Joshua Humphries

Also, it's not clear why the `MytestServiceClientInterface` would reference the `Mytest` message _at all_, since the request and response types point to `ListMytestsRequest` and `ListsMytestsResponse`, not to `Mytest`. Maybe you...

@akosyakov, can you provide a small repro case? I had questions above about what you are saying and how to reproduce it.

Connect does not use server-sent-events for a few reasons: 1. SSE is a text format. It doesn't play nicely with binary payloads, such as binary-encoded protobuf messages. Such messages would...

This has been something we've discussed a lot. The main challenge was the potential for drastically increasing the surface area of the API and the extra surface area for bugs...

@XSAM, let's try to reach consensus in the issue (#186) before proceeding further with this PR.

@dfawley, I think stale bot got it wrong on this one, too. I see this is "on hold", but probably should not be closed -- AFAICT, it is not pending...

@replu, that is not the proper idiom. Code should examine if the error is nil, and if it is not _assume_ the other value is. So in that case, instead...

Re-opening just to serve as a doc bug: we don't intend to change the current behavior, but we definitely want to be more clear in the docs and in Go...

Perhaps it would be possible to move some of the Arrow integration stuff into a sub-package, so that depending on the main "iceberg" package (which defines types like the manifests,...

> Would your solution in https://github.com/connectrpc/connect-go/pull/851#discussion_r2135765436 solve this case nicely too and we can pass RequestInfo to the handler? Oh, good idea. That would also completely eliminate any `atomic.Pointer`/`atomic.Value` shenanigans...