[petsc-dev] SuperLU build error

Xiaoye S. Li xsli at lbl.gov
Fri Dec 18 14:48:29 CST 2015


Thanks.  I updated the superlu_5.1.tar.gz tar ball with this fix.

Sherry

On Fri, Dec 18, 2015 at 10:37 AM, Satish Balay <balay at mcs.anl.gov> wrote:

> Sorry - looks like the superlu issue sliped through the cracks.
>
> However supleru_dist was fixed in upstream tarball [so its patched
> in tarballs used by petsc aswell]
>
> I've fixed the superlu issue in petsc 'maint' and 'master'- and
> verified superlu+superlu_dist build with:
>
> CFLAGS="-Wformat -Wformat-security -Werror=format-security"
>
> Sherry,
>
> Could you fix superlu with the following patch?
>
>
> https://bitbucket.org/petsc/pkg-superlu/commits/0342e0bb6526fa61e40db44bd70bd0802e9b32a2
>
> Thanks,
> Satish
>
> On Fri, 18 Dec 2015, Garth N. Wells wrote:
>
> > On 18 December 2015 at 16:45, Barry Smith <bsmith at mcs.anl.gov> wrote:
> >
> > >
> > >    Garth,
> > >
> > >      Yes this has been an on going nuisance. Do you have this problem
> with
> > > the master development version or some previous release?
> > >
> > >
> > master (updated earlier today).
> >
> > Garth
> >
> >
> > >    Satish,
> > >
> > >       I thought this was now fully resolved?
> > >
> > >    Barry
> > >
> > > > On Dec 18, 2015, at 2:52 AM, Garth N. Wells <gnw20 at cam.ac.uk> wrote:
> > > >
> > > > I've reported before that SuperLU has a bug that causes compile
> failures:
> > > >
> > > >   util.c:32:18: error: format not a string literal and no format
> > > arguments [-Werror=format-security]
> > > >        fprintf(stderr, msg);
> > > >
> > > > I usually fix this by hand, but the way PETSc since very recently
> resets
> > > user changes in cloned external packages is overwriting the change I
> make.
> > > >
> > > > There is a similar bug in SuperLU_dist that I also fix by hand.
> > > >
> > > > Garth
> > >
> > >
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20151218/383249aa/attachment.html>


More information about the petsc-dev mailing list