Fil wrote:
> Hi,
>
> I've just been asking for help on the irc #svn channel, and Erik Huelsmann
> told me to fwd my thoughts to this list as they can be relevant to the
> "takeover" command under development.
>
> Basically, we've developed our project (www.spip.net) with CVS, and are
> switching to SVN.
>
> The issue is that it's a web application, and many users have installed it
> with a simple command : cvs co, directly in their web directory. Then some
> have patched the script, but more importantly the user files (added
> documents, modules, configuration files) are mixed within the same
> directory
> structure. So it's a bit tough to say "remove the CVS working copy and
> install a fresh SVN working copy".
>
> So maybe the "takeover" command would make sense here.
>
> Of course I can't wait for it to be developed and released, so my problem
> stays, but I'll find a dirty way to do it.
>
> My guess for the moment is that I should have an exact point where all
> files
> are exactly the same in the CVS and in the SVN repo, then checkout the SVN
> repo at that point, remove all files (except the .svn/ directories), and
> redistribute these .svn/ directories in the existing (CVS) working copy.
> If
> you know any script that does this kind of ugly stuff, pls let me know.
cd a-new-svn-checkout
find . -name .svn -print0 | tar -T - --null -cv | tar -C
../my-existing-cvs-working-copy -x
Max.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Aug 21 12:26:09 2005