The monthly password reminder e-mail sent by Mailman to subscribers has broken encoding in (at least) the Czech version.
The message is received (with Content-Type: text/plain; charset="utf-8" and Content-Transfer-Encoding: base64) as
VGF0byB6cHLvv712YSBqZSB6YXPvv71s77+9bmEgamVkbm91IG3vv71z77+977+9bu+/vSwgYWJ5[…] which is “Tato zpr�va je zas�l�na jednou m�s��n�, aby”[…].
AFAICT, the problem is that, in 711af11b1b87c0cc7732ad93776aa8c8a6d4089b (T261031), the configuration of Mailman was changed so that the text encoding for all languages was set to utf-8. While we all love Unicode and UTF-8, the problem is that the real contents of all the encoded text files in Mailman was not changed. And as this configured encoding is used to convert the contents of the e-mail text template file (/templates/cs/cronpass.txt), which is, in fact, still in iso-8859-2, to Unicode, it means the final e-mail text is full of Unicode replacement characters.
I guess the problem might not be limited to a single template in a single language. (And it is probably the reason why e.g. https://lists.wikimedia.org/mailman/listinfo/wikics-l is broken as well.)