Copyright (C) 2001,2002 Python Software Foundation

Here's a brief to do list.

- The header parsing algorithm should be exposed in the API.  This is
  because other programs might want to search for headers without
  sucking them into Message objects.

- Parser should have a non-strict mode, where bogus header lines are
  ignored (or captured for debugging) without raising an exception.
  See tests/test_nonconformant.py for a message that has
  non-conformant MIME boundary separators.  This currently raises a
  BoundaryError, which is appropriate for strict parsing.  The
  non-strict mode would parse this into the intended message tree.

- There should be a "SmartParser" class that can guess about common
  misformatted messages (i.e. missing a MIME end boundary).

- We should support RFC 2231.  Oleg Broytmann has patches to add this
  support, so we'll look into it for email-1.3.

- Parser should create the most concrete class instance that it can
  (this may require some sort of factory registry).

- There should be some way to reduce the memory footprint by
  optionally storing message parts (or the whole message a la
  multifile) on disk.



Local Variables:
mode: indented-text
indent-tabs-mode: nil
End:
