Zach Daniel
Zach Daniel
If the relationship is not a belongs to relationship, most likely yes.
@nallwhy its hard to reason about what we might do about this w/o a reproduction, or at least some information about how or why it failed. "it fails" is not...
Something definitely seems strange there. From what I can tell, `ash_archival` allows all strategies, and I can't find a reason we'd be resetting it or altering it. You're using `ash_postgres`...
Manual actions are now skipped by that check as of 867a84bad7f81765326cc7b8d5ea4754c3014141
The rest will be a bigger job, though, and will require more than a few changes so we'll have to leave it for now :)
This one is a bit complex, as right now validations are in control of setting their own values, and ultimately the only solution is to just add branching logic inside...
Interesting. I can see a case why we may not want to validate them, but there is a bit of a complexity here which is that casting input in Ash...
🤔 that is an interesting point. So in reality it sounds like our search case we should not be using the binary casting at all, but should be doing some...
Could you try cloning ash_graphql down and pointing its ash dependency at this branch and running tests? If they pass then I think we're good here.
🚀 Thank you for your contribution! 🚀