feat(db): support PostgreSQL migration schemas

[onmax]

Allow PostgreSQL migration history to live alongside application tables, with atomic migration execution.

hub: {
  db: {
    dialect: 'postgresql',
    migrationsSchema: 'auth'
  }
}

prepare history under advisory lock
  move public._hub_migrations into the configured schema
apply each pending migration under the same lock
  recheck history → run SQL + record migration atomically

The runner and CLI use the configured history table. Existing rows survive relocation; if both history tables exist, execution stops for manual reconciliation. Omitting migrationsSchema keeps the existing search-path behavior. Application tables keep the schemas declared in Drizzle.

This branch also includes the PostgreSQL runtime environment-variable support from #923, allowing deployments to use PG* variables without a connection URL.

Compatibility

• The configured schema exists before migrations run. SQL that creates it must use CREATE SCHEMA IF NOT EXISTS.
• Stop old runners before relocating history. Move history back before restarting old code during rollback.
• PostgreSQL migrations run in one transaction and require PL/pgSQL. Commands such as CREATE INDEX CONCURRENTLY cannot run in this transaction.
• Other dialects retain their existing behavior. Neon HTTP has not been verified against a live endpoint.

[vercel]

@onmax is attempting to deploy a commit to the NuxtLabs Team on Vercel.

A member of the Team first needs to authorize it.

[pkg-pr-new]

npm i https://pkg.pr.new/@nuxthub/core@921

commit: 783861d