[CORE-134] Fix daylight savings time test bug

[trholdridge]

Ticket: https://broadworkbench.atlassian.net/browse/CORE-134

Previously, if the local timezone had an offset change in the next 30 days, certain tests would fail. I didn't figure out the full root cause, but it probably has to do with parts of the actual implementation using seconds (NihLink(eraCommonsUsername, iat + 60 * 60 * 24 * 30)) and the tests using days. Changing just minusDays to minusSeconds made it work, but I converted the plusDays as well for good measure.

The tests always passed in GH actions, so to replicate the problem and verify the solution (only within 30 days of a time changeover!):

git checkout 1646e20daaeca31c11e40fcc59e49986aa8833b;
TZ="America/New_York" sbt "testOnly *NihServiceUnitSpec";
git checkout 1e297af8e4fdc5b85f4faecab4f57fa43c1bdd81;
TZ="America/New_York" sbt "testOnly *NihServiceUnitSpec";

https://xkcd.com/2867/

[calypsomatic]

I think this makes sense!

[davidangb]

nice, thank you!