Keynote: Asynchronous Dependency Injection
summary
Covers migrating slow dependencies from eager injection to Dagger Producers vs RxJava + Dagger, swapping @Provides for @Produces, wiring an executor module and using listenable future callbacks so dependencies build on a cache thread pool in parallel. Shows concrete workarounds: deferred-logger wrapper, converting futures to Lazy, and how producers collapse serial graphs. Flags concrete limits: libraries that force main-thread init (Fresco + OkHttp), deep serial dependency chains, high mental overhead and months of work for modest wins — typically 100–200 ms — so treat as a last resort.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 2:23What is dependency injection
- 7:07Why async dependency injection
- 11:46Code approach with RX Java
- 16:54Dagger Producers overview
- 23:18Crossing thread boundaries
- 26:52Clever implementation tricks
- 31:59Should you use async DI
- 35:09Conclusion and Q&A
