Esta página ainda não foi traduzida, por isso é exibida em inglês. Ajude a traduzir

Dart vs FxDart › #15 · Fetch profiles, two at a time

Fetch profiles, two at a time

FxDart wins async

• toAsync
• map
• concurrent

Requirement

Fetch six user profiles from a (simulated) API, but never more than two requests in flight at once — the API rate-limits. Results must come back in the original order. To prove the limit held, the fake fetch counts how many requests overlap and both versions print the maximum observed.

This is the task plain Dart has no primitive for. Future.wait runs everything at once; batching into pairs wastes time waiting for the slower of each pair; doing it right means writing a worker pool by hand — index bookkeeping, a shared cursor, pre-sized result slots. FxDart's .concurrent(2) is that worker pool, as one word: as each request finishes the next one starts, and order is preserved.

user#1, user#2, user#3, user#4, user#5, user#6
max requests in flight: 2

Side by side

Why they differ

The two versions print the same thing — the difference is what you had to write and what you now have to maintain. The native worker pool is real production boilerplate (and easy to get subtly wrong: off-by-one on the shared cursor, forgetting to pre-size the results list, losing ordering). In the FxDart version the concurrency policy is a single chain step, so changing the limit — or removing it — touches one number instead of the function's whole shape.