This time I will show you how promise replaces the callback function in the code. What are the precautions for promise to replace the callback function in the code. The following is a practical case. Let’s take a look. one time.
In the process of learning Node.js, I came into contact with how to use async to control concurrency (use async to control concurrency)
The essence of async is a process control. In fact, in asynchronous programming, there is a more classic model called the Promise/Deferred model (of course there are more related solutions, such as eventproxy, co, etc., you will be digging holes when you encounter it)
First, let's think about a typical asynchronous programming model, consider this topic: read a file, and output the content of the file on the console
var fs = require('fs'); fs.readFile('1.txt', 'utf8', function (err, data) { console.log(data); });
It seems very simple, go one step further: read two files, in The console outputs the contents of these two files
var fs = require('fs'); fs.readFile('1.txt', 'utf8', function (err, data) { console.log(data); fs.readFile('2.txt', 'utf8', function (err, data) { console.log(data); }); });
What if more files are read
var fs = require('fs'); fs.readFile('1.txt', 'utf8', function (err, data) { fs.readFile('2.txt', 'utf8', function (err, data) { fs.readFile('3.txt', 'utf8', function (err, data) { fs.readFile('4.txt', 'utf8', function (err, data) { // ... }); }); }); });
This is the legendary callback hell. You can use async to improve this code, but in this case In the example we need to use promise/defer to improve it
promise Basic concept
First of all, it is an object. It is no different from ordinary JavaScript objects. At the same time, it is also a specification. It agrees with a unified interface for asynchronous operations and represents the final result of an asynchronous operation. The code is written in a synchronous manner and the operations performed are asynchronous, but It also ensures that the order of program execution is synchronous
1. Promise has only three states, unfulfilled, completed (fulfilled) and failed (rejected)
2. The state of promise can be unfulfilled Convert to completed, or incomplete to failure
3. The state transition of promise only occurs once
promise has a then method, and the then method can accept 3 functions as parameters. The first two functions correspond to the callback functions of the two states of promise: fulfilled and rejected. The third function is used to process progress information
In order to understand it, some important principles must be remembered: .then() always returns a new promise, such as the following code:
var promise = readFile() var promise2 = promise.then(readAnotherFile, console.error)
Here The parameters readAnotherFile and console.error of then represent the action onFulfilled after the asynchronous operation succeeds or the action OnRejected after the failure. That is to say, the readAnotherFile function is executed after the file is successfully read, otherwise an error is printed and recorded upon failure. This implementation is only one of the two possible
It can also be understood as:
promiseSomething().then(function (fulfilled) { // 当 promise 状态变成 fulfilled 时,调用此函数 }, function (rejected) { // 当 promise 状态变成 rejected 时,调用此函数 }, function (progress) { // 当返回进度信息时,调用此函数 });
The two parts of the Promise law must be separated:
1. then() is always Returns a new promise every time you call it, it doesn't matter what the callback does, because .then() has given you a promise before the callback is called, the behavior of the callback only affects the implementation of the promise, if the callback returns a value, then the promise will use that value. If the value is a promise, return the value after the promise is implemented to this value. If the callback throws an error, the promise will reject the error
2. By .then() The returned promise is a new promise, which is different from the promise on which .then() was called. Long promise chains sometimes hide the fact that each time .then() is called, a new promise is generated. , it must be noted here that what you really need to consider is that your last call to .then() may represent a failure. If you do not catch this failure, it will easily cause your error exception to disappear.
Let’s look at an exploit q A simple example to handle this problem:
var Q = require('q'); var defer = Q.defer(); /** * 获取初始 promise * @private */ function getInitialPromise() { return defer.promise; } /** * 为 promise 设置三种状态的回调函数 */ getInitialPromise().then(function (success) { console.log(success); }, function (error) { console.log(error); }, function (progress) { console.log(progress); }); defer.notify('in progress'); // 控制台打印 in progress defer.resolve('resolve'); // 控制台打印 resolve defer.reject('reject'); // 没有输出。promise 的状态只能改变一次
Passing of promise
#then method will return a promise, in the following example , we use outputPromise to point to the promise returned by then.
var outputPromise = getInputPromise().then(function (fulfilled) { }, function (rejected) { });
Now outputPromise becomes a promise controlled by function(fulfilled) or function(rejected). The straightforward meaning is: when function(fulfilled) or function(rejected) returns a value, such as a string, array, object, etc., then the status of outputPromise will become fulfilled.
