core: remove envoy_stream_options

[rebello95]

As of #616, this is no longer used or necessary since it's handled by the router.

Resolves #643.

Signed-off-by: Michael Rebello me@michaelrebello.com

[junr03]

I had left this behind after I realized about #679. I was going to set the option up in this struct. Given @goaway took #679, up to him if he wants to use as is or do something else.

[rebello95]

Leaving the struct makes sense to me if that's how we're going to set up HTTP/2 preferences per-stream (though a lot of the deletions in this PR will still be relevant). WDYT @goaway?

[goaway]

Honestly, I'm ambivalent. Headers can work for almost every potential use-case I can think of, and there's already precedent (retry policies, timeouts). The only downside of headers is that they aren't typed, but this is all internal anyways. This is a potentially useful feature for cases where there wouldn't be headers (redis connections?), but it's not like we can't revisit it when/if that's relevant.

I guess I'm (ever so) slightly in favor removing it for now, simply because it adds overhead to all the function signatures.

[rebello95]

Sounds good. I agree that this could be used internally in the future, but we can always revert/revisit when that time comes (rather than keeping dead code on master).