Eliza Weisman

Got it, yeah, thanks for clarifying that. I saw the phrase "periodic maintenance loop" and assumed this was background maintenance work that would run less frequently.

> * How is `JoinMap` implemented? A layer on top of `JoinSet` or a completely separate type? I'm prototyping a `JoinMap` impl in `tokio-util`, should be up for review shortly...

`JoinMap` sketch: #4538

One major deficiency of the current `JoinSet` (and `JoinMap` as of #4538) APIs is that there isn't any practical way to use them in a polling-based context. I noticed this...

> See the implementation of `BroadcastStream` for how the same problem was solved for another collection. Yeah, I considered an approach like that, where the `JoinSet`/`JoinMap` is moved in and...

Merged an implementation of a `JoinMap` type in `tokio-util` in #4640. We should probably do a release of both `tokio` (with the latest `JoinSet` changes) and `tokio-util` (with `JoinMap`) with...

> Regarding making `JoinSet::insert` public, there are some concerns that we might want to change how the future is allocated when it is inserted into a `JoinSet`, which we wouldn't...

Here's my proposal for making `JoinSet` play nice with the task builder API: https://github.com/tokio-rs/tokio/pull/4687 What do others think?

You might want to check out something like [`capsule`](https://lib.rs/crates/capsule) for packet-level use cases. It's built on top of Tokio (although it looks like it has yet to be updated to...

Of course, if it's possible for `quanta` to ever generate `Instant`s with the `u64` value 0, this won't work, so we'd need to know for sure that this is the...