[CI] Determine why codecov.io is reporting ~50% coverage instead of ~70%

[unlikelyzero]

Summary

After #4561 we're reporting at 53%. Let's investigate why.

Also, need to learn about #4605 (comment)

[akhenry]

Istanbul reports line coverage of ~ 70%

[akhenry]

@scottbell to investigate

[scottbell]

This PR in particular reduced coverage ~13%:
#4550

[scottbell]

I wonder if just adding some code coverage here:
https://app.codecov.io/gh/nasa/openmct/compare/4550/changes
would help things

[scottbell]

One thing @unlikelyzero would like me to check is if istanbul is busted, and what it is doing with test files themselves for test coverage.

[unlikelyzero]

Our istanbul config may need to be modified to ignore our spec files
https://github.com/istanbuljs/nyc#using-include-and-exclude-arrays

[scottbell]

Our istanbul config may need to be modified to ignore our spec files
https://github.com/istanbuljs/nyc#using-include-and-exclude-arrays

Good find!

[scottbell]

@unlikelyzero found the answer. We were excluding *Spec.js files in Codecov (#4550) but not in our istanbul reporter. Because of this, there was a big drop in code coverage as we removed our ~99% code covered test files from the considered coverage set.

To remedy, I think we should make exclusions & inclusions primarily in istanbul, and use Codecov primarily a dumb report viewer. I've got a PR opened here to do this:
#4714

[unlikelyzero]

@scottbell can you add some testing note? Thanks!

[scottbell]

To test, ensure Codecov is reporting the lower 56% number for code coverage, which should be what istanbul is reporting too.

[shefalijoshi]

Verified Fixed