[Scons-users] SCons vs Meson for Python projects

Russel Winder russel at winder.org.uk
Fri Nov 25 10:47:25 EST 2016


On Fri, 2016-11-25 at 11:50 +0100, Saša Janiška wrote:

> […]
> Probably…still I do not want to go (back) to C++.

Fair enough.

> […]
> Numba might be interesting, indeed. However, for that kind of
> optimizations it is still too early, so the usage of Cython is more
> geared towards providing Pythonic API for 3rd party C library,
> iow. using Cython as bindings generator.

Is SWIG a better way of constructing API adapters?

>[…]
> Chapel is probablly an overkill for my (current) needs. Otoh, I
> looked
> at e.g. D, but for GUI app, PyQt is still much better (considering I
> do
> not want to use Qt via C++).

D doesn't have a viable Qt binding yet, but something may be happening
on that front. Qt is obviously the main way of using Qt but PyQt is far
superior in my view. But I am doing GTK+3 just now as I only have to
deal with Linux. D beats C++ for this, and I am dithering over D vs
Python.

> […]
> PyChapel is interesting, thank you for the pointer.

Chapel isn't a general purpose programming language, it is very much a
language for describing parallel and concurrent computations across
multi-multicore-processor clusters.
 
> […]
> Just wonder how would PyChapel work with PyQt…

I think the "vision" is for PyChapel to allow people to use Matplotlib
to render visualisations of the results of a Chapel computation. And as
a vehicle for slowly replacing NumPy with Chapel for grunt computation.

However I am not on the strategy committee. We will get a workshop and
keynote on Chapel at ACCU 2017 (http://conference.accu.org ) from the
Chapel project lead, so he would be the person to quiz on this.
 
> […]
> I’m still trying not to enter into Java/JVM and keep my life simple
> with
> the Python, but on the long run I consider SCons along with enscons
> (https://pypi.python.org/pypi/enscons/) better than fiddling with
> distutils/setuptools.

Mixing native code and JVM code is probably still not a really viable
thing despite that many, many people do it. Python + native code works
much much better.

> > Cython is handled, and there are Qt tools.
> 
> OK.

SWIG is also handled.

> > There was a move to get RST and Sphinx support but contributions
> > faded
> > away.
> 
> That’s sad…even CMake has it.

I have given up on ReStructuredText, and now use Asciidoc when I do not
use LaTeX. I am unlikely ever to go back to ReStruturedText as the
source relies on being rendered in a monospace font to be usable, and I
only use proportional ones. Clearly this is my problem, but it stops me
being interested in Sphinx, and in the build frameworks for it. Sorry.


-- 
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: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: This is a digitally signed message part
URL: <https://pairlist4.pair.net/pipermail/scons-users/attachments/20161125/52cb87d3/attachment-0001.pgp>


More information about the Scons-users mailing list