Mon, 23 Aug 2010 21:15:55 +0000, Pauli Virtanen wrote: [clip] > in the history to have the wrong content -- so to be sure, we have to do > a brute-force comparison of the tree against SVN for each commit. The > particular bug here was fixed in the conversion tool, but better safe > than sorry.
Also this is now done. The repository at http://github.com/numpy/numpy_svn_test4 is essentially fine. You can grab it via git clone git://github.com/numpy/numpy_svn_test4.git The deleted SVN branches can also be obtained: git fetch origin refs/svn/*:refs/remotes/svn/* They are not pulled by default. *** svn-all-fast-export still has some bugs, so revisions that in SVN are r2478, r2491, r2534, r2602, r2603, r4113, r4179, r4180 contain the wrong tree in Git. The errors are - Wrong branch point for branches/arraykit (an abandoned and deleted branch that was never merged back to trunk) - The commits r4113, r4179, r4180 contain changes in wrong order (the subsequent commits are OK). I'd be willing just to ignore these. Such small errors should not matter. *** So as far as I see, we would be ready to go with this. Pauli _______________________________________________ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion