Re: RSR-merge clarification

Geoffrey M. Clemm (gclemm@tantalum.atria.com)
Tue, 20 Jul 1999 22:42:29 -0400 

Date: Tue, 20 Jul 1999 22:42:29 -0400
Message-Id: <9907210242.AA18594@tantalum>
From: "Geoffrey M. Clemm" <gclemm@tantalum.atria.com>
To: Tim_Ellison@oti.com
Cc: ietf-dav-versioning@w3.org
In-Reply-To: <1999Jul20.135000.1250.1260324@otismtp.ott.oti.com>
Subject: Re: RSR-merge clarification

Good catch, Tim.

The versioning-01.1 RSR-merge conflict semantics was rather
arbitrary and had the potential of masking the conflicts.
In the versioning-02 spec, this section has been modified to state
"in case of a conflict, a conflict status is returned and no revision
is selected".

And yes, the trivial case of the merge rule constituents 
selecting only a single revision results in a successful
selection of that revision, without conflict.

Cheers,
Geoff

   From: Tim_Ellison@oti.com (Tim Ellison OTT)

   Perhaps you can clarify the description of the 'RSR-merge' element for me.

   I understand that merge rule takes the 'ultimate' descendent of the 
   versioned resource revisions selected, but the conflict case confuses me. 
    The doc (Clemm et al. versioning-01.1 May 17, 1999, 7.1.7) says:
	"If no selected revision is a descendent of all the
	other selected revisions, then a conflict has occurred,
	and the DAV:rsr-merge selects the first revision
	selected by an element in the sequence."

   Do I choose ANY element in the sequence and pick the first revision 
   selected?
   If so this seems somewhat arbitrary.

   I'd appreciate any insight to the intention.

   Thanks
   Tim

   p.s.  I'm assuming that the trivial case of the merge rule constituents 
   selecting only a single revision is a successful selection of that revision, 
   without conflict.