• 同步异步执行问题


    多个网络请求同时执行,等所有网络请求完成,再统一做其他操作,我们可能会想到dispatch_group_async、dispatch_group_notify结合使用。

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    dispatch_queue_t queue = dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0);
        dispatch_group_t group = dispatch_group_create();
        dispatch_group_async(group, queue, ^{
            NSLog(@"任务一完成");
        });
         
        dispatch_group_async(group, queue, ^{
            NSLog(@"任务二完成");
        });
         
        dispatch_group_async(group, queue, ^{
            NSLog(@"任务三完成");
        });
        //在分组的所有任务完成后触发
        dispatch_group_notify(group, queue, ^{
            NSLog(@"所有任务完成");
        });

    或者使用栅栏

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    - (void)barrier {
    //    dispatch_queue_t queue = dispatch_queue_create("com.lai.www", DISPATCH_QUEUE_CONCURRENT);
        dispatch_queue_t queue =dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT,0);
         
        dispatch_async(queue, ^{
            NSLog(@"任务1-1完成");
        });
        dispatch_async(queue, ^{
            NSLog(@"任务1-2完成");
        });
         
        dispatch_async(queue, ^{
            NSLog(@"任务1-3完成");
        });
         
        dispatch_barrier_async(queue, ^{
            NSLog(@"以上任务都完成 dispatch_barrie完成");
        });
    }

    比如上述写法,内部执行的是同步操作没有问题,如果以上三个任务都是异步的,比如是网络请求,那么就达不到我们想要的效果。因为异步,请求没有回来,dispatch_group_notify或者dispatch_barrier_async已经执行了。

    我们可以采用信号量或者dispatch_group_enter、dispatch_group_leave实现。

    核心思想:将异步变成同步

    下述描述了常见场景下的代码实现:包括顺序执行和同时执行异步操作

    顺序执行 :

    方式 : 信号量semaphore (必须放在子线程 dispatch_semaphore_wait会卡死主线程) (例:1执行完了执行2)
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    - (void)serialBySemaphore {
        dispatch_async(dispatch_get_global_queue(0, 0), ^{
     
            dispatch_semaphore_t semaphore = dispatch_semaphore_create(0);
     
            [self requestOneWithSuccessBlock:^{
                dispatch_semaphore_signal(semaphore);
            }];
     
            dispatch_semaphore_wait(semaphore, DISPATCH_TIME_FOREVER);
     
            [self requestTwoWithBlock:^{
            }];
        });
    }
    执行到dispatch_semaphore_wait时,由于信号量为0,进行等待,请求1完成后调用dispatch_semaphore_signal ,信号量不再为0,接着执行请求2

    方式: GCD dispatch_group_enter/leave (例: 1, 2 同时执行 执行完了再执行3)
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    -(void)serialByGroupWait {
         
        dispatch_group_t group = dispatch_group_create();
         
        dispatch_group_enter(group);
        [self requestOneWithSuccessBlock:^{
            dispatch_group_leave(group);
        }];
         
        dispatch_group_enter(group);
        [self requestTwoWithBlock:^{
            dispatch_group_leave(group);
        }];
      // 1  2同时执行
         
        dispatch_group_wait(group, DISPATCH_TIME_FOREVER);// 1 2 执行完 下面才会执行
         
        dispatch_group_enter(group);
        [self requestThreeWithBlock:^{
            dispatch_group_leave(group);
        }];
         
      // 1 2 3 都完成 才会执行
        dispatch_group_notify(group, dispatch_get_global_queue(0, 0), ^{
            NSLog(@"all request  done!");
        });
    }
    执行到dispatch_group_wait时,由于enter数不等于leave数,进行等待,请求1,2都完成后调用dispatch_group_leave ,enter数等于leave数,接着执行请求3。 请求1,2,3都执行后,dispatch_group_notify执行

