You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by "Stuckenschneider, Steve (Mission Systems)" <St...@ngc.com> on 2005/02/15 21:25:38 UTC
running SVN clients in VMS
Hello,
I am part of a software support group that works primarily in VMS, although some
of our software is on Linux platform. We also use a special software environment
called X-Midas. For our CM, we currently use CVS over the web; our repository is
currently at another location; it's basically CVS over our own special version of
SourceForge.
We are considering switching to an in-house repository using subversion. The repository
and server would most likely reside on a Linux machine, but we must be able to
access from VMS, probably using netscape or mozilla web browser (similar to what we
do now to access our CVS repository).
Is it possible to do this?
Thanks in advance,
Steve
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: running SVN clients in VMS
Posted by Ben Collins-Sussman <su...@collab.net>.
On Feb 15, 2005, at 3:25 PM, Stuckenschneider, Steve ((Mission
Systems)) wrote:
>
>
> Hello,
>
> I am part of a software support group that works primarily in VMS,
> although some
> of our software is on Linux platform. We also use a special software
> environment
> called X-Midas. For our CM, we currently use CVS over the web; our
> repository is
> currently at another location; it's basically CVS over our own special
> version of
> SourceForge.
>
> We are considering switching to an in-house repository using
> subversion. The repository
> and server would most likely reside on a Linux machine, but we must be
> able to
> access from VMS, probably using netscape or mozilla web browser
> (similar to what we
> do now to access our CVS repository).
>
> Is it possible to do this?
>
When you say, "access via web browser", do you simply mean "browse the
repository through a VMS web browser"? If so, that's a very different
thing than running a subversion client directly on VMS. And it's easy
to do. There are number of CGI programs for making a subversion
repository web-browsable, such as viewcvs and websvn. Take a look at
the 'links' page on the subversion website to see them all.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: SVN and htp/https cookies?
Posted by Ben Collins-Sussman <su...@collab.net>.
On Feb 15, 2005, at 4:42 PM, Stephen Willey wrote:
>
> Thanks for the quick response.
>
> I understand the credential caching that svn does. What's additionally
> needed is authentication state caching. Most svn user actions require
> multiple http/https requests. Without any caching of the
> authentication
> state each request has to be authenticated separately. If the authn
> backend is a networked resource then the otherwise spiffy svn request
> turns
> into molasses.
>
> Perhaps the best solution for us is to use svn+ssh://
>
I don't understand: are you actually experiencing a problem? Or are
you just predicting one? Yes, you're right that svn does multiple http
requests per action. And I can assure you that it *doesn't* re-prompt
the user for every request. :-)
Neon (http://www.webdav.org/neon) is the library the svn client uses to
send and receive all HTTP traffic. Neon has full support for openssl
and certs, and yes, it's smart enough to remember authentication
between requests. It sends a request, gets a 401 back, which causes
the svn client to prompt for authentication. Neon then remembers and
uses that authentication for every subsequent request. When the
connection to the server is finally closed, the svn client itself
(optionally) caches the authentication info in ~/.subversion/auth/, for
the next time 'svn' is invoked.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
RE: SVN and htp/https cookies?
Posted by Stephen Willey <wi...@cac.washington.edu>.
Thanks for the quick response.
I understand the credential caching that svn does. What's additionally
needed is authentication state caching. Most svn user actions require
multiple http/https requests. Without any caching of the authentication
state each request has to be authenticated separately. If the authn
backend is a networked resource then the otherwise spiffy svn request turns
into molasses.
Perhaps the best solution for us is to use svn+ssh://
Thanks,
steve
> -----Original Message-----
> From: Ben Collins-Sussman [mailto:sussman@collab.net]
> Sent: 15 February, 2005 14:19
> To: Stephen Willey
> Cc: users@subversion.tigris.org
> Subject: Re: SVN and htp/https cookies?
>
>
> On Feb 15, 2005, at 4:06 PM, Stephen Willey wrote:
>
> >
> > To save me from doing the test myself does anyone know if
> the svn unix
> > client will handle http/https cookies in a browser-ish way?
> In other
> > words
> > can I give the initial https response a cookie and hav it returned
> > with the
> > subsequent requests from the instance of the client?
> >
> > Do to the stateless nature of http/https and the latency in network
> > based
> > authentication mechanisms it is necessary to hav
> authentication state
> > cached
> > somewhere. Tying cached state to user sessions is commonly
> done with
> > cookies.
> >
> >
>
> There are no cookies. But svn does have a generic authentication
> caching mechanism, both for normal username/passwords and SSL certs.
> Read all about it in chapter 6 of the book.
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: SVN and htp/https cookies?
Posted by Ben Collins-Sussman <su...@collab.net>.
On Feb 15, 2005, at 4:06 PM, Stephen Willey wrote:
>
> To save me from doing the test myself does anyone know if the svn unix
> client will handle http/https cookies in a browser-ish way? In other
> words
> can I give the initial https response a cookie and hav it returned
> with the
> subsequent requests from the instance of the client?
>
> Do to the stateless nature of http/https and the latency in network
> based
> authentication mechanisms it is necessary to hav authentication state
> cached
> somewhere. Tying cached state to user sessions is commonly done with
> cookies.
>
>
There are no cookies. But svn does have a generic authentication
caching mechanism, both for normal username/passwords and SSL certs.
Read all about it in chapter 6 of the book.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
SVN and htp/https cookies?
Posted by Stephen Willey <wi...@cac.washington.edu>.
To save me from doing the test myself does anyone know if the svn unix
client will handle http/https cookies in a browser-ish way? In other words
can I give the initial https response a cookie and hav it returned with the
subsequent requests from the instance of the client?
Do to the stateless nature of http/https and the latency in network based
authentication mechanisms it is necessary to hav authentication state cached
somewhere. Tying cached state to user sessions is commonly done with
cookies.
I hope this isn't answered in a FAQ, although I wouldn't be surprised if it
was.
Thanks,
steve
Stephen S. Willey email: willey@washington.edu
Software Engineer uspo: 4545 15th Ave NE
Networks and Distributed Computing Seattle, WA 98105-4527
University of Washington office: 206-616-7240
Seattle, Washington fax: 206-685-4045
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: running SVN clients in VMS
Posted by Garrett Rooney <ro...@electricjellyfish.net>.
Stuckenschneider, Steve (Mission Systems) wrote:
>
> Hello,
>
> I am part of a software support group that works primarily in VMS, although some
> of our software is on Linux platform. We also use a special software environment
> called X-Midas. For our CM, we currently use CVS over the web; our repository is
> currently at another location; it's basically CVS over our own special version of
> SourceForge.
>
> We are considering switching to an in-house repository using subversion. The repository
> and server would most likely reside on a Linux machine, but we must be able to
> access from VMS, probably using netscape or mozilla web browser (similar to what we
> do now to access our CVS repository).
I don't know of anyone who's gotten Subversion (even just the client) to
work on VMS. It's not impossible to make it happen, but I don't think
anyone has actually done it yet. The first step would be obtaining a
working version of APR for VMS, the stock APR from Apache doesn't
support VMS out of the box, but I've heard rumors of an Apache2 port to
VMS by HP, which would have to include an APR port, so that might be a
place to look.
There are a few people who have expressed interest in such a port in the
past though, myself included, so if you do give it a shot I'd love to
hear about any problems you run into.
-garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org