RE: "2xx: Partial Merge" ?

From: Clemm, Geoff (gclemm@rational.com)
Date: Mon, Sep 11 2000

  • Next message: Clemm, Geoff: "RE: Versioning TeleConf Agenda, 8/11/00 (Monday) 2pm-3pm EST"

    Message-ID: <3906C56A7BD1F54593344C05BD1374B10D9E5C@SUS-MA1IT01>
    From: "Clemm, Geoff" <gclemm@rational.com>
    To: ietf-dav-versioning@w3.org
    Date: Mon, 11 Sep 2000 13:09:57 -0400
    Subject: RE: "2xx: Partial Merge" ?
    
    With Greg's suggested extension to MERGE, this all works (thanks Greg!).
    In particular, you would ask for the DAV:resourcetype and DAV:merge-set
    with your MERGE request, and that would tell you which updated resources
    were collections with pending merges.
    
    Also note that if the ignored-set is empty, there can't be any pending
    merges.  But in any case, I believe you can find out all the information you
    need
    to know in the response to the MERGE request.
    
    Cheers,
    Geoff
    
    -----Original Message-----
    From: Vasta, John [mailto:jvasta@rational.com]
    Sent: Friday, September 08, 2000 6:02 PM
    To: Clemm, Geoff; ietf-dav-versioning@w3.org
    Subject: RE: "2xx: Partial Merge" ?
    
    
    The server cannot know what to add to the ignored-set for an unmerged
    directory, because it does not yet know the members of the directory. The
    directory itself would be added to the update-set, if I am interpreting the
    postconditions correctly.
    
    But I guess a client could look at all members of the update-set, and if any
    of them are collections which did not get automatically merged, merge them
    and reinvoke MERGE on them. I don't see how the ignored-set is relevant in
    this case.
    
    John
    
    > -----Original Message-----
    > From: Geoffrey M. Clemm [mailto:geoffrey.clemm@Rational.Com]
    > Sent: Friday, September 08, 2000 5:02 PM
    > To: ietf-dav-versioning@w3.org
    > Subject: Re: "2xx: Partial Merge" ?
    > 
    > 
    > 
    > But this situation is handled by the "ignored-set" element, 
    > isn't it?  In
    > particular, a client would know to re-invoke the merge if there are
    > some "ignored" versions, and if there are any collections that require
    > merging (in case the collection merge made one of the "ignored"
    > versions visible).
    > 
    > Cheers,
    > Geoff
    > 
    >    From: "Vasta, John" <jvasta@rational.com>
    >    Date: Fri, 8 Sep 2000 16:30:07 -0400 
    > 
    >    For a recursive merge over versioned collections, the 
    > server may not be able
    >    to recurse into a collection which it cannot merge 
    > automatically (because it
    >    cannot know the members of the collection until after it 
    > is merged). So a
    >    client must be prepared to repeatedly invoke the MERGE 
    > method, and perform
    >    merges on at least collections, until it no longer gets a 
    > Partial Merge
    >    response.
    > 
    >    At least that's what I was hoping it was for!
    > 
    >    John
    > 
    >    > -----Original Message-----
    >    > From: Geoffrey M. Clemm [mailto:geoffrey.clemm@Rational.Com]
    >    > Sent: Thursday, September 07, 2000 4:21 PM
    >    > To: ietf-dav-versioning@w3.org
    >    > Subject: "2xx: Partial Merge" ?
    >    > 
    >    > 
    >    > 
    >    > I have a "2xx: Partial Merge" response status for MERGE in 
    >    > the 7.0 draft.
    >    > I'm not sure what this was for, so I'll delete it unless someone
    >    > remembers what it was supposed to mean.
    >    > 
    >    > Cheers,
    >    > Geoff
    >    > 
    >