Jacob Magnusson
Jacob Magnusson
Still can't come up with any other negative than that publishing crates that depend on cornucopia won't work. Just go with the crate solution already 😁 (Just kidding, it's great...
You are thinking things through, I like that. Good luck with the implementation! Can't wait to see the end result 😀
My vote goes to number two. And I disagree that a commit hook should be optional as that would mean human error is still in play. Or as an option...
Sounds reasonable 👍🏻
I didn't know SystemTime can be used to store historical dates..? Either way I think the feature approach I mentioned might make more sense though, as postgres-types does it this...
Okay, sounds good 👍
Isn't schemars 1.0 basically the same as 0.9? Edit: [Yup, basically the same](https://github.com/GREsau/schemars/releases/tag/v1.0.0)
@jplatte what about releasing 0.16 with only the schemars version change? Then do a 0.17 release when all the other stuff is ready?
I think that's a great idea. @Virgiel @LouisGariepy What do you think?
> which would imply that all occurrences of a postgres type would use the same user-specified type This is a fair assumption to make IMO. I can't come up with...