wg-async VS delimited

I believe the solution mentioned in rust-lang/wg-async is the way to go: Add traits like AsyncWrite, AsyncRead, Executor/Runtime to the std so that tokio/async-std can implement them.

This is a consequence of runtimes relying on global variables that their core future types are dependent on. Creating abstractions to solve this problem is one of the main goals of the the async working group [0].

[0]: https://github.com/rust-lang/wg-async

Provided the futures you're selecting over are cancellation safe the plain loop over select! should be fine. Multiple channels in particular are safe to select over - if you have futures that aren't cancellation safe, you can just wrap them up in a task on the end of a channel and then select on that.

I'd like to be able to write runtime agnostic async libs.

You might be interested in taking a look at and potentially participating in the "Async Vision Document"[1] which is an exercise the team is going through to collect feedback about the current state of the ecosystem and what the pain points are, as well as a way to lay doing what the desired future state of async Rust should be[2]. The process is happening, as you would expect, in the open and there's still time to influence it[3] if your concerns aren't yet addressed or even mentioned[4].

[1]: https://rust-lang.github.io/wg-async-foundations/vision.html

[2]: https://blog.rust-lang.org/2021/03/18/async-vision-doc.html

[3]: https://github.com/rust-lang/wg-async-foundations/pulls

[4]: https://github.com/rust-lang/wg-async-foundations/issues

Thanks for the feedback. I posted this comment to a relevant github issue, fyi.

Oh, I have written my own share of userspace C context switching libraries, I know all the gory the details :). For example see my minimalist [1] stackful coroutine library: the full context switching logic is three inline asm instructions (99% of the complexity in that code is to transparently support throwing exceptions across coroutine boundaries with no overhead in the happy path).

You need compiler help for the custom calling convention support and possibly to optimize away the context switching overhead for stackful coroutines, which is something that compilers can already do for stackless coroutines.

The duff device is just a way to simulate stackless coroutines (i.e. async/await or whateverer) in plain C, in a way that the compiler can still optimize quite well.

[1] https://github.com/gpderetta/delimited/blob/master/delimited...

Using gcc extended asm you can pass literal constants to the asm and they will be expanded textually (or at least their address will). I don't think the details are fully documented anywhere and I had to use intel syntax to make it work, but it might be possible even wit AT&T syntax.

Take a look a this[1] for example. See how trampoline, the destructor and the size are passed in with the 'i' constraint and are referred to their value with the %cX constraint (yes, the code is write only and even with a lot of comments I have only the most vague idea of what I was trying to do here).

Probably more work is require for PIC though.

[1] https://github.com/gpderetta/delimited/blob/7e755d643ee45897...