Modify w3c import script to be able to run over a whole repo.
This change allows us to run the import script over a whole repo, rather
than just the 'submitted' and 'accepted' sub directories. This is another
step towards being able to import and run the repos wholesale, rather than
Most of the changes in this patch involve changes to handle malformed input files, log them, and continue. We need to continue to make things more robust here, and should start to add tests for this, but I'll defer this for now.
Given that ultimately my goal is to make the import processes simpler rather than more complex, I want to defer writing tests for stuff I'm trying to delete. Once I get closer to have the script actually in the form we'll want long-term, and we'll want to support this properly, I'll make sure we have tests (right now, I'm the only one playing around with this).