[Mimedefang] filtering on locale info

[Philip Prindeville]

Dave Helton wrote:
> I'm getting a lot of spam from non-english speaking, or typing, sources.
>
> Can anyone tell me their method of dealing with non UTF-8 or us-ascii
> such as iso-2033-jp and GB2312.
>
> Dave
>
>   

Assuming you've got SpamAssassin installed as well, just try:

ok_locales en
ok_languages en


score UNWANTED_LANGUAGE_BODY    4.95



or so (4.95 is a little high for most folks). Though if you ask me...
the rule is somewhat buggered (or more accurately, the function
is_charset_ok_for_locales() is painfully naive).

If someone sends a message in windows-1254 (for instance), but only
uses the lower half (i.e. the ascii portion)... I'd still want to
reject it. The requirements for mail are that messages that fit
into USASCII be tagged as such. (You really only need 3 encodings
anyway, right? USASCII, ISO-8859-1, and UTF-8...)

Mailers that don't do this are broken and should be fixed.

-Philip