V3 patch causes EH failures in ILP32 mode

[Gabriel Dos Reis]

Benjamin Kosnik:

> I'm trying to figure out how to machine-check this. So, you're saying that
>
>
> class nonpod
> {
> 	int a;
> }
>
> will be different than
>
> class nonpod
> {
> 	public:
> 	int a;
> nonpod(const nonpod& o) : a(o) {}
> }
>
> for:
>
> void foo(nonpod a) // not by ref?
> { }

Yes.  I believe I sent you a (private) message to that
effect.  You probably won't see the difference on x86, but
you'll see it on targets like PowerPC.  On such machine,
you'll get the address of the object in the first register
instead of its value.  You can probably do a diff
on the generated assemblies.

-- Gaby