Conda is well suited to this. I use it to bundle all sorts of stuff on Windows. (You write recipes (see https://github.com/conda/conda-recipes for examples), then 'conda build' them, which produces a package that can be subsequently installed with conda install. Can sign up to anaconda.org and then upload the package into your own channel, such that a plain 'conda install -c janssen foobar' will install your package and all the deps (which were specified in the recipe/meta.yaml).
Sent from my iPhone > On Dec 16, 2015, at 13:00, Bill Janssen <jans...@parc.com> wrote: > > I'd like to build a Python-based deliverable for Windows. It includes > many gnarly packages, like numpy, scipy, statsmodel, ggplot, kivy, ZODB, > ZEO, etc. They include Cython modules (and scipy may even require > Fortran, for all I know). > > On OS X, I build this all from source by starting with Kivy, which is > packaged as a venv inside an OS X application, and add in the other > stuff. But I'm not sure this is the best way to proceed on Windows (7, > 8, and 10). I'm also used to using mingw on Windows, but again, I'm > not sure that's appropriate. > > Any advice would be appreciated... > > Bill > _______________________________________________ > python-win32 mailing list > python-win32@python.org > https://mail.python.org/mailman/listinfo/python-win32 _______________________________________________ python-win32 mailing list python-win32@python.org https://mail.python.org/mailman/listinfo/python-win32