I'm afraid so.  The python3 uptake process was expected to take five years overall, and we are only up to about the second year at this point.  So while you may have been away from Python for 6 years, you came back right in the middle of an unprecedented transition period.

I agree that it is unfortunate that a shipping library is not functioning correctly with respect to the Python3 bytes/string separation, but no one had tried to use mailbox in python3 enough to have encountered this problem.  You will note that that this bug report was a *performance* bug report initially, and as such had lower priority.  The encoding issue was recognized much more recently.  And before that could be fixed correctly, the email package had to be fixed to handle bytes input.  That happened only just before the end of the beta phase for 3.2.  Now it is too late to make further API changes for 3.3, and in any case it seems counter-productive to make an API change that we don't really want in the library long term.

You could work up a patch to fix this, use it locally, and contribute it so that it makes it in to 3.3.  Perhaps if you and/or someone else can come up with a patch before RC2 it could even go in to 3.2.  I haven't looked at it (yet), but I'm hoping that the patch isn't actually that hard to fix the encoding issues (as opposed to the performance issue, which may take more work).

Or perhaps you could monkey-patch in your encoding fix until 3.3 comes out.  Of course, right now using 2.7 with an eye to staying compatible with python3 is also a perfectly sensible option.