@binaryseed @bruce Let's take a beat and sort out what we want in place for a 1.5 release, and what we want to communicate with subsequent release numbers.
In my view 1.5 is basically good to go. We achieved a complete rework of the schema internals, and we have covered the ground we need to support certain new use cases.
It is entirely possible that we'll need more changes to support fancier SDL or extend or w/e that forces some minor breaking changes, and for that we can do a 1.6.
In particular, I really don't mind if we need to do a 1.6 in a month or two because handling extend or handling a :persistent_term schema backend required some meaningful changes. We aimed for too much with 1.5 I think.
I agree that 1.5 is ready to go!
Agreed. I'm gonna try to handle this https://github.com/absinthe-graphql/absinthe/issues/887 today, and I want to provide some clarity WRT hydrate in the docs, and then I think we cut a release.
Related to the release discussion, I wanted to make sure my question about the use_parent breaking change was seen before making a decision on the release: https://github.com/absinthe-graphql/absinthe/pull/882#issuecomment-605300660
Edit: this is now resolved by #891, :shipit:
I've added two other bugs that have been around for a while to the 1.5 milestone. Working on them now.
@benwilson512 Seems like all that's left is a little documentation for the hydrate additions, would you say that's right?
@binaryseed I'm trying to finish https://github.com/absinthe-graphql/absinthe/issues/683. In theory it could be a breaking change for anyone who was sending in the right data but accidentally had the wrong variable type specified.
I just pushed up PRs to fix two issues that have been recently reported:
Ok! I think Absinthe itself is good to go. Absinthe.Phoenix however needs updates to work with Phoenix 1.5. Technically, we could just go ahead and cut Absinthe 1.5 right now and say that we still only support Phoenix 1.4. I don't really like that though.
I'm working on Absinthe.Phoenix 2.0 which will support Phoenix 1.5. That should be done today. If it isn't done this week we should just go ahead and release Absinthe 1.5 regardless.
We did it, it's out!
Amazing!!
Most helpful comment
Ok! I think Absinthe itself is good to go. Absinthe.Phoenix however needs updates to work with Phoenix 1.5. Technically, we could just go ahead and cut Absinthe 1.5 right now and say that we still only support Phoenix 1.4. I don't really like that though.
I'm working on Absinthe.Phoenix 2.0 which will support Phoenix 1.5. That should be done today. If it isn't done this week we should just go ahead and release Absinthe 1.5 regardless.