ScalaFix linters

[bplommer]

I raised in #5484 the possibility of using scalafix to enforce naming conventions. Other rules that could be useful to improve consistency and reduce maintenance burden:

• organize imports - this will introduce a lot of merge conflicts when introduced but should subsequently reduce them
• Disallow use of fs2's sync compiler - Add scalafix linter for use of fs2 Sync compiler #5536
• Disallow use of throw in parsers not prefixed with unsafe
• No non-final case classes (added but not yet linting at error level)
• No final case objects
• Recurring headers must have a semigroup instance (unless they're called Set-Cookie)
• Contents of internal packages must be package-private
• case classes with private constructors leaking public copy constructors
• Replace assert(a == b) with assertEquals(a, b), assert(a != b) with assertNotEquals(a, b)

[rossabaker]

All of these are wonderful ideas, except I don't know why final case objects are bad. Is it just redundancy? Redundancy is bad.

[armanbilge]

Scala 3 complains about them actually. We don't have fatal warnings in 0.22 but they get removed up in 0.23 anyway.

https://github.com/http4s/http4s/pull/5474/files#diff-e372aa8a2995d9def1db9dc2b9274a92d0fdbf3203e57087f1e0caee1486178b