On Nov 20, 2018 9:15 AM, Gilles Chehade <gil...@poolp.org> wrote: > > On Sun, Nov 18, 2018 at 08:32:47AM -0600, Edgar Pettijohn III wrote: > > Use new syntax. > > > > Sorry was on the road. > > Comment inlined: > > > > Index: makemap.8 > > > > =================================================================== > > RCS file: /cvs/src/usr.sbin/smtpd/makemap.8,v > > retrieving revision 1.29 > > diff -u -p -u -r1.29 makemap.8 > > --- makemap.8???? ??13 Feb 2016 08:53:18 -0000???? ??1.29 > > +++ makemap.8???? ??18 Nov 2018 14:29:33 -0000 > > @@ -105,8 +105,11 @@ In addition to adding an entry to the pr > > ??one must add a filter rule that accepts mail for the domain > > ??map, for example: > > ??.Bd -literal -offset indent > > -table domains "/etc/mail/domains" > > -accept for domain <domains> deliver to mbox > > +table domains db:/etc/mail/domains.db > > + > > why db ?
Do you need makemap for file backend? > > > +action "local" mbox > > + > > +match for domain <domains> action "local" > > ??.Ed > > ??.Sh VIRTUAL DOMAINS > > ??Virtual domains may also be kept in tables. > > @@ -140,11 +143,13 @@ In addition to adding an entry to the vi > > ??one must add a filter rule that accepts mail for virtual domains, > > ??for example: > > ??.Bd -literal -offset indent > > -table vdomains "/etc/mail/vdomains" > > -table vusers "/etc/mail/users" > > +table vdomains db:/etc/mail/vdomains.db > > +table vusers db:/etc/mail/users.db > > + > > why db ? > > > +action "local" mbox virtual <vusers> > > > > -accept for domain <vdomains> virtual <vusers> deliver to mbox > > -accept for domain example.org virtual <vusers> deliver to mbox > > +match for domain <vdomains> action "local" > > +match for domain "example.org" action "local" > > ??.Ed > > ??.Sh FILES > > ??.Bl -tag -width "/etc/mail/aliasesXXX" -compact > > > > Documentation should stick to the file backend which is the best one for > the general case. > > The db backend is an extension of the file backend and unless you have a > very specific use case, it brings no benefit whatsoever. It ISN'T faster > than the file backend and unless you have a good rationale for using it, > there's actually no good reason to. > > The only reason we still support it is because some corner cases do make > sense, and even in those cases I'd argue there are better backends. > > > -- > Gilles Chehade @poolpOrg > > https://www.poolp.org tip me: https://paypal.me/poolpOrg