gpt4 book ai didi

c++ - C++ 中的 CountDownLatch 使用 Boost Mutexes 和 Condition

转载 作者:太空宇宙 更新时间:2023-11-04 12:19:40 25 4
gpt4 key购买 nike

我尝试使用增强互斥体和条件变量来实现 CountDownLatch。下面是代码,想知道我是否需要添加任何其他内容。

如何对这段代码进行单元测试?

template< class TypeVal >
class AtomicCounter
{
public:
AtomicCounter( TypeVal val ) : m_typeVal( val )
{
}

AtomicCounter() : m_typeVal(0)
{
}

AtomicCounter(const AtomicCounter& cpy) : m_typeVal(cpy.m_typeVal)
{
boost::mutex::scoped_lock scoped_lock(cpy.m_atomicMutex);
m_typeVal = cpy.m_typeVal;
}

const AtomicCounter& operator=(const AtomicCounter& other)
{
if (this == &other)
return *this;
boost::mutex::scoped_lock lock1(&m_atomicMutex < &other.m_atomicMutex ? m_atomicMutex : other.m_atomicMutex);
boost::mutex::scoped_lock lock2(&m_atomicMutex > &other.m_atomicMutex ? m_atomicMutex : other.m_atomicMutex);
m_typeVal = other.m_typeVal;
return *this;
}

virtual ~AtomicCounter()
{
}

const TypeVal& getCount() const
{
boost::mutex::scoped_lock lock( m_atomicMutex );
return m_typeVal;
}

const TypeVal& setCount( const TypeVal &val )
{
boost::mutex::scoped_lock lock( m_atomicMutex );
m_typeVal = val;
return m_typeVal;
}

const TypeVal& increment()
{
boost::mutex::scoped_lock lock( m_atomicMutex );
m_typeVal++ ;
return m_typeVal;
}

const TypeVal& decrement()
{
boost::mutex::scoped_lock lock( m_atomicMutex );
m_typeVal-- ;
return m_typeVal;
}

const TypeVal& increment(const TypeVal& t)
{
boost::mutex::scoped_lock lock( m_atomicMutex );
m_typeVal+=t ;
return m_typeVal;
}

const TypeVal& decrement(const TypeVal& t)
{
boost::mutex::scoped_lock lock( m_atomicMutex );
m_typeVal-=t ;
return m_typeVal;
}


private:
mutable boost::mutex m_atomicMutex;
TypeVal m_typeVal;
};

class CountDownLatch
{
public:
CountDownLatch( int count ): m_cdlCount( count )
{
}

CountDownLatch(const CountDownLatch& cpy)
{
boost::unique_lock<boost::mutex>::unique_lock(const_cast<boost::mutex&>(cpy.m_cdlMutex));
m_cdlCount = cpy.m_cdlCount;
}

const CountDownLatch& operator=(const CountDownLatch& other)
{
if (this == &other)
return *this;
boost::mutex::scoped_lock lock1(const_cast<boost::mutex&>(&m_cdlMutex < &other.m_cdlMutex ? m_cdlMutex : other.m_cdlMutex));
boost::mutex::scoped_lock lock2(const_cast<boost::mutex&>(&m_cdlMutex > &other.m_cdlMutex ? m_cdlMutex : other.m_cdlMutex));
m_cdlCount = other.m_cdlCount;
return *this;
}

virtual ~CountDownLatch()
{
}
void wait()
{
boost::mutex::scoped_lock lock( m_cdlMutex );
if( m_cdlCount.getCount() > 0 )
{
m_cdlCondition.wait( lock );
}
}
void wait( uint64_t timeoutMicros )
{
boost::mutex::scoped_lock lock( m_cdlMutex );
if( m_cdlCount.getCount() > 0 )
{
boost::posix_time::time_duration td = boost::posix_time::milliseconds( timeoutMicros );
m_cdlCondition.timed_wait( lock, td );
}
}
void countDown()
{
boost::mutex::scoped_lock lock( m_cdlMutex );
if( m_cdlCount.decrement() == 0 )
{
m_cdlCondition.notify_all();
}
}

int getCount()
{
return m_cdlCount.getCount();
}


private:
boost::mutex m_cdlMutex;
boost::condition_variable m_cdlCondition;
AtomicCounter< int > m_cdlCount;
};

最佳答案

对于单元测试,您可以尝试压力测试。例如,对于 CountDownLatch,创建 25 个同时调用 CountDownLatch::countDown() 的测试线程、同时调用 CountDownLatch::getCount() 的其他 25 个线程,以及另外 25 个线程调用 CountDownLatch::wait() 的线程。为了让事情更同步,使用屏障,或者让线程休眠直到相同的绝对时间。通过加入所有线程,确保所有线程正确终止(无死锁)。确保 CountDownLatch::m_cdlCount 最终为零。

多次运行相同的测试(合理的时间)。

您可以对 AtomicCounter 使用相同的基本思想。

可能还有其他测试多头代码的技术,但这是我最熟悉的一种。

关于c++ - C++ 中的 CountDownLatch 使用 Boost Mutexes 和 Condition,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/5825565/

25 4 0
Copyright 2021 - 2024 cfsdn All Rights Reserved 蜀ICP备2022000587号
广告合作:1813099741@qq.com 6ren.com