pool: Break on failed pool tests.

[davecgh]

Any calls to t.Fatal inside of a t.Run will only fatal that specific instance of run. This adds logic to stop running the remaining pool tests when there is a failed test.

[jholdstock]

There is already a built-in flag -failfast which you can pass to go test for this behaviour. Hard-coding it and removing the togglability seems a bit heavy handed. Personally I like to know how many tests are broken so I can review them all in one pass, rather than having to repeatedly "fix one test, run tests, fix one test, run tests, etc"

-failfast
    Do not start new tests after the first test failure.

from https://pkg.go.dev/cmd/go/internal/test

[davecgh]

I'm aware of the flag, but I'm not sure how else we can avoid purging the database on a test failure (which #389 does) without adding these.

In general, I like to see all failing tests too, but the way these tests are written is not at all conducive to that. It cost me quite a lot of time not being able to see the data left in the database.

I'm open to another method, but after having just suffered through it, I really don't think we want the database purged on test failures. It makes it exceedingly difficult to analyze test failures when you nuke the data.

[davecgh]

Alright, I thought of a way to allow -failfast to work and still allow #389 to retain the data but not otherwise. Closing and will update #389 accordingly.