Stephan Michels wrote:
Am Do, den 08.04.2004 schrieb Guido Casper um 20:18:

Stephan Michels wrote:

What the status of the WebDAVSource? It seems that it doesn't implement
any versioning. Or am I wrong?

No. I (vaguely) remember some discussion considering your VersionableSource vs. Sylvain's VersionedSource from:
http://www.cocoondev.org/projects/cvssource.html


However ...


Do you have plan to do something in this direction?

I like the fact that WebDAV makes versions available via seperate URIs. So there is no need (and no value IMO) to have things like get/setVersion() since there is no way to anticipate the versioning scheme (i.e. what's the name of the previous version?). All you need is a way to get at an ordered list of versions which might as well be complete version URIs.


Do you have an example to get older versions via URI?

The DeltaV spec has a REPORT method which allows for a version-tree report listing the URIs of all version resources.
http://www.webdav.org/deltav/protocol/rfc3253.html#rfc.section.3.7


Slide's WebdavResource has a report() method accounting for that.



I wonder if it's a good idea to expand this concept to the Repository interface so any implementation can make up its own or find a common URI scheme that includes versioning/branching/etc. information. It is quite easy for an application to manage things like "versioning state" on its own and the Repository stays more generic.


Okay, what do you use as client? The WebDAV client of Slide? Is it good
enough?

Well, Slide's "old" WebDAV client seems to be the most complete right now.


Slide's CVS contains a JSR147 implementation as well. JSR147 strangely is geared towards DeltaV-only. Slide's implementation therefore has some proprietary extensions (I never understood why there is no JSR for a general WebDAV API). AFAIU these extension has been proposed to the expert group (the creator of slide's implementation is a member thereof).

TBH I don't know what your best bet right now is. But Slide's "old" WebDAV client is quite stable, already part of Cocoon's CVS and "good enough" for anything I can think of right now.

Guido



getVersions() (and maybe getBranches()) objectively has to be extended to manage branches.

WDYT about that?


Yes, I think you're right. It only feels a little odd to bypass the
Source IF.

Stephan Michels.




--
Guido Casper
-------------------------------------------------
S&N AG, Competence Center Open Source
                    Tel.: +49-5251-1581-87
Klingenderstr. 5    mailto:[EMAIL PROTECTED]
D-33100 Paderborn   http://www.s-und-n.de
-------------------------------------------------



Reply via email to