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

Re: #540: "jumbo" frames

From: Michael Sweet <msweet@apple.com>
Date: Fri, 27 Jun 2014 14:05:57 -0400
Cc: Greg Wilkins <gregw@intalio.com>, Mark Nottingham <mnot@mnot.net>, HTTP Working Group <ietf-http-wg@w3.org>
Message-id: <6D1A46FB-F81A-4CDB-B2E0-AF9C31F56898@apple.com>
To: Martin Thomson <martin.thomson@gmail.com>
Martin,

On Jun 27, 2014, at 1:05 PM, Martin Thomson <martin.thomson@gmail.com> wrote:
> On 27 June 2014 09:45, Michael Sweet <msweet@apple.com> wrote:
>> FWIW, my current HTTP/2 implementation does not support continuation
>> headers, for the simple reason that 99.9999% of our users do not need
>> extremely large header support (for ActiveDirectory/RFC 4559 deployments).
> 
> Just to clarify behaviour: what do you do when:
> a) you are asked to transmit more headers than will fit in a single
> frame through your API?

The API will return an error.

For IPP requests the likelihood of going over 16k-1 is basically 0 unless Kerberos authentication (RFC 4559) is being used.  If it becomes an issue I can always force HTTP/1.1 operation when Kerberos authentication is used.  Of I could decide to support CONTINUATION frames in the future.  But since 99.99% of our users do not use Kerberos for printing (and those that do are generally using SMB or IPP to a Windows server where I don't expect IPP over HTTP/2 support), I have little incentive to implement support for something I'll never use.

> b) you receive a CONTINUATION frame?

413

_________________________________________________________
Michael Sweet, Senior Printing System Engineer, PWG Chair


Received on Friday, 27 June 2014 18:06:24 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:14:31 UTC