W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 2007

Re: Redirection of a POST as a GET

From: Henrik Nordstrom <henrik@henriknordstrom.net>
Date: Wed, 07 Mar 2007 20:09:10 +0100
To: Julian Reschke <julian.reschke@gmx.de>, Anne van Kesteren <annevk@opera.com>
Cc: "ietf-http-wg@w3.org Group" <ietf-http-wg@w3.org>
Message-Id: <1173294550.13906.128.camel@henriknordstrom.net>
tis 2007-03-06 klockan 16:18 +0100 skrev Julian Reschke:

> > redirection of a POST as a GET etc. I'm not aware of a full list though, 
> > but such a list would probably become apparent once you start doing 
> > actual testing.
> Is that about status 302 vs 303? Is there an open issue around here?

Well, kind of. The specs and real world is still not aligned, just as in
the case of Content-Location.

RFC2616 added the 303 and 307 status codes trying to solve the problem
by defining new status codes replacing the earlier. But it did not
change 302 to reflect what the majority of the implementations actualy
do other than adding a note that many implementations is broken and uses
GET instead, and neither did it mark 301/302 as unusable.. And with most
implementers (all sides) still following the majority instead of

And as result

In response to a POST

      2616    Most User-Agents
301   POST*   GET
302   POST*   GET
303   GET     GET?
307   POST    POST?

* with a note that there is implementations using GET.
? Possibly not implemented

As a result very few expects implementations to follow the RFC wrt
301/302 responses, and the new status codes is not used very much as few
see a need for them..

Here is some stats from a sample of 3xx responses to little more than 3
million POST requests:

301 1100
302 78793
303 664
304 25209
307 31

the complete breakdown in status codes in response to POST can be found
at http://www.henriknordstrom.net/code/http_post_status-20070307.txt


Received on Wednesday, 7 March 2007 19:09:24 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 11:10:41 UTC