[Mimedefang] graphdefang cores with large amounts of data

[Graham Dunn]

On Wed, Oct 15, 2003 at 08:00:34PM -0400, Lee Dilkie wrote:
> I was having a problem with graphdefang just "losing" history. It was
> happening at irregular intervals but I couldn't keep more than a day and a
> half before everything would reset.
> 
> On a hunch, I modified the md_graphlog function in mimedefang.pl to replace
> the $Subject that is syslogged with just "Dummy Subject". This was a hunch
> because I noticed that a number of the syslogs had not-so-nice characters in
> them, extracted from the subject lines of the emails and the syslog parser
> might have gotten confused. I say "might" because this is my first
> experience with Perl and it looks a lot like APL so far...
> 
> Anyway, won't know for sure if this worked until tomorrow but things are
> still holding together.
> 
> Anyway, if it doesn't work, has anyone else seen this behaviour? I'm not
> processing millions of messages, maybe a couple hundred a day.

This sounds like what I was first experiencing when I installed
graphdefang. The solution was to download and unpack a clean copy of db
from the sleepycat site. I forget if I needed a configure in the main
directory, but make / make install in the perl/DB_File directory fixed
the problem.

This was on FreeBSD-4.8 stable.

Graham