Max Inden

Is this item still actionable? > Requiring a reference to the connection can be problematic, if self is already borrowed. Is this still an issue, given that Rust's borrow checker...

Nothing comes to mind. I will try to reproduce locally.

Hypothesis: - We no longer pad with `0x00`s, but `0xc8` instead. - `0x00` padding is filtered out early in our processing pipeline https://github.com/mozilla/neqo/blob/4ad2f9a3a98db0b164b6106d79959831799dd84b/neqo-transport/src/packet/mod.rs#L619-L628 - `0xc8` is interpreted as a long...

For what it is worth, I too am unable to reproduce the regression locally with the latest patches. Here using the `Download` benchmark: ``` critcmp main scone group main scone...

> the second connection seems to run into idle timeouts on both client and server. I am assuming that you are basing this on the following log line? ``` 0s269ms...

@larseggert with neqo-neqo passing the resumption and zerortt interop test, can this issue be closed?

> 1-conn/1-100mb-resp (aka. Download)/client: 💚 Performance has improved. > > time: [98.680 ms 98.918 ms 99.154 ms] > thrpt: [1008.5 MiB/s 1010.9 MiB/s 1013.4 MiB/s] > change: > time: [-12.936%...

This pull request is ready for a first review. Due to its substantial size, I recommend beginning with the (updated) description of the pull request.

> > Long-lived send and receive buffers come with a bit of complexity. Do you think that complexity is worth the benefit? In other words, are you in favor of...

@larseggert I did the renames suggested above.