c# TPL中的数据竞争-Data Race

例一：

Parallel.For(1, arraySize, i =>
{
ary[i] = ary[i] + ary[i - 1];
});
如下：
object _lock = new object();
Parallel.For(1, arraySize, i =>
{
lock(_lock)
ary[i] = ary[i] + ary[i - 1];
});

 

解决方案:lock 确保当一个线程位于代码的临界区时，另一个线程不进入临界区，如果其他线程试图进入锁定的代码，则它将一直等待，直到该对象被释放

例二：
private static readonly AsyncLocal<LocalLifeTineScopelrapper> AsynclLocallLifeScope = new AsyncLocal<LocalLifeTimeScopelrapper>();

private static ILifetimeScope GetCurrentUow()
{
    var uow = AsyncLocalLifeScope.Value ? .LifetimeScope;
    if(uow == null)
    {
        return null;
    }
    return uow;
}

解决如下

private static ILifetimeScope GetCurrentUow()
{
    lock(AsynclLocallLifeScope)
    {
        var uow = AsyncLocalLifeScope.Value ? .LifetimeScope;
        if(uow == null)
        {
            return null;
        }
        return uow;
    }
}