`read_mo` does not decode `msgctxt`

[tomasr8]

When reading an MO file, the message context is kept as bytes unlike the msgid and msgstr which are decoded to utf8. This is surprising because in order to use for example Catalog.get, one must pass the context encoded as bytes. I wonder if this is intentional or just an omission?

Example: When I read an MO file compiled from this PO file:

msgctxt "fooctxt"
msgid "foo"
msgstr "bar"

The message key is ('foo', b'fooctxt') (notice the that fooctxt is a bytes object):

>>> f = open('messages.mo', 'rb')
>>> cat = read_mo(f)
>>> cat._messages
OrderedDict({('foo', b'fooctxt'): <Message 'foo' (flags: [])>})