John Wells

I modified `examples/bindless.rs` to [use hassle-rs (DXC)](https://github.com/attackgoat/screen-13/commit/5dfaa55a9ef487070a9f1111906a79fdc4663fa7) and had some success: ```sh RUST_LOG=info cargo run --example bindless -- --hlsl --hassle-rs ``` When I don't pass the `--hassle-rs` flag it uses...

I'm interested in finding the minimal HLSL that you think should work, but does not. Hopefully that would help figure out if there is anything this crate could change which...

> If it helps at all ... Yes - it very much does. Between this codebase being somewhat sizable and my other projects being my main focus I very much...

I think I have a solution, but I'm going to have to write a few more advanced examples to make sure it handles the edge cases (multi-layer image rendering where...

TODO: - [x] `Resolver::lease_render_pass()`: Look through secondary passes when figuring out attachments if the attachment is unused in the first pass(es)

TL;DR: I had to make a number of fixes and roll-back a few of the merge-check changes (because they were not the issue), but this still needs additional tests and...

I ran into this recently with a depth pass and was able to resolve it by changing `src/graph/resolver.rs:1971` to use the "late" access instead of "early": ```rust let next_access =...

The logic in #89 may be the correct fix for this case. If possible, could you test against this branch?

This is fixed by #81 The PR allows `multipass.zip` to not throw any validation errors but the logic results in a blank screen. I think this is the valid for...

Synchronization works at the level of stage and access, where atomic operations like `atomicAdd(..)` happen within some compute, fragment, _etc_ stage. `AccessType::General` waits for all stages and ensures all access...