Phoenix Kahlo

Procedurally modifying a URI with a regex concerns me. I think an approach based on using `url.parse` and other built-in URL-parsing machinery would be less likely to introduce bugs. I...

> Does this fix [rms-support-letter/rms-support-letter.github.io#5037](https://github.com/rms-support-letter/rms-support-letter.github.io/pull/5037)? I assume you're responding to my code suggestion? Yeah, I ran the string `"https://[email protected]/sticks-stuff"` through my suggested function and it successfully returned `null`. -- Edit...

No, I only register the object on the object space it's intended to be invoked on. If I registered it with all object spaces, how would the invocation be applied...

Thanks for the review! Splitting this into more incremental commits and making the other changes you mentioned makes sense to me. Just so you know, though, it might be a...

Should be ready for re-review now. I split this up into various commits which should make this a lot easier to see what I'm doing if you go through them...

> Should we introduce a new public future (perhaps Handshaking) returned by accept, and move this and into_0rtt there? > > Pros: > > - Reduced side effects here and...

> After sleeping on it I think the current naming is probably the sweet spot. Huh alright. I mean I'd be fine renaming it to `IncomingConnection::connect` on both the quinn...

> Could we simplify the API by returning Connecting unconditionally, but constructing it such that it immediately yields a ConnectionError when awaited if the proto-layer accept failed? That seems like...

I re-added the `RetryPolicy` callback to the quinn-proto testing utilities, due to me re-added the quinn-proto `reject_remote_address` test (now renamed to `reject_manually`). With regards to your previous comment on this...

All comments should be resolved, and I just refactored the commit history and force-pushed again so it should be more clean now.