@Zgarb another bug!

SomS+↔

the first S should take 3 arguments, o, m and S+↔

(note: ‼oTS+↔ does the exact thing SomS+↔ should do too)

Wait, ಠ_ಠ

oh you didn't copy it from here?

I didn't even realise you had the same versioon

then alright you don't need to credit me

I came here to ask whether Husk has a a[::b] function.

Ċ

(and specifically gaps)

Ah, thanks!

but Ċ0 doesn't behave like Jelly's m0$

Yeah however @Zgarb feature-request: ^ make Ċ0 equivalent to Jelly's m0 (make it append the list to its reverse), or at least make it do something useful.

(just saying because Husk has borrowed a lot from Jelly and why else would one use gaps instead of rev)

@EriktheOutgolfer This is expected (altough annoying) behavior. The type of S+↔ can't be both [Int] -> [Int] and [[Int]] -> [[Int]], so the type check fails.

@Mr.Xcoder That's a good idea, especially since it currently just hangs...

@Zgarb but is it ever the latter? I thought the m handles that stuff

@EriktheOutgolfer It is, since you're also trying to call it on the full matrix.

If M is the matrix, then SomS+↔M == oS+↔mS+↔M == S+↔mS+↔M, except that the types don't work out.

do you intend to fix that in a later version? it might be a hindrance sometime

No, it really can't be fixed without a complete overhaul of the type system and inference engine. :/

Or some hacky special casing for S and similar combinators, which I don't really want to do.

special casing? eugh