Postgres supports various authentication flows as part of the startup phase. We should implement a minimal subset of these:
For this, we should either delegate the real auth work to the engine, or create a new trait entirely to handle auth (selecting which sub-flow to use, and verifying presented data). This process should return some kind of user principal to the engine, so it can e.g. adjust which tables are available for a given user.
Postgres supports various authentication flows as part of the startup phase. We should implement a minimal subset of these:
For this, we should either delegate the real auth work to the engine, or create a new trait entirely to handle auth (selecting which sub-flow to use, and verifying presented data). This process should return some kind of user principal to the engine, so it can e.g. adjust which tables are available for a given user.