Subject: Re: CVS anon stuff
To: Grey Wolf <>
From: Jeff Rizzo <>
List: current-users
Date: 03/29/2000 14:50:39
On Wed, Mar 29, 2000 at 01:47:12PM -0800, Grey Wolf wrote:
> I can't seem to do this right.  I have
> CVS_RSH=ssh
> and I can log in via cvs login, but when I go to do a 'cvs co syssrc/sys'
> as requested, it insists on creating a syssrc dir under /usr/src.

Yes, because you're checking something new out... if you're updating an
existing tree, you want to use "cvs update" instead.  (See below)

> So I have some questions:
> 1.  How do I update /usr/src directly using CVS?

Assuming you've extracted from tarfiles (or used sup), you need to make
sure the files named "Root" in the CVS directories have the correct 
information. (By default, the have incorrect info for anon CVS)

To put the correct repository info into the /usr/src tree, do this:

% echo >/tmp/Root
% cd /usr/src
% find . -name Root -exec cp /tmp/Root {} \;
% cvs -q update -d -P

The "find" will correct the tree's idea of CVSROOT, and the update will
update the files.  (the -d -P means to create new directories as needed,
and to prune empty ones.)

> 2.  Do I *really* have to have a repository in addition to my /usr/src
>     tree?  (It's lame if this is the case.)

Not sure if I follow, but I don't think so.  One thing you may need to
do (if you don't already) is to build in obj directories, otherwise
you may need to do a "make cleandir" before every "cvs update".

> 3.  How do I retrieve the equivalents of ksrc-common and ksrc-i386 without
>     retrieving the rest of the architectures under /sys?  Is there an easy
>     way to do this?

Not that I know of.  :-{

> 				--*greywolf;

Hope this helps.


Jeff Rizzo