Thomas Eizinger

I think in principle, I am okay with this direction. > The idea of doing this, arose from me working on trying to integrate rust-libp2p with [DataDog/glommio](https://github.com/DataDog/glommio/). With that executor,...

> Re: boxing and cancellation-safety: now that GATs are possible, traits can include lifetime-bearing associated futures, so it's possible to abstract over futures like `Accept` without boxing. That said, a...

> The associated types can be `Unpin`! `pin-project-lite` makes working with this type of thing very easy, as demonstrated by quinn's own wrapping of `Notified`. Unless I am missing something,...

> You could refactor the traits that must be implemented to propagate `!Unpin` futures upwards rather than requiring implementations to erase them behind a poll-style API. The extra indirection can't...

> > [flatten StreamMuxer interface](https://github.com/libp2p/rust-libp2p/pull/2622/commits/cf2c11d213a3aba40faf92f5c9aa2c433399073a) > > Thanks for adjusting to the change so quickly. //CC @thomaseizinger to see your change in action on a different transport. Cool! I'll definitely...

> [melekes#1](https://github.com/melekes/rust-libp2p/pull/1) broke the smoke test. trying to figure out why. Did it? I noticed that the test wasn't passing but last time I checked, it also wasn't passing on...

This is obsolete with #2996.

> I am fine with either solution. I like that test input generation is specified at the type level, that is why I didn't use proptest thus far. Proptest on...

> > without a dependency to proptest in the original crate and thus without orphan-rule issues. > In this case, we control all crates and we would depend on proptest...

This seems to have been an issue with the dependencies installed through my package manager. A wash of updates came in today, including one for electron19 which fixed the issue.