This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
fixing PR32765
- From: Jack Howarth <howarth at bromo dot msbb dot uc dot edu>
- To: fortran at gcc dot gnu dot org
- Date: Fri, 10 Aug 2007 17:51:52 -0400
- Subject: fixing PR32765
Mike Stump looked at PR32765 and suggested several
ways to fix it. On the link line for the failing tests
libgomp.fortran/pr32550.f90 and libgomp.fortran/crayptr2.f90
we can either use -static-libgcc, -lgcc_s or libgomp.a.
I prefer using -static-libgcc and verified that both
testcases pass when that flag is passed for darwin. The
pr32550.f90 testcase will be trivial to patch. The change...
ndex: libgomp/testsuite/libgomp.fortran/pr32550.f90
===================================================================
--- libgomp/testsuite/libgomp.fortran/pr32550.f90 (revision 127350)
+++ libgomp/testsuite/libgomp.fortran/pr32550.f90 (working copy)
@@ -1,5 +1,6 @@
! PR fortran/32550
! { dg-do run }
+! { dg-options "-static-libgcc" { target *-*-darwin* } }
integer, pointer, save :: ptr
integer, target :: targ
will suffice. However the crayptr2.f90 testcase is more
problematic since we already have a dg-options comment line.
In c/c++ sources, multiple comment lines for dg-options seem
to co-exist with each being executed. However, I am finding
that when I add...
Index: libgomp/testsuite/libgomp.fortran/crayptr2.f90
===================================================================
--- libgomp/testsuite/libgomp.fortran/crayptr2.f90 (revision 127350)
+++ libgomp/testsuite/libgomp.fortran/crayptr2.f90 (working copy)
@@ -1,5 +1,6 @@
! { dg-do run }
! { dg-options "-fopenmp -fcray-pointer" }
+! { dg-options "-static-libgcc" { target *-*-darwin* } }
use omp_lib
integer :: a, b, c, d, p
Only the second dg-options line seems to be used. This makes me wonder
if the dejagnu harnesses for fortran are broken for such multiple
dg-options lines. If so, we will need to find out how to craft a
single dg-options line that achieves both. In any case, if I add...
! { dg-options "-fopenmp -fcray-pointer -static-libgcc" { target *-*-darwin* } }
to the crayptr2.f90 testcase it passes so we have the fix as soon
as we figure out the dejagnu syntax appropriate for these fortran
source files.
Jack