PostgreSQL returns time.Time in UTC with different nanoseconds

[erizocosmico]

Not sure we can do much, but it's worth adding on the README as known limitations and keeping it here until we figure out a workaround.

[erizocosmico]

Tried a PoC of using the unix timestamp. But that does not contain milliseconds as well, so... same thing.

[Serabe]

More information about this lib/pq#227

I agree with the proposed solution and would love to see it coming from pq. In the meantime, can we hack something on top?

[erizocosmico]

The problem is not pq, is postgres. There's just no space in a timestamptz field to store nanoseconds. Don't think there's gonna be a solution for that any time soon.

So I think it's better to just leave this open as a reminder (or document it on the readme as known issues).

[erizocosmico]

Discussed this IRL with @Serabe. Solution will be to truncate all time fields before saving.