Christian Legnitto

Awesome, thanks for the pointers! I Hope to look into this more after the holidays.

I'm not too familiar with this stuff, but it looks like the example in the wgpu repo calls `surface.resume()` when `start_condition()` is met: https://github.com/gfx-rs/wgpu/blob/0dc9dd6becdfd69ab241169171c37428fefaa1da/examples/src/framework.rs#L395 It looks like `start_condition()` is only...

There is a RFC to add `f16` to Rust: https://rust-lang.github.io/rfcs/3453-f16-and-f128.html

Hmmm, from the tests it looks like this might already be supported? https://github.com/EmbarkStudios/rust-gpu/blob/main/tests/ui/lang/core/mem/create_unitialized_memory.rs

You might want to use something like https://github.com/TimelyDataflow/differential-dataflow or https://github.com/vmware/database-stream-processor to make the calculations incremental. An example of using the former is in https://github.com/dtolnay/cargo-tally and the latter has some good...

How does this interact with drop and cancelation? Is it better or worse than the current model? This reminds me a lot of completion based io for some reason (https://www.ncameron.org/blog/async-io-with-completion-model-io-systems/).

Perhaps dumb questions: 1. Are the stubs a maintenance burden? 2. Is it strictly better to have stubs than nothing at all? 3. Does having the stubs and test machinery...

> To some extend. We already had a case of someone adding a new shim modeled after one of the stubs. But it is unclear whether that stub can even...

Makes sense to me!

I had commented in discord about having a filesystem abstraction, which is basically the end state of this issue for fs isolation. Fwiw, there are crates that purport to have...