[Numpy-discussion] Guidance regarding build and testing

Chris Barker chris.barker at noaa.gov
Mon Dec 22 12:11:12 EST 2014


On Mon, Dec 22, 2014 at 7:40 AM, Maniteja Nandana <
maniteja.modesty067 at gmail.com> wrote:
>
> Thank you for the suggestion. I will remember this as a viable option. As
> of now, I have a virtual machine of ubuntu running on Windows. So, I wanted
> to have least overhead while running the VM.
>

As a note, virtual environments and conda environments are NOT like virtual
machines -- they add no runtime overhead whatsoever -- they do add some
disk space overhead if that's an issue (though conda environments use hard
linking wherever possible, so very little of that, either)

If you have a production use of python/numpy then it would be very helpful
to use a *environment system to keep your development work separate from
your production work. If not, then no reason not to work in the "
main" python environment.

A few other notes:

* You can certainly do numpy development on Windows, too, if you're more
comfortable there.

* I highly recommend "develop mode" when working on packages:

   python setup.py develop.

   What that does is put links in to the package source location, rather
than copying it into the python installation. That way, it looks like it's
installed (you don't need to do any path manipulations or change imports),
but changes to the source will be immediately available (if it's compiled
extensiun, they do need to be re-compiled, of course, which running
setup.py develop again will do.

* You are right, you can't accidentally mess up the main master branch when
you are working in a clone -- you don't have the permissions to do that,
and your clone is completely independent anyway. However, it's a good idea
to make a brach for a particular feature experiment, as it:
  - keeps thing cleaner an easier to keep tack of for your own work flow.
  - make it much easier to create a merg-able pull request if you decide
you's like to contribute your work back to the main project.

I'd read up on the git workflow to get an idea how to do all this. This:

https://sandofsky.com/blog/git-workflow.html

Is a pretty good intro -- and there are many others.

HTH.
 -Chris

-- 

Christopher Barker, Ph.D.
Oceanographer

Emergency Response Division
NOAA/NOS/OR&R            (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chris.Barker at noaa.gov
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20141222/e063e7da/attachment.html>


More information about the NumPy-Discussion mailing list