info-cvs
[Top][All Lists]
Advanced

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

Re: cvs on unix simple security issue


From: Mark D. Baushke
Subject: Re: cvs on unix simple security issue
Date: Wed, 08 Mar 2006 02:31:50 -0800

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

address@hidden writes:

> Arthur Barrett wrote:
> 
> > Don't allow interactive login to the unix box runnin the CVS server?
> >
> > CVSNT also allows you to run the service as a specific user (so there is
> > no context(?) switch - then all the files are "owned" by that user -
> > preventing other users from rm'ing it, there is probably some option to
> > do that on traditional CVS).
> 
> Someone inside suggested WINCVS, I've not looked at it yet. I'm curious
> about how authentication works. Are you saying we would not be able to
> tell who did the last commit on an object by looking at history? Big
> deal to configure and use pserver?

WINCVS is just a client GUI for CVSNT. There are a number of different
authentication methods you can use with CVSNT. However, this list does
not really cover CVSNT or WINCVS.

The :pserver: and :ext: methods are common between CVS and CVSNT.

You can google (or yahoo! search or whatevery your favorite search
engine might be) for the various artiles in site:lists.gnu.org where the
relative merits of :pserver: are discussed.

I am personally not in favor of it, but even I would not go so far as to
say that it would not be able to tell you who said they did the commit
on an object by looking at the log of commits. 

CVS (and the fork that has become CVSNT) was developed back when the
world was a more friendly and collaborative environment and you could
trust people to be honest about changes to the source code being
controlled by an SCM. If your committers are mostly honest, then you are
likely to get mostly honest results regardless of which authentication
mechanism you use. However, if your network has any miscreants (or their
botnet agents) on it, then you may wish to think about using a bit
stronger authentication and you might want to let the vendor worry about
getting that part of the system implemented correctly. I do not trust
that :pserver: is free of bugs and it may even be possible to forge or
spoof changes from another users using it.

        Good luck,
        -- Mark
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQFEDrKWCg7APGsDnFERAoQwAJ92KC3Jne4D3pXeR2EYkIumCWd7LACfYujs
4tLVGZzFfisiAZ3maW2H66s=
=yJS7
-----END PGP SIGNATURE-----




reply via email to

[Prev in Thread] Current Thread [Next in Thread]