On Fri, Apr 20, 2001 at 03:15:30AM -0400, Tim Peters wrote:
> [Guido]
> > I've got a fairly complete implementation of iterators along the lines
> > of Ping's PEP (slightly updated).
> > ...
> > My question is: should I just merge this code onto the trunk (making
> > it part of 2.2), or should we review the design more before committing
> > to this implementation?

> My answer is both! *Most* of what you described is no longer controversial;
> 2.2 is mondo pre-alpha (so we're not "stuck" with anything you check in now);
> and it's much more convenient (for me - heh) to try out if it's in the
> regular build tree. I bet Greg Wilson would like it for his Set PEP work
> too, as abusing the __getitem__ protocol for set iteration is giving him
> headaches. WRT what may still be controversial points, there's no substitute
> for trying a thing.

I don't totally agree. Removing something from the trunk is not as easy as
not adding it ;) But I agree that, since the *concept* of iterators, and the
basic implementation, all are good things, they should be checked in. I
still don't like:

> > ...
> > - The test "key in dict" is implemented as "dict.has_key(key)". (This
> > was done by implementing the sq_contains slot.

> That's probably controversial, but also easy to rip out (sounds approximately
> self-contained) if the peasants storm your castle with flaming dungballs
> <wink>.

Fetchez-la-vache!-ly y'rs

(Oh, now I get it... Iterators are Guido's wooden rabbit, with 'key-in-dict'
hidden inside... I just hope it's Sir Bedevere that's building it ;)

--
Thomas Wouters <thomas@...>

Hi! I'm a .signature virus! copy me into your .signature file to help me spread!