Newconstraints phase3, Add newtypes: Size, SizeSpec and class Sized.

[TimSheard]

This is a follow on PR. Originally rebased on top of PR #4120, it is now rebased on the current master, and 4120 has been closed.

Spec does 3 things.

1. Adds generic size functions on numerous types with HasSpec instances, by adding Sized instances. Current
   instances include Int, Set, List, Map, and Size itself.
2. Retracts the And opertor from BoolFn, which turned out to have some technical difficulties.
3. Adds the methods 'cardinalTypeSpec' and 'cardinalTrueSpec', these allow any HasSpec instance to
   use the function cardinality :: HasSpec fn t => Spec fn t -> Spec fn Int.
   It computes bounds on the total number of possible solutions generated by 'genFromTypeSpec'
   This supports tighter (more accurate) bounds in a number of places.

Checklist

• Commit sequence broadly makes sense and commits have useful messages
• New tests are added if needed and existing tests are updated
• When applicable, versions are updated in .cabal and CHANGELOG.md files according to the
  versioning process.
• The version bounds in .cabal files for all affected packages are updated. If you change the bounds in a cabal file, that package itself must have a version increase. (See RELEASING.md)
• All visible changes are prepended to the latest section of a CHANGELOG.md for the affected packages. New section is never added with the code changes. (See RELEASING.md)
• Code is formatted with fourmolu (use scripts/fourmolize.sh)
• Cabal files are formatted (use scripts/cabal-format.sh)
• hie.yaml has been updated (use scripts/gen-hie.sh)
• Self-reviewed the diff

[MaximilianAlgehed]

I have left some comments, but I haven't gone over all the changes to Base in detail yet. There is a lot of stuff around the Size stuff that already exists for Integer and the other numeric types. I think it would be wise to drop Size as a type, in favour of Integer (to avoid wrapping issues when computing cardinality for e.g. Word64).

[MaximilianAlgehed]

After we've revamped Size I can take another look at the changes to Constrained.Base - they should be simpler then.

[MaximilianAlgehed]

@TimSheard how do you feel about s/Size/Natural instead of Integer? Then the issues with subtraction etc. should be resolved? Possibly we might want to have another look at the Num instance for Term fn Natural if there is an issue with negate vs (-) there?

[lehins]

Possibly we might want to have another look at the Num instance for Term fn Natural if there is an issue with negate vs (-) there?

There is ugly issue with Num for Natural, which is why I usually suggest to avoid it. For example if f is a polymorphic reusable function that is potentially defined in another library:

ghci> let f = subtract (-1)
ghci> :t f
f :: Num a => a -> a
ghci> map f [10::Natural]
[*** Exception: arithmetic underflow

[TimSheard]

I am thinking that we make Size = Integer.
Now we have an instance of genFromSpec
genFromSpec :: Spec fn Integer -> GenT m Integer
But we write our own function
genFromSize :: Spec fn Integer -> GenT m Integer, that never generates an Integer less than 0,
and we use this function at appropriate places in our code.

[MaximilianAlgehed]

With that solution genFromSize = genFromSpec . geqSpec 0?

[TimSheard]

Yes, that would be right. Any take a look, I have swithced to Size = Integer