dispatch group.h

版本：iOS13.5

group.h

dispatch其他文件通道

索引

• dispatch_group_create
  创建可以与block关联的组
• dispatch_group_async
  将block提交到调度队列queue并异步执行，并将block与给定的调度组group关联。
• dispatch_group_async_f
  将函数work提交给调度队列queue并异步执行，并将work与给定的调度组group关联。
• dispatch_group_wait
  同步等待，直到与group关联的所有block都已完成，或者直到超时为止。（会堵塞当前线程）
• dispatch_group_notify
  与group关联的所有块均已完成后，将block提交给队列queue执行。
• dispatch_group_notify_f
  与group关联的所有函数均已完成后，将work提交给队列queue执行。
• dispatch_group_enter
  手动将group的计数加1
• dispatch_group_leave
  手动将group的计数减1
  group的计数为0时才会告诉dispatch_group_wait和dispatch_group_notifygroup中所有关联的block已完成

详解

• 创建可以与block关联的组。

dispatch_group_t dispatch_group_create(void);

• 将block提交到调度队列queue并异步执行，并将block与给定的调度组group关联。

void dispatch_group_async(dispatch_group_t group, dispatch_queue_t queue, 
                           dispatch_block_t block);

例：
    dispatch_queue_t queue = dispatch_queue_create("Dispatch_group", DISPATCH_QUEUE_CONCURRENT);
    dispatch_group_t group = dispatch_group_create();
    dispatch_group_async(group, queue, ^{
        NSLog(@"block1");
        sleep(1);
    });
    dispatch_group_async(group, queue, ^{
        NSLog(@"block2");
        sleep(1);
    });
    dispatch_group_async(group, queue, ^{
        NSLog(@"block3");
        sleep(1);
    });
输出：因为queue是并行的 所以结果为132 也会出现123
block1
block3
block2

• 将函数work提交给调度队列queue并异步执行，并将work与给定的调度组group关联。

void dispatch_group_async_f(dispatch_group_t group, dispatch_queue_t queue,
                            void *_Nullable context, dispatch_function_t work);

work 函数typedef void (*dispatch_function_t)(void *_Nullable)
context work传入的参数（传该参数的地址）
dispatch_group_async_f相较于dispatch_group_async就是将block变成work函数

例：
    dispatch_queue_t queue = dispatch_queue_create("Dispatch_group", DISPATCH_QUEUE_CONCURRENT);
    dispatch_group_t group = dispatch_group_create();
    dispatch_group_async_f(group, queue, nil, test);

void test(void *context) {
    NSLog(@"test");
}
输出：
test

• 同步等待，直到与group关联的所有block都已完成，或者直到超时为止。（会堵塞当前线程）

long dispatch_group_wait(dispatch_group_t group, dispatch_time_t timeout);

timeout DISPATCH_TIME_FOREVER表示永不超时 DISPATCH_TIME_NOW表示立即超时 或dispatch_time(DISPATCH_TIME_NOW, 1 * NSEC_PER_SEC)指定时间
long 与该组block关联的所有block在超时时间内完成则返回0，超时则返回非0

例：
    dispatch_queue_t queue = dispatch_queue_create("Dispatch_group", DISPATCH_QUEUE_CONCURRENT);
    dispatch_group_t group = dispatch_group_create();
    dispatch_group_async(group, queue, ^{
        sleep(1);
        NSLog(@"block1");
    });
    dispatch_group_async(group, queue, ^{
        sleep(1);
        NSLog(@"block2");
    });
    dispatch_group_async(group, queue, ^{
        sleep(1);
        NSLog(@"block3");
    });
    
    dispatch_async(queue, ^{
        long count = dispatch_group_wait(group, dispatch_time(DISPATCH_TIME_NOW, 2 * NSEC_PER_SEC));
        NSLog(@"count = %ld", count);
    });
输出：因为sleep(1)，block只执行了1秒，超时时间为2秒，所以count = 0表示未超时
block1
block3
block2
count = 0
若block2 sleep(3)，则输出
block1
block3
count = 49
block2
超时时间2秒内block2未执行完，所有会超时 count = 49

• 与group关联的所有块均已完成后，将block提交给队列queue执行。

void dispatch_group_notify(dispatch_group_t group, dispatch_queue_t queue,
    dispatch_block_t block);

当group中没有关联的块时 会直接执行block

例：
    dispatch_queue_t queue = dispatch_queue_create("Dispatch_group", DISPATCH_QUEUE_CONCURRENT);
    dispatch_group_t group = dispatch_group_create();
    dispatch_group_async(group, queue, ^{
        sleep(1);
        NSLog(@"block1");
    });
    dispatch_group_async(group, queue, ^{
        sleep(3);
        NSLog(@"block2");
    });
    dispatch_group_async(group, queue, ^{
        sleep(1);
        NSLog(@"block3");
    });
  
    dispatch_group_notify(group, queue, ^{
        NSLog(@"所有block都完成");
    });
输出：
block1
block3
block2
所有block都完成

• 与group关联的所有函数均已完成后，将work提交给队列queue执行。

void dispatch_group_notify_f(dispatch_group_t group, dispatch_queue_t queue,
    void *_Nullable context, dispatch_function_t work);

相关释义可查看dispatch_group_async_f
dispatch_group_notify_f相较于dispatch_group_notify就是将block变成了函数work

• 手动将group的计数加1

void dispatch_group_enter(dispatch_group_t group)

• 手动将group的计数减1

void dispatch_group_leave(dispatch_group_t group);

注意：dispatch_group_enter和dispatch_group_leave必须成对出现
dispatch_group_async就是先让group的计数加1，待block执行完成后group的计数减1

dispatch_group_async(group, queue, ^{

});
相当于
dispatch_group_enter(group);
dispatch_async(queue, ^{
    dispatch_group_leave(group);
});

group的计数为0时才会告诉dispatch_group_wait和dispatch_group_notify组group中所有关联的block已完成

例：
    dispatch_queue_t queue = dispatch_queue_create("Dispatch_group", DISPATCH_QUEUE_CONCURRENT);
    dispatch_group_t group = dispatch_group_create();
    dispatch_group_async(group, queue, ^{
        sleep(1);
        NSLog(@"block1");
    });
    dispatch_group_enter(group);
    dispatch_async(queue, ^{
       sleep(3);
       NSLog(@"block4");
       dispatch_group_leave(group);
    });
    
    dispatch_group_notify(group, queue, ^{
        NSLog(@"所有block都完成");
    });
输出：
block1
block4
所有block都完成