[Buildroot] [PATCH 1/1] package/elfutils: fix build with uclibc

[Peter Seiderer <ps.report@gmx.net>]

Hello Fabrice,

On Thu, 18 Mar 2021 18:48:24 +0100, Fabrice Fontaine <fontaine.fabrice@gmail.com> wrote:

> Hello Peter,
> 
> Le jeu. 18 mars 2021 ? 18:33, Peter Seiderer <ps.report@gmx.net> a ?crit :
> >
> > Hello Fabrice,
> >
> > On Thu, 18 Mar 2021 08:42:19 +0100, Fabrice Fontaine <fontaine.fabrice@gmail.com> wrote:
> >  
> > > Update first patch to fix a build failure with uclibc and elfutils in
> > > version 0.183.0 which errors out if obstack is not found since
> > > https://sourceware.org/git/?p=elfutils.git;a=commit;h=da855fc9cd415c288bfcb9de1f3d5eb329de0557
> > >
> > > Indeed, obstack has been removed from uclibc-ng since version 1.0.21 and
> > > https://github.com/wbx-github/uclibc-ng/commit/0bd6bfb2b643ea2b4b1440dfd917ba752f0c0d15
> > >
> > > Fixes:
> > >  - http://autobuild.buildroot.org/results/003ae73c7b48ad6b9837fb19ca197061d20eaf6d
> > >  
> >
> > Thanks for taking care of the build failure, but...
> >  
> > > Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
> > > ---
> > >  ...-Add-a-enable-disable-progs-configure-option.patch | 11 +++++++++++
> > >  1 file changed, 11 insertions(+)
> > >
> > > diff --git a/package/elfutils/0001-Add-a-enable-disable-progs-configure-option.patch b/package/elfutils/0001-Add-a-enable-disable-progs-configure-option.patch
> > > index e51f749cf6..bac9bc1245 100644
> > > --- a/package/elfutils/0001-Add-a-enable-disable-progs-configure-option.patch
> > > +++ b/package/elfutils/0001-Add-a-enable-disable-progs-configure-option.patch
> > > @@ -20,6 +20,8 @@ Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
> > >  Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
> > >  [Rebased on 0.183]
> > >  Signed-off-by: Peter Seiderer <ps.report@gmx.net>
> > > +[Fabrice: don't error out if obstack is not found]
> > > +Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
> > >  ---
> > >   Makefile.am  | 6 +++++-
> > >   configure.ac | 6 ++++++
> > > @@ -60,6 +62,15 @@ index d345495..0dd4b1e 100644
> > >   dnl zlib is mandatory.
> > >   save_LIBS="$LIBS"
> > >   LIBS=
> > > +@@ -579,7 +585,7 @@ saved_LIBS="$LIBS"
> > > + AC_SEARCH_LIBS([_obstack_free], [obstack])
> > > + LIBS="$saved_LIBS"
> > > + case "$ac_cv_search__obstack_free" in
> > > +-        no) AC_MSG_FAILURE([failed to find _obstack_free]) ;;
> > > ++        no) obstack_LIBS= ;;
> > > +         -l*) obstack_LIBS="$ac_cv_search__obstack_free" ;;
> > > +         *) obstack_LIBS= ;;
> > > + esac
> > >  --
> > >  2.30.1
> > >  
> >
> > Is this not a complete unrelated/new matter (and so belongs to an extra patch)?  
> From my perspective, this is not unrelated as obstack has "always"
> been used by some of the elfutils program but I'm open to discussion.

O.k, good point ;-), maybe this reasoning is worth to be more emphasized in
the commit message/patch-commit-message?

> I can create a dedicated patch but it doesn't make sense to send it
> upstream without first patch.

For upstream the check/error-out on obstack should depend on the
--{enable,disable}-progs option...

> I don't know if first patch was sent and if so, what was upstream reaction?
> An other option would be to disable elfutils on uclibc as I don't
> think we should maintain not upstreamable patch forever.

Give it a upstream chance... ;-)

Nevertheless you can add my:

Reviewed-by: Peter Seiderer <ps.report@gmx.net>

Regards,
Peter

> >
> > Regards,
> > Peter
> >
> >
> >  
> Best Regards,
> 
> Fabrice