On Thu, 19 Dec 2002, Jim Pazarena wrote:
> I presume exim is dying within my local_scan routine. I see 10 to 20
> leftover files each day among the 1000's my server receives each day.
>
> What method would be the most efficient to trace this problem on a
> _live_ system?
A sledgehammer approach would be to modify your local_scan to insert
calls to log_write() to see where the crash is happening. Start with one
on entry and one on exit. This will, of course, increase the size of
your log somewhat...
An alternative is to provide a signal handler for the common crash
errors such as SIGSEGV and write to the log from inside the handler.
Hmm. I wonder if Exim itself should set up such a signal handler before
it calls local_scan? Perhaps it should. I'll think about that one.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
