Drift: typed SQL for an offline-first Flutter app

drift is useful because it focuses on typed tables, migrations, reactive queries, and transactions. The important engineering move is to place the package behind a clear boundary, so your Flutter UI depends on a stable capability rather than a vendor-shaped API.

What the library should own

• Keep package calls inside a named adapter or feature boundary.
• Make lifecycle, errors, and loading state part of a testable contract.
• Expose only the capability the app needs; hide implementation details from the whole tree.

A focused starting point

@DriftDatabase(tables: [Todos])
class AppDatabase extends _$AppDatabase {
  AppDatabase(super.e);
  @override
  int get schemaVersion => 1;

  Stream<List<Todo>> watchOpenTodos() =>
      (select(todos)..where((t) => t.done.equals(false))).watch();
}

Production checklist

1. Read the README and changelog for the exact version you pin.
2. Add one test for lifecycle, failure, and app background/foreground behavior.
3. Verify every platform your product supports, not only the developer machine.
4. Record ownership, upgrade cadence, and rollback notes in the repository.

Common pitfall

Treat migrations as production code and test old schemas before shipping an upgrade.

Takeaway

A good open-source library does not replace architecture. It makes one difficult boundary clearer, observable, and easier to replace.