> But the expression
>
> __ < 2 + __ * sin(__ / 2) or die __
>
> curries to
>
> sub { $_[0] < 2 + $_[1] * sin(sub { $_[0] / 2 }) or die $_[2] }
>
> That's very odd.
No. You got it exactly right. :-)
> I really hope I'm missing something, but __ the way you use it
> is too magical. It's not DWIM but DWISHMIIKWIWD (do what I should
> have meant if I knew what I was doing).
Perhaps.
> Your original statement that an ad hoc curried expression stops
> at a simple assignment sounds fine to me. (It actually fits in
> very well with functional programming which avoids assignment.)
>
> If that rule is modified by an explicit subroutine prototype that
> establishes a curry context then I think we can make this stuff
> work.
Okay. I'll try that in the next version of the RFC.
Damian