Ruin programs, benchmark your testing.
Ideally, software projects contain large test suites that verify the majority of its behavior.
Is this ideal met in practice? Who knows.
Enter West, a tool which aims to provide benchmarking capabilities to your test suites by determining how well they detect changes to your program.
Specifically, West works by mutating Javascript programs and then confirms whether your test suites detect these changes as failures.
A tenet of this tool is the following (unproven) claim:
"Assume a monkey hits a keyboard for a while and suddenly finds itself with a copy of your software project. If the monkey then makes a bunch of random changes to your program, those changes are likely unwanted."
west /path/to/working/dir "cmds --needed to --run tests"
For now it assumes the tests run locally, and simply copying the files in the working directory and running the test commands will trigger the tests.
If the test command returns 0, we assume tests have passed. Else we assume some kind of failure.