dispatch_async 與 dispatch_get_global_queue

#define kBgQueue dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0)

- (void)viewDidLoad

{

    [super viewDidLoad];

    dispatch_async(kBgQueue, ^{

        NSData* data = [NSData dataWithContentsOfURL: kLatestKivaLoansURL];

        [self performSelectorOnMainThread:@selector(fetchedData:) withObject:datawaitUntilDone:YES];

    });

}

dispatch_async會向kBgQueue隊列中添加新的任務去執行，這裏kBgQueue隊列使用dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0)獲得。

Dispatch Queues

Dispatch Queues從使用的角度將更象另一種形式的Operation Queues只是 Operation Queuse是用ObjectC的Dispatch Queues是C的

dispatch Queues有serial Queues 也被稱爲私有dispatch Queues，一個時間只能運行一個task，順序運行

dispatch_queue_t queue;
queue = dispatch_queue_create("myQueue", NULL);  

dispatch_async(queue, ^{
        printf("Do some work here.\n");
    });
    printf("The first block may or may not have run.\n");
    dispatch_sync(queue, ^{
        printf("Do some more work here.\n");
    });
    printf("Both blocks have completed.\n");

這裏使用了同步dispatch和異步dispatch，推薦使用dispatch_async這樣才能真正體現其中的優勢同步相當於WaitUntil ＝ YES

 

 

還有一種就是Concurrent Queues每個程序系統自動提供了3個Concurrent Queues

dispatch_queue_t aQueue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0);

dispatch_queue_t aHQueue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_HIGH, 0);
 dispatch_queue_t aLQueue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_LOW, 0);

啥意思一看就明白，3個優先級別的concurrent queues

 

最後一個特殊的Dispatch Queue就是main dispatch Queue 也是程序啓動自動生成

dispatch_queue_t mainQueue = dispatch_get_main_queue();

 

concurrent queues和main queue 都是由系統生成而且 dispatch_suspend, dispatch_resume, dispatch_set_context,這些函數對他們無效

 

但是我們的應用不是簡單的同步也異步的運行，應用經常是混合的

比如我們要task1 task2 task3 都運行完成後才能異步運行task4 task5 task6我們該怎麼做呢？這裏我們可以引入group的概念

 

    dispatch_queue_t aDQueue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0);
    dispatch_group_t group = dispatch_group_create();
    // Add a task to the group
    dispatch_group_async(group, aDQueue, ^{
        printf("task 1 \n");
    });
    dispatch_group_async(group, aDQueue, ^{
        printf("task 2 \n");
    });
    dispatch_group_async(group, aDQueue, ^{
        printf("task 3 \n");
    });
    printf("wait 1 2 3 \n");
    dispatch_group_wait(group, DISPATCH_TIME_FOREVER);
    printf("task 1 2 3 finished \n");
    dispatch_release(group);
    group = dispatch_group_create();
    // Add a task to the group
    dispatch_group_async(group, aDQueue, ^{
        printf("task 4 \n");
    });
    dispatch_group_async(group, aDQueue, ^{
        printf("task 5 \n");
    });
    dispatch_group_async(group, aDQueue, ^{
        printf("task 6 \n");
    });
    printf("wait 4 5 6 \n");
    dispatch_group_wait(group, DISPATCH_TIME_FOREVER);
    printf("task 4 5 6 finished \n");
    dispatch_release(group);

 

以上的運行結果：

The first block may or may not have run.

Do some work here.

Do some more work here.

Both blocks have completed.

wait to go task 1 2 3.

task1;

task2;

task3;

finished task 1 2 3.

wait 4 5 6 

task 4 

task 5 

task 6 

task 4 5 6 finished 

、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、、

有時候我們也可以將設定一個數據在queue中 也可以定義一個結束函數

dispatch_set_finalizer_f 是在dispatch_release時候被調用

    dispatch_queue_t serialQueue = dispatch_queue_create("com.example.CriticalTaskQueue", NULL);
    if (serialQueue)
    {
        dispatch_set_context(serialQueue, self);
        dispatch_set_finalizer_f(serialQueue, &myFinalizerFunction);
    }
    
    dispatch_group_t group = dispatch_group_create();
    
    // Add a task to the group
    dispatch_group_async(group, serialQueue, ^{
        printf("task 1 \n");
    });
    
    dispatch_group_async(group, serialQueue, ^{
        printf("task 2 \n");
    });
    
    dispatch_group_async(group, serialQueue, ^{
        printf("task 3 \n");
    });
    printf("wait 1 2 3 \n");
    dispatch_group_wait(group, DISPATCH_TIME_FOREVER);  
    dispatch_release(group);
    dispatch_release(serialQueue);