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

Re: h2 header field names

From: Mark Nottingham <mnot@mnot.net>
Date: Thu, 4 Sep 2014 14:42:23 +0300
Cc: Martin Thomson <martin.thomson@gmail.com>, Roy Fielding <fielding@gbiv.com>, HTTP Working Group <ietf-http-wg@w3.org>
Message-Id: <4CFA7625-24C4-49CB-BCED-32598C181ACC@mnot.net>
To: Poul-Henning Kamp <phk@phk.freebsd.dk>
Because chunk boundaries have no semantics in HTTP/2, whilst headers do. 


On 4 Sep 2014, at 11:30 am, Poul-Henning Kamp <phk@phk.freebsd.dk> wrote:

> --------
> In message <843A253B-27A6-474A-B0DD-55DE2D8CA988@mnot.net>, Mark Nottingham wri
> tes:
> 
>> I was responding to your question about the 'architectural decision'
>> of a character set.
> 
> Yes, and I'm asking why that that particular decision is out of
> bounds for HTTP/2.0, when we can kill other mis-uses of HTTP/1.1
> without trouble ?
> 
> How is restricting the charset in a way which is compatible with
> what the HTTP/1.1 spec says different from throwing out chunked
> boundaries in a way which is compatible with what HTTP/1.1 says ?
> 
> Both a architectural decisions which means that some tiny subset of
> HTTP/1.1 traffic won't tunnel through HTTP/2.0.
> 
> But one decision could be made with no trouble, the other is out of bounds ?
> 
> Why ?
> 
> -- 
> Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
> phk@FreeBSD.ORG         | TCP/IP since RFC 956
> FreeBSD committer       | BSD since 4.3-tahoe    
> Never attribute to malice what can adequately be explained by incompetence.
> 

--
Mark Nottingham   http://www.mnot.net/
Received on Thursday, 4 September 2014 11:42:52 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 30 March 2016 09:57:10 UTC