On Fri, Oct 02, 1998 at 04:22:35PM +0100, Nigel Metheringham wrote:
> By the description of the bug, you can work round this by never doing
> lookups on hostnames... which is exactly what we do (only the IP addresses
> are reliable so why bother reverse mapping them - it takes significant
> extra time).
Yep I've seen the same thing, the backtrace from gdb shows:
#0 0x371c8 in host_scan_for_local_hosts ()
#1 0x37764 in host_find_byname ()
#2 0x374dc in host_find_byaddr ()
#3 0x482c8 in smtp_start_session ()
#4 0x1e180 in handle_smtp_call ()
#5 0x1f818 in daemon_go ()
#6 0x2dfc4 in main ()
Your workaround described above explains why I don't see this
behaviour on certain machines (they never do lookups as you say)...
Setting up a test host with reverse but not forward:
%exim -bh 194.217.125.254
**** SMTP testing session as if from host 194.217.125.254
**** Not for real!
>>> looking up host name for 194.217.125.254
>>> IP address lookup yielded eximtest.frost.net
Segmentation fault
Every time.
Regards, Matthew
--
Matthew Frost http://www.frost.org/
email: matthew@???
"My feet, my arms and my ears. And, your feet." dfax: +44 870 160 0853
--
*** Exim information can be found at
http://www.exim.org/ ***
