Re: avoiding forward declarations in stencil.hh

[David Kastrup]

David Kastrup <address@hidden> writes:

> "address@hidden" <address@hidden> writes:
>
>> On Jan 30, 2012, at 2:49 PM, David Kastrup wrote:
>>
>>> logical impossibility
>>
>> This is all I needed to know - thanks!
>
> Well, it forces
> sizeof(x) <= 2*sizeof(x)

sizeof(x) >= 2*sizeof(x) of course.

> and we know sizeof(x)>= 0, so it is not logically impossible as long as
> the structure does not contain anything else beside two structures of
> its own kind.
>
> But there won't be room for payload then.
>
> Using a pointer is ok, and SCM is kind of a pointer.

-- 
David Kastrup