[Python-Dev] RFC: PEP 460: Add bytes % args and bytes.format(args) to Python 3.5

Georg Brandl g.brandl at gmx.net
Tue Jan 7 11:33:55 CET 2014


Am 07.01.2014 10:59, schrieb Paul Moore:
> On 7 January 2014 09:40, Georg Brandl <g.brandl at gmx.net> wrote:
>> Very nice, thanks.  If I was to make a blasphemous suggestion I would
>> even target it for Python 3.4.  (No, seriously, this is a big issue
>> - see the recent discussion by Armin - and the big names involved show
>> that it is a major holdup of 3.x uptake.)  It would of course depend
>> a lot on how much code from unicode formatting can be retained or
>> adapted as opposed to a rewrite from scratch.
> 
> Will the relevant projects actually support only 2.X and 3.4/5+? If
> they expect to or have to support 3.2 or 3.3, then this change isn't
> actually going to help them much. If they will only support versions
> of Python 3 containing this change, then it may well be worth
> considering the impact of delaying it till 3.5.

Yes, exactly.

Another, and probably better, proposal would be to make 3.5 the
"ultimate" viable porting target: we now know pretty well what the
major remaining roadblocks (real and perceived) are for our developers
and users.

The proposal would be to focus entirely on addressing these roadblocks
in the 3.5 version, and no other new features -- the release cycle
needn't be 18 months for this one.  This is similar to the moratorium
for 3.2, but that one came too early for 3.x porting to really profit.

In short, I am increasingly concerned that although we are going a
pretty good way (and Nick's FAQ list makes that much clearer than
anything else I've read), but it is not perceived as such, and could
be better.  We have brought Python 3 on the community, and as such
we need to make it very very clear that we are working with them, not
against them.  A minor release dedicated to that end should be a
very direct representation of that.

I know about the "release everything to PyPI" strategy, but it just
doesn't have the same impact.

It would be very cool to have multiple projects working together with
us for this, and at the release of 3.5 final, present (say) a Mercurial
that works on 2.5 and 3.5.

Mostly pipe-dreams though...
Georg





More information about the Python-Dev mailing list