Clear all output should also delete execution count

[chrislill]

Issue Type: Bug

Clear all output is an important feature when using source control on a long running project. We use it so that a diff only identifies code changes, rather than minor changes in output or execution count. Standard functionality in Jupyter is that execution count is cleared when clear all output is selected. Execution count is considered to be part of the output.

Steps to repro:

1. Run some cells in a notebook in VS Code
2. Click Clear All Output
3. Inspect Execution Count
4. Check in to git
5. Add a cell in the middle of the code
6. Rerun the notebook
7. Inspect Execution Count
8. Do a diff

Expected results:

1. Execution count is empty
2. The diff only picks up the new cell

Actual results:

1. Execution count is still populated
2. The diff picks up the changed execution counts

Extension version: 2019.10.41019
VS Code version: Code 1.38.1 (b37e54c98e1a74ba89e03073e5a3761284e3ffb0, 2019-09-11T13:35:15.005Z)
OS version: Windows_NT x64 10.0.18362

System Info

Item	Value
CPUs	Intel(R) Core(TM) i7-6600U CPU @ 2.60GHz (4 x 2808)
GPU Status	2d_canvas: enabled flash_3d: enabled flash_stage3d: enabled flash_stage3d_baseline: enabled gpu_compositing: enabled multiple_raster_threads: enabled_on native_gpu_memory_buffers: disabled_software oop_rasterization: disabled_off protected_video_decode: unavailable_off rasterization: enabled skia_deferred_display_list: disabled_off skia_renderer: disabled_off surface_synchronization: enabled_on video_decode: enabled viz_display_compositor: disabled_off webgl: enabled webgl2: enabled
Load (avg)	undefined
Memory (System)	15.93GB (6.75GB free)
Process Argv	C:\Users\clill\Source\Repos\handson\housing.ipynb
Screen Reader	no
VM	0%

.

[rchiodo]

Thanks for the feedback. Seems like execution count should be cleared to me too.

[DonJayamanne]

@DavidKutu Please continue working on this.
I think the saved notebook file still has the execution count.
I.e. currently the UI gets updated, however the raw ipynb file still has the execution count. I didn't mean to commit the changes to remove the cell exec count (as mentioned inthe PR review comments, it was accidental)

[DonJayamanne]

The execution_count is still in the ipynb file (open in regular editor).
Based on bug report, this shouldn't exist.

Actual results:
1. Execution count is still populated
2. The diff picks up the changed execution counts

``

[DonJayamanne]

Moving back to back log to ensure we add tests for this.

[DonJayamanne]

Tests have been added