jrhee17

Hi, I think I agree with you > A sender MUST NOT generate multiple header fields with the same field name in a message unless either the entire field value...

Talked with the others, instead of enforcing the validation of a single content-type in `ResponseHeaders`, it's probably enough to not set a `content-type` when calling `content`. How we can do...

It's been a while since I created this issue - looking at this issue again, I agree it's probably reasonable to make this the default behavior 👍

Looking at the logs, it seems like the connection was reset for some reason: ``` 07:51:22.179 [armeria-common-worker-epoll-3-4] WARN c.l.a.i.c.AbstractHttp2ConnectionHandler - [id: 0xfa43be94, L:/127.0.0.1:34782 - R:127.0.0.1/127.0.0.1:43087] HTTP/2 connection error: io.netty.handler.codec.http2.Http2Exception: failed...

It looks like the CI is still failing due to a blocking call. Can you check this?

Sorry, still looks like the blockhound CI is failing

> I understand that how TlsClientSpec will be integrated with ClientTlsConfig is still undecided. The rough idea I had was: - `ClientTlsConfig`, `ServerTlsConfig` is removed and replaced with solely `TlsProvider`...

> Ideally, the client should: > Check the connection pool for an existing connection, ie using the hostname Only resolve DNS if no pooled connection exists This would minimize DNS...

Note: I understood that if users are not interested in the server request or client response (fire-and-forget pattern), then the window frame won't be sent. ref: https://github.com/jrhee17/armeria/tree/poc/fire-and-forget I think most...

I think it's best to start with the direct issue you would like to solve - specifying `GracefulShutdown` for `ClientFactory#workerGroup`. What do you think of starting with adding the following...