PostgresOrd derived entities should be schema-qualified like other PostgresType entities

[daamien]

Let's say I have the really dumb idea of creating a data type with the exact same name that an existing Postgres Type.

mod foo {
    use pgrx::prelude::*;

    #[derive(PostgresType, Serialize, Deserialize, Debug )]
    #[derive(Eq, PartialEq, Ord, Hash, PartialOrd)]
    #[derive(PostgresEq)]
    #[derive(PostgresOrd)]
    #[derive(PostgresHash)]
    #[inoutfuncs]
    pub struct Text {
    // [...]
    }
}

The #[derive(PostgresType)] related entities will be written correctly

• CREATE TYPE foo.Text;
• CREATE FUNCTION foo."text_lt"(/* .... */)

However the #[derive(PostgresOrd)] will be written without the schema prefix

CREATE OPERATOR FAMILY Text_btree_ops USING btree;
CREATE OPERATOR CLASS Text_btree_ops DEFAULT FOR TYPE Text USING btree FAMILY Text_btree_ops AS
OPERATOR 1 <,
OPERATOR 2 <=,
OPERATOR 3 =,
OPERATOR 4 >=,
OPERATOR 5 >,
FUNCTION 1 text_cmp(Text, Text)

Which gives the following error:

create extension foo;
ERROR:  function text_cmp(text, text) does not exist

My understanding is that the PostgresOrd derived entities should be schema qualified like other entities:

CREATE OPERATOR FAMILY foo.Text_btree_ops USING btree;
CREATE OPERATOR CLASS foo.Text_btree_ops DEFAULT FOR TYPE foo.Text USING btree FAMILY foo.Text_btree_ops AS
OPERATOR 1 <,
OPERATOR 2 <=,
OPERATOR 3 =,
OPERATOR 4 >=,
OPERATOR 5 >,
FUNCTION 1 foo.text_cmp(foo.Text, foo.Text);

See code above:
https://github.com/pgcentralfoundation/pgrx/blob/develop/pgrx-sql-entity-graph/src/postgres_ord/entity.rs#L84

I understand that this a very narrow use case and a good practice is to avoid naming new types with the names of regular Postgres types.

Nonetheless it seems like a good idea to avoid those kind of collisions.