Peter Scott writes:
> >You're right. If RFC 45 is implemented they would however be inconsistent.
>
> No, || is half-consistent at the moment: the left hand side is forced into
> scalar context but the result context propagates down the right hand
> side. I challenge anyone to come up with a rationalization for this that
> does not invoke implementation expediency.
As Damien observed, it's by the very definition of a compound assignment:
X OP= Y
is defined to be
X = X OP Y
So the only thing tested for truth with ||= is the LHS. Nothing to do
with implementation.
Nat
(PS, loved the obscenity analogy :-)
