chkpkg and soname-unused

[Yann Rouillard]

Hmm, that's really strange.
The linker detects that the library is not required but still keeps it.
It doesn't happen for every library. I added -lssl on the command line and
the libssl library was properly removed from the dependencies.

The debug messages from ld are:

debug: file=libnsl.so.1  unused: does not satisfy any references; retained:
compensating for insufficient dependencies
debug: file=libintl.so.8  unused: does not satisfy any references;
retained: compensating for insufficient dependencies
debug: file=libssl.so.1.0.0  unused: does not satisfy any references;
discarded

I am still looking for the reason.

Yann



2013/11/4 Rafael Ostertag <raos at opencsw.org>

> Hi Yann
>
> On Mon, Nov 04, 2013 at 06:10:30PM +0100, Yann Rouillard wrote:
> > Hi Rafi and slowfranklin,
> >
> > Tell me how to reproduce the problem and I will have a look (for
> > slowfranklin, I suppose I just have to recompile tracker).
> > I am surprised that a patch could have such a bad side effect as screwing
> > the "-z ignore option" and I would prefer to dig first into the problem.
>
> You can take `gamin'. Remove both CHKPKG_OVERRIDES for CSWgamin and build a
> package on sparc and x86. That should do the trick. Let me know if I can
> be of
> assistance.
>
> cheers
> rafi
>
> >
> > Yann
> >
> >
> > 2013/11/4 Rafael Ostertag <raos at opencsw.org>
> >
> > > Hi Dago
> > >
> > > On Mon, Nov 04, 2013 at 05:44:09PM +0100, Dagobert Michelsen wrote:
> > > > Hi Rafi,
> > > >
> > > > > Could it be related to the buildfarm update?
> > > >
> > > >
> > > > Probably, the T5220 is now running 105400-04 whereas unstable10x is
> > > still running 147441-19.
> > > > Should I update the x86 machines also? My impression is to patch as
> less
> > > as possible to
> > > > not introduce new linker symbol anomalies. The patch on the farm was
> > > needed to fix an ugly
> > > > bug in zfs which prevented us from doing backups, so no option of not
> > > installing ;-)
> > >
> > > If it is related to Solaris patches, patching x86 would mean that `-z
> > > ignore'
> > > might not work on x86 anymore, as well. So, wouldn't it be smarter, to
> > > figure
> > > out why it stopped working on sparc? I mean, -z ignore ain't such an
> > > esotheric
> > > switch and has been put in place for good reasons, hasn't it?
> > >
> > > cheers
> > > rafi
> > >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opencsw.org/pipermail/maintainers/attachments/20131104/d0b5d1a6/attachment.html>