Exim (was: List of bugs that *must* be fixed before freezing Hamm)
On Thu 12 Mar 1998, Paul Slootman wrote:
> On Thu 12 Mar 1998, Mark Baker wrote:
> > On Thu, Mar 12, 1998 at 11:01:22AM +0100, Paul Slootman wrote:
> > > I myself have tried out exim recently after being a smail user for a
> > > long time (9 years, I believe). It's great! I.e., another vote for
> > > making it the standard debian MTA.
> > Did you use eximconfig? If so, how did you get on with it? As I already have
> > a working, and moderately customised, configuration, I haven't actually run
> > it myself except to test, so I'd very much like to hear of any problems with
> > it, particularly if exim is going to be the standard in 2.0.
> Hmmm, I think it went OK; I had to hack the generated config file anyway
> because I needed to do some rewriting. AFAIK the "normal" stuff didn't
> need any extra hacking.
A couple of additional comments:
- The /etc/aliases file is not terminated by a newline. The same bug is
in smail (see "#18516: /etc/aliases does not end with a newline").
This means that when e.g. cnews appends its lines, the mailer-daemon
alias gets mangled:
mailer-daemon: postmaster#-- cnews begin
- If /etc/aliases doesn't exist, exim refuses to deliver any mail. After
the upgrade from smail, I got error messages from cron about not being
able to find runq. I then purged smail (hoping that
/var/spool/cron/crontabs/mail was a conffile), and that removed
/etc/aliases... I then didn't get _any_ mail at all:
1998-03-12 16:15:01 0yD9hF-0008S9-00 <= email@example.com U=root P=local S=741
1998-03-12 16:15:01 0yD9hF-0008S9-00 system_aliases director: failed to open /etc/aliases for linear search: No such file or directory
1998-03-12 16:15:02 0yD9hF-0008S9-00 == firstname.lastname@example.org D=system_aliases defer (-18): failed to open /etc/aliases for linear search: No such file or directory
1998-03-12 16:19:17 0yD9lN-0008T4-00 <= email@example.com U=paul P=local S=1315 id=19980312161917.03281@hamm
1998-03-12 16:19:18 0yD9lN-0008T4-00 == firstname.lastname@example.org R=smarthost defer (-1):
Especially the last one, which was outbound to another system, should
simply have been delivered; that shouldn't depend on a local
/etc/aliases, right? I'd suggest that the absence of an alias file
shouldn't prevent the delivery of mail, unless of course the address
can't be resolved without a relevant line in aliases.
home: email@example.com | work: firstname.lastname@example.org
http://www.wurtel.demon.nl | Murphy Software, Enschede, the Netherlands
Support Randal Schwartz!
See http://www.lightlink.com/fors/ or send empty email to email@example.com
E-mail the word "unsubscribe" to firstname.lastname@example.org
TO UNSUBSCRIBE FROM THIS MAILING LIST. Trouble? E-mail to email@example.com