Adding ancestorTitles property to JSON test output

[mikehedman]

Summary
This addresses the need brought up in #2479

The motivation I have is that in order to properly report our test runs to regulatory agencies, we are planning on using the JSON output to feed reports. But our unit tests can be nested a couple of layers deep in 'describe' and 'in' blocks. Currently, the JSON output only includes the title of the nearest test block, thus, losing the context of its parents.
For example, we want to see:

• passwords
  • size
    • too short
    • too long
      and not just:
• too long

Test plan
I've altered the sum.test.js file to create tests with zero, one, and many ancestors, and added checks for the new ancestorTitles property.

[cpojer]

Fine with me. Thanks!

[cpojer]

What kind of regulatory agencies do you report results from Jest to?

[mikehedman]

@cpojer Thanks for merging this in.

To answer your question about regulatory agencies, I work primarily in the medical device field. So the agency that would ultimately get my results would be the FDA here in America, and the analogous agencies overseas.
Here's a link to a high level FDA guidance that applies to work I do: https://www.fda.gov/MedicalDevices/ucm085281.htm#_Toc517237962

In practical terms, generally raw unit test results are not required. But there are cases where unit testing (as opposed to what I would call "functional" testing, such as hitting a web page via Selenium) is really the only way to properly exercise certain functionality. In these instances, a company's Regulatory/QA group might supplement their standard testing with specific unit testing results. Companies can also augment their standard reporting with a higher level "oh yeah, and we also continually run 6000 unit tests on this product".

I've also worked a fair bit in the "payments" industry, which also cares a lot about software doing what it's supposed to ;) Here's a relevant guidance document: https://www.pcisecuritystandards.org/documents/terminal_security_best_practices.pdf
See section B8 where it expressly states: "Low-level unit testing of the software should be performed and recorded. This may require writing specific test software or modifying calling functions to specifically cover boundary checking and error handling."

I hope that helps.

[cpojer]

Thanks for the added context, that's really exciting!

[github-actions]

This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.
Please note this issue tracker is not a help forum. We recommend using StackOverflow or our discord channel for questions.