    方式三:回调中执行
    1
    2
    3
    4
    5
    6
    - (void) serialByCallBack {
        [self requestOneWithSuccessBlock:^{
            [self requestTwoWithBlock:^{
            }];
        }];
    }
    low方法,请求一多,嵌套恶心


    同时执行 :

    方式:信号量
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    -(void)concurrentBySemaphore {
     dispatch_group_t group = dispatch_group_create();
         
        dispatch_group_async(group, dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{
            dispatch_semaphore_t sema = dispatch_semaphore_create(0);
            [self requestOneWithSuccessBlock:^{
                dispatch_semaphore_signal(sema);
            }];
            dispatch_semaphore_wait(sema, DISPATCH_TIME_FOREVER);
        });
         
        dispatch_group_async(group, dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{
            dispatch_semaphore_t sema = dispatch_semaphore_create(0);
            [self requestTwoWithSuccessBlock:^{
                dispatch_semaphore_signal(sema);
            }];
            dispatch_semaphore_wait(sema, DISPATCH_TIME_FOREVER);
        });
        dispatch_group_notify(group, dispatch_get_main_queue(), ^{
            NSLog(@"全部搞完了");
        });
    }

    方式:dispatch_group_enter
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    -(void)concurrentByGroup {
         
        dispatch_group_t group = dispatch_group_create();
         
        dispatch_group_enter(group);
        [self requestOneWithSuccessBlock:^{
            dispatch_group_leave(group);
        }];
         
        dispatch_group_enter(group);
        [self requestTwoWithBlock:^{
            dispatch_group_leave(group);
        }];
         
      // 1 2  都完成 才会执行
        dispatch_group_notify(group, dispatch_get_global_queue(0, 0), ^{
            NSLog(@"all request  done!");
        });
    }


    扩充:循环请求情况 顺序请求/同时请求


    *模拟循环网络请求 同时进行 统一回调 (GCD + 信号量方式)
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    - (void)concurrentTest1 {
        dispatch_group_t group = dispatch_group_create();
        for (int i = 0 ; i < 5; i++) {
            dispatch_group_async(group, dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{
                dispatch_semaphore_t sema = dispatch_semaphore_create(0);
                 // 模拟请求 ↓
                dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{
                    sleep(3);
                    NSLog(@"任务%d完成",i);
                     dispatch_semaphore_signal(sema);
                });
                 // 模拟请求 上
                dispatch_semaphore_wait(sema, DISPATCH_TIME_FOREVER);
            });
        }
        dispatch_group_notify(group, dispatch_get_main_queue(), ^{
            NSLog(@"全部搞完了");
        });
    }

    执行结果:

    1
    2
    3
    4
    5
    6
    00:21:38.413337+0800 mmmm[4595:182516] 任务1完成
    00:21:38.413337+0800 mmmm[4595:182518] 任务3完成
    00:21:38.413358+0800 mmmm[4595:182517] 任务2完成
    00:21:38.413358+0800 mmmm[4595:182515] 任务0完成
    00:21:38.413447+0800 mmmm[4595:182519] 任务4完成
    00:21:38.413843+0800 mmmm[4595:182428] 全部搞完了



    *模拟循环网络请求 同时进行 统一回调 (GCD + group enter/leave 方式)
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    - (void)concurrentTest2 {
        dispatch_group_t group = dispatch_group_create();
        for (int i = 0 ; i < 5; i++) {
            dispatch_group_enter(group);
              // 模拟请求 ↓
            dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{
                sleep(3);
                NSLog(@"任务%d完成",i);
                dispatch_group_leave(group);
            });
              // 模拟请求 ↑
        }
        dispatch_group_notify(group, dispatch_get_main_queue(), ^{
            NSLog(@"全部搞完了");
        });
    }

    执行结果:

