W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2013

Re: Issue # 164: HTTP Method Case

From: James M Snell <jasnell@gmail.com>
Date: Wed, 3 Jul 2013 11:25:48 -0700
Message-ID: <CABP7RbdOtmgnMYe0CnnXNz3dUx2qY-yhQGBqNzegNcLFvPXhMA@mail.gmail.com>
To: Roberto Peon <grmocg@gmail.com>
Cc: Martin Thomson <martin.thomson@gmail.com>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
Appendix A.1 of the current header compression draft shows "get" and
not "GET" as the pre-filled value...


Regardless, this is not a compressor issue. It's a http semantic layer issue.

On Wed, Jul 3, 2013 at 11:22 AM, Roberto Peon <grmocg@gmail.com> wrote:
> I agree with Martin-- this data should be as it was in HTTP/1.1
> Reasoning:
> 1) It works.
> 2) The methods are so very extremely likely to already be represented in the
> compressor, that we shouldn't much care about what is in them anyway.
> -=R
> On Wed, Jul 3, 2013 at 11:18 AM, Martin Thomson <martin.thomson@gmail.com>
> wrote:
>> On 3 July 2013 10:43, James M Snell <jasnell@gmail.com> wrote:
>> > Need to clarify whether or not HTTP methods in 2.0
>> > are case sensitive or not.
>> HTTP/2.0 is actually silent on the issue.  And I believe that it can
>> remain so.
>> HTTP/2.0 makes header field names lowercase (and mandates that), which
>> is fine because that has a performance and compression advantage.  A
>> similar argument could be made for any header field value, :method not
>> being particularly special in this regard.
>> In this specific case, if we were to mandate a particular case, I'd
>> advocate uppercase.  I believe that some code breaks if it receives a
>> 'get' instead of a 'GET', so uppercase would have to be the choice.
Received on Wednesday, 3 July 2013 18:26:35 UTC

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