On Mon, Sep 01, 2014 at 08:54:36PM +0200, Jérémie Courrèges-Anglas wrote: > j...@wxcvbn.org (Jérémie Courrèges-Anglas) writes: > > > Vadim Zhukov <persg...@gmail.com> writes: > > > >> 2014-08-20 2:47 GMT+04:00 Jérémie Courrèges-Anglas <j...@wxcvbn.org>: > >>> > >>> Sendmail will probably leave the base source tree during this release. > >>> Here's the port I use since several months - but with a very limited use > >>> case. The rc script is (for now) named esendmail. > >>> > >>> Test reports and comments welcome. > >> > >> Two small questions here, otherwise reads fine. I didn't test running > >> it yet, though. > >> > >> 1. It looks like missing a README on how to enable sendmail. > > > > Yup, I'd first mention the root crontab entry for the queue runner. > > First stab at it. Maybe it's too verbose, maybe there are important > missing items. I moved most hints from sendmail-enable/disable in the > README, while here. > > >> Or it > >> could run sendmail-enable and sendmail-disable right in the > >> rcscript... > > > > This sounds a bit too automated to my taste. :) > > But if other MTAs in the ports tree behave like this, why not. > > Other MTAs don't behave so. I think that providing reasonable > documentation and configuration templates is nicer in the end. > > >> 2. Why do hide the "mv ${PREFIX}/sbin/makemap > >> ${PREFIX}/libexec/sendmail" command? > > > > No good reason. I was propably bored when I added this and the @echo > > "Installing..." stuff. > > I removed a bit more of those. > > Other changes: > - added @newuser/@newgroup incantations, since smmsp from base will > probably die soon, just like the named user for bind. > - use _smmsp in submit.mc > - disable ident queries, per ajacoutot@ in src/ > - correct the messed up @sample stuff, the cf files ended up in /etc/ > instead of /etc/mail/. oops > - replace documentation in sendmail-enable with more error checking > > Next step could be splitting it in -main and -libmilter, and build > a shared lib in the latter. > > Comments / feedback still welcome. :) >
I've been running this version on one of my systems with no issue so far. I'm going to try the sasl+ldap version on my test setup at LAAS in the next days if I find time for that. -- Matthieu Herrb