On 12/01/2014 08:45 AM, Aurelien Bompard wrote:

I'm really interested in any insight on this issue. Thanks for reading all that :-)

I just took a quick look, but I think the problem is likely that mailman.email.message.Message overrides email.message.Message.__getitem__() as follows:

def __getitem__(self, key):
    # Ensure that header values are unicodes.
    value = email.message.Message.__getitem__(self, key)
    if isinstance(value, str):
        return unicode(value, 'ascii')
    return value

If we trace back a bit in the code, email.message.get_param(...) does

    for k, v in self._get_params_preserve(failobj, header):

and _get_params_preserve(...) does

    value = self.get(header, missing)

which because of our override decodes the entire "attachment; filename*=UTF-8''d%C3%A9jeuner.txt" string as ascii into u"attachment; filename*=UTF-8''d%C3%A9jeuner.txt" which is then parsed into the tuple (u'UTF-8', u'', u'd\xc3\xa9jeuner.txt').

It seems to me that any way to fix this is a horrible kludge. In particular, the value portion of the tuple has already been garbled, and I don't see a good way to fix it once it's been returned.

I've looked, again briefly because after a short time my head spun out of control, at trying to fix it by overriding additional methods like _get_params_preserve(), but I didn't get far.

-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan