It sounds like the problem is that as dumped revisions are inserted into the
repository, they receive new revision numbers. (Unless this is an
incremental dump being used to create an identical copy of the original
repository.) And like the load process doesn't maintain the mapping between
the dumped revision numbers and the loaded revision numbers. It doesn't
sound like that would be hard to do in principle, though it might be messy
in practice.
Dale
-----Original Message-----
From: Walden Mathews [mailto:walden@eqwality.com]
| This would seem to have the advantage that renames and moves in the
| source repository would be automatically reflected correctly, and all
| the intermediate states would be visible in the internal repository.
You may have a problem with operations that are based on
svn copy, because copies are from objects at a particular revision,
and that "copy from" revision is encoded in the dump. The
repository you are playing this into will almost certainly have
its own revision history, and you will get errors to the effect
that "object does not exist in that revision".
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Thu Nov 4 20:38:54 2004