"Daniel P. Berrange" <berrange(a)redhat.com> wrote:
On Wed, Dec 17, 2008 at 09:57:23AM +0100, Jim Meyering wrote:
> john.levon(a)sun.com wrote:
> > Fix GCC hard-coding in python/
> ...
> > diff --git a/acinclude.m4 b/acinclude.m4
> > --- a/acinclude.m4
> > +++ b/acinclude.m4
> > @@ -75,6 +75,11 @@ AC_DEFUN([LIBVIRT_COMPILE_WARNINGS],[
> >
> > WARN_CFLAGS="$COMPILER_FLAGS $complCFLAGS"
> > AC_SUBST(WARN_CFLAGS)
> > +
> > + COMPILER_FLAGS=
> > + gl_COMPILER_FLAGS(-Wno-redundant-decls)
> > + NO_RDECLS_CFLAGS=$COMPILER_FLAGS
> > + AC_SUBST(NO_RDECLS_CFLAGS)
> > ])
> >
> > diff --git a/python/Makefile.am b/python/Makefile.am
> > --- a/python/Makefile.am
> > +++ b/python/Makefile.am
> > @@ -35,7 +35,7 @@ python_LTLIBRARIES = libvirtmod.la
> >
> > libvirtmod_la_SOURCES = libvir.c types.c libvirt-py.c libvirt-py.h
> > # Python header files contain a redundant decl, hence:
> > -libvirtmod_la_CFLAGS = -Wno-redundant-decls
> > +libvirtmod_la_CFLAGS = @NO_RDECLS_CFLAGS@
>
> That -Wno-redundant-decls option is no longer necessary, at least
> on rawhide. When I remove it, compiling with "make -C python
> CFLAGS='-Wredundant-decls -Werror'" still succeeds.
What about with older RHEL-5 python though ? Python 2.4 had rather a large
number of flaws in its public header file that have caused plenty of
compile warning problems in past.
It's no trouble for me either way.
I find that it is a losing battle to try to enable aggressive
compiler warnings on older systems, but don't feel strongly
about this particular case.
I mentioned it because there are new gcc-warning-related macros in gnulib,
and it'd be slightly easier to use them if the ones libvirt uses don't
have tendrils reaching into relatively distant Makefile.am files.