Step 1 - Payload + worker + main

import 'dart:convert';
import 'dart:isolate';

String buildPayload(int rows) {
  final list = <Map<String, dynamic>>[];
  for (var i = 0; i < rows; i++) {
    list.add({'id': i, 'name': 'item-$i', 'flags': [i % 3, i % 5]});
  }
  return jsonEncode(list);
}

int parseAndWalk(String raw) {
  final decoded = jsonDecode(raw);
  if (decoded is! List<dynamic>) return 0;
  var sum = 0;
  for (final row in decoded) {
    if (row is Map<String, dynamic>) {
      final id = row['id'];
      if (id is int) sum += id;
    }
  }
  return sum;
}

Future<void> main() async {
  final raw = buildPayload(80000);
  final sum = await Isolate.run(() => parseAndWalk(raw));
  print(sum);
}

Step 2 - Tune size

• If work finishes too fast, raise rows. If allocations dominate, lower rows and add a deliberate CPU loop inside parseAndWalk for simulation only.

Practice tasks

• Time parseAndWalk(raw) on the main isolate vs Isolate.run with Stopwatch and print both durations.
• Return a record ({int sum, int count}) from the worker instead of a single int.
• Estimate copy cost vs compute time by trying several rows values.