OT: procmailrc question

[Robert Holtzman]

On Wed, Jan 08, 2014 at 03:38:50PM +1100, Cameron Simpson wrote:
> On 07Jan2014 22:04, Ranjan Maitra <maitra.mbox.ignored at inbox.com> wrote:
> > I am new to making my own procmail recipes and am a bit stumped with
> > this one.
> > 
> > I use sylfilter as my spam-filtering tool: this tool is available at
> > http://sylpheed.sraoss.jp/sylfilter/
> [...]
> > Being perhaps foolhardy, I have been trying to get syfilter to work
> > with procmail using my .procmailrc. I was trying to mimic the example
> > here:
> > http://www.math.tohoku.ac.jp/~kuroki/keijiban/procmailrc.txt
> > 
> > However, the issue is that I do not quite see how to get this to work
> > because unlike bogofilter, sylfilter does not add anything to the
> > header. If it returns a value 0, the mail  is classified as junk,
> [...]
> > (which the user does not see):
> > Return values:
> >   0   junk (spam)
> 
> As you say, it does not modify the email.
> 
> Instead, you want a recipe shaped like this one:
> 
>   # Get rid of duplicates
>   :0 Whc: .msgid.lock
>   | formail -D 16384 .msgid.cache
>   :0 a
>   /dev/null
> 
> This recipe discards duplicate emails, but by replacing the "formail" command
> with your "sylfilter" command you can have it discard (or file) junk email.
> 
> Example (untested, needed adapting):
> 
>   # Get rid of duplicates
>   :0 Wc
>   | sylfilter ...
>   :0 a
>   /dev/null
> 

Dropping messages to /dev/null is dangerous. It's a good way to lose
legit mail that was misclassified. Better to create a spam directory and
put mail classified as spam in there. It's easy to skim thru that and
pick out any mail that was misclassified.

-- 
Bob Holtzman
Your mail is being read by tight lipped 
NSA agents who fail to see humor in Doctor 
Strangelove 
Key ID 8D549279
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.fedoraproject.org/pipermail/users/attachments/20140107/65fb8f42/attachment.sig>