Steve Dignam

yeah we don't currently support linting of function bodies from a postgres parser perspective, the function body is just a string, in this case a dollar delimited string but linting...

Can you give a small example of what you're seeing? Wondering if we can make something more generic

Thanks for the report! We don't currently parse PL/pgSQL but [it's on the roadmap](https://github.com/sbdchd/squawk/blob/3d72eafe67179870dfab06d4a85b23187891afdb/PLAN.md#plpgsql-strings). I was looking at the grammar for it and it's thankfully only 4k LOC: https://github.com/postgres/postgres/blob/f9b1192190aac56a319a1e3e93cf3c29d9f722dd/src/pl/plpgsql/src/pl_gram.y#L4244 Instead...

Oh nice find, looks like we're not checking for the postgres version in the rule: https://github.com/sbdchd/squawk/blob/412466cb8648125ab2a5c1206ac97aec8354224a/crates/squawk_linter/src/rules/adding_not_null_field.rs#L18C55-L18C55 Edit: Oh we used to not have a check, then we added one, then...