If you see odd characters in the message bodies inside your list's digests, set the digest-charset to the character set most often used by your subscribers when posting. UTF-8 is usually a good choice.

Mailing list digests are a funny thing. It's a single email where the bodies of at least one mailing list post are concatenated together. There's a RFC that describes how they're supposed to work and we follow that standard, but the problem is in those message bodies.

If you allow HTML on your list, for example, this can cause issues with Digests: We can't append blobs of HTML one after the other so we have to use the text/plain version of the message that usually accompanies the HTML. (HTML email is usually a MIME attachment and a second attachment contains a non-formatted version for readers who can't or don't want to read the HTML version.) If a plaintext version isn't provided, that message body will appear empty in the digest.

Another complexity arises when posts have varied character sets: If one sender posts in 7-bit ASCII and another in 8-bit UTF-8, some message bodies will appear correctly and others might appear garbled. That's why setting a higher-order character set via digest-charset that allows both sets to be rendered correctly can help.