W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > April to June 1999

RE: New status code: unordered collection

From: lisal <lisal@microsoft.com>
Date: Mon, 17 May 1999 17:54:16 -0700
To: "Yaron Goland" <yarong@microsoft.com>, <jamsden@us.ibm.com>, <w3c-dist-auth@w3.org>
Message-ID: <38CFE5072909D311830C00A0C9C74EB925BB@PTPUP.dfpt.extest.microsoft.com>
That sounds incredibly ambitious, but I like the general idea of a
root-element for errors.  It would be sufficient to limit ourselves to
defining a root just for errors, so that we could do:

450 Misc. Error HTTP/1.1
blah: blah

<DAV:errors>
  <DAV:advancedcollections><DAV:unorderedcollection/></>
  <foo:bar>baz</>
</DAV:errors>

Of course, defining what error codes this would work on, sufficiently
carefully in order to preserve true interoperability, sounds
difficult... ah, if we could just do everything.

Lisa

> -----Original Message-----
> From: Yaron Goland 
> Sent: Wednesday, May 12, 1999 8:42 PM
> To: jamsden@us.ibm.com; w3c-dist-auth@w3.org
> Subject: RE: New status code: unordered collection
> 
> 
> There are relatively few error codes and we should be very 
> hesitant before
> handing out new ones. I think 409 is a good error code for 
> this situation
> but obviously more data is needed. We should provide that 
> additional data
> either as a header in the 409 response or in the body.
> 
> If we are going to use the body then we do the world a great 
> favor if we can
> come up with a single format so that multiple, independent, 
> error conditions
> can be described. The most likely choice is XML but XML has a 
> problem. It is
> illegal for an XML document to have more than one root. This 
> means that if I
> want to return two error conditions in a single response 
> which were created
> by two unrelated groups I can't return them in a single XML 
> document because
> they have different roots. Therefore we seem to have three choices:
> 
> 1 - Don't use XML.
> 	Nice thought but probably impractical for reasons that market
> marketoids heart's glow bright red.
> 
> 2 - Use MIME Multipart to include multiple independent XML documents
> 	This will work but doesn't it just seem such a waste to have to
> throw in MIME multipart processing just because the XML guys 
> made a silly
> mistake?
> 
> 3 - Invent a global root element
> 	I'm a big fan of this solution. Let's just invent some universal
> root element (how about <root>?) and declare that ALL XML 
> returned in WebDAV
> error codes MUST go inside this element. Now we can throw in as many
> independent XML documents as we want and not have to worry 
> about the single
> root problem. If we feel like being really neighborly we can 
> even present
> this solution to other groups and maybe get all IETF XML to 
> be put inside
> this element.
> 
> 	Just a thought,
> 			Yaron	
> 
> 
> > -----Original Message-----
> > From: jamsden@us.ibm.com [mailto:jamsden@us.ibm.com]
> > Sent: Wednesday, May 12, 1999 12:27 PM
> > To: w3c-dist-auth@w3.org
> > Subject: Re: New status code: unordered collection
> > 
> > 
> > 
> > 
> > sounds OK to me.
> > 
> > 
> > 
> > 
> > Jim Whitehead <ejw@ics.uci.edu> on 05/12/99 02:08:43 PM
> > 
> > Please respond to ejw@ics.uci.edu
> > 
> > To:   WEBDAV WG <w3c-dist-auth@w3.org>
> > cc:    (bcc: Jim Amsden/Raleigh/IBM)
> > 
> > Subject:  New status code: unordered collection
> > 
> > 
> > 
> > 
> > In the current Advanced Collections specification, the 409 
> > (Conflict) status
> > code is used for cases where the server cannot perform the 
> > client's request
> > to place a resource at a specific position in a collection 
> because the
> > collection is unordered.  In my opinion, this is a good case 
> > for introducing
> > a new status code, say 425 (Unordered Collection).
> > 
> > 409 is a sub-optimal choice because this status code is used by many
> > existing WebDAV methods, as well as new methods in the 
> > advanced collections
> > specification, for non-ordering related problems.  Since a 
> > client can, if it
> > knows the error is the result of the collection not being 
> > ordered, simply
> > re-submit the request without the Position header, having 
> > this case clearly
> > separated from other 409 cases would give the client this option.
> > 
> > I'm not sure whether this code should also be used for other 
> > positioning
> > errors (such as trying to place a resource after one that 
> > doesn't exist).
> > I'm tending to think it should not.
> > 
> > - Jim
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> 


Received on Monday, 17 May 1999 20:54:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:43:49 GMT