[Scons-users] [scons-users] Support for Python 3?

[Russel Winder]

On Sat, 2012-06-09 at 18:16 -0400, Gary Oberbrunner wrote:
[...]

> A lot of the initial work was done a while ago; search the commit messages

> (or dev list archives) for 2to3 for some info.  Probably the best thing is

> to start a branch for it; from what I've heard I don't think it's possible

> to continue to support 2.4 (our current baseline) and 3 at the same time,

> but there's a lot I at least don't know (Unicode, how far back can we still

> support -- old python versions are still important, and probably other

> things).


A number of things for me to chip in:

— I want to ditch Python 2 as soon as possible, but my timetable has to
be aligned with the PyPy folk getting Python 3 compatibility so I can
choose between CPython and PyPy as implementation.

— Supporting Python 2 and Python 3 in a single code base is not easy but
it is possible, just.

— Everyone I am presenting Python training courses to is at least on
Python 2.5, though almost all are on Python 2.6, 2.7 and 3.2. I think
switching SCons base version from 2.4 to 2.6 is viable and doable.

— If the SCons codebase can have all the pre-2.6 bits removed only then
does it become viable to support 2 and 3 with the same codebase.

— SCons core needs a rewrite.

— SCons 3 could be a Python 3 only thing.

— Using 2to3 as a tool to keep a Python 3 codebase from a Python 2.6/2.7
codebase isn't viable, 2to3 is a one shot transform tool to switch from
Python 2 to Python 3.

— Keeping separate Python 2 and Python 3 codebases in sync is painful.

— Python 2.7 has almost all the Python 3 extras back ported, not all of
them are in 2.6.

— Moving the base of SCons to 2.7 might make it feasible to maintain a
single codebase usable with Python 2 and Python 3.

— Having a Python 3 codebase usable with Python 2.7 is not trivial but
is feasible.  Not sure about 2.6.

— All of this could happen if there was a bit of money to have someone
work on this for a few months full time.

— The current state of volunteer labour enables us to bug fix and do
small tinkerings, it isn't going to see the large development needed to
progress SCons properly.

— SCons needs a bit of resource, not necessarily money, a couple of
companies could assign people to work on SCons as their job for a few
months.

— The lack of significant progress is irritating a lot of people who are
not being vocal.

— Presenting Python training courses I am finding more and more people
who use SCons but never tell anyone they do.

— I wish I was in a position to do more than just send in emails and do
few things with plugins such as the D plugin. 

-- 
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/20120610/99640e01/attachment.pgp>