Hacker News Reader

If you're running hundreds of very small tests, that one time initialization could take up a large portion of time.

So...maybe...consolidate the process of executing hundreds of little things?

How tiny do the steps in this thought process need to be? We can go on all night.

It's cleanest if each test runs in a new execution context, so that tests can't effect one another.

That's a naive solution to a tiny problem.