两种调用方法的方法有何区别？

Question

In the method below what is the difference between using

ListThreads.Add(new Task(() => item.Execute()));

and

ListThreads.Add(new Task(item.Execute));

private void Execute()
{
    for (int i = 0; i < ThreadNumber; i++)
    {
        ListObjects.Add(new MyClass(i + 1, ThreadNumber));
    }
    foreach (MyClass item in ListObjects)
    {
        ListThreads.Add(new Task(() => item.Execute()));
        ListThreads[ListThreads.Count - 1].Start();
    }
    Task.WaitAll(ListThreads.ToArray());
}

Answer 1

You ask the difference between

() => item.Execute()

and

item.Execute

The former is a lambda that calls item.Execute . The, item.Execute , is a method group. When they are passed to the constructor of Task they are both converted to a delegate of type Action .

There is quite a difference though. The lambda captures the variable item . And the method group does not. This means that when the lambda is executed, the value of the variable item may be different from its value when you passed the lambda to the constructor of Task .

To make the lambda version equivalent to the method group version you could introduce a local variable:

foreach (MyClass item in ListObjects)
{
    MyClass tmpItem = item;
    ListThreads.Add(new Task(() => tmpItem.Execute()));
    ListThreads[ListThreads.Count - 1].Start();
}

Do note that the language has been modified between C# 4.0 and C# 5.0. In C# 5.0 the code in your question behaves in exactly the same way as does the code above in this answer. For more details see:

• Closing over the loop variable considered harmful, part one
• Closing over the loop variable considered harmful, part two
• Has foreach's use of variables been changed in C# 5?
• Is "Access to modified closure" resolved by comprehension syntax?

Answer 2

First of, it's a bad idea to use foreach variable in lambda expression. So, in this particular case the correct way is to write ListThreads.Add(new Task(item.Execute));