Skip Montanaro writes:
While I agree Apple should fix their software, shouldn't Postel's Law apply here, assuming there is a reasonable interpretation of multiple C-T-E headers?
The thing about lists and the Postel Principle is that the Principle applies to both ends of the mailing list: we're supposed to accept garbage coming in to the mailing list, but we're not supposed to send it back out!
There isn't a consistent interpretation of these headers on multiparts with value quoted-printable, which is why they're specifically excluded by RFC 2045. I suspect that the reason they don't screw things up globally for messages produced by Apple Mail is that the entire message is ASCII anyway, so they're actually no-ops as long as the MUA doesn't explicit flag them (which is what Python 3's email package does, and that's why this thread is happening).