Date: Tue, 29 Dec 2020 15:25:36 +0100 From: "Ronald Klop" <ronald-lists@klop.ws> To: "Dan Mahoney (Gushi)" <freebsd@gushi.org> Cc: freebsd-ports@freebsd.org Subject: Re: DMA -- difference between base and port? Message-ID: <op.0wd40ykqkndu52@sjakie> In-Reply-To: <e3965857-8ba1-62ba-9f39-dbb5945251f@prime.gushi.org> References: <27a72fde-d96c-25e6-ff62-85767da510b7@prime.gushi.org> <op.0vzr21pukndu52@sjakie> <e3965857-8ba1-62ba-9f39-dbb5945251f@prime.gushi.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Some questions below. On Tue, 22 Dec 2020 01:56:04 +0100, Dan Mahoney (Gushi) <freebsd@gushi.org> wrote: > On Mon, 21 Dec 2020, Ronald Klop wrote: > >> On Mon, 21 Dec 2020 12:12:02 +0100, Dan Mahoney (Gushi) >> <freebsd@gushi.org> wrote: >> >>> Hey there, >>> At the day job we've been using mail/dma port for a number of years >>> now, and the rollout and config of files in /usr/local/etc/dma is part >>> of our deploy process. >>> It only recently occurred to us that there was a "dma" in base since >>> probably 11.0 (whomever wrote the release notes missed that -- and the >>> manpage doesn't mention when it was added to FreeBSD). >>> We notice that the "newaliases" function in /etc/mail/mailer.conf is >>> missing from the port version -- which means if you're using ports >>> dma, you probably want to set newaliases to something like >>> /usr/bin/true (dma doesn't use an aliases db, so there's no need to >>> rebuild one, as newaliases would). Again. something we noticed in our >>> deployment process with puppet. Why are you calling newaliases if dma does not use an aliases db? >>> I can't find a feature-by-feature comparison for what one would >>> install the port for (other than inertia, like we have). >>> There's no "version" command that I can find in DMA. (tried -h, -?, >>> -v --version, -V). >>> Does "our" DMA track the Dragonfly version (like the base sendmail or >>> openssl track world) or is it completely forked and unlikely to >>> incorporate changes? This would be useful in feature comparison. >>> Is it worth mentioning this in the pkg-message for mail/dma? >>> -Dan >>> >> >> On 13-CURRENT I have: >> # more /usr/src/contrib/dma/VERSION >> v0.11 >> >> But the version nr doesn't tell the whole story. There are some code >> syncs after 0.11. >> https://github.com/freebsd/freebsd/commits/master/contrib/dma >> >> There is a PR to upgrade base to 0.13 already. >> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=244630 >> >> Would love to see a sync with even newer code. I made a fix upstream >> myself. > > Yes, my issue with "newaliases" failing actually comes down to two open > issues: > > 1) No released version will work with "newaliases" unless "newaliases" > is called as a bareword (so calling /usr/bin/newaliases, as puppet does, > causes you to get a "no recipients" error) This is fixed in dma head, > but not in either the current port version, or the freebsd base version. > I'm a bit confused. You say "this is fixed in dma head". I see no matching commit about this fix after 0.13 release in https://github.com/corecode/dma/commits/master . > We've tweaked it by telling our deploy tools (puppet) to call newaliases > and handing it a path, but we prefer to hand exec's full paths to > binaries. > > 2) DMA's still broken if you've got an alternate alias file defined, see > https://github.com/corecode/dma/issues/90 > > (Given, this is the ports mailing list, but those should also be fixed > in the ports version, with the latter perhaps being fixed one dma hits > 0.14 or whatever version number has that fix, rather than manually > patching 0.13.) > > -Dan > If I read it correctly I have the idea that your issue is in FreeBSD base, ports and the dma github code? So it seems it needs to be solved upstream first. I'm asking this, because I would like to have the 0.13 update in base and make sure that your issue is not a showstopper for this but a separate issue independend of the version of contrib/dma. Regards, Ronald.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?op.0wd40ykqkndu52>