Eitan Chatav

Ok, I haven't written out proofs but here's how `Applicative`, `Alternative`, `Divisible` and `Decidable` are with "monoids all the way down". Consider the following types; ```Haskell data Day f g...

I've tried at times to make a `params` term that could be used in this type of situation but I never found a way to get it to work right...

```Haskell class All KnownNat ixs => HasParams (ixs :: [Nat]) (params :: [NullType]) (tys :: [NullType]) where parameters :: NP (TypeExpression db) tys -> NP (Expression grp lat with db...

```Haskell (Values_ (hmap ((`as` Alias) . Set) (params @'[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17]))) ```

Oh those label types aren't thrown away. > In my case the names are exactly like the ones specified with the overloaded labels so it should be possible to just...

We cannot avoid natural numbers because these actually correspond to PostgreSQL [positional parameter](https://www.postgresql.org/docs/15/sql-expressions.html#SQL-EXPRESSIONS-PARAMETERS-POSITIONAL) numbers.

`make cffi` seems to work as does `cabal build`. No errors at least. I do use homebrew but could not find the `crt0` library to install. I'll hunt for it.

> @echatav do you think it's possible to reproduce your issue to the point where it does not require a database? @sol You don't need the database to see the...

Here is the doctest failure in [Squeal's CI for GHC 9.0](https://github.com/morphismtech/squeal/runs/4646823670?check_suite_focus=true#step:6:5789)

@sol I've made a [minimal example branch](https://github.com/morphismtech/squeal/tree/dev-ghc9-doctest-minimal) which [demonstrates the test error](https://github.com/morphismtech/squeal/pull/311/checks).