RE: Copying DAV:comment on CHECKIN and VERSION-CONTROL

Requiring that the value of DAV:comment of a checked-out
resource be copied to the version created by CHECKIN sounds
right to me (and I'll bet it's what all the implementors
end up doing).  Note that some systems allow you to modify
the comment on a version, so we probably don't want to require
that it is immutable.  Note that CHECKOUT does not change the
state of a VCR, so a server would never copy anything at that
time.  Probably what you want here is for the DAV:comment
property to be *removed* as a side-effect of the CHECKOUT
operation.

Just goes to show that it is a good thing we didn't try to
define general behavior of live properties under versioning!

Cheers,
Geoff

-----Original Message-----
From: Roy Seto [mailto:Roy.Seto@oracle.com]
Sent: Friday, November 02, 2001 6:51 PM
To: ietf-dav-versioning@w3.org
Subject: Re: Copying DAV:comment on CHECKIN and VERSION-CONTROL


That sounds like a good process to me.

On DAV:comment, I think the common use case is for authors to use a
new comment most times they CHECKIN a new version, to document the
changes specific to that version.

To support this common case, I think the property list should say
that servers MUST copy DAV:comment from the checked-out resource to
the new version on CHECKIN. This allows clients to consolidate the
PROPPATCH of DAV:comment with the PROPPATCH of any dead properties
of the VCR (and other copied live properties) before CHECKIN, and
makes the behavior atomic. I also think the property list should
say that servers SHOULD NOT copy DAV:comment from the predecessor
version to the VCR on CHECKOUT, so that clients do not need to do a
PROPPATCH to guarantee that DAV:comment was not incorrectly carried
forward from the previous version.

Thanks,
Roy

"Clemm, Geoff" wrote:

> Currently, the server gets to decide which live properties
> of a VCR are captured by versions.
>
> It would be reasonable to identify a set of live
> properties that SHOULD/MUST be captured by versions,
> and publish this list in an internet draft and on
> the DeltaV web site.  This list would then be added
> to the protocol document when we go to the next standard level.
>
> Cheers,
> Geoff
>
> -----Original Message-----
> From: Roy Seto [mailto:Roy.Seto@oracle.com]
> Sent: Thursday, November 01, 2001 8:50 PM
> To: ietf-dav-versioning@w3.org
> Subject: Copying DAV:comment on CHECKIN and VERSION-CONTROL
>
> I don't see anything in the spec that requires DAV:comment
> to be copied from checked-out resources to versions on
> CHECKIN, or from unversioned resources to the initial
> version on VERSION-CONTROL.
>
> In particular, since DAV:comment is a live property, it
> doesn't get copied with the dead properties, and it is not
> mentioned explicitly in postcondition
> DAV:initialize-version-content-and-properties of Section 4.4
> or postcondition DAV:put-under-version-conrtrol of Section
> 3.5.
>
> Is this the intended behavior?
>
> Thanks,
> Roy

Received on Friday, 2 November 2001 22:52:30 UTC