In the following example, we can see that when we change the status of inputPromise to fulfilled through defer.resovle(), the console outputs fulfilled.
When we change the status of inputPromise to fulfilled The status changes to rejected through defer.reject(), and the console outputs rejected
var Q = require('q'); var defer = Q.defer(); /** * 通过 defer 获得 promise * @private */ function getInputPromise() { return defer.promise; } /** * 当 inputPromise 状态由未完成变成 fulfil 时,调用 function(fulfilled) * 当 inputPromise 状态由未完成变成 rejected 时,调用 function(rejected) * 将 then 返回的 promise 赋给 outputPromise * function(fulfilled) 和 function(rejected) 通过返回字符串将 outputPromise 的状态由 * 未完成改变为 fulfilled * @private */ var outputPromise = getInputPromise().then(function (fulfilled) { return 'fulfilled'; }, function (rejected) { return 'rejected'; }); /** * 当 outputPromise 状态由未完成变成 fulfil 时,调用 function(fulfilled),控制台打印 'fulfilled: fulfilled'。 * 当 outputPromise 状态由未完成变成 rejected, 调用 function(rejected), 控制台打印 'rejected: rejected'。 */ outputPromise.then(function (fulfilled) { console.log('fulfilled: ' + fulfilled); }, function (rejected) { console.log('rejected: ' + rejected); }); /** * 将 inputPromise 的状态由未完成变成 rejected */ defer.reject(); // 输出 fulfilled: rejected /** * 将 inputPromise 的状态由未完成变成 fulfilled */ //defer.resolve(); // 输出 fulfilled: fulfilled
当 function(fulfilled) 或者 function(rejected) 抛出异常时,那么 outputPromise 的状态就会变成 rejected
var Q = require('q'); var fs = require('fs'); var defer = Q.defer(); /** * 通过 defer 获得 promise * @private */ function getInputPromise() { return defer.promise; } /** * 当 inputPromise 状态由未完成变成 fulfil 时,调用 function(fulfilled) * 当 inputPromise 状态由未完成变成 rejected 时,调用 function(rejected) * 将 then 返回的 promise 赋给 outputPromise * function(fulfilled) 和 function(rejected) 通过抛出异常将 outputPromise 的状态由 * 未完成改变为 reject * @private */ var outputPromise = getInputPromise().then(function (fulfilled) { throw new Error('fulfilled'); }, function (rejected) { throw new Error('rejected'); }); /** * 当 outputPromise 状态由未完成变成 fulfil 时,调用 function(fulfilled)。 * 当 outputPromise 状态由未完成变成 rejected, 调用 function(rejected)。 */ outputPromise.then(function (fulfilled) { console.log('fulfilled: ' + fulfilled); }, function (rejected) { console.log('rejected: ' + rejected); }); /** * 将 inputPromise 的状态由未完成变成 rejected */ defer.reject(); // 控制台打印 rejected [Error:rejected] /** * 将 inputPromise 的状态由未完成变成 fulfilled */ //defer.resolve(); // 控制台打印 rejected [Error:fulfilled]
当 function(fulfilled) 或者 function(rejected) 返回一个 promise 时,outputPromise 就会成为这个新的 promise.
