Symbol signatures diverge from README example

[matejcerny]

The README's "Example output" for get-external org.typelevel:cats-core_3:2.10.0 cats.Monad no longer matches actual output.

1. Unbounded type parameters gain explicit bounds. [A, B] renders as [A >: Nothing <: Any, B >: Nothing <: Any]

2. Curried parameter lists are joined with :. Multiple parameter lists render as (fa: F[A]): (f: ...) instead of adjacent (fa: F[A])(f: ...), making methods look like they return a function.

flatMap — README vs. actual:

// README claims:
def flatMap[A, B](fa: F[A])(f: Function1[A, F[B]]): F[B]

// actual:
def flatMap[A >: Nothing <: Any, B >: Nothing <: Any](fa: F[A]): (f: Function1[A, F[B]]): F[B]

Same pattern on pure, flatten, iterateWhile, etc.

[rochala]

Another variant of the explicit-bounds problem (point 1): higher-kinded type parameters render their bounds as TypeLambda:

// cellar get-external org.typelevel:cats-effect_3:latest cats.effect.std.Semaphore
def mapK[G >: TypeLambda <: TypeLambda](f: package.~>[F, G]): (G: MonadCancel[G, ? >: Nothing <: Any]): Semaphore[G]
def apply[F >: TypeLambda <: TypeLambda](n: Long): (F: GenConcurrent[F, ? >: Nothing <: Any]): F[Semaphore[F]]

Expected roughly: def mapK[G[_]](f: F ~> G)(using G: MonadCancel[G, ?]): Semaphore[G].

[rochala]

AI generated comment above ^

[matejcerny]

AI generated comment above ^

Great example, thanks! This is how it currently looks like on my latest branch def mapK[G[_]](f: ~>[F, G])(implicit G: MonadCancel[G, ? >: Nothing <: Any]): Semaphore[G], I'll improve it.

One question though, should the renderer be aware of the Scala version? Ie should the wildcard be printed as ? or _? (imho it would be much easier to start with the modern syntax and add the distinction later)