This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Testsuite fixes for xstormy16
- From: Geoffrey Keating <geoffk at geoffk dot org>
- To: Nick Clifton <nickc at redhat dot com>
- Cc: gcc-patches at Gcc dot gnu dot org
- Date: 12 Nov 2004 13:38:14 -0800
- Subject: Re: Testsuite fixes for xstormy16
- References: <m3k6st91mj.fsf@redhat.com>
Nick Clifton <nickc@redhat.com> writes:
> Hi Guys,
>
> Please could I have permission to apply the following patch ? It
> fixes several of the tests in the GCC and G++ testsuite so that they
> will work with the xstormy16 toolchain (a 16-bit target) or else
> labels them with an expected failure if they are testing a feature
> that the xstormy16 port does not support.
>
> Cheers
> Nick
>
> gcc/testsuite/ChangeLog
> 2004-11-10 Nick Clifton <nickc@redhat.com>
>
> * g++.dg/opt/strength-reduce.C: Replace 0x7fffffff with
> __INT_MAX__ so that the test can be compiled by toolchains
> targeted at 16-bit processors.
>
> * gcc.dg/20040625-1.c: Use long constants rather than integers
> when performing arithmetic that will overflow 16-bits.
>
> * gcc.dg/i386-sse-8.c: Expect this test to fail for the
> xstormy16. It does not support vectors.
This description is not right; the problem is not that xstormy16
doesn't support vectors, at least it should support, say, V2QI fine;
the problem is that it doesn't support 128-bit operations like V2DI.
> * gcc.dg/ultrasp4.c: Replace all occurrences of the type 'long'
> with the macro PTR_TYPE. Define this macro to an integer type
> suitable for holding a pointer.
>
> * gcc.dg/compat/struct-align-1_x.c: Use shorter integers if
> the target is a 16-bit processor.
>
> * gcc.dg/cpp/if-2.c: Expect a different warning message from
> 16-bit targets where the character constant will be too long.
>
> * g++.dg/opt/asm1.C: Expect this test to fail on the xstormy16
> where the double will not be directly addressable.
Otherwise OK, go ahead and commit with that change.