Potential fix for rdar://4658012

[Richard Guenther]

On 8/28/06, Richard Guenther <richard.guenther@gmail.com> wrote:
> On 8/28/06, Josh Conner <jconner@apple.com> wrote:
> > Richard Guenther wrote:
> > > On 8/26/06, Richard Guenther <richard.guenther@gmail.com> wrote:
> > >> On 26 Aug 2006 11:12:03 -0700, Ian Lance Taylor <iant@google.com> wrote:
> > >> > "Richard Guenther" <richard.guenther@gmail.com> writes:
> > >> >
> > >> > > Instead the following patch makes the C testcase above share the
> > >> stack
> > >> > > slot.  (completely untested)
> >
> > Richard -
> >
> > Thanks for looking into this, and creating the test case.  Could you
> > tell me in what way it fails, and on what platform?  I don't see
> > incorrect behavior reading through the assembly that is generated, at
> > least for ppc, and it at least compiles for linux-x86.
>
> It ICEd for me on i686-pc-linux-gnu.  Unfortunately it doesn't at
> work, but only at
> home(!?).

Btw., the basic idea was to create a testcase that passes/returns
gimple registers
which are forced to stack on the target ABI.  Which is why I chose
_Complex double.

Richard.