- From: Martin Thomson <martin.thomson@gmail.com>
- Date: Mon, 2 Dec 2013 11:20:10 -0800
- To: Jeff Pinner <jpinner@twitter.com>
- Cc: Alan Shreve <alan@inconshreveable.com>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
On 2 December 2013 11:12, Jeff Pinner <jpinner@twitter.com> wrote: > The reason it's not stream closed in section 6.1 is the stream may be in the > idle state when you receive the DATA frame. You know, I think that I'm OK with STREAM_CLOSED there too. Though treating that as a connection error (as opposed to a stream error) is probably best. But that's a discretionary thing. But then, I think that it's perfectly valid for an implementation to upgrade all their stream errors to connection errors.
Received on Monday, 2 December 2013 19:20:37 UTC