Re: -r & -b options combined

[Joe Maimon]

Dan Nelson wrote:

> In the last episode (Feb 14), Ronan Waide said:
>  
>
> > On February 14, address@hidden said:
> >    
> >
> > > I just switched on the -r option on spamass-milter in combination with
> > > -b and I'm slightly dismayed to discover that this means I get a copy
> > > of all the rejected spam as well as the "> spam, < spam-threshold"
> > > mail. I'd kinda expected that 'reject mail' wouldn't go any further
> > >      
>
> Yes, everyone's expectations are different regarding what -r/b/B
> combinations should do.  Joe Maimon has a preliminary patch that
> implements one of my TODO items so that you can specify the exact score
> ranges that trigger each flag's action.  I haven't looked closely at it
> though.
>  
Since I reused a lot of homegrown code, you probably should :-)

The patch also supports multiple invocations of the arguments and 
rejection messages and also allows you to discard messages as well.

To the best of my knowledge, I am the only one using it....but it works 
for me. I havent tested it on the official 0.3.0 yet.

>  
>
> > I've just found a much bigger problem with this: it works by spawning
> > a new sendmail process (yes, I've read the comments on why this is
> > necessary) which seems to be resulting in an infinite loop as the new
> > sendmail process also trips the spam filter which in turn starts a
> > new sendmail process...
> >
> > Cheers,
> > Waider. How do I get all these magic brooms back in the closet?
> >    
>
> Easiest way is to make sure that all your internal mailservers are
> listed in -i.  That won't work if your -b address gets routed in a way
> that it exits your mail system and comes back in, though (-b
> address@hidden, who has a rule that CCs an internal address, for
> example).
>
>  
The archives should contain a bit on both of these topics.

The patches I have up on the project site also include a 
"loop-detection" fix.