On Wednesday, August 5, 2015 at 10:29:21 AM UTC-6, Chris Angelico wrote:
> On Thu, Aug 6, 2015 at 2:10 AM, Rustom Mody <rustompm...@gmail.com> wrote:
> > 1 + x
> > does not *call* 1 .__add__(x)
> > It *is* that
> > [Barring corner cases of radd etc]
> > IOW I am desugaring the syntax into explicit method-calls so you can see
> > all the calls explicitly
> > Then it becomes evident -- visibly and in fact --that the tail call is the
> > __add__ method not the solderdiersVsDefenders
> 
> Except that it *isn't* that, precisely because of those other cases.
> When Python sees an expression like "1 + x" and doesn't yet know what
> x is, it can't do anything other than record the fact that there'll be
> a BINARY_ADD of the integer 1 and whatever that thing is. That object
> might well define __radd__, so the call is most definitely not
> equivalent to the operator.
> 
> And the ultimate result of that addition might not even be a function
> call at all, if it's implemented in C. Or if you're running in PyPy
> and the optimizer turned it into machine code. So no, even though you
> can define addition for *your own classes* using __add__ or __radd__,
> you can't reinterpret every addition as a function call.
> 
> ChrisA

Good (intricate) point. 
-- 
https://mail.python.org/mailman/listinfo/python-list

Reply via email to