[Ironpython-users] Guidelines for embedding Python libraries in installers

[Cesar Mello]

Thank you very much for the explanation Jeff!

For my current prototype I have added the IronPython's MSI to the
bootstrapper; and the NumPy's \DLLs and \Lib folders to my products's dir;
and then I setup the host library path with both IronPython's folders and
the app's \DLLs and \Lib. It seems to be working nicely.

Just as a curiosity, a small weird thing I saw with NumPy was this issue:
http://stackoverflow.com/questions/5526151/ironpython-invocation-from-c-sharp-with-scipy-fails-with-importexception-no
.
But I added the reference and it works now.

Thanks a lot!

Best regards!
Mello





On Tue, May 8, 2012 at 12:47 PM, Jeff Hardy <jdhardy at gmail.com> wrote:

> On Tue, May 8, 2012 at 6:11 AM, Cesar Mello <cmello at gmail.com> wrote:
> > Hi!
> >
> > Please I would like to know if there is any recommendation for embedding
> > Python libraries together with IronPython in applications. I'm asking
> this
> > just to know if there is a path to follow towards some "ecosystem"
> regarding
> > Python libraries distribution.
>
> The standard home for Python packages is PyPI (aka "the Cheeseshop") -
> http://pypi.python.org. Python packaging in general is handled the
> standard distutils library, and the setuptools/distribute extensions
> (most of which are folded into distutils2, now known as 'packaging' in
> Python 3.3). Installation is done with easy_install (part of
> setuptools/distribute) or pip.
>
> One thing I'll give the Ruby guys - they got the gem ecosystem in
> place early and stuck with it.
>
> Now, IronPython doesn't support a lot of this right now. distribute
> and pip mostly work (there are a couple of tiny patches needed to them
> that I need to work on), but most of the pypi packages are bz2, which
> was a problem until I added bz2 a couple of days ago. Getting
> distribute and pip (and maybe virtualenv) working is one of my goals
> for 2.7.3.
>
> >
> > I have seen there is an installer for NumPy based on ironpkg here:
> > http://www.enthought.com/repo/.iron/
> >
> > Should that be a recommended approach to follow?
>
> That was a one-off created by Enthought for distributing NumPy. I
> don't think there's much reason to use it.
>
> Since I know you want to include the libraries, the easiest way (for
> now) would be to use a clean CPython installation and use easy_install
> to add the necessary packages (they are stored in Lib/site-packages).
> Then copy the site-packages directory into the IronPython stdlib so
> that it can all be deployed together.
>
> - Jeff
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/ironpython-users/attachments/20120508/f820239b/attachment.html>