Term appears in its own list of ancestors (in some cases)

[russell-rozenbaum]

An example case is:

type x = (
  a = Int, 
  b = Float
)
in

The ancestors of the tuple type (a = Int, b = Float) has itself at index 0. This is also observed if the term is a basic type, and potentially more cases.

Likely an issue with implementations not using the helpers and using the helpers to add stuff to the info map... if they do not use the helpers, then they will be using the shadowed version of ancestors, which has the current node already included.

[7h3kk1d]

I added #1864 which is a pbt that tests this. We should instead add one that fully asserts that the children of a term contain it's id. We need a children function on expressions to do that though.