Re: LOCK error marshalling (lockdiscovery property included?)

Lisa Dusseault wrote:

> Hello client implementors! -- please respond on this thread and let us 
> know if you use the 'lockdiscovery' property value, which is returned in 
> failed LOCK requests.
> 
> (If no clients use the property value as returned in failed LOCK 
> requests -- maybe they all do another round-trip to get the value -- 
> then we might as well simplify, rather than optimize for this failure case)

OK,

I just tested Adobe Golive (the client besides our own I'm aware of 
using shared locks) with Apache/moddav, trying to get an exclusive lock 
on a resource that already has a shared lock. The response from 
Apache/moddav is a 423 Locked, so no Multistatus.

Unless I'm missing something this means that the marshalling shown in 
section 8.10.10 of RFC2518 
(<http://greenbytes.de/tech/webdav/rfc2518.html#rfc.section.8.10.10>) 
indeed won't help any clients at all; and as it is underspecified we 
should just get rid of it.

Proposed changes:

1) Change section 8.10.2 
(<http://greenbytes.de/tech/webdav/rfc2518.html#rfc.section.8.10.1.p.2>) 
to say: "The response for a successful LOCK creation request MUST 
contain the value of the lockdiscovery property in a prop XML element."

2) When we define named preconditions, define a content model for the 
precondition that will reveal the lockdiscovery property.

Jason, could you please add this to the issues list? (I'm adding this to 
my list as 
<http://greenbytes.de/tech/webdav/draft-reschke-webdav-locking-latest.html#rfc.issue.8.10.1_lockdiscovery_on_failure>, 
please let me know if should rephrase it).


Best regards, Julian


-- 
<green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760