Bug#423171: possible-missing-colon-in-closes should not check ancient entires
Thijs Kinkhorst <thijs@debian.org> writes:
> Package: lintian
> Version: 1.23.30
> Hi,
> I got these warnings on checking squirrelmail:
> W: squirrelmail: possible-missing-colon-in-closes l436:Closes #143277
> W: squirrelmail: possible-missing-colon-in-closes l438:Closes #136612
> W: squirrelmail: possible-missing-colon-in-closes l530:Closes #125866
> This concerns changelog entries from more than 4 years old. I don't
> think it's right for lintian to be expecting me to change those entries,
> especially because it brings no benefit since that changelog entry will
> not be processe by dak to create close-messages.
> I suggest to only check the topmost changelog entry, as some other tests
> already do aswell if I remember correctly.
None that I could find, and I couldn't see a way of doing that without
some structural changes, although I could look again. However, it's not
correct that only the top-most changelog entries are relevant. If someone
builds the package with the -v option to dpkg-buildpackage, they can
include historical changelog entries. That's actually the best way of
including entries from unreleased versions, experimental versions, NMUs,
and the like. So it's unclear how far to go back.
(Although I suppose one could do the same check against the changelog
entry included in the *.changes file, which would be an interesting route
to take. I'll have to think about doing it that way.)
--
Russ Allbery (rra@debian.org) <http://www.eyrie.org/~eagle/>
Reply to: