Bausch, Jean wrote:
I have now found who swallows the Umlauts in message bodies:
I had inserted a 'fmt -s -w 80' in the mail aliases. Unfortunately on Solaris 7 this command erases the Umlauts.
Interesting. You can try changing that to
LANG=de_DE fmt -s -w 80
to make sure Solaris can handle the 8 bit characters. But this assumes
every email you put through will be in iso-8859-1, which is increasingly
not true in today's international world.
If you don't specify a LANG, or specify the wrong one, fmt doesn't know
what character set the input file is in, and if it's a multibyte
character set (UTF-8, UTF-16, or any Japanese/Chinese/Korean character
set) it might end up splitting a character between bytes when it inserts
a newline. So they do the "safe" thing and strip all 8-bit characters
when LANG is not set.
Hence my next question:
Is there an option in mailman 2.1.2 to autowrap the messages after 80 characters or so?
I don't think there's really a safe way to do this. It would certainly
screw up signatures on emails.
All MIME emails are required to be 76 characters or less per line; you
could write a Handler (look at Mailman/Handlers/Decorate.py for an
example in the Mailman source code) to take any non-MIME emails and
automatically make them a single text/plain attachment encoded with
quoted-printable, but you'll need to be careful about character set
issues for the UTF-8 and CJK users.
The email.Message and email.Charset modules make this really easy; all
the work is already done for you.
Ben
