Thomas Eizinger

Interesting, so that means the existing checks for GSO don't work. Can you log this error? https://github.com/quinn-rs/quinn/blob/4f8a0f13cf7931ef9be573af5089c7a4a49387ae/quinn-udp/src/unix.rs#L339

Actually, the fact that it fails on the comparison means we never get to this error path? That is odd. It would mean that Android silently fails despite it not...

> The value of `n` in the `if` condition is 8192. Which is obviously way too big for a single UDP packet on the public Internet. I am surprised that...

> Maybe something like this? [inetic@26e3950](https://github.com/inetic/quinn/commit/26e3950bfd37048777b9674edf45b05bb6e59abd) I think this is a good start. The fallback value IMO when we have a parsing error should perhaps be `true` though and we...

> > I wonder if it is specific to Android and stock Linux 4.14 does error? If so, then I think it would make sense to guard such a check...

> > I don't love parsing a string for it, but if that's the best option, so be it. > > One other option that comes to my mind is...

Perhaps it might be worth picking up https://github.com/libp2p/rust-libp2p/issues/1533 again. If we were to integrate with tracing, you would be able to see the span-stack of each message. So instead of...

CI is green! Ready for review @conectado !

Now that #6941 is merged and deployed to the AWS gateway which tracks latest `main`, this can be tested using clients built from this branch. I've done that with the...

> When connecting to wildcard DNS resources, each new connection that matches the wildcard drops the first packet. Is that expected? This makes, for example, loading GitHub.com a bit slower...