This certainly does look like a bug. However, AC_MSG_RESULT should not
be needed because the AC_CACHE_CHECK should already have printed out the
result. I would suggest the following simpler patch instead:
2001-03-13 Steven G. Johnson <[EMAIL PROTECTED]>
* aclang.m4 (AC_PROG_F77_C_O): define F77_NO_MINUS_C_MINUS_O
when test fails, not when it succeeds(!).
===================================================================
RCS file: /cvs/autoconf/aclang.m4,v
retrieving revision 1.123
diff -u -r1.123 aclang.m4
--- aclang.m4 2001/02/19 15:11:52 1.123
+++ aclang.m4 2001/03/13 06:10:44
@@ -1330,7 +1330,7 @@
ac_cv_prog_f77_c_o=no
fi
rm -f conftest*])
-if test $ac_cv_prog_f77_c_o = yes; then
+if test $ac_cv_prog_f77_c_o = no; then
AC_DEFINE(F77_NO_MINUS_C_MINUS_O, 1,
[Define if your Fortran 77 compiler doesn't accept -c and -o together.])
fi
