Re: Alternatives for remote access?

David Summers wrote:
> Maybe I'm missing something but I don't see why setting up an apache
> server should take more than about 10-15 minutes if it is packaged
> correctly (like my apache RPMs) (I do it very
> frequently especially when testing multiple versions of Subversion).
>
> With SSH port-forwarding, how is CVS via ssh port-forwarding significantly
> different from Subversion via ssh port-forwarding? It seems to me that
> they both do the same thing.

CVS+ssh doesn't use port forwarding. IIRC it runs something like "ssh
user@host cvs server". In that case, cvs has no open port at all; it's
just an executable. So there's no setup at all to go from local access
to remote access, assuming ssh is already running. It definitely is
easier. (Plus cvs has an unfair advantage - it's already installed on
most people's machines. No need to track down dependencies, etc.)

On the other hand, this has always struck me as a kludge:

- the users must have a local Unix uid with write access to the
repository. Although as someone on slashdot pointed out to me, you can
limit that account to run only the command "cvs server" (or whatever the
subcommand is.)

- it doesn't play well with non-Unix stuff. I believe there's no pure
Java CVS library that can do this without calling an external ssh
executable.

If this method existed for Subversion (ra_pipe is supposed to be that,
right?), probably a fair number of people would use it. I wouldn't be
one of them; I'm happier with ra_dav.

-- 
Scott Lamb
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org