Martijn Faassen: The Call of Python 2.8

[Chris Angelico]

On Mon, Apr 14, 2014 at 11:51 PM, Marko Rauhamaa <marko at pacujo.net> wrote:
> Chris Angelico <rosuav at gmail.com>:
>
>> If you're going to do that, why not just port your code to 3.x and be
>> done with it? Who has the resources to put hours and hours of dev time
>> into a 2.8?
>
> Somewhat related. Only yesterday I ported/reimplemented a software
> package to python3. On the finish line, I ran into a problem: xlwt
> only supports 2.6, 2.7 and 3.3. My system has python3.2.
>
> So I backtracked to python2.7.
>
> So not only do we have a schism between python2 and python3 but there's
> one between 3.0 and 3.3. I can't help but wonder if PEP 414 was a
> mistake.
>
> Serves me right for being an "early adopter."

So get Python 3.3 for your system, then. It's not that hard. You might
need to build it from source (not hard at all), or grab packages from
a newer version of Debian/RHEL/etc (also not hard, although there
might be additional consequential package requirements). The two
should happily coexist.

Also, the EOL for Python 3.2 is way *way* nearer than EOL of the 2.x
line. If you declare that your package requires 2.6/2.7/3.3
(preferably also support 3.4), so be it. It won't be long before all
supported systems can get 3.3+, so that won't be a problem. PEP 414
was useful because we can confidently target a newer 3.3 and expect
that people will be able to get there before long.

ChrisA