Re: Window pointer

>
> The question is whether an implementation that does not store a Window
> pointer is non-conforming even if it resolves URIs as required. E.g.,
> XHR objects may have some private baseURI property which is updated if
> the associated Document's baseURI changes. That would ensure URIs are
> resolved correctly while allowing implementations to free the resources
> associated with the original Window. You would not be able to tell the
> difference from the outside, so the "MUST store Window pointer" is not
> generally testable, and untestable requirements create confusion as we
> have it here. Clearly such an implementation should be conforming.
>

Thanks. That's precisely my point. The storage is an implementation detail.

Subbu