Check Randomness in Phaser Tests

The phaser tests use partly fixed partly time-based seeds, so the choice of tests will affect behaviour.
@cameel wanted to look into them on a per-case basis.

I'd suggest to give all of them a fixed seed by default and add an soltest option with a nightly run that changes that behaviour to random seeds (and maybe higher sample sizes to avoid failures then, since runtime will be less of a concern).

Start by locating the phaser tests and the soltest option handling. Trace how their seeds and sample sizes are selected, then determine how a nightly run could enable random seeds; done means fixed seeds by default with a separate random-seed mode for nightly testing.

Written by the indexing model from the issue text.