Benjamin Saunders

> For Rust-based layer implementations this may be somewhat cumbersome, for which we could re-expose the setters behind a cfg(feature = "returnedonly-setters") of sorts? Sounds like a good idea to...

+1 to surprise that Linux doesn't generate an error on unknown cmsg. If that's happening, then it's totally reasonable for quinn-udp to check the kernel version. I don't love parsing...

The reason behind our batched high level APIs isn't to avoid multiple hash-table lookups -- it's to avoid churning the connection Mutex, which might be contended, at the `quinn` layer....

Looks generally attractive. I wonder if we could move the errors from `write_limit` to `mark_blocked`, and rename the latter to something like `fn begin_write(&mut self) -> Result` to simplify things...

My thinking is that caller logic is simplified if the branch on "is the limit 0" gets pushed down into shared logic, so you can write `stream.begin_write()?` instead of `if...

Makes sense. I don't feel that strongly, either way is a big improvement.

> I saw only 80% of packets were received on the server Are you losing IP packets or application-layer datagrams? If there's no IP packet loss, but you are losing...

> I still experienced only receiving 70% of data in less than a minute. I'd recommend trying to trace the individual lost packets through the stack. If you're not seeing...

> Do I have to set both on the client and server side? Congestion control affects outgoing data, so you should adjust the congestion control config wherever you're sending data...

I also encountered this with a simple horizontal line with a horizontal gradient stroke style in cetz.