Re: New check: ahead-of-time-standards-version
Raphael Geissert <atomo64+debian@gmail.com> writes:
> Russ Allbery wrote:
>> Raphael Geissert <atomo64+debian@gmail.com> writes:
>>> Because there were maintainers who had set a standards version which
>>> was higher than the $current one (e.g. 3.8.0.0 in a package uploaded
>>> two years ago). And now that 3.8.0 is $current, the package appears
>>> not to checked for the new standards.
>> This really doesn't seem useful to me, sorry. I'm not willing to apply
>> it, although I won't veto it if someone else sees it as useful enough
>> to warrant the maintenance.
> The thing is that such packages are completely _wrong_ and should not be
> distributed as is. The more the number of packages Debian provides, the
> challenge of maintaining an overall good QA level becomes harder.
I think this is far too strong. The most likely cause of this problem is
the perfectly benign error of forgetting to update the timestamp in
debian/changelog that was left over from the first change one added after
the previous upload.
However, I came back to this while preparing the next release and decided
that you're correct, it's a valid mistake to find, even if I'm not sure
it's going to be particularly useful. It's something that the maintainer
should fix regardless, although I expect that just updating the changelog
date is the fix in most cases.
I've added this to the Git repository and updated the test suite
accordingly, and also fixed all of the Policy release dates to be in UTC
and changed the date comparison code to use UTC so that we don't have
problems around border cases.
--
Russ Allbery (rra@debian.org) <http://www.eyrie.org/~eagle/>
Reply to: