Robert Graf-Waczenski wrote:
> The currently available migration tools create a dump file that i can
> import to subversion, potentially piped and filtered via svndumpfilter.
>
> Is there any chance to achieve the structure sketched above (with the
> "trunk" and "branches" subfolders below each component's own folder
> and with branches created as correct svn cheap copies) and still have
> the same history information that we have in VSS?
I'm not sure if you're referring to vss2svn with "currently available
migration tools", but vss2svn will correctly translate a VSS "branch"
action into a SVN cheap copy operation. So if you need to maintain
history then I would suggest importing the whole thing to SVN as is,
then moving stuff around afterwards, either manually or using a script
as mentioned by others.
However, whenever there is a situation such as this that requires such
deep restructuring of your repository, my advice is always to simply
start fresh. Export all your stuff out of VSS, arrange it how you want,
then import it to SVN as one step and start from there. Either keep your
VSS repo around as read-only for reference, or if you want to remove VSS
completely from your computers (and who doesn't?), convert that to a
separate, read-only SVN repo that you can use for lookups. After a short
while, you probably won't even need it at all anymore.
If you have more specific questions regarding vss2svn, there is a
separate mailing list; see:
<http://www.pumacode.org/projects/vss2svn>
toby
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Fri Sep 15 19:59:23 2006