Remove the Generic instance for LicenseId

[Kleidukos]

This was one of the main culprits when analysing the resources taken
by GHC when compiling lib:Cabal.

Is related to #8074.

---

This is what the timings look with, and without the instance:

then

---

Please include the following checklist in your PR:

• Patches conform to the coding conventions.
• Any changes that could be relevant to users have been recorded in the changelog.
• The documentation has been updated, if necessary.

Please also shortly describe how you tested your change. Bonus points for added tests!

[Mikolaj]

Why is Setup trying to make up for the LicenseId speedup?

[Kleidukos]

Good question, this is indeed quite puzzling.

[andreasabel]

Upstream (GHC) issue: https://gitlab.haskell.org/ghc/ghc/-/issues/16577

[andreasabel]

@Kleidukos @gbaz: I just had another idea: If a large enum-type is problematic for GHC's deriving mechanism, how about breaking it up into smaller bits, and restore the original interface using PatternSynonyms?
This way maybe we can have the cake and eat it, too.

I suppose since GHC's stock deriving mechanism does not support nested enum types, one would have to code the Enum and Bounded instances for the composite type manually from the derived instances of the components.

[Kleidukos]

@andreasabel It's an interesting solution, could you show us some benchmarks? I used https://github.com/codedownio/time-ghc-modules

[gbaz]

i think that idea is a lot of work for an instance that nobody really seems to need that much :-)

[andreasabel]

@andreasabel It's an interesting solution, could you show us some benchmarks?

Na. It is an idea, but I don't have the resources to follow up on it.

[Mikolaj]

LGTM. Let's merge.