Concurrency practice

Opening your lesson

A release review mixes workload choices, worker payloads and coroutine results. Each item asks a different question about concurrency, so identify the boundary before choosing a technique. A lock trace describes shared state, while a process payload describes isolated state.

You reuse the earlier models on fresh fictional records and run the coroutine tasks with new inputs. Don't infer a speedup from these small examples. The checked output concerns ordering, ownership and correctness. Keep the limitations visible: thread and process scheduling are modeled here, while async functions execute in the workspace. Finish every task you own and retain the original order when the caller asks for it. Empty work is a valid completed batch.

import json

def process_payload(payload: str) -> str:
    data = json.loads(payload)
    shared = data['start']
    local = {}
    for command in data['trace']:
        op, worker = command.split(':')
        if op == 'read':
            local[worker] = shared
        else:
            shared = local[worker] + 1
    return json.dumps(shared)

A write uses its worker's saved read, so overlapping increments can be lost. Keep one shared value and a dictionary of each worker's most recent read. A single pass through the supplied trace exposes lost updates without starting any thread.

Choose your character

Your Work Gym account