On 7 Sep 2005 at 12:10, Jason Morehouse wrote about
"[exim] bounce per lang":
| We're getting a large amount of Russian spam. Well... I think it's
| spam... but can't really read it! :)
|
| Anyway, what's the best way to bounce message per language.. or
| character set I would guess. We're using exim 4.43.
"Bounce" is a dirty word. I'll assume you mean "reject". ;)
The easiest way to do this is with the exiscan patch (if you haven't
already applied the patch, best to upgrade to 4.52). For example, I
have the following stanza in my MIME ACL:
deny message = Sorry, emails must be in english \
(detected charset=$mime_charset)\n \
Add +bypass to username to bypass this restriction
!sender_domains = lsearch;/usr/exim/domain_whitelist
!authenticated = *
condition = ${if match {$recipients} {.*\\+bypass} {0}{1}}
condition = $mime_is_coverletter
condition = ${if match {${lc:$mime_charset}} \
{\N(gb2312|big5|big-5)\N} \
{1}{0}}
mime_regex = \N[\x80-\xff]\N
logwrite = :reject: ASIAN REJECT ($mime_charset), F=$sender_address, S=$header_subject:, T=$recipients
The mime_regex condition is to check for there actually being some
non-ASCII characters in the message body, as we get legitimate email
from asian sources written in english.
Again, to make sense of this example you'll need the exiscan-acl
patch applied (or V4.5x), and refer to exiscan-acl-spec.txt (or
spec.txt).
- Fred