feat: port the shared-memory transport onto the WorkerChannel abstraction

[mdashti]

This PR ports our shared-memory ("shm") transport onto the WorkerChannel abstraction.

It stacks on the transport-completion PR (#27). The base is that branch, so the diff is only src/shm/* plus the mod shm declaration and its two Cargo deps. The transport plugs in with no core changes, which is the point.

What

src/shm/* (rings, DSM, mesh, framing) moves over verbatim. The old ShmMqWorkerTransport: WorkerTransport + ShmMqWorkerConnection: WorkerConnection become ShmChannelResolver: ChannelResolver + a ShmWorkerChannel: WorkerChannel: execute_task reads the rings keyed off the task; coordinator_channel routes each SetPlanRequest to the destination worker's inbox over the mesh, so the coordinator's dispatch is the plan-delivery path.

Workers are passive executors, same as the gRPC model: they run the ready-to-run per-task plans the coordinator dispatches (task-specialized, nested stages Remote), whose boundary leaves read the mesh. Nothing on the worker re-plans, converts, or dispatches.

Verified

The in-crate in_process test runs real distributed queries through the shm rings (gather, shuffle, early-cancel, producer-loss), green in both feature configs, under the no-gRPC CI job inherited from #27. It exercises #27's extension points end to end: the dispatch handoff through DispatchPlanSource, the push-side placement through route_partition, and the metrics drain through metrics_store + decode_task_metrics. self_hosted is deferred and gated out; its no-gRPC-default role is the in-process transport's and its ring safety-net role is this test.

Consumed by

pg_search runs its MPP on this transport (paradedb#5516), which is what exercises it end to end.

[mdashti]

Incorporated into main by the rebase sync, as a3416a7 and 925ff31. Closing; the branch stays for the pg pin until it repins to main.