Enable test parallelization.

[theengineear]

Currently, each test runs in an iframe one-by-one. There’s no real reason why we couldn’t parallelize them. The hardest part would be figuring out how to conceptually merge the tests streams so that the output is legible as a single test stream.

It should be possible to reliably report all tests in the exact same order in the output based on the subtest discovery algorithm. The potential downside is that a test could finish, but won’t appear in the output until it’s turn is up. This may or may not be a bad experience.

[theengineear]

This isn’t super high-priority, but I wanted to get the idea written down so that it can be revisited later if needed.