Michael Colvin wrote:
> After looking into QMT, which has recipient validation built in, I'm not
> sure Spamdyke really needs it...  The implementation in QMT allows for
> VPOPmail and non-VPOPmail qmail servers to easily validate recipients.  If
> Spamdyke implemented a version based on cdb files, with VPOPmail servers,
> something would have to be put in place to build those cdb files from the
> database.
> 
> Spamdyke is fantastic at what it does.  I'm not sure that it needs to be
> complicated.  Of course, as long as the validation is easy enough to
> disable, then I guess it wouldn't matter, and non-VPOPmail users could
> enable it and use the cdb files...  If Spamdyke included the ability to
> validate against the VPOPmail database, I'm not sure it would be any more or
> less efficient than the patch that's included in QMT.  Eric?

My guess is that performance would be about the same whether spamdyke or 
chkuser does the validation. I don't see the issue as being performance 
related though. I'm more interested in having configuration options in a 
simple, manageable place. I'd like to see spamdyke handle whatever 
configuration variables are practical, even if spamdyke were to simply 
set an environment variable for some other code to pick up. The fewer 
number of patches to qmail source, the better.

Which makes me wonder about chkuser. That patch is implemented in a 
non-invasive fashion, as most of the code sits outside of qmail proper. 
  Most if not all of the chkuser configuration parameters can be altered 
with environment variables.

Sam, have you looked at bringing chkuser functionality into the spamdyke 
realm? I would expect that you could probably find a way to integrate 
chkuser into spamdyke, eliminating the need for the chkuser patch to 
qmail. This would simply QMT a bit as well.

Thanks for bringing this up Michael.

-- 
-Eric 'shubes'

_______________________________________________
spamdyke-users mailing list
spamdyke-users@spamdyke.org
http://www.spamdyke.org/mailman/listinfo/spamdyke-users

Reply via email to