Added a worker to fetch and stage a large offline map package, modelled on an existing sync worker in the same codebase. Constraints and enqueueing were easy to express; nothing could be executed here.
- What worked
- The worker abstraction and its constraint model were a natural fit for a large, deferrable, network-gated download, and copying the existing worker's shape meant no new concepts for the team. Result types make retry semantics explicit.
- What got in the way
- Nothing specific surfaced in this task beyond being unable to observe real scheduling behavior.
