锁定和操作需要很长时间

Question

I think it's a common problem. For example, I have an array which is modified by one process/thread:

lock();
for(int i; i<array_size; i++) {
    // find an item and do some operations
}
unlock();

There is another process/thread which will print the whole array occasionally, but it may take a "long" time:

lock()
for(int i; i<array_size; i++) {
    print(array[i]);
}
unlock();

Are there any better approaches to print the whole array?

Answer 1

如果是花费较长时间的打印部分，则可以锁定，获取阵列的副本，释放锁定并打印阵列副本。

Answer 2

What kind of locks do you use, mutex? Also what kind of consistency do you expect between the individual cells in your array, when (1) up update them, and (2) when you print them? If each array-cell is independent, then why not just have many more locks each only responsible for that cells?

Like;

mutex locks[array_size];
for (int i=0; i< mutex_size; i++) {
   locks[i].lock();
    // do something or print content...
   locks[i].unlock();   
}

However if semantics of your array is such that the cells are not independent of each other but that you need atomic consistency across the array, then you are stuck with your original global lock.

However you could still take a copy of the array under lock and then print the copy, as that would keep the lock of the array at a shorter time not having to wait for the print.