    1
    2
    3
    4
    5
    6
    2019-04-12 00:26:38.607040+0800 mmmm[4641:184613] 任务0完成
    2019-04-12 00:26:38.607043+0800 mmmm[4641:184612] 任务2完成
    2019-04-12 00:26:38.607059+0800 mmmm[4641:184611] 任务3完成
    2019-04-12 00:26:38.607067+0800 mmmm[4641:184610] 任务1完成
    2019-04-12 00:26:38.607088+0800 mmmm[4641:184625] 任务4完成
    2019-04-12 00:26:38.607353+0800 mmmm[4641:184559] 全部搞完了

    *模拟循环网络请求 顺序进行 (GCD + 信号量方式)
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    - (void)serialTest1 {
        dispatch_semaphore_t sema = dispatch_semaphore_create(0);
        dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{
            for (int i = 0 ; i < 5; i++) {
                NSLog(@"开始%d",i);
                // 模拟请求 ↓
                dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{
                    sleep(3);
                    NSLog(@"任务%d完成",i);
                    dispatch_semaphore_signal(sema);
                });
                // 模拟请求 上
               dispatch_semaphore_wait(sema, DISPATCH_TIME_FOREVER);
            }
             
           NSLog(@"全部搞完了");
        });
    }

    执行结果

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    2019-04-12 01:31:45.291818+0800 mmmm[5417:215007] 开始0
    2019-04-12 01:31:48.297299+0800 mmmm[5417:215008] 任务0完成
    2019-04-12 01:31:48.297746+0800 mmmm[5417:215007] 开始1
    2019-04-12 01:31:51.298592+0800 mmmm[5417:215008] 任务1完成
    2019-04-12 01:31:51.298841+0800 mmmm[5417:215007] 开始2
    2019-04-12 01:31:54.300477+0800 mmmm[5417:215008] 任务2完成
    2019-04-12 01:31:54.300908+0800 mmmm[5417:215007] 开始3
    2019-04-12 01:31:57.305197+0800 mmmm[5417:215008] 任务3完成
    2019-04-12 01:31:57.305623+0800 mmmm[5417:215007] 开始4
    2019-04-12 01:32:00.311062+0800 mmmm[5417:215008] 任务4完成
    2019-04-12 01:32:00.311407+0800 mmmm[5417:215007] 全部搞完了

    *模拟循环网络请求 顺序进行 (GCD + group enter/leave 方式)
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    - (void)serialTest2 {
        dispatch_group_t group = dispatch_group_create();
        for (int i = 0 ; i < 5; i++) {
            dispatch_group_enter(group);
            // 模拟请求 ↓
            dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{
                sleep(6 - i);
                NSLog(@"任务%d完成",i);
                dispatch_group_leave(group);
            });
            // 模拟请求 ↑
            dispatch_group_wait(group, DISPATCH_TIME_FOREVER); // 顺序执行与同步执行的不同点
        }
        dispatch_group_notify(group, dispatch_get_main_queue(), ^{
            NSLog(@"全部搞完了");
        });
    }

    执行结果

    1
    2
    3
    4
    5
    6
    2019-04-12 00:33:46.020376+0800 mmmm[4748:188409] 任务0完成
    2019-04-12 00:33:51.021098+0800 mmmm[4748:188409] 任务1完成
    2019-04-12 00:33:55.022758+0800 mmmm[4748:188409] 任务2完成
    2019-04-12 00:33:58.023783+0800 mmmm[4748:188409] 任务3完成
    2019-04-12 00:34:00.027929+0800 mmmm[4748:188409] 任务4完成
    2019-04-12 00:34:00.028444+0800 mmmm[4748:188374] 全部搞完了
  • 相关阅读:
    网络基础
    python之bootstrap(组件,SweetAlert插件)
    python之网页小图标的使用
    python之bootstrap(基本)
    python之JQuery(批量操作,监听按键)
    python之JQuery(文档操作,事件委托,事件冒泡)
    python之JQuery(动画点赞实例)
    python之JQuery(hover,input值动态变化的实现)
    Less-1
    python100例 21-30
  • 原文地址:https://www.cnblogs.com/Mr-zyh/p/12191759.html
Copyright © 2020-2023  润新知