On Sun, Apr 20, 2014 at 07:27:08PM +0000, Kristján Valur Jónsson wrote:
> Well, "for i in x" and other iteration constructs already call "iter
> ()" on their iterable. That's the point. Unless you want to manually
> iterate using "next ()" then the distinction between an iterable and
> an iterator is academic.
Do you think that for loops are the only way that iteritems and
friends get used? You can't know that. I know from my own code that it
isn't true. Iterators are first class objects like any other object, and
they get put into tuples, passed around as arguments to functions,
stored in variables, and so on. They get tested to be iterators, e.g.:
assert the_items is iter(the_items)
and, yes, sometimes people call next() on them directly.
There is a reason that viewitems() etc. were added to Python 2.7 rather
than just changing iteritems() to return a view. It would break backward
compatibility, and break people's code. How much code? I don't know, but
*any* breakage defeats the purpose of this suggestion. The whole point
of this PEP is for the iter* methods to work the same way in Python 2
and Python 3, not "nearly the same, sorry if it breaks your code".
If we wanted "nearly the same", we already have that: dict.items() in
Python 2 and 3 is already nearly the same, they are both iterable.
Nearly the same is not good enough.
If this is a cunning plan to make Nick's suggestion sound better by
suggesting an even worse alternative, it's working :-)
--
Steven
_______________________________________________
Python-Dev mailing list
Python-Dev@python.org
https://mail.python.org/mailman/listinfo/python-dev
Unsubscribe:
https://mail.python.org/mailman/options/python-dev/archive%40mail-archive.com
