[Code Coverage] Codecov.io reports appear to be mostly broken

[unlikelyzero]

Summary

I noticed that we had a codecov.io report for release/3.0.0 which had the following calculations:

• 48% total
• 57% for the e2e-stable flag
• 44% for e2e-full flag
• 37% for unit flag

How is it possible that our total is 48% but the e2e-stable tests cover 57%?

I switched back to 'master' and noticed something our lowest coverage was for 'plot/plugins/mctPlot.vue'. We which shouldn't near-zero (0.27%) as this area is heavily exercised in our e2e suites.

Opening the plot/plugins/MctPlot.vue revealed that the two lines which were covered appeared to be randomly distributed

We'll likely need to do a deeper investigation into why the code coverage is mostly inaccurate.

Probable causes:

• Sourcemaps mismatch between test suites and code coverage tooling (karma istanbul)
• Issues in npm run start:coverage
• Could be the same root cause as [Build] Code Coverage in codecov.io appears to have random noise #5811

[evenstensberg]

@unlikelyzero Could you explain the process of generating a coverage report? Reason for asking is that this PR is probably better to be done with someone that has CI and codecov permissions.

[unlikelyzero]

@evenstensberg I'm not ready to blame codecov.io quite yet. It's possible that we're generating the coverage incorreclty and codecov.io is just the bearer of bad news. Please see this documention update and provide feedback https://github.com/nasa/openmct/pull/7027/files

[msf-caesar]

Try with:

1. Double-checking the sourcemaps configuration for test suites and code coverage tooling (karma istanbul).
2. Verifying that the npm run start:coverage script is correctly set up.
3. Investigating if the issue is related to the known root cause mentioned in the GitHub issue.
4. Running additional tests and analyzing test results to identify any areas that are not being properly covered.
5. Considering updating the code coverage tooling or configuration if necessary.

[evenstensberg]

@unlikelyzero okay, I'll have a look at the build.