Nathan Faubion

The compiler version is part of the externs themselves and is validated. However if we don’t even read the externs in some cases due to cache-db, that could be an...

It would need to carry an `IsProp` constraint. ```purescript value :: forall a r i. IsProp a => a -> IProp (value :: a | r) i ``` This can...

https://github.com/purescript/purescript/blob/fcf0acbcf11ad113820d32ce9ee82365a6015c9c/src/Language/PureScript/TypeChecker/Kinds.hs#L613-L614 This is a bad check here. We should zip the type variables with the signature if it's there, like is done in https://github.com/purescript/purescript/blob/f9f75f43bd1bdbc1e866fdba876582385e3f4c65/src/Language/PureScript/TypeChecker.hs#L566-L571

I'm personally not in favor of this change. We don't restrict attempts at impredicativity, so I think it would be weird to restrict this for records/rows (essentially because they are...

See https://github.com/purescript/purescript/issues/2748 and the discussion. This feature was removed some time ago.

Downsides to existential encodings: * You need a variation for every kind and number of variables, as long as you always keep the existentials at the end of the variable...

I think `forall` is appropriate in the data declaration. I know it seems wrong, but it corresponds to the CPS encoding of a datatype.

You can't have existential constraints with newtypes because they have to be bundled in the constructor. Constraints on functions are just arguments, and constraints on constructors are just values. You...

FWIW, regarding `spago install`, my experience mirrors @thomashoneyman's. I personally would be disappointed to see it gone. In more complex situations I've setup to encode targets, I've either constructed it...

It's that way "because". See https://github.com/purescript/purescript/pull/3200. Basically multiway-if would be more general, but there's nothing fundamentally preventing it's inclusion.