Re: lintian.debian.org -> git
On Sun, 2008-08-10 at 15:57 +0200, Frank Lichtenheld wrote:
> On Sun, Aug 10, 2008 at 01:55:28PM +0100, Adam D. Barratt wrote:
> > One thing I noticed is that the results for
> > apparently-corrupted-elf-binary displayed on lintian.d.o still include
> > packages shipping 64-bit libraries on i386.
> >
> > I've installed lintian 1.24.3 on an i386 etch system and checking the
> > i386 package of fakeroot 1.9.6 does not raise
> > any apparently-corrupted-elf-binary tags but does correctly issue
> > shlib-with-executable-bit for the lib64 binaries, so I'm not sure why
> > it's not working on gluck.
>
> Makes it any difference if you have binutils-multiarch installed or not?
> (just wild guessing...)
Thanks for the suggestion, but unfortunately not; binutils-multiarch
wasn't installed initially and after installing it the results are still
correct.
The only other thing I can think of at the moment is that the version
check I guarded the readelf fallback code with isn't working on gluck
for some reason - i.e. "dpkg-query -W -f='${Version}\n' binutils" isn't
returning a string starting "2.17".
> Anyway, yesterday Chrsistoph Berg and I started to make ourself familiar
> with the workings of Mole and I hope that we can reenable the lintian
> worker so that we can take the checks off gluck's aching back and run
> it in a sid chroot somewhere.
>
> If we get that to work nicely it would make sense to replace the input
> for lintian.debian.org with that.
That certainly seems like a much better idea longer term than trying to
work around issues with the checks running on etch (and running on a
less-loaded box certainly wouldn't hurt).
Adam
Reply to: