Open cloutiertyler opened 4 months ago
I agree is desirable to have concrete types. The issue is that dates have several:
https://docs.rs/chrono/latest/chrono/
This probably require a proposal
I'm going to leave this open, but I suspect this should really be defined by: https://github.com/clockworklabs/SpacetimeDBPrivate/issues/839
Wherein we would just choose to use a u64 or something so as to not impose our own opinion about how you should represent timestamps in a given language.
Mario: Then does this means adding chrono? Trying of use u64 instead internally is a footgun similar to how bool in SQLite is 1/0 (ie: Lets add proper date types, it also matters to match sQL spec)
If you use a timestamp in a table or as a reducer argument, then its going to generate a u64 which is unintuitive.