Run all |> within a fact irrelevant of if a previous one failed

[josephwilk]

This requires some serious macro magic which may not be worth it...

fact "test" do
  10 |> 11

  10 |> 22
end

Would result in 2 failures rather than 1.

[josephwilk]

Only used for internal testing but the idea is realised here in the fail predicate:

6dbeac5

In fail its a little simpler since we only need to wrap asserts in exception catchers.

[josephwilk]

First prototype for using this for all facts:

https://github.com/josephwilk/amrita/compare/multiple_runs_per_test?expand=1#L0R203

[josephwilk]

Completed in: https://github.com/josephwilk/amrita/tree/multiple_runs_per_test

One core problem is the alternative provided syntax and the errors mocks make.

Now we can, should we :)