Unicode in main body can cause duplicates when overlapping archiver with imports

When you have emails that have been archiver through the archiver.py process and then later re-import with import-mbox.py, emails that have unicode in the main body may cause the MID (message ID) to be different due to different SHA digests. This should be fixed by detecting such issues and converting to ASCII in a uniform and consistent way.

Currently, it _does_ attempt that, but the result is not the same...

Compare Unicode handling in archiver.py and import-mbox.py, starting by reproducing an archive-then-reimport workflow with a Unicode message body. Trace how each process derives the MID and SHA digest; done means the same message receives the same MID and digest in both paths.

Written by the indexing model from the issue text.