Gavin King
Gavin King
> This is a bug, the method ` Join join(String attributeName);` should be ` Join join(String attributeName);` instead. I agree. We should fix this in the next major version.
I've added the license header, and finished the Javadoc. I would still like to know if anyone thinks we should make this more typesafe, so that, for example, you can't...
> See https://github.com/jakartaee/persistence/issues/377 Yeah, I think this and #377 are essentially the same issue. Or at least the solution would be the same.
> Can you sketch out how the type safety could look like? Well, I would either - overload `extract`, and have `DateField` and `TimeField` instead of `TemporalField`, or - add...
(And yes, I agree that the value of the leverage of doing this is pretty marginal, which is precisely why I didn't do it in the initial proposal.)
I mean, if everyone is happy with it, and wants it, then I think it's fine.
Folks, I've finally gotten round to making a decision on this, after chatting with @lukasj, and I've updated the PR to add a more typesafe version of the API.
> `OffsetDateTime` Ideally yes, that would be nice, except that then we would be opening the whole incredible can of worms of what is the actual _semantics_ of `OffsetDateTime` in...
> If you think this is important, then the thing I guess I would prefer to do is specify that you can cast from `Instant` to `LocalDateTime` using the `cast()`...
> Of course it is platform dependent, but that IMO shouldn't stop us from providing these useful fields for a commonly used data type. Right, but the question is: if...