EmailMessage.message().as_string() incorrectly base64 encoded on Python 3

[aRkadeFR]

I have a different behavior on body encoding on python 2.7.12 and python 3.5.2 with Django 1.8.15:

# -*- coding: utf-8 -*-
from __future__ import unicode_literals, absolute_import

from django.conf import settings
settings.configure(DATABASES={}, SECRET_KEY='not empty')

from django.core.mail import EmailMessage


email = EmailMessage('Chère chère', 'Super \' fort très fort.',
                     'from@example.com', ['to@example.com'])
print(email.message().as_string())

On python 3, the body will be encoded as base64, but not on python 2.

I've found couple of tickets relating to the same subject (#3472, and #11212 among others), and I've found in the Django code 1.8.15 in the file django/core/mail/message.py:

# Don't BASE64-encode UTF-8 messages so that we avoid unwanted attention from
# some spam filters.
utf8_charset = Charset.Charset('utf-8')
utf8_charset.body_encoding = None  # Python defaults to BASE64

I would think from the code and the tickets that Django tries to never base64 the body of EmailMessage.

Is it the desired behavior?

[Claude Paroz]

Looks like a legitimate bug (I reproduced on master). The charset dance we're doing in SafeMIMEText.__init__ isn't sufficient for all cases, apparently.

[Claude Paroz]

​Pull request

[Claude Paroz <claude@…>]

In 458e2fbf:

Fixed #27333 -- Prevented BASE64 encoding in message.as_string() on Python 3

Thanks Tim Graham for the review.