[IPython-dev] Poll on Deprecating %load_ext
Nicholas Bollweg
nick.bollweg at gmail.com
Thu Jul 9 14:50:14 EDT 2015
I've never used install_ext, so I am all for killing it!
Since someone else raised the entry_point thing as a drive-by: I do support
using this mechanism, vs. a homerolled solution, but see the limitations it
imposes. Namely, it would be unreasonable to load all installed
entry_points at runtime, so you would still need some mechanism to do so
selectively, and a magic probably still makes the most sense. It wouldn't
be crazy though, if after failing to find a magic, it went ahead and tried
to load that magic as an entry point :)
Bundled assets are another place where entry_points would make things
better... for python people. Can't speak to other kernels. For example,
jupyter-pip is the easiest way to do this today, but presently doesn't
(won't ever) work with wheels:
https://github.com/jdfreder/jupyter-pip/issues/15
so it probably wouldn't work with flit, which looks nice... though YAML
please :)
On Thu, Jul 9, 2015 at 2:25 PM Wes Turner <wes.turner at gmail.com> wrote:
> * http://mail.scipy.org/pipermail/ipython-dev/2014-April/013880.html
> (SEC: "%install_ext security and reproducibility")
>
> * [ ] setuptools entry_points for extensions
> * [ ]
> https://wrdrd.com/docs/consulting/education-technology#jupyter-and-reproducibility
> On Jul 9, 2015 11:07 AM, "Brian Granger" <ellisonbg at gmail.com> wrote:
>
>> I am +1 on deprecating it.
>>
>> On Thu, Jul 9, 2015 at 11:04 AM, Matthias Bussonnier
>> <bussonniermatthias at gmail.com> wrote:
>> > Hello list,
>> >
>> >
>> > TL;DR: We are considering deprecating the %install_ext magic[1] do you
>> have
>> > usages that are not covered by doing actual packages?
>> >
>> > Longer version:
>> >
>> > As we are refactoring and cleaning IPython for a 4.0 release we come
>> across
>> > lots of
>> > small pieces of code that have been around for a long time, and for
>> which
>> > the presence
>> > into IPython might only be historical reason.
>> >
>> > I recently came across new magics projects recommending to be install
>> using
>> > the %install_ext magic.
>> > As we all know, Python packaging system, have been improving the last
>> few
>> > years, and it is now relatively simple[2]
>> > to put a python package on PyPI. As you all know, or might learn, you
>> do not
>> > need to install a magic with %install_ext,
>> > you can just make it a python module.
>> >
>> > There is then no apparent reasons to keep %install_ext around, as we do
>> not
>> > want to reinvent the wheel, and %install_ext
>> > does not cover many areas that pip/conda/apt does. For example,
>> security,
>> > update, uninstall, venv ...
>> >
>> > So the question is: Do you use %install_ext for usage that cannot be
>> covered
>> > by pip or other package manager tools ?
>> > Do you have any opposition of us deprecating %install-ext ? Do you
>> approve
>> > of us getting rid of %install_ext ?
>> >
>> > You can reply to this thread or on the PR to give us your input.
>> >
>> > Thanks,
>> > --
>> > M
>> > [1]: https://github.com/ipython/ipython/pull/8601
>> > [2]: shameless plug for thomas project:
>> https://github.com/takluyver/flit
>> >
>> > _______________________________________________
>> > IPython-dev mailing list
>> > IPython-dev at scipy.org
>> > http://mail.scipy.org/mailman/listinfo/ipython-dev
>> >
>>
>>
>>
>> --
>> Brian E. Granger
>> Cal Poly State University, San Luis Obispo
>> @ellisonbg on Twitter and GitHub
>> bgranger at calpoly.edu and ellisonbg at gmail.com
>> _______________________________________________
>> IPython-dev mailing list
>> IPython-dev at scipy.org
>> http://mail.scipy.org/mailman/listinfo/ipython-dev
>>
> _______________________________________________
> IPython-dev mailing list
> IPython-dev at scipy.org
> http://mail.scipy.org/mailman/listinfo/ipython-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/ipython-dev/attachments/20150709/65c1caf7/attachment.html>
More information about the IPython-dev
mailing list