spamassassin-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rejaine Monteiro <>
Subject Re: extract eml forwarded attached mail and sa-learn
Date Thu, 16 Mar 2017 12:58:42 GMT
Thank you all for the tips.

I instruct all users to always forward spam as an attachment (mta -> 
forward as attachment option), but obviously there is always that user 
who sends up forwarding the message purely.

I agree that there are problems with users who will end up forwarding 
any unwanted email as spam and I also see it as a big problem to solve 
because certain exaggerations will surely appear ... initially I am 
thinking of visually analyzing the messages that will be forwarded and 
try  filter what really deserves to be learned by sa-learn (laborious 
but necessary)

My problem is aggravated, because I do not have the user accounts 
centralized in the spam gateway (it only analyzes spam and forward to 
internal servers, I do not use user_prefs, for example), but I'll try ...

Ps: sorry for my bad English, I hope you can understand, I'm using 
google translator to write;)

Em 15-03-2017 19:00, Kris Deugau escreveu:
> Rejaine Monteiro wrote:
>> Does anyone know of command or script in to extract an "forwarded
>> attached" email (eg: Forwarded.eml file attached email) on linux command
>> line/shell script/perl/pyton etc.. I  am trying to reformime, ripmime,
>> but I'm not succeeding yet....
>> The idea is to make the user, when receiving spam, redirect mail  to a
>> spam account (ie. spam@mydomain),  like attached email, run a scritp to
>> extract the attached EML file and run sa-learn on it...
>> Thanks any tips!!
> I posted the code I wrote for our spam report handling address here, 
> in December 2013:

> (It took a while to find because the Apache archives don't have a 
> search function, and the other public list-archive sites all seem to 
> be variously broken or awkward to search;  I ended up looking back in 
> the folder I file this list in.)
> Note this is a blob integrated in our local mail delivery handling 
> chain, but that should cover the basics.
> Fair warning:  That was the easy part.  The hard part is in getting 
> users to correctly forward messages as an RFC822 attachment 
> (message/rfc822), if you don't have a handy webmail system with a 
> "report as spam" button that Does It Right.
> You'll also get users (mis)reporting legitimate mail of all kinds as 
> spam (deliberately or otherwise);  hand-sort the reports *before* 
> feeding the messages to sa-learn.  We have a few users that regularly 
> report pretty much "all the mail in my Inbox", most of which is ham, 
> and which would seriously break Bayes if we just passed those reports 
> through unsorted.
> -kgd

Rejaine da Silveira Monteiro
Tel: (31) 2102-8854
Jamef Encomendas Urgentes - Matriz - Belo Horizonte/MG

View raw message