Re: Dependencies on shared libs, take 2

[Russ Allbery <rra@debian.org>]

Peter Samuelson <peter@p12n.org> writes:
> [Russ Allbery]

>> They *usually* do, but not all E tags are certain problems.  Of course,
>> maintainers could use overrides.

> I'm opposed to adding overrides to my packages for cases where, in my
> view, lintian should somehow have enough information to see the case as
> a false positive.

Yes, there is that too.  I don't think people should have to add an
override if it's a bug in lintian.  And if dak ever starts using lintian
results, we'll have to figure out a way to handle that.

But of course if that happens, please do report the bug if it isn't
already reported.  There are no false-positive bugs currently open against
lintian; every false positive that's been reported has either been fixed
or is sufficiently difficult to fix for one reason or another that we've
decided overrides are the correct way of handling exceptions.  Also, in
the latter case, if overrides are the right way to handle exceptions, the
lintian tag description, with -i, should say so, and if it doesn't, please
report a minor bug.

> I suppose others feel that being lintian-clean is so important that it's
> worth muddying the waters by overriding all false positives, not only
> the exceptional situations but the lintian bugs.  If most people feel
> that way, I guess it would be reasonable to add lintian to the set of
> dak sanity checks.  But I don't believe I'm the only one who disagrees.

I think it makes sense to add certain lintian checks to the dak sanity
checks.  I'm certain that you don't want all of them, and I'm fairly
certain that you don't even want all of the current Es.

-- 
Russ Allbery (rra@debian.org)               <http://www.eyrie.org/~eagle/>