Marcel van Lohuizen
Marcel van Lohuizen
We are anticipating proposing a new model for closedness where embedding itself has no semantics regarding closedness, but rather we require the `x...` notation to recursively open up a reference....
Paul's comment of the original intent still stands. We first need a way to remove `Lookup` first, before introducing it as something else. This is quite arduous, as there are...
> I agree that the first step here is likely to fully phase out the existing Lookup method first, which I describe in Ah. Missed that. I think that phasing...
I've updated the testing setup for TestValues. It now tests for the particular error message. You can run ``` CUE_UPDATE=1 go test ./internal/core/subsume ``` To update the generated error messages....
It really should be top. And the API should be fixed if it is inconsistent. We probably should start coming up with a better API alltogheter. But for now it...
Brief update on slices: We are still quite undecided about this one. - we should have some form of slice notation - the current semantics is not ideal, so it...
I'm not entirely sure what the issue was there, so adding a test seems prudent.
FTR: a simplified reproducer: ``` d: d.a d: a: a: 2|1 ``` and ``` a a: a: 2|1 ``` Also related ``` b: a b: (1|2) | 2 a: 1...
BTW, why do you think this should be a cycle error?
See also #4089