Mailing List Archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [tlug] SVN: Your Environment, BerkleyDB/Web or FSFS?



>>>>> "Ian" == Ian Wells <ijw@example.com> writes:

    >> I run SVN with the regular svnserve, not as a WebDAV module. As
    >> I mentioned in my presentation, this is mainly because WebDAV
    >> was a bear to setup and I did not have the time to get it
    >> working. I imagine that WebDAV is the way to go, since it gives
    >> you all the super-fly Apache access controls for free.

    Ian> Is there any other reason to do this?

Future-proofing.  WebDAV is going to happen, and many repositories
will want to take advantage of the network effects of standardization.
That may not apply to yours or Josh's usage, of course.

    Ian> I've never really worked out what advantages you get with
    Ian> webdav - but, to be honest, I've never really worked out why
    Ian> the svn developers would be interested in supporting 2
    Ian> protocols, either.

They need something simple and reliable for basic usage, that's
svnserve.  If you expand the acronym "DAV", you'll see why subversion
has to be interested in WebDAV.  WebDAV promises integration with all
kinds of things.  Zope speaks WebDAV, for example ... how about a wiki
backed by subversion?  Emacs has some WebDAV functionality (it's
idiosyncratic, low-level, and definitely not integrated into the UI
the way TRAMP/EFS/ange-ftp are, but it's there).  The real competition
to subversion (GNU Arch, Darcs, Monotone) all speak WebDAV or have
WebDAV planned.

Of course they'd prefer to only support one, and that would need to be
WebDAV.

Unfortunately, OSS WebDAV is a crock at the moment.  It took me all of
four hours to get libcurl's "easy" API integrated into XEmacs, and the
only change I made to the API was to omit the initialization and
finalization functions (they're taken care of by the Lisp allocation
and garbage collection machinery).  Most of that time was spent
understanding the libcurl API and the XEmacs FFI---the actual coding
took something like 90 minutes, and my first attempt to fetch a web
object _worked_ ... recovering from _that_ surprise took about 15
minutes. ;-)  That's great ... but it's not WebDAV.

libneon is (a) a much more crockish API, (b) very low-level from the
standpoint of an editor application, and (c) unusable by itself.  You
have to add libxml or expat support, and even then you don't have
anything usable, because the SAX parsers don't know anything about
WebDAV semantics.  Totally unlike the libcurl "easy" API.  To get to
that level, you need to use cadaver, but cadaver's UI is gratuitously
different from an FTP client or smbclient, so you have to write expect
scripts from scratch.  There is no libcadaver.

Of course, I could easily have missed more appropriate library
support; I'd *love* to hear about it!  But I'm not a total newbie, if
a bit of googling didn't show anything recognizable to me, there are
probably lots of other wannabe webdavvers in the same boat.

So I think WebDAV is still a couple of years away from full
integration into the open source community.

****************************************************************

For those of you who are interested in being lectured, I've finally
got around to publishing a blog (in fact, several).  I think I've got
permissions for anonymous users to read them enabled.  Let me know if
I think wrong  :-) The one of immediate relevance is

http://turnbull.sk.tsukuba.ac.jp/Zope/Software/WebDAVSupport

There's also potential interest in

http://turnbull.sk.tsukuba.ac.jp/Zope/OSSEcology

and for the totally masochistic and lacking in anything resembling a
life

http://turnbull.sk.tsukuba.ac.jp/Zope/Research

There's not much there yet and the organization isn't great.  Quantity
will change rapidly over the next few weeks, I think.  We'll see about
quality.  Oh yeah, although the engine is Zwiki, editing (including
comments) by Anonymous Cowards is not yet enabled (and may never be,
depending on whether I find time to do a security audit).  Feel free
to send comments and flames directly to me, though.

Oh, one more thing ... these are URLs, not URNs.  They may disappear
at any time, to reappear elsewhere in my webspace.

-- 
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.


Home | Main Index | Thread Index

Home Page Mailing List Linux and Japan TLUG Members Links