[Scons-users] How important is python 3?

[Russel Winder]

On Wed, 2012-12-05 at 09:52 -0500, Robert Lupton the Good wrote:
[…]

> As of last August, Rob Pike told me that google had no plans to move...


But isn't his plan to replace all Python use in Google with Go use?
Vested interest ;-)


> I take your point, but my reading is that the python world is still

> largely thinking of python 2.7 as being the version to target for the

> next couple of years.  I suppose that we'll all have to move to

> python3 eventually (I just wish it offered a larger improvement over

> python2), but not until all systems I deploy on have python3 as the at

> least almost-default python, with all the usual scientific packages

> such as numpy and scipy pre-installed.


Well the Python world treats Python 2 as legacy and is focussed entirely
on Python 3 with backports to 2.7. The problems are the few things that
are not yet Python 3 compatible like Twisted Internet and SCons.

The most important changes in Python 3 are internal to the PVM and the
MOP, well apart from various large scale rationalizations of some of the
libraries.

There is no Python 2.8, so yes, Python 2.7 is the end of the Python 2
line, the only way forward is Python 3. We have started the "2 versions"
period. I guess it will be 3 or 4 years before people will have shifted.

In presenting my Python Workshop I find out quite a lot about who is on
which version of Python. Some folks are stuck with Python 2.5 as their
infrastructure suppliers say they have no intention of changing the
Python version ever.  Some (very large) organizations are about half-way
through the 2 → 3 shift already.

Ubuntu will be Python 3 even if RHEL is Python 2.

[…]

> Absolutely; that's why I haven't raised this point on the list before!

> However, it does mean that scons doesn't have spare developer effort.

> Being a volunteer project, if the active developers want to move to

> python3 that's going to happen and my desires won't change that.


The problem is how to gain traction on SCons user organizations putting
some resource into SCons development.  PyPy put out a begging bowl and
get funding. Given there is a SCons Foundation a legal entity
incorporated somewhere in the USA, SCons could do something similar.


> I have looked at the code base, and it's not obvious for how an

> experienced C/C++/python programmer would understand enough to fix

> things like this without a significant learning curve.   Yes; this is

> an excuse.


Not an excuse at all, a very good reason. This is why SCons needs
resource so as to put 2 or 3 man-days per week focused on SCons
development. This happening with Groovy and Gradle did massive things to
the quality of those systems. An interested party with money changes the
whole game.

-- 
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: sip:russel.winder at ekiga.net
41 Buckmaster Road    m: +44 7770 465 077   xmpp: russel at winder.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
Url : <http://four.pairlist.net/pipermail/scons-users/attachments/20121205/a5a0ecbc/attachment.pgp>