Unfoldable.generate

[matthewleon]

Construct an Unfoldable by applying a function to an incremented index.

[paf31]

Isn't this the same as map f [0..] in Haskell. Maybe the simplest thing would be to just export nats :: Unfoldable f => f Int

[matthewleon]

This PR avoids creating an ephemeral data structure to map over. I imagine that GHC is smart enough to stop that from happening, but we will be unfolding strict data structures in PureScript (and the lazy lists we do have retain a reference to the head).

That said, exporting nats seems like a reasonable idea as well.

[paf31]

I think there are a lot of cases where we could do map fusion on behalf of the compiler, in libraries, but I don't think we should. If someone needs fusion to happen enough, they can write the definition inline in a where clause, right? The core libraries shouldn't do everything, just enough, in my opinion at least.

[matthewleon]

Sounds reasonable. I'll get around to a PR for nats soon then.

[matthewleon]

Now that I reconsider, though, the issue with exporting nats is that it only works on lazy structures. I guess that's the issue for me... Either we do fusion ourselves, or the function won't terminate.

Am I overlooking a simple solution?

[paf31]

In that case, I think what we need is a range function.