On Tue, 28 Jan 2003, Nico Erfurth wrote:
> > http://www.bgl.nu/~glouis/exim.out
>
> Hmmm, strange, it crashed VERY early in the transport.
> But it could be inside of the retry handling (databasestuff).
> I guess we will have to wait for Phil to wake up ;)
<fx> alarm-clock </fx>
Hmm. There aren't a huge number of debugging prints in the smtp
transport at the start, so it isn't narrowed down very much. Like Nico,
my suspicion is on the database stuff, because that's where such crashes
have been in the past, but see my comment below.
I think what you will have to do is to insert lots of lines like
debug_printf("+++1\n");
...
debug_printf("+++2\n);
...
into the source of transports/smtp.c so as to get additional debugging
output about where it is getting to before it crashes. It doesn't seem
to have got as far as checking the retry information for the host,
because that has already got some debugging statements in the function
retry_check_address().
Put some debugging around line 1937, to see if it is successfully
getting the port number.
Philip
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
