Awesome, thanks! I update it whenever I work with it, which has ramped up in the last few weeks (added cast support a few days ago). It uses pg_analyse to parse the SQL and uses a few tricks to get everything updated.

The best thing (IMO) is that there is basically no funny stuff, no filename conventions, no funny delimiters. It’s just regular Postgres SQL, and a couple of very small config files. Works perfectly with git.