On Wed, Apr 15, 2009 at 7:27 PM, Tiziano Zito <opossumnano@gmail.com> wrote:
On Wed 15 Apr, 18:25, Charles R Harris wrote:
> > > The Git repository indicates that something like this occurred:
> > >
> > >       5655 5656 5657 5658 5659 5660
> > >        o----o----o----o----o----o----o 5661
> > >         \___________________________/
> > >
> > I think that's my fault, but I haven't figured out how to revert this.
> >
> > Can you undo whatever happened in
> > http://projects.scipy.org/scipy/changeset/5661, and I won't try this
> > anymore if it messes up the repository and not just the commit
> > message.
>
> I found reverting the central svn repository to be a total PITA. If someone
> still has the correct versions of the files the easiest thing to do is just
> commit them.

I had just checked out revision 5660, so I could commit it back again if
it's needed, but one would probably need to change all relevant
files in order to convince svn that they are new, and then manually
solve all conflicts by accepting my own version. Let me know, I
could do it around 6:30 AM UTC on Apr 16!

Rename the relevent good files,  check out the latest svn revision, replace the files with your good versions and commit. If someone knows an easier way please let me know. Fixing svn after a corrupted commit almost drove me crazy when I had to do it.

Chuck