On Thu, Jan 16, 2003 at 12:28:31AM +0100, Branko Cibej wrote:
> Brandon Ehle wrote:
> >ra_svn is working great for us even on large repositories. There's
> >couple of bugs here & there but on the whole its far more stable than
> >ra_dav and faster too.
>
> This is very bad news. To hear that ra_dav, which is more than two years
> old, behaves worse than ra_svn, a relative newcomer --- well, it makes
> one wonder if we didn't make the wrong decision about our initial remote
> protocol after all. ra_dav has taken up a lot of our time, an it looks
> like we still haven't even started working on the most basic stability
> problem.
Wrong decision? Nah. The ability to use WebDAV against the repository is
still just as cool as two years ago. Even cooler with the autoversioning
stuff.
Problems? Sure. Our code? Apache? Unknown.
ra_svn had a *ton* of prior experience to work with, and a lot of code to
leverage. Did it have to revamp its editor usage a few times? Nope. Change
the reporter interface? Nope. Heck... deal with the introduction of the
reporter interface? Nah. etc etc. Of course, it also had the benefit of a
single purpose protocol to radically simplify things. No question there, but
do I have a Python or Perl library that can talk to svnserve? Nope. To
mod_dav_svn? Hell ya. And a C library, and a Ruby library, and ...
Each has their pros and cons. But you really can't compare them simply based
on their lifetimes. You can certainly compare them on complexity, and I can
definitely agree that it causes us pain. Worth it? I believe so, yes.
Cheers,
-g
--
Greg Stein, http://www.lyra.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Jan 16 01:29:38 2003