On Sat, Oct 05, 2019 at 06:27:08AM -0700, Steve Kargl wrote:
> On Sat, Oct 05, 2019 at 03:18:45PM +0200, Rainer Orth wrote:
> > Hi Steve,
> >
> > >> the new testcases FAILs on sparc and a couple of other targets:
> > >>
> > >> +FAIL: gfortran.dg/pr91497.f90 -O (test for excess errors)
> > >>
> > >> Excess errors:
> > >> /vol/gcc/src/hg/trunk/local/gcc/testsuite/gfortran.dg/pr91497.f90:13:13:
> > >> Error: Old-style type declaration REAL*10 not supported at (1)
> > >> /vol/gcc/src/hg/trunk/local/gcc/testsuite/gfortran.dg/pr91497.f90:20:30:
> > >> Error: Invalid real kind 10 at (1)
> > >> /vol/gcc/src/hg/trunk/local/gcc/testsuite/gfortran.dg/pr91497.f90:55:22:
> > >> Error: Invalid kind for REAL at (1)
> > >>
> > >
> > > Does the attach patch fix targets without REAL(10) and REAL(16)?
> >
> > unfortunately not: I get
> >
>
> The failures are in a test with
>
> +! { dg-require-effective-target fortran_large_real }
>
> so that directive does not do what I thought. Guess I'll
> simply stripe out the testing of large REAL kinds.
>
Should be fixed with
Index: gcc/testsuite/ChangeLog
===================================================================
--- gcc/testsuite/ChangeLog (revision 276625)
+++ gcc/testsuite/ChangeLog (working copy)
@@ -1,3 +1,8 @@
+2019-10-05 Steven G. Kargl <ka...@gcc.gnu.org>
+
+ PR fortran/91497
+ * gfortran.dg/pr91497.f90: Run on i?86-*-* and x86_64-*-* only.
+
2019-10-05 Paul Thomas <pa...@gcc.gnu.org>
PR fortran/91926
Index: gcc/testsuite/gfortran.dg/pr91497.f90
===================================================================
--- gcc/testsuite/gfortran.dg/pr91497.f90 (revision 276625)
+++ gcc/testsuite/gfortran.dg/pr91497.f90 (working copy)
@@ -1,4 +1,4 @@
-! { dg-do compile }
+! { dg-do compile { target { i?86-*-* x86_64-*-* } } }
! { dg-options "-Wall" }
! Code contributed by Manfred Schwarb <manfred99 at gmx dot ch>
! PR fortran/91497
--
Steve
