Taiki Endo
Taiki Endo
> Would it make sense to depend on `either` instead? In addition to it being a breaking change, we are also implementing stream and async io traits on our `Either`...
Hmm, if the stream, future, and output are `Send`, `buffered` should also be `Send`. https://github.com/rust-lang/futures-rs/blob/183f8c61371d4c3891683ca51c72e579891bfa79/futures/tests/auto_traits.rs#L1117-L1120
> $ cargo build --no-default-features This is a mistake in the current readme (#174). Could you try `cargo install urdf-viz --no-default-features`?
If this can be implemented without spinlock, I tend to accept this.
See https://github.com/tokio-rs/bytes/issues/461 for previous discussions.
> Is there anything really blocking https://github.com/tokio-rs/bytes/pull/467 from being merged? The situation has changed since that PR was opened and I'm wondering about how to update it. The main issue...
How does `rand` handle this case? It would be preferable to provide the same behavior as they do.
Hmm, looking at the PR that added Clone impl (https://github.com/smol-rs/fastrand/pull/4 by @cloudhead), it seems that the current behavior is intentional?
Since windows_process_extensions_raw_arg was stabilized on Rust 1.62, I would accept a patch to support this on 1.62+ using the build script. It would also be nice to do #6 at...
(Some code still refers `feature = "io"`...)