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

Re: NEW ISSUE: Clarify whether request must be processed before responding with redirection codes

From: cowwoc <cowwoc@bbs.darktech.org>
Date: Thu, 04 Jul 2013 10:05:40 -0400
Message-ID: <51D58134.8060109@bbs.darktech.org>
To: Julian Reschke <julian.reschke@gmx.de>
CC: ietf-http-wg@w3.org
On 04/07/2013 10:01 AM, Julian Reschke wrote:
> On 2013-07-04 15:44, cowwoc wrote:
>> On 04/07/2013 5:36 AM, Julian Reschke wrote:
>>> On 2013-07-03 21:16, cowwoc wrote:
>>>> Hi,
>>>>
>>>> Now that status code 303 and others no longer imply GET, we need to
>>> > ...
>>>
>>> How so?
>>>
>>> 303 always implied GET. 301, 302, and 307 never implied GET.
>>>
>>> The only change we made is to allow 301 and 302 to rewrite POST to GET
>>> (because that's what UAs do).
>>>
>>> Best regards, Julian
>>>
>>
>>      I think you misunderstood what I meant. In HTTP 2.0, the following
>> sentence was added to the description of HTTP 303: "This status code is
>> applicable to any HTTP method."
>
> You are referring to the revision of HTTP/1.1, not HTTP/2.0, right?

     Yes. Comparing 
http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html#sec10.3.4 to 
https://svn.tools.ietf.org/svn/wg/httpbis/draft-ietf-httpbis/latest/p2-semantics.html#status.303 
I see this new sentence.

>
>>      Meaning, although it was originally meant to be returned in
>> response to HTTP POST, it is now legal to return from HTTP
>> GET/PUT/DELETE as well. At least, that is my interpretation.
>
> It was always legal to return it for something different than POST. 
> The change that we made was in order to clarify that
>
> "This method exists primarily to allow the output of a POST-activated 
> script to redirect the user agent to a selected resource." (RFC 2616)
>
> does not imply that it's for POST only. See 
> <http://lists.w3.org/Archives/Public/ietf-http-wg/2007JulSep/0048.html> for 
> context.
>
> Best regards, Julian
>

     I understand. I'm saying that the 
http://en.wikipedia.org/wiki/Post/Redirect/Get paradigm is well 
understood but once you open the door to HTTP GET/PUT/DELETE it brings 
up the question of what it all means. It's not that this wasn't possible 
in the past. It's that returning HTTP 303 from these other methods isn't 
common or well understood yet.

Gili
Received on Thursday, 4 July 2013 14:06:37 UTC

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