create sendmail directories only from BSD.sendmail.dist
ClosedPublic
Actions

Authored by emaste on Mar 22 2020, 1:17 PM.

Details

Reviewers

bapt
manu
gshapiro

Commits

rS359215: pkgbase: create sendmail directories only from BSD.sendmail.dist

Summary

Else when WITHOUT_SENDMAIL is set with pkgbase we create a sendmail package containing nothing but two directories.

Diff Detail

Lint

Lint Skipped

Unit

Tests Skipped

Event Timeline

emaste created this revision.Mar 22 2020, 1:17 PM

manu accepted this revision.Mar 22 2020, 1:22 PM

This revision is now accepted and ready to land.Mar 22 2020, 1:22 PM

We might need to look at a more holistic approach to this, e.g. adding an mtree option to include or exclude entries based on tags, or perhaps preprocessing the mtree files before stageworld. However, we already had BSD.sendmail.dist split out (although rS348981 re-added clientmqueue to BSD.var.dist, duplicating the entry in BSD.sendmail.dist) so it seemed sensible to use it for now.

BSD.var.dist still has at, dma, and unbound entries.

emaste added a reviewer: gshapiro.Mar 22 2020, 1:53 PM

emaste added inline comments.Mar 22 2020, 1:55 PM

etc/mtree/BSD.sendmail.dist
15	accidentally deleted a space on this line, will revert that before commit.

It is fine to commit as is. One thing to consider would be simplifying:

/set gname=daemon
            mqueue          tags=package=sendmail

mqueue          gname=daemon tags=package=sendmail

(note that the mqueue line above should be indented but the comment block doesn't allow leading spaces)

Closed by commit rS359215: pkgbase: create sendmail directories only from BSD.sendmail.dist (authored by emaste). · Explain WhyMar 22 2020, 3:37 PM

This revision was automatically updated to reflect the committed changes.

emaste added a commit: rS359215: pkgbase: create sendmail directories only from BSD.sendmail.dist.

Herald added a subscriber: imp. · View Herald TranscriptMar 22 2020, 3:37 PM

In D24149#531114, @gshapiro wrote:

It is fine to commit as is. One thing to consider would be simplifying:

Thanks, sorry I didn't see your comment before committing. I agree this is preferable and committed in rS359216 - although I don't think it matters too much for this specific instance there are are couple more cases in BSD.var.dist that we'll have to address and it's good to have this as a clean example.