You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Just a thought, if we got rid of the ipython sphinx directives, how will the docs build on systems without ipython? They do exist, for example build machines, or someone just simply looking at the pre-reqs.
lib/matplotlib/sphinxext/ipython_directive.py depends on IPython. Though I admit that lib/matplotlib/sphinxext/ipython_console_highlighting.py doesn't.
It just feels like we're stamping on some toes by having IPython helpers in the matplotlib source. Their home really should be in the IPython core (or some spin-off package).
Yeah -- the ipython_directive was written by John Hunter, so I think it just started life in matplotlib, though it probably shouldn't have. It isn't a requirement of matplotlib's documentation build.
ipython_console_highlighting doesn't depend on IPython, but it probably logically belongs in IPython. Though I understand the need to keep the matplotlib build from depending on IPython. In that case, perhaps we should move it to our private sphinx extension directory (i.e. not install it), and encourage third-party users of the extension to get it from IPython.
Once ipython/ipython#3352 is merged and has made it into a release, we should deprecate our copies of the ipython sphinx directives.
The text was updated successfully, but these errors were encountered: