[Ironpython-users] Adoption of IronPython

Jeff Hardy jdhardy at gmail.com
Fri Jul 18 11:27:24 CEST 2014


On Thu, Jul 17, 2014 at 3:38 PM, Pawel Jasinski <pawel.jasinski at gmail.com>
wrote:

> Recently I had 2 ideas or perhaps suggestions which may help
> IronPython to get better.
> I asked myself, what would help us to get more general python packages
> to get IP specific patches and keep things alive. Most of them use
> travis as a test framework and cover most of the python versions,
> including pypy, under linux. So what is stopping us from testing the
> package under IronPython? Travis does not have windows support.
>
> So the idea 1, help good people of github to get windows support for
> travis.
> Or if you prefer, get mono in shape so it can be used as test
> environment for IP under linux.
>

I noticed this while working on virtualenv. They run tests on CPython and
PyPy but not Jython or IronPython. Making sure IP is part of their
regression suites is important, but the burden falls on us to get it setup
(which is fair IMO).

One of my goals for the B3 -> RC 2.7.5 release is much better Mono support.
I'm going to restrict myself to working on my Mac and see what happens (and
then try it in an Ubuntu VM just to double check). Supporting it on travis
could get some of the key projects that drive adoption.


>
> Second observation from porting exercise of bytes vs. chars vs.
> unicode. Every package keeps its own platform.py or similar which
> deals with 2.x/3.x differences. After n-th package, I thought "if
> there was one package for all to deal with it, someone could update it
> to handle not only 2.x and 3.x but also ip2.x". Guess what, there is
> already a package which does it and it appears to be used by a few
> projects. It is six.
>
> The idea 2, check what can be done about our notorious encode problem
> with help of six.
>

Also a good idea. If six-ified libraries also work on IronPython, that's a
big win for us. There's another, similar library whose name escapes me
right now.

The third thing (and the hardest) is that IronPython needs better docs.
Ideally we'd have a copy of the Python stdlib docs that called out
compatibility issues, and also guides on how to embed IronPython, which is
a critical use case.

- Jeff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/ironpython-users/attachments/20140718/f29efc39/attachment.html>


More information about the Ironpython-users mailing list