Re: getting rid of the .changes requirements
Niels Thykier <niels@thykier.net> writes:
> So, someone could write a third-party Lintian check to check for a
> missing -v in the changes file for use with backports[1]?
> ~Niels
> [1] Technically, it would check for only having "one" entry, which
> seems to be a fairly reliable indicator of a problem (but can have
> false-negatives).
Oh, hey, that's a good point. It's going to be very rare that there
should be only one changes entry for a backport-targetted upload in
*.changes, and all of those cases should be cases where the build number
(after the +) is >1. So, check if the version ends in ~bpoNN+1 and is
targetted at *-backports and, if so, warn if the *.changes file only has
one version of history.
--
Russ Allbery (rra@debian.org) <http://www.eyrie.org/~eagle/>
Reply to: