Native Notebooks are creating a new kernel on any execution after moving to a different document

[IanMatthewHuff]

Pretty bad issue right now, especially as I don't think that the old kernel is getting shutdown. It will do this again and again for me when moving to another document.

Have one suspicion that I'm looking into right now.

[greazer]

Not able to repro this on shipping extension. May be a regression that was recently caused (but thankfully not shipped)

[DonJayamanne]

I tried, & cannot repro this issue.

[DavidKutu]

I'm also unable to repro this.

[IanMatthewHuff]

@DavidKutu @DonJayamanne I'll take over investigation on this.

[IanMatthewHuff]

Did some more work to pin down at least one repro. Scenario is as follows. Python extension required for this scenario.

1. Create a new jupyter notebook
2. Run this notebook with a selected installed kernelspec, not just an interpreter (we need 'startUsingKernelSpec' as our connection type)
3. Change the active interpreter
4. Do something to retrigger vscode kernel provider fetching swap to a different document and back
5. Execute a cell again (it will reconnect to a new kernel instance)

[rchiodo]

Validated. Does not create a new kernel after following the steps above.