Character Encoding on EmailMessage when sending

[Helmut]

I'm not 100% sure if this is bug, or whether my understand of unicode is wrong. Nevertheless Django 1.11 behaves differently than 1.10.7. Using Django 1.10.7, if I send an Email using EmailMessage, but I encode the subject and body with UTF-8, it sends successfully. For example:

email = EmailMessage("asd’asd".encode("utf-8"), "asd’asd".encode("utf-8"), "blah@foo.com", ["yup@yadda.com"])
email.send()

But if I run the same code in Django 1.11, it throws an exception because of the unicode character that was used:
'ascii' codec can't decode byte 0xe2 in position 92: ordinal not in range(128)

So, might there be a bug in 1.11, or should I not be using the ".encode"? (Was using that because of weird user input)

And furthermore, why does pre-1.11 not throw an exception?

I thank you for you attention.

[Tim Graham]

To help answer your question, you could ​bisect to find the commit where the behavior changed.

Generally, you should use our support channels to ask "is it a bug?" questions rather than seeking help through the ticket tracker.

[Claude Paroz]

There's absolutely no reason to encode the content passed to EmailMessage, so even if it was not failing before, I would be against restoring the previous behavior.
Nonetheless the EmailMessage __init__ docstring talks about accepting UTF-8 bytestrings, which should be removed IMO.
Unless someone comes with a compelling use case to allow bytestrings for EmailMessage subject and body. Binary content should be left for attachments.

[Claude Paroz <claude@…>]

In d4d812c:

Refs #28196 -- Removed mentions of bytestrings for EmailMessage

With Python 3, there are no more reasons to special-case EmailMessage
arguments which should be plain strings.