Committed: fix g++.dg/init/array16.C, skip for mmix-*-*. PRtarget/18911. Testsuite issues.

[Hans-Peter Nilsson <hp at bitrange dot com>]

See the PR, where it says the purpose of the test is compilation
anyway. It timed out for the MMIX simulator, and scaling down
the figures in the PR it seems it would time out for the CRIS
simulator (CGEN-based; quite fast) if the host was e.g. a 133MHz
P4.  Which would be a bit underdimensioned these days, perhaps,
but IMHO too close to the limit.

I had hoped to do this:
-// { dg-do run }
+// { dg-do run { target native } }
+// { dg-do compile { target { ! native } } } Causes timeout for slow simulators.

but that doesn't work: apparently the *keyword* from the last
dg-do is used with the *condition* from the first.  (At least it
seems so when swapping those two lines.)  Not that this
contradicts the documentation (I see nothing about multiple
dg-do:s), but at least it's a bit surprising.

It can be argued to make the test compile-only for all targets,
but as long as it can run anywhere, I guess it should.  Instead
I committed this, tested on mmixware-sim, cris-elf and native
powerpc-linux.

	PR target/18911
	* g++.dg/init/array16.C: Skip for mmix-*-*.


Index: g++.dg/init/array16.C
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/g++.dg/init/array16.C,v
retrieving revision 1.2
diff -p -u -r1.2 array16.C
--- g++.dg/init/array16.C	31 Aug 2005 17:08:39 -0000	1.2
+++ g++.dg/init/array16.C	23 Oct 2005 20:47:02 -0000
@@ -1,4 +1,6 @@
-// { dg-do run }
+// Causes timeout for the MMIX simulator on a 3GHz P4 and we can't
+// have "compile" for some targets and "run" for others.
+// { dg-do run { target { ! mmix-*-* } } }
 // { dg-options "-mstructure-size-boundary=8" { target arm-*-* } }

 // Copyright (C) 2004 Free Software Foundation, Inc.

brgds, H-P