Ok, I'm solving a little problem that is specific to the way that I
am mis-treating CVS ! :)
I have two web developers that share so much of what they do, that
they only have one working copy between them rather than one each.
They co-ordinate between themselves and this also saves me setting
up a multitude of webservers for each working copy of the web site.
So I remove the Root files because we use a remote ssh connection and
the CVSROOT stores the username for the repository machine.. which is
different for the two different people....
Of course you will think that I shouldn't be having two developers in
one working copy of the source and I agree... but this works for now
and so I am sharing what I have learnt with other CVS people....
Also I needed to know about this, for when I decided to change the
directory name for the repository... I didn't think it would matter
to the existing working copy that I had moved it, because I did the
right thing by changing my CVSROOT environment variable - of course
some how it remembered the old path... despite my attempts to make it
forget.
I understand the convenience you have provided the CVS user by implementing
the "Root" file memory... my only criticism is that this behaviour is
not mentioned in the documentation anywhere.
cheers
jb.
On Wed, May 31, 2000 at 01:36:50PM -0400, Larry Jones wrote:
> Jarrod Douglas writes:
> >
> > CVS stores the location of the repository used in a "get" or "update"
> > operation in the CVS/Root file of the working directories.
> > When you then do subsequent cvs operations, it uses the contents
> > of this file in preference to the CVSROOT envrionment variable or
> > what is specified in the dialogs of WinCVS (when using that front end).
>
> That is exactly right -- CVS remembers where the stuff came from and
> does all subsequent operations to the same repository. This is intended
> to be a user convenience -- if you're working with multiple
> repositories, you don't have to constantly be resetting $CVSROOT or
> using -d to get things to work right, it all happens automagically.
>
> > The solution I am using, is to delete all the "Root" files via a
> > simple Unix find command and then set the CVS_IGNORE_REMOTE_ROOT
> > enviromnent variable for the shell of all subsequent cvs operations...
> > This prevents cvs from writing out a new value into the Root file
> > again... and the absence of the Root file, forces CVS to use your
> > envronment variable or settings you have set.
>
> What problem are you trying to solve?!? It seems to me that you're just
> making your life more difficult...
>
> -Larry Jones
>
> Fortunately, that was our plan from the start. -- Calvin
