W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2014

Re: #549: END_STREAM flag on CONTINUATION

From: Mark Nottingham <mnot@mnot.net>
Date: Thu, 24 Jul 2014 16:40:33 -0700
Message-Id: <F33C3095-B2E6-4957-9EF9-A2B63231FCB3@mnot.net>
To: HTTP Working Group <ietf-http-wg@w3.org>
Im not hearing any consensus to make a change here, and the issue itself seems more about how people perceive the flags than any concrete problem.

If someone can suggest text to clarify here, please suggest it to Martin; Im closing the underlying issue as WONTFIX.

Regards,


On 20 Jul 2014, at 12:47 pm, Mark Nottingham <mnot@mnot.net> wrote:

> <https://github.com/http2/http2-spec/issues/549>
> 
> In the issue comments, Martin points out the reasoning behind the current design:
> 
>> To articulate the reasons for the current design:
>> 
>> 	 END_STREAM has no place on PUSH_PROMISE, or continuations thereof.
>> 	 Placing END_STREAM on HEADERS ensures that there is no possibility for stupid errors where you have END_STREAM but not END_HEADERS.
>> 	 CONTINUATIONS are in most respects a way to create a single frame from many. Logically, they are part of the preceding HEADERS/PUSH_PROMISE. Adding some flags from the preceding frame but not others is conceptually muddy.
>> I don't recall any complaint from people actually implementing the protocol.
> 
> The first point is the one Im focusing on; it seems like there are arguments on both sides as to how this could be confusing, making it a toss-up (and thus inclining me to close the issue with no action).
> 
> Any further discussion? Could we address the issue that people have by clarifying the text, perhaps?
> 
> --
> Mark Nottingham   http://www.mnot.net/
> 
> 
> 
> 

--
Mark Nottingham   http://www.mnot.net/
Received on Thursday, 24 July 2014 23:40:59 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 30 March 2016 09:57:09 UTC