Fix method signature rendering fidelity in TypePrinter

[matejcerny]

Fixes #101.

• Curried parameter lists no longer joined with :
  • Consecutive parameter lists now render adjacently ((fa: F[A])(f: ...)) instead of (fa: F[A]): (f: ...)
  • The : separator is now emitted only before a genuine (non-methodic) result type.
• Unbounded type parameters no longer gain explicit bounds
  • [A, B] renders as [A, B] again rather than [A >: Nothing <: Any, B >: Nothing <: Any]
  • Bounds are now elided by comparing the printed bound against Nothing/Any rather than its raw toString.

Additional rendering fixes

• Function-type sugar
  • Function1[A, B] now renders as A => B and ContextFunctionN as ?=> ..., with parentheses added where required - around a function-typed left operand ((A => B) => C) and around arrow sugar in extends position.
• implicit parameter lists
  • Legacy implicit param lists now render with the implicit keyword (previously only using was handled).

Tests

Added TypePrinter unit tests plus Scala 2 and Scala 3 fixtures (CellarCurried, CellarSugar) covering:

• curried lists
• function/context-function sugar
• using/implicit parameters

Updated the README example output to match.

[matejcerny]

cc @rochala 🙏

[rochala]

Thanks for contribution. I'll take a look at it during the weekend. Sorry for the delay 🙈

[matejcerny]

Thanks for contribution. I'll take a look at it during the weekend. Sorry for the delay 🙈

No worries! I would really like to help more with Cellar, as there are quite a few places that could be improved. Also, it's currently the backbone of agentic programming in Scala, so I guess it deserves more love 🙂

[matejcerny]

FYI I've already spent some time and tokens on:

• HKT

  • previously def compose[G >: TypeLambda <: TypeLambda](evidence$1: Applicative[G]): Applicative[TypeLambda]
  • now def compose[G[_]](implicit evidence$1: Applicative[G]): Applicative[[α] =>> F[G[α]]]

• Functions

  • previously p: Function1[A, Boolean]
  • now p: A => Boolean

• Tuples

  • previously Tuple2[F[Boolean], F[A]]
  • now (F[Boolean], F[A])

I'll create separate PR's in case you agree that "this is the way" 🙂