Thomas Eizinger

> These super-verbose replies aren't helpful, I think. My question is very simple: you wrote at the top of this issue that removing default-members works in your workspace, but you...

I am glad it is helpful! It would be nice to have this in the book in some form so it is more discoverable. In my experience, feature-gating the attributes...

Sure! As far as I remember from https://github.com/aya-rs/aya/pull/1229, the summary is that the aya maintainers considered it better / more correct to: - Ensure an eBPF kernel crate only ever...

To do this, we'll need to figure out how to make browsers trust the self-signed certificate during start-up, right?

> I'm not sure exposing more errors is useful. The motivation behind the current behavior is that UDP network errors are almost invariably one of two cases: those that can...

> Further, if your socket is bound to the wildcard address, as typical, there's no point recreating it in response to network changes. There is. On MacOS, a socket bound...

To be clear, if we decide to change this, then I'll obviously move the current behaviour up into `quinn` so for those users, nothing changes.

> > There is. On MacOS, a socket bound to the wildcard address will not be able to use newly attached interfaces. > > What, really? Is that documented anywhere?...

> > If there is a route (to a link-local peer), then the syscall isn't going to fail with NetworkUnreachable, right? > > It will if you try to communicate...

> > Being able to discard that server immediately gives a much better UX than having to rely on a (series of) timeouts before it is marked as unresponsive. >...