Bug#486682: lintian: warn if wildcards in .install could let a SONAME bump go unnoticed?
* Frank Lichtenheld [Tue, 17 Jun 2008 21:13:23 +0200]:
> On Tue, Jun 17, 2008 at 11:13:44AM -0700, Russ Allbery wrote:
> > Adeodato Simó <dato@net.com.org.es> writes:
> > > * Russ Allbery [Tue, 17 Jun 2008 09:27:23 -0700]:
> > >> Wouldn't Lintian already detect that problem and warn that the SONAME of
> > >> the library doesn't match the package name?
> > > Well, that works when lintian is run against the upload that introduces
> > > the bug. This check would act as a "style checker" for the practice that
> > > could eventually cause the bug (and works at any point between the
> > > adding of the wildcard and the fatal upload happening). So, it's a bit
> > > of a "prevention" mechanism.
> > Ah, I see. If you list the specific SONAME, does dh_install then fail
> > when it can't find any matching files? (I seem to have had it
> > successfully produce empty packages for me in the past, but it may be that
> > the mistake I made then was different than this one.)
> Only if you use --fail-missing, I think.
No, --fail-missing is something different: fail if some files in --sourcedir
are not marked for install somewhere.
I just checked, and dh_install effectively fails if a wildcard does not
match any pattern:
dh_install
dh_install: mlocate missing files (debian/tmp/usr/bin/foo*), aborting
make: *** [install-stamp] Error 1
Cheers,
--
Adeodato Simó dato at net.com.org.es
Debian Developer adeodato at debian.org
Faced with the choice between changing one's mind and proving that there
is no need to do so, almost everyone gets busy with the proof.
-- J.K. Galbraith
Reply to: