Jupyter extension doesn't need to notify user of extension running code in their kernel nearly as often as it does.
greazer opened this issue · comments
When my extension calls executeCodeInKernel, the jupyter extension always notifies me. While I understand the thought behind this, it seems like overkill and can make for a bad experience if an extension consistently executes code in the kernel.
Seems like a one-time notification the first time this occurs after a kernel has been started would be sufficient. This is enough to tell the user that an extension is indeed running code in their kernel.
In addition, just making use of a kernel busy state indicator for additional extension executions, would be called for as well. The user could hover over this indicator and be given the reason that their kernel is busy.
The stop button on the notebook is appropriate for user-invoked cell executions, but isn't appropriate for a "system-invoked" execution.