- From: Daniel Sommermann <dcsommer@fb.com>
- Date: Tue, 22 Apr 2014 10:35:32 -0700
- To: Martin Thomson <martin.thomson@gmail.com>
- CC: "William Chan (ιζΊζ)" <willchan@chromium.org>, HTTP Working Group <ietf-http-wg@w3.org>
On 04/22/2014 10:01 AM, Martin Thomson wrote: > I'm not sure that the problem is any different. You still have to > ensure that you wait long enough after sending DRAINING before you > send the GOAWAY (the above math is exactly the same). The difference > is that for a client, two GOAWAY frames are treated exactly the same, > i.e., clients don't need two code paths. OK that sounds reasonable. I'm sure you were planning this already, but we should be explicit that the first GOAWAY with extra acknowledged streams does not imply the client can continue to open new streams up to that stream id. I couldn't find any text in the spec for how to process multiple GOAWAY frames. This can actually happen today, apart from this proposal, if some connection level error happens on the final stream(s) after a GOAWAY NO_ERROR. Just to keep in mind while drafting the new text.
Received on Tuesday, 22 April 2014 17:36:02 UTC