Re: dump and load question

        Just one quick bit of advice: in theory, using the dump+load technique is
just fine. In practice, however it's much different. Take this for example:

        Your repository has been pretty stable for the first 3000 commits. A user
commits a .proj file by mistake, but you don't catch it for 30 revisions. In
the meantime, several users merge back & forth between different branches,
each time noting something like "merged branches/my_rewrite from r3031-3033
into the trunk". If the accidental addition of the ".proj" file happened in
r3015... well, you can see now how the off-by-one errors will start flooding
in.

        I would highly recommend just removing files through "svn rm .proj; svn ci
proj". This way the existence of the mistake remains, and no problems with
version numbers in the logs are created.

On Friday 30 December 2005 11:56 am, Shawn Matlock wrote:
> Ryan wrote:
> The point being an administrator can clean the repository of commits that
> were not supposed to happen (importing a .proj file), and users can update
> a mostly valid working copy.

-- 
Best Regards,
Dan Falconer
"Head Geek",
AvSupport, Inc. (http://www.partslogistics.com)
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org