It turns out, TAJO-1191 is slightly more complicated than I thought. (https://issues.apache.org/jira/browse/TAJO-1191)
Basically, it's about whether TimeDatum may have a timezone tied with it. I **believe** TimeDatum is originally designed to hold a time period value, not an instant. (TimestampDatum seems to be the canonical container for instants.) So, it doesn't make sense to apply any timezones to TimeDatum values, but it's being done in a few places. And, that's why the test is failing on my machines. I'm going to try to fix it by removing all timezone-related references around the class, but I want to check my assumption with you before I proceed. What do you think about it?
