I believe their utility has been superseded by several things:
- Failing tests are detailed in the build output, without navigating through JUnit a report.
- If there are any deficiencies there, we should identify and fix them because afaik most if not all developers look solely at the phpunit build output to understand failing tests. Both locally during development, and in WMF CI. Also in other CI (GitLab, Travis, GitHub) focus has been for many years on improving the console formatting. There is very little momentum behind facciliating HTML-based GUIs for continuous integration.
- Slow tests are reported in the build output (thanks to SpeedTrapListener).
- I know a few of us used to find these by sorting the JUnit report tables, but the SpeedTrapListener report seems to have superseded this.
- We run tests pre-merge and disallow merging of failing tests.
- This means much of Jenkins' JUnit viewer's capabilities don't work or produce incorrect/deceptive information. E.g. about a test having recovered or improved to run faster, because it is comparing them to an unrelated previous CI build for another repo/branch/commit.