RE: HTTP/2.0 -04 candidate

From http://tools.ietf.org/html/draft-unicorn-httpbis-http2-00#section-6.6:

   The state of promised streams is bound to the state of the original
   associated stream on which the PUSH_PROMISE frame were sent.  If the
   originating stream state changes to fully closed, all associated
   promised streams fully close as well. [[anchor10: Ed.  Note: We need
   clarification on this point.  How synchronized are the lifecycles of
   streams and associated promised streams?]]

I thought previous on-list discussion had agreed to remove this tie between stream lifecycles and just send a stack of RST_STREAMs in a single packet if need be.  Is that still the consensus?

-----Original Message-----
From: Martin Thomson [mailto:martin.thomson@gmail.com] 
Sent: Monday, July 1, 2013 5:23 PM
To: HTTP Working Group
Subject: HTTP/2.0 -04 candidate

Those people who volunteered to contribute to the layering work in the SF interim have come up with something.  This includes a restructuring of the content.

Since the changes are large in scope, we're not submitting this as draft-ietf-httpbis-http2-04.  I've put this up as an individual submission so that people can comment on structure, text, omissions:

 http://tools.ietf.org/html/draft-unicorn-httpbis-http2-00


Note: This is a proposal for the content of the draft-ietf-httpbis-http2-04.  Please let us know - as soon as possible
- if the idea of this becoming a -04 offends you somehow.

This draft includes resolutions to all the issues on our milestone, with the exception of two (#75: default priorities, #17: opaque data in GOAWAY and RST_STREAM), which I plan to address tomorrow.

Now, most of the final pass is my fault (with a little help from the github unicorn), so blame me for all the bad stuff and praise Jeff and James for providing all the good stuff.

Received on Tuesday, 2 July 2013 15:46:25 UTC