Bug#681410: false postives for arch-dependent-file-not-in-arch-specific-directory diagnose
Jakub Wilk <jwilk@debian.org> writes:
> * Matthias Klose <doko@debian.org>, 2012-07-12, 23:58:
>> I noticed a lot of false positives:
>> E: libpython3.3-dev: arch-dependent-file-not-in-arch-specific-directory
>> usr/lib/python3.3/config-3.3m-x86_64-linux-gnu/libpython3.3m-pic.a
> Move it to /usr/lib/<multiarch-triplet>/python3.3/config-<abitag>/ and add
> a symlink if really needed. This will make lintian happy.
>> So maybe better check if the path either contains the multiarch name or
>> the debian architecture name.
> I think we should rather remove existing exceptions (for ruby and openjdk,
> after they are fixed) than to add more.
What (besides minor improvements in ease of maintenance of Lintian) would
be gained from doing so? It looks to me like this is an arch-qualified
path, so it satisfies the requirements of multiarch, and the only issue is
that Lintian has to be slightly more complex to detect that.
I'm not very active at the moment, so weigh my opinion accordingly, but I
always tried to apply the general principle that if Lintian *could* figure
out the right thing to do, fixing Lintian rather than changing packages
just to make life easier for Lintian is the right approach unless the
package behavior introduces an actual bug or Policy violation. If people
have do things just to make Lintian happy, it discourages people from
using Lintian and paying attention to it.
--
Russ Allbery (rra@debian.org) <http://www.eyrie.org/~eagle/>
Reply to: