Sounds like a bug, but I don't think we'll be able to help without seeing your project. If you're concerned about attaching it here, you can also send it just to Apple using https://bugreport.apple.com.

I have uploaded a zip file containing some of the code in my project.

Thanks! Reproduced, and still seems to occur on master.

@swift-ci create

There's definitely something really strange going on here.

If I make delimiterItem a parameter of the appropriate type (Parser<Substring, ((TableDataAlignment) -> ([InlineNode]) -> BlockNode) -> ([InlineNode]) -> BlockNode>), I do not hit the error. If I use an explicit type on the local variable, I do still hit the error.

I've attached a minimized reproducer.

The diagnostic here is not good, but I think the change in behavior here (of disallowing this) may be correct. I'll have to take a closer look to confirm.

The issue is that if the curry function is used in a context where U is a non-@escaping function type, we previously would allow that function to escape, but no longer do.

public func curry<T, U, V>(_ f: @escaping (T, U) -> V) -> (T) -> (U) -> V {
    return { x in { y in f(x, y) } }
}

Ahh, so the diagnostics issue is SR-5184 about not printing @escaping when relevant. Good catch.

@octree Looks like we produce a much better message now in Xcode 11.4 which points to mismatch in result type of the `apply`:

error: cannot convert return expression of type '(Y, [(@escaping (W) -> ([X]) -> R) -> ([X]) -> R])' to return type '(Y, [((W) -> ([X]) -> R) -> ([X]) -> R])'
    return apply(
           ^

What do you think about that error message?

Since it's been a while I'm going to resolve this issue. Please feel free to re-open if the problem still persists.