helo,
> There was a problem with this in the 2.00 release which is fixed in 2.01.
yes test_host compiles on 2.01 no problem: I can give results on three
platforms.
On Irix 6.2 on R8000 machines (with an extra XPI interface) , I get
| physig3 [exim-2.01/build-IRIX6-iris4d64] ./test_host
| Exim stand-alone host functions test
| Actual local interface address is 255.255.255.255
| Actual local interface address is 255.255.255.255
| Actual local interface address is 0.0.0.0
| Testing host lookup
On Irix 6.3 on O2 machines (no XPI), I get
| phoxy2 [exim-2.01/build-IRIX632-iris4d] ./test_host
| Exim stand-alone host functions test
| Actual local interface address is 255.255.255.255
| Actual local interface address is 0.0.0.0
| Testing host lookup
and on a Linux machine I get the re-assuring
| phalanger1 [exim-2.01/build-Linux-i386-linux] ./test_host
| Exim stand-alone host functions test
| Actual local interface address is 127.0.0.1
| Actual local interface address is 192.168.2.168
| Testing host lookup
so this just confirms what we thought we knew.
> > Where does exim look to resolve the local interface addresses?
> It calls fcntl() with the SIOCGIFCONF option. The code is in the source
> file called host.c.
Looking at host.c, I would say that Philip meant 'ioctl' rather then 'fcntl',
but anyway: is there any more elementary way of testing this function than what
is performed in 'test_host'? I doubt it. What is really odd is that 'ifconfig
ec0' gives a sensible answer:
| ec0: flags=c63<UP,BROADCAST,NOTRAILERS,RUNNING,FILTMULTI,MULTICAST>
| inet 192.168.2.75 netmask 0xffffff00 broadcast 192.168.2.255
the machines round here are a bit 'quirky' from time to time as we don't have
any systems support and I'm not trained or employed in this capacity, just find
it interesting. If anyone has any ideas on what else I could test or change I
would be grateful to try it.
cheers than,
James
--
*** Exim information can be found at
http://www.exim.org/ ***
