[m-rev.] for post-commit review: more useful type mismatch diagnostics

[Julien Fischer]

On Fri, 14 Feb 2025 at 00:07, Zoltan Somogyi <zoltan.somogyi at runbox.com> wrote:

> Print more useful actual/expected type diffs.

...


> @@ -2331,12 +2388,14 @@ arg_type_list_diff_pieces_loop(ContextPieces, TypeCtorPieces,
>      ( if ActualArgType = ExpectedArgType then
>          DiffPieces = TailDiffPieces
>      else
> +        % XXX If the type constructor has arity 1, should we print
> +        % "the only argument of ..." instead of "the first argument of ..."?

I think so.

The rest of the diff looks fine.

Julien.