这样做的意义在于聚合结果 (Q.all),管理延时,异常恢复等等
比如说我们想要读取一个文件的内容,然后把这些内容打印出来。可能会写出这样的代码:
// 错误的写法 var outputPromise = getInputPromise().then(function (fulfilled) { fs.readFile('test.txt', 'utf8', function (err, data) { return data; }); });
然而这样写是错误的,因为 function(fulfilled) 并没有返回任何值。需要下面的方式:
var Q = require('q'); var fs = require('fs'); var defer = Q.defer(); /** * 通过 defer 获得promise * @private */ function getInputPromise() { return defer.promise; } /** * 当 inputPromise 状态由未完成变成 fulfil时,调用 function(fulfilled) * 当 inputPromise 状态由未完成变成 rejected时,调用 function(rejected) * 将 then 返回的 promise 赋给 outputPromise * function(fulfilled) 将新的 promise 赋给 outputPromise * 未完成改变为 reject * @private */ var outputPromise = getInputPromise().then(function (fulfilled) { var myDefer = Q.defer(); fs.readFile('test.txt', 'utf8', function (err, data) { if (!err && data) { myDefer.resolve(data); } }); return myDefer.promise; }, function (rejected) { throw new Error('rejected'); }); /** * 当 outputPromise 状态由未完成变成 fulfil 时,调用 function(fulfilled),控制台打印 test.txt 文件内容。 * */ outputPromise.then(function (fulfilled) { console.log(fulfilled); }, function (rejected) { console.log(rejected); }); /** * 将 inputPromise 的状态由未完成变成 rejected */ //defer.reject(); /** * 将 inputPromise 的状态由未完成变成 fulfilled */ defer.resolve(); // 控制台打印出 test.txt 的内容
方法传递
方法传递有些类似于 Java 中的 try 和 catch。当一个异常没有响应的捕获时,这个异常会接着往下传递
方法传递的含义是当一个状态没有响应的回调函数,就会沿着 then 往下找
没有提供 function(rejected)
var outputPromise = getInputPromise().then(function (fulfilled) { })
如果 inputPromise 的状态由未完成变成 rejected, 此时对 rejected 的处理会由 outputPromise 来完成
var Q = require('q'); var fs = require('fs'); var defer = Q.defer(); /** * 通过defer获得promise * @private */ function getInputPromise() { return defer.promise; } /** * 当 inputPromise 状态由未完成变成 fulfil 时,调用 function(fulfilled) * 当 inputPromise 状态由未完成变成 rejected 时,这个 rejected 会传向 outputPromise */ var outputPromise = getInputPromise().then(function (fulfilled) { return 'fulfilled' }); outputPromise.then(function (fulfilled) { console.log('fulfilled: ' + fulfilled); }, function (rejected) { console.log('rejected: ' + rejected); }); /** * 将 inputPromise 的状态由未完成变成 rejected */ defer.reject('inputpromise rejected'); // 控制台打印 rejected: inputpromise rejected /** * 将 inputPromise的状态由未完成变成fulfilled */ //defer.resolve();
没有提供 function(fulfilled)
var outputPromise = getInputPromise().then(null, function (rejected) { })
如果 inputPromise 的状态由未完成变成 fulfilled, 此时对 fulfil 的处理会由 outputPromise 来完成
var Q = require('q'); var fs = require('fs'); var defer = Q.defer(); /** * 通过defer获得promise * @private */ function getInputPromise() { return defer.promise; } /** * 当 inputPromise 状态由未完成变成 fulfil时,传递给 outputPromise * 当 inputPromise 状态由未完成变成 rejected时,调用 function(rejected) * function(fulfilled) 将新的 promise 赋给 outputPromise * 未完成改变为 reject * @private */ var outputPromise = getInputPromise().then(null, function (rejected) { return 'rejected'; }); outputPromise.then(function (fulfilled) { console.log('fulfilled: ' + fulfilled); }, function (rejected) { console.log('rejected: ' + rejected); }); /** * 将 inputPromise 的状态由未完成变成 rejected */ // defer.reject('inputpromise rejected'); /** * 将 inputPromise 的状态由未完成变成fulfilled */ defer.resolve('inputpromise fulfilled'); // 控制台打印fulfilled: inputpromise fulfilled
可以使用 fail(function(error)) 来专门针对错误处理,而不是使用 then(null,function(error))
var outputPromise = getInputPromise().fail(function (error) { })
看这个例子:
var Q = require('q'); var fs = require('fs'); var defer = Q.defer(); /** * 通过defer获得promise * @private */ function getInputPromise() { return defer.promise; } /** * 当 inputPromise 状态由未完成变成 fulfil 时,调用 then(function(fulfilled)) * 当 inputPromise 状态由未完成变成 rejected 时,调用 fail(function(error)) * function(fulfilled) 将新的 promise 赋给 outputPromise * 未完成改变为reject * @private */ var outputPromise = getInputPromise().then(function (fulfilled) { return fulfilled; }).fail(function (error) { console.log('fail: ' + error); }); /** