For testing of functionality behind feature flags, we currently rely on a script to generate the list of all feature flags, turn all of them on, then run all tests with those feature flags. While this does provide sufficient test coverage, it is also quite inefficient, especially as we are also running the same set of tests without any feature flags.

It would be really nice to map a certain code change to affected feature flags.

Shorter term, we can generate a list of feature flags that have definitively not been modified so we don't have to run any tests against them.

Longer term, it would be nice to enforce mapping of tests to either feature flags and/or some other basic unit of functionality in the Server codebase so that a direct dependency chart of JS tests and C++ code changes can be drawn.