What is fair in testing and what is being tested? If I can stub `secure_hash` and `bar`, that’s a pretty easy test to write. Assuming the symbols are visible at

What is fair in testing and what is being tested? If I can stub `secure_hash` and `bar`, that’s a pretty easy test to write. Assuming the symbols are visible at test time, stubbing those is relatively straight forward in C, Java, Bash, and Ruby, JavaScript, etc. Even in the above examples, a linter should point out that the second `p == null` is inaccessible. I’m very pro test coverage, but static analysis shouldn’t be thrown out just because unit tests are available.