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

RE: Should a DELETE against a non-existent resource return a 2xx code?

From: Yves Lafon <ylafon@w3.org>
Date: Fri, 24 Aug 2007 02:57:17 -0400 (EDT)
To: Mike Dierken <dierken@gmail.com>
Cc: 'Yaron Goland' <yarong@microsoft.com>, 'Henrik Nordstrom' <henrik@henriknordstrom.net>, ietf-http-wg@w3.org
Message-ID: <Pine.LNX.4.64.0708240255460.29297@ubzre.j3.bet>

On Thu, 23 Aug 2007, Mike Dierken wrote:

>
> Hmm, I've always preferred a 200 Success, on the basis that the state of
> 'being gone' is now successfully transferred. I've also thought it was good

If the resource was not there in first place, a 404 indicates that.
If the resource is present, a 200 for the first DELETE, then 410 for the 
subsequent ones leave no room for ambiguity.

> to have multiple DELETE requests respond with the same status code. For
> example, if the first DELETE failed to respond (but work was actually
> performed), the second request (which could have been auto-retried by a
> client library or intermediary) would return a 4xx code, even though the
> resource existed at the time of the first request.
>
> But there is a definite lack of clarity on the 'right' answer.



>
>
>
>> -----Original Message-----
>> From: ietf-http-wg-request@w3.org
>> [mailto:ietf-http-wg-request@w3.org] On Behalf Of Yaron Goland
>> Sent: Thursday, August 23, 2007 9:40 AM
>> To: Henrik Nordstrom
>> Cc: ietf-http-wg@w3.org
>> Subject: RE: Should a DELETE against a non-existent resource
>> return a 2xx code?
>>
>> Yeah it seems pretty clear now that 404 is the right answer.
>> I think my internal spec lawyer just went looney.
>>
>>> -----Original Message-----
>>> From: Henrik Nordstrom [mailto:henrik@henriknordstrom.net]
>>> Sent: Wednesday, August 22, 2007 4:12 PM
>>> To: Yaron Goland
>>> Cc: ietf-http-wg@w3.org
>>> Subject: Re: Should a DELETE against a non-existent
>> resource return a
>>> 2xx code?
>>>
>>> On mn, 2007-08-20 at 10:05 -0700, Yaron Goland wrote:
>>>
>>>> What’s the opinion of the working group?
>>>
>>> I would prefer 404. It's quite likely useful to the client to know
>>> that the object wasn't found by the server.
>>>
>>> 404 is also in line with how servers respond to other requests for
>>> non-existing resources.
>>>
>>> Regards
>>> Henrik
>>
>>
>
>

-- 
Baroula que barouleras, au tiu toujou t'entourneras.

         ~~Yves
Received on Friday, 24 August 2007 06:57:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:15 GMT