Tobski
Tobski
@forenoonwatch I suppose we could make an exception for existing type identifiers, little awkward but I suppose it should work fine. The main issue at this point though is that...
@gnl21 I made changes along the lines of your feedback - please take a look and let me know if you still have issues.
@Try yep what @nhaehnle said. It's an invalid shader. spirv-val is supposed to catch that after compilation and I _thought_ that it would run that automatically as part of glslang,...
@s-perron looks like this was assigned to you for review, are you able to take a look? @ShchchowAMD can you rebase this again please?
@ShchchowAMD in case you haven't seen this, can you look at/respond to @llvm-beanz's comments?
Hi @jkoenen, while a layout transition might actually perform a write in reality, from the spec POV it does not. If an implementation does access an image as part of...
Oh my apologies, I misread the example - been too long since I've looked at this clearly 😬 Let me take another look at this..
Okay now I'm very confused. I disagree with the comments I made in the examples (I wrote those too!). So two things - firstly, as a rule of thumb, you...
Ack, hold up, this problem might run deeper than I thought, don't raise a validation issue yet (If you have nearly written it out feel free to post it though)...
> I think the issue is that the layout transition only makes the memory available and not visible? Yes, the issue is that writes are only supposed to need previous...