use naive Lazy List foldr

[matthewleon]

Currently, Data.List.Lazy defines foldr in terms of foldl on a reversed list. This preserves stack safety.

Might it be worth sacrificing stack safety to allow foldr to work with potentially very large or infinite lists by using the naive definition (trusting the user to use a lazy operation)? If not, is that worth having as a separate function?

Quick example of a consequence of the current def: a function like traverse_ has to wait until the entire list is traversed before actions can begin.

[paf31]

Having a variant of foldr which looks like (a -> Lazy b -> b) -> List a -> b -> b can be very useful for modularity, or if you're using infinite or lazy lists. Let's add it!

[matthewleon]

Having a variant of foldr which looks like (a -> Lazy b -> b) -> List a -> b -> b can be very useful for modularity, or if you're using infinite or lazy lists. Let's add it!

In the above are arguments flipped on purpose or did you intend (a -> Lazy b -> b) -> b -> List a -> b?

[paf31]

@matthewleon Shall