Conflict Resolution, another try

Hey,

I went through conflict resolution again. The resulting system is unfortunately 
not entierly trivial, but implementable IMO and not unecessarily complex.

http://wiki.kolab.org/Conflict_Resolution#Kolab_IMAP_Conflict_Resolution

The system is built around the assumption that it is impossible to avoid 
conflicts on the server (which it is with IMAP AFAIK).

For this reason:
* There is always only a single valid object instance of a single kolab 
object, which can be identified by the procedure here:
http://wiki.kolab.org/Conflict_Resolution#Identifying_the_valid_object_instance

* Writing clients must check that they didn't create a conflict, by ensuring 
their written object became a valid object according to the procedure above.
For clients to be able to identify their own written version,
http://wiki.kolab.org/Conflict_Resolution#X-Kolab-Version-UID was introduced.

While X-Kolab-Version-UID is currently stored as a MIME-Header, I think it 
should rather go to the XML object, as noted in the wiki.

I think the primary question for now is if you approve of the need of X-Kolab-
Version-UID, in which case I would make it part of KEP 17 and generate it when 
writing an object using libkolabxml.

I would then probalby also start a proof of concept implementation of the 
identification process.

And please tell me if parts are not understandable in the wiki, I realize it's 
sometimes complex and found it hard to write that stuff down in a 
comprehensible way.

Cheers,
Christian

_______________________________________________
Kolab-devel mailing list
[email protected]
https://www.intevation.de/mailman/listinfo/kolab-devel

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)

iEYEABECAAYFAk/N+iEACgkQF4TVOOpldADqagCdHVcJYijkFRVRaxSqg4Bwn8q5
z0QAoJpNE11c83oG0LxPlViRS4g2Z4qB
=bX7T
-----END PGP SIGNATURE-----