gap-packages/JupyterKernel

Use py-gap to do the jupyter communication in python

Opened this issue · 3 comments

With py-gap working reasonably well, use the Jupyter implementation of the Jupyter protocol and py-gap to execute the GAP bits.

This removes the maintenance load for the Jupyter protocol, and gets rid of some ugly hacks that were necessary to implement the protocol (such as using IO_fork). See also #45,

First steps here: markuspf/py-gap@5b7b0a4

This might also address any issues that exist on CoCalc (or at least make them an SEP)

py-gap is AFAICT dead, but there is gappy which is actively maintained and may (eventually?) be used by SageMath.

yet another alternative would be to use Xeus, the official C++ bindings for Jupyter. They don't look that hard to use (famous last words 😂)