Allan Hansen writes:
LookupError: unknown encoding: iso-8859-8-i
Finally, in the captured message source (the last which was Cc'd to me) I do see this:
Content-Type: text/plain; charset="iso-8859-8-i"
Yes, that's the problem. Python doesn't define a codec for that encoding.
It's generally preferable that mail senders use UTF-8, as that is likely to be implemented correctly in almost all modern mail clients, while the ISO-8859/8 series are likely to be implemented only in mail clients that specifically support Hebrew.
That said, if the users are unable or unwilling to use UTF-8, it would be possible to create a codec for "iso-8859-8-i", I don't know how to do that, and I doubt Mark does. You could ask on python-list@python.org, or file an RFE on Python's GitHub tracker. For Mailman's purpose, there's no need to handle bidi, just codepoint for codepoint translation from "iso-8859-8-i" to Python internal encoding. So I guess the encoder and decoder from "iso-8859-8" (which does exist) can be used.
If it's just code-for-code, then we can figure out how to have it used only for Mailman.
Steve