Pretty printer: ability requests in ability handler (and ability def'n) that are symboly are rendered incorrectly

[kylegoetz]

Describe and demonstrate the bug

ability Test where
  (+) : Nat -> Nat -> Nat

Now add it

> update Test
> view Test

Screenshots
Screenshots of both viewing the ability definition and an ability handler. Observe <> is not wrapped in parens despite being in prefix position.

Environment (please complete the following information):

• ucm --version release/0.5.49'
• OS/Architecture: macOS 15.5

I would suggest either rendering them in infix position or wrapping parens but keeping prefix.

The reason this is a problem is that when you edit the definition or the handler, the unaltered code is malformed when dropped into the scratch file because the symboly op is unwrapped but in prefix position.

[kylegoetz]

Relatedly,

fp.Foldable.foldMap = Foldable.foldr ((Semigroup.<>) << f) Monoid.mempty

didn't round-trip. It became

fp.Foldable.foldMap f =
  use Semigroup <>
  fp.Foldable.foldr (<> << f) Monoid.mempty

where the symboly <> wasn't wrapped in parentheses and thus isn't valid syntax.

[kylegoetz]

And relatedly, edit mconcat gave me this in my scratch file:

fp.Monoid.mconcat : [a] ->{fp.Monoid a, Semigroup a} a
fp.Monoid.mconcat as =
  use fp.Semigroup <>
  List.foldRight <> fp.Monoid.mempty as

The <> in the bottom-most line should be wrapped in parentheses. It's an ability request, not a regular function.

[aryairani]

Ah, yes. Thanks for the report.