Re: Locking URIs rather than Resources

I agree with all aspects of Yaron's interpretation of the "URL locking"
proposal.  It illustrates the fact that in this proposal, a LOCK (and UNLOCK)
request never fails due to the absence of a Lock-Token header.

A benefit of this approach can be seen if you invert the order in which the
locks are issued, namely, instead of:
  LOCK /x; Depth:0
  LOCK /x/y
you request:
  LOCK /x/y
  LOCK /x; Depth:0

In the current 2518 model, the first sequence would fail (for the reason
cited by Yaron) while the second would succeed (I believe ... with 2518,
you probably could make good arguments for both failure and success of the
second sequence).

In contrast, the proposed URI locking model treats these sequences
simply and identically: they both succeed, and they both end up in the
same state - with a Depth:0 lock on /x and a Depth:infinity lock on
/x/y.  In this state, if you want to do a PUT to /x/y and /x/y does
not identify a resource, you need both lock tokens (reasonably enough,
since you are modifying the state of both /x and /x/y).

Cheers,
Geoff

   From: "Yaron Goland (Exchange)" <yarong@Exchange.Microsoft.com>

   Before I jump into the fray I want to make sure that I actually understand
   the proposal. One sentence in particular gave me pause for thought. Geoff
   stated that "...a LOCK request *never* requires a lock token". I though hard
   about this, trying to understand exactly what it meant. Below I lay out my
   thinking process so that if I have misunderstood the proposal my error can
   be found and corrected.

   Imagine user U takes out a depth 0 lock on collection C. Collection C has no
   children. The ramification is that no one but U may add a new child to C.
   User A then shows up and wants to create a new resource with the URL C/R.
   Before creating the resource, however, A would like to first reserve the
   name C/R. To this end A takes out a lock on C/R.

   In the old system A's request would be rejected. The reason is that in the
   old system locks are on resources. Therefore for A to lock the name C/R it
   has to create some resource which 'owns' C/R and then lock that resource.
   This is where lock-null resources came from. They are the 'stand-in'
   resource who holds the name C/R until A is ready to transfer ownership of
   the name to its final destination. However creating the lock-null resource
   to hold C/R means that the lock-null resource would automatically become a
   child of C. U, through its lock on C, has reserved the exclusive right to
   create children of C. Therefore A's request to lock C/R is rejected because
   the creation of the lock-null resource to hold the name C/R would violate
   U's lock on C. (Is it just me or does the previous sound like an Abbott and
   Castello routine? Castello: Who has the lock? Abbott: Who has the lock.
   Castello: No, who has the lock?!? Abott: Exactly.)

   But in the new proposal locks aren't on resources, they are on URIs.
   Therefore when A locks C/R it is not locking a resource, it is locking a
   name. Names aren't members of collections, resources are. Therefore A can
   lock C/R without running afoul of U's lock on C because C/R, even though it
   is locked, is still not a member of the collection. When A will run into
   trouble is when it attempts, through a COPY, MOVE, PROPPATCH, PUT, MKCOL,
   etc. to associate the URL C/R with a resource. By associating a resource
   with C/R A will cause that resource to become a member of C. Of course U has
   reserved the right to create new members of C, therefore A's request will be
   rejected.

   The result is that A can lock C/R without submitting a lock token. However
   there isn't much A can do with the lock on C/R because doing anything to it
   would require associating it with a resource, which U's lock prevents.

   Geoff, did I get it right?

			   Yaron

   > From: Geoffrey M. Clemm [mailto:geoffrey.clemm@rational.com]
   > 
   > A lock can be placed on a URL with a LOCK request.  This URL is called
   > the "lock base".  If the lock is Depth:N, then a lock is induced on
   > any URL that consists of the lock base extended by N or fewer
   > segments.  Only an authorized holder of a lock token for a locked URL
   > can modify either the state of the resource identified by the URL, or
   > which resource is identified by the URL.  A URL MAY NOT be locked
   > with two exclusive locks with the same locktype.

Received on Thursday, 30 December 1999 22:26:25 UTC