In-code creation of raids causing identical timestamp to differ due to microsecond differences.

[GabeStah]

In DummyDataWorker for example, creation of raid occurs as expected in the database, rounding the timestamp to the nearest second:

after.to_f                   = 1412446435.3014085
participation.timestamp.to_f = 1412446435.0

This causes filter checks to fail because the values do not match.

Possible solution:

1. Force rounding before checks occur.
2. Use PostgreSQL to allow microsecond data saved in database.