W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > April to June 2002

Re: Issue: COPYMOVE_LOCKED_STATUS_CODE_CLARIFICATION

From: Elias Sinderson <elias@cse.ucsc.edu>
Date: Sat, 04 May 2002 15:44:48 -0700
Cc: WebDAV <w3c-dist-auth@w3.org>
Message-id: <3CD46460.1080005@cse.ucsc.edu>
It seems reasonable to me, although the details are still to be worked 
out for a few small things.

Elias



Jason Crawford wrote:

> The recent thread entitled
> RE: need clarification about COPY to locked resource response cod       e
>   and
> RE: need clarification about COPY to locked resource response cod e
> recently died down with the following posting.
> http://lists.w3.org/Archives/Public/w3c-dist-auth/2002AprJun/0043.html
> 
> I just want to bring it to conclusion and get it on the ISSUES list.
> 
> 
> So we have a proposal on the table regarding error reporting for lock
> conflicts
> during COPY/MOVE.  I'll try to describe it below.
> 
> If a portion of the destination is locked and preventing a portion of the
> COPY/MOVE,
> then a 207 Multistatus MUST be returned by the server.   That MultiStatus
> response should
> indicate (with 409 Conflict status) what source resource could not be
> copied/moved and
> optionally say something about the locked destination resource the locked
> that prevented
> the copy.
> 
> <D:response>
>     <D:href>/bla/source</D:href>
>     <D:status>HTTP/1.1 409 CONFLICT</D:status>
>     <D:responsedescription>
>       <D:error>
>          <D:locked-destination>
>            <D:href>/other/destref...</D:href>
>          </D:locked-source>
>       </D:error>
>     </D:responsedescription>
>  </D:response>
> 
> That D:error tag is optional.  And clearly we'd need the add definitions
> for the
> D:error tag and the descendent tags listed in the example above.
> We should work out if the dest URL listed is the root of the lock or the
> locked resource
> that was directly affecting the operation.  We should possibly clarify
> "protected URL"
> status codes and what URL is depicted.
> 
> If none of the COPY/MOVE could be done due to a locked destination, then it
> is acceptable,
> although obviously not optimal, to simply return a 409 CONFLICT rather than
> a
> 207 MULTISTATUS.
> 
> As for *source* locks...
> 
> Source lock error don't occur for COPY, just MOVE.  And the 423 LOCKED
> status
> code will be used instead of 409 CONFLICT.  More specifically...
> 
> In the case of a MOVE operations where a portion of the source is locked,
> the server
> is to respond with 207 MULTISTATUS and the body of that will simply
> indicate the
> locked source resources with a 423 LOCKED status.
> 
> <D:response>
>    <D:href>/bla/...</D:href>
>    <D:status>HTTP/1.1 423 LOCKED</D:status>
> </D:response>
> 
> 
> In the case that the request (source) URL is locked in a way that prevents
> any other
> MOVEs to occur, the server is free to simply reply with a 423 LOCKED rather
> than
> a 407 MULTISTATUS.
> 
> In all cases, minimization rules apply.  I will start a seperate thread for
> Minimization
> issues if necessary.
> 
> So does this sound acceptable to everyone?
> 
> J.
> 
> 
> 
> 
Received on Saturday, 4 May 2002 18:51:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:44:00 GMT