It looks like something in the GHC 9 / Lens 5 combination yields
unexpected results when using view on Lens', with GHC complaining that
it's passed a Lens' instead of a Getting (which should still be
compatible).  Commenting out the signatures of Ganeti.Network.poolLens,
Ganeti.Network.poolArrayLens and Ganeti.Utils.MultiMap.multiMapL makes
Ganeti build again, so I'd keep this ugly hack until we figure out
what's actually wrong.

Signed-off-by: Apollon Oikonomopoulos <apoikos@dmesg.gr>

Citing the GHC 9.0.1 changelog[1]:

  Breaking change: Template Haskell splices now act as separation points
  between constraint solving passes. It is no longer possible to use an
  instance of a class before a splice and define that instance after a
  splice. For example, this code now reports a missing instance for C
  Bool:

    ```
    class C a where foo :: a
    bar :: Bool
    bar = foo
    $(return [])
    instance C Bool where foo = True
    ```

In other words, code order matters in Template Haskell files, which
breaks the Haskell test build with GHC >= 9.0.1. Fix this by reordering
various Arbitrary instances so that they are defined before being used.

Also the Arbitrary instance for IAllocatorParams was silently acting as
a generic instance for (Container JSValue), providing instances for
DiskParams, HypervisorState and DiskState as well. Change the Arbitrary
instance declaration to Arbitrary (Container JSValue) to better reflect
its generic purpose.

[1] https://downloads.haskell.org/ghc/9.0.1/docs/html/users_guide/9.0.1-notes.html#highlights

Signed-off-by: Apollon Oikonomopoulos <apoikos@dmesg.gr>