Nick Coghlan <[EMAIL PROTECTED]> wrote: ... > > _temp = x.y > > x.y = type(temp).__irebind__(temp, z) ... > I was thinking of something simpler: > > x.y > x.y = z > > That is, before the assignment attempt, x.y has to resolve to *something*, but > the interpreter isn't particularly fussy about what that something is.
OK, I guess this makes sense. I just feel a tad apprehensive at thinking that the semantics differ so drastically from that of every other augmented assignment, I guess. But probably it's preferable to NOT let a type override what this one augmented assignment means; that looks like an "attractive nuisance" tempting people to be too clever. Still, if you write a PEP, I would mention the possible alternative and why it's being rejected in favor of this simpler one. Alex -- http://mail.python.org/mailman/listinfo/python-list
