[SOLVED] Spam-filter
- Thread starter Grogory
- Start date
hm - haven't tried it with cyrillic letters in particular, but the following howto works in general:
https://wiki.apache.org/spamassassin/WritingRules (first example)
this stackoverflow thread could also help:
https://superuser.com/questions/118...cyrillic-russian-utf-8-letters-base64-encoded
however - keep in mind that trying to block 'sex' can easily also block Sussex - so keep an eye open for false positives!
hope this helps!
https://wiki.apache.org/spamassassin/WritingRules (first example)
this stackoverflow thread could also help:
https://superuser.com/questions/118...cyrillic-russian-utf-8-letters-base64-encoded
however - keep in mind that trying to block 'sex' can easily also block Sussex - so keep an eye open for false positives!
hope this helps!
hm - you can test if it works (send an email which contains the word you want to block - and check the spam-report in the header) - hope this helps!
I tested, but it does not working with russian sentenses... The problem, I fink, that the body of russian email can have utf-8, win1251 and koi8_r encoding. If I understand, emal in this cases is transfered with base64 encoding. I fink spamassasin do not recognise body content in this cases. It will be good, if spamassassin convert e-mal to the utf-8, analise body, and then convert mail to encoding in witch it was received and send it. But I fink it is impossible.... But I can not undestand, how it works in Barracuda Spam filter. In Barracuda I can add russian phrases (regexp) from web interface and it work very good...
In theory, if the e-mail is decoded correctly all of this should end up in perl-strings - in that case you'd need to make sure that the regex matches the correct perl-chars - maybe https://perldoc.perl.org/perlrebackslash.html and http://sa-russian.narod.ru/no_russian.html can provide some starting point
Hi, very usefull link for russian phrases is http://sa-russian.narod.ru/downloads.html. I download 6 files from this link, then edit the file tokens_utf-8.txt - create minimal test rule with regexp, then I generate *.cf rule files by the script make_rules.pl, after that I combited all *.cf files to the file cutom.cf and put the file to the /etc/mail/spamassassin/. After that spamassassin work with russian phrases. It seems to me
a good idea would be to include the posibility to add custom phrases for spamassassin to the web interface of PMG (as implemented in Barracuda Spam Firewall), is not it?
a good idea would be to include the posibility to add custom phrases for spamassassin to the web interface of PMG (as implemented in Barracuda Spam Firewall), is not it?