Re: spamd vs. 'make test'

[Sidney Markowitz]

If we are really lucky, this is the same problem as bug #3443 that I 
recently entered. I think Murphy's law says that we are not that lucky. 
My bug is that that tests with spamd consistently do not kill processes 
when running under Cygwin.

I did see the problem under Redhat 9 when I control-C'd out of a test 
and it left some spamd processes running.

Can anyone think of a more robust way to kill running spamd processes 
without introducing a vulnerability that would let unauthorized people 
from killing a real spamd?

 -- sidney