added Normalizable impl for Option<DateTime..

[Dowwie]

No description provided.

[sfackler]

Huh, how does a null bound compare to the bound not being present? Are they equivalent?

[sfackler]

I don't understand - it doesn't look like that code is ever creating a RangeBound with a None in it.

The key question is if the Postgres server understands a present-but-null bound and if it will ever produce one of those.

[sfackler]

The Ok(Range::new(None, None))? That constructs an unbounded range, not a range bounded by null on both ends.

[Dowwie]

@sfackler A RangeBound(None)? Is that even a range type in postgres? testing in psql, it returns an exception: DETAIL: Missing left parenthesis or bracket.

Bounds are requirements in Postgres

[sfackler]

Ok so it sounds like this PR isn't needed then, right?