[darcs-users] Re: Darcs cgi-push, preliminary implementation

[Stephen J. Turnbull]

>>>>> "Juliusz" == Juliusz Chroboczek <Juliusz.Chroboczek at pps.jussieu.fr> writes:

    >> Out of curiosity, why not just use DAV

    Juliusz> Warning: lengthy anti-WebDAV rant follows.  Please
    Juliusz> discard this message straight away, and read no further.

*chuckle*

    Juliusz> There's no such thing as WebDAV. 

That's a bit strong.  There is a WebDAV (admittedly it's defined over
several RFCs, but then so are most useful Internet protocols), and
whether it will stand up to practical use is a good question; that's
why they call them "RFCs".  Odds ain't great---compare the RFC count
(currently around 4000, I guess) vs the STD count (maybe 150?).

    Juliusz> DeltaV (RFC 3253) adds a form of versioning to WebDA.  If
    Juliusz> memory serves, the model it uses for versioning is
    Juliusz> sequences of trees with explicit locking, but not much
    Juliusz> beyond that; not even delta computation can be done
    Juliusz> server-side.  I do not see how DeltaV can be useful in
    Juliusz> practice.

DeltaV is already good enough to support wikis and Zope and the like.
I don't see why it wouldn't allow you to use an httpd as the backing
store for Darcs instead of a file system, with somewhat better
features for that use than raw HTTP 1.1.  Heck, you admit as much
yourself, by saying you could use it for locking.

    >> like Subversion does?

    Juliusz> AFAIK, SVN uses a fragment of WebDA+V with a bunch of
    Juliusz> non-standard extensions, which they implement using
    Juliusz> non-standard WebDA properties.

Right.  And if they work, they can be written up in an RFC.

I don't disagree with any of the factual statements in Juliusz's
rant.  I just take the position that it means that a WebDAV
implementation of a Darcs server will require substantial extension of
the WebDAV protocol, and a lot more work than the "why not use WebDAV
like Subversion does?" way of asking the question suggests.

And, of course, it runs the risk that Juliusz is right in the sense
that WebDAV can't/won't be extended in ways that serve Darcs's needs.
But we don't know that yet.

If somebody wants to look at WebDAV, I think it would be a great
service to the SCM community.  You should try to use the SVN
extensions to support Darcs.  You should also look at GNU Arch's use
of WebDAV, AFAIK Arch uses only the WebDAV locking extensions to the
basic HTTP 1.1 protocol.  If the semantics of those extensions are
fine for Darcs as far as they go, but need further extension, that's
good---progress toward that RFC.  If not, it's back to the drawing board.

I'm moving in the direction of checking out these issues myself, but I
need to build a flexible, programmable client first (cadaver doesn't cut it).

-- 
School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.