[cvsnt] Re: Mixing :local: and client/server protocols (was:Whenyou gotta....)

[Bo Berglund]

First of all the webserver is a development server that sits behind a 
Symantec Raptor firewall with no connection whatever to the real world. 
So storing the root there is not that much of a problem.

But nevertheless I use a specially created account for this purpose that 
is only allowed read access to the CVS repository.

I will look to see if I can get it to work also without any user:password 
entries. This would of course be better. OTOH it is really not a problem
for me, I just supplied the info on how I did it to help the original
poster....

/Bo

-----Original Message-----
From: cvsnt-bounces at cvsnt.org [mailto:cvsnt-bounces at cvsnt.org]On Behalf
Of Oliver Giesen
Sent: den 15 juni 2004 11:21
To: cvsnt at cvsnt.org
Subject: Re: [cvsnt] Re: Mixing :local: and client/server protocols
(was:Whenyou gotta....)


Bo Berglund wrote:

> I didn't realize that the initial checkout actually store the user AND
> password in the CVS/Root files.  I thought that the password was stored
 > in a location in the Registry

That's only if you use the Login command. The CVSROOT you specify on 
checkout is always stored literally in ./CVS/Root exactly as you 
specified it. However, using Login is no alternative in your case if you 
want to run the command as a different user than the one doing the commit...


 > I thought that the password was stored
> in a location in the Registry and since the checkout was done by a real
> user to begin with I had to set this -d parameter to get it to work when
> the update is launched from the postcommit script, since that would run
> in a different context and thus have access to different User hive in
> the registry.
> 
> But I just tested and it really stores both user and password so the only
> thing that needs to be done is initially to check out using the correct
> user/password combination to override the Windows login.

Do you really need this to be run from a different user account? Why not 
simply let it run under the account that did the commit? That way you 
could eliminate username and password info from the CVSROOT completely. 
I personally wouldn't feel too good knowing a valid username/password 
combination to my CVS server was stored in plain-text on a publicly 
accessible web server...

Cheers,

-- 
Oliver
----	------------------
JID:	ogiesen at jabber.org
ICQ:	18777742	(http://wwp.icq.com/18777742)
_______________________________________________
cvsnt mailing list
cvsnt at cvsnt.org
http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt