Add option to stop whenever a test fails

When debugging and trying to look at failed tests one by one, it can be cumbersome to always have to wait for all tests to finish or to have to scroll all the back up to the failed test you were looking at.

So when debugging test per test, this feature would be awesome, while being quite easy to implement.

The issue names no files, tests, or entry points. Start by tracing the test runner's existing failure handling and option parsing; done means users can enable a documented option that stops the run after a test failure, with coverage for the resulting behavior.

Written by the indexing model from the issue text.