Open expipiplus1 opened 1 year ago
Oof. Can you see a way to implement it? Might have to be removed.
As it is, the constraint production can't remain at the same type, because without a Prod which is positive in the eventual Constraint
type we have nowhere to push <*>
into.
The constraint
function however wouldn't need to change as it could easily forward to Constraint myProd (Pure myPredicate)
Ahh! This is actually annoying for passing in monadic/user state to the productions because if we put the constraining function inside a
Prod
(as is done for Many and Alts) then we can't discover the nulls without the user state, and hence can't use user state for avoiding rerunning the grammar!