don't see an immediate problem.
I don't know root cause yet, but I don't think you want to - to call preprocess(); you probably want do just do a full parsing with parse() - to...
Understood, and I don't intend to use git commits to track executables. The current intention is to keep the longstanding Khronos location (see link at the beginning of the Readme)...
@jekstrand, agreed. Commit ebb7bf37fd4681dd33e6dabdb3401187f7eef59f removes them. @xorgy, leaving issue open for future use of GitHub Releases. (They are currently hosted by Khronos.)
I just want to add a few notes of how this might by trickier than it first appears, though that also depends on how other parts of WGSL are defined,...
Yes, you could. Forbidding scalar.x, scalar._1, etc. is consistent with an HLL that has many features with many limitations and needing a complex negative testing system to make sure no...
My point was to be cautious adding this feature, because it seems you don't want to add it all the way, which is a root cause of portability issues with...