[Python-Dev] [Python-checkins] r64424 - in python/trunk:Include/object.h Lib/test/test_sys.py Misc/NEWSObjects/intobject.c Objects/longobject.c Objects/typeobject.cPython/bltinmodule.c

[Guido van Rossum]

On Thu, Jun 26, 2008 at 8:21 AM, Nick Coghlan <ncoghlan at gmail.com> wrote:
> Barry Warsaw wrote:
>>
>> On Jun 26, 2008, at 10:40 AM, Nick Coghlan wrote:
>>
>>> Well, that and the beta deadline (we have to draw the line somewhere, or
>>> we'll be stuck in an eternal spiral of "just one more feature")
>>
>> Guido wanted to get the beta out when we did exactly so we could draw this
>> line in the sand.  I'd much rather people be spending time getting what
>> features we do have tested, stabilized, bug fixed, and turning the buildbots
>> green across the board.
>
> Having been caught by the 2.5 beta deadline with the changes that eventually
> became PEP 361 (and I think were significantly improved by the additional
> attention that was available due to the delay) I understand completely.
>
> (And to everyone with features that get bumped to 2.7/3.1 because of this...
> while a number of you no doubt know this already, it really is astonishing
> how soon the next release seems to roll around, even with our fairly
> leisurely release schedules!)

I'd like to separate the concerns though.

I personally don't want to see the feature in its current form added
to 2.7/3.1 either. As others pointed out, it's a wart on the
bin/oct/hex functions.

So as far as the feature design goes, I offer some suggestions: a new
module; or a new function in math; or a new method on float. Since
Raymond is the champion for the feature let him choose the API from
those alternatives.

Regarding the addition to 2.6/3.0 post beta 1, I think a new module
has the most chance of success, especially if it's written in Python
in such a way as to need minimal changes between 2.6 and 3.0.

-- 
--Guido van Rossum (home page: http://www.python.org/~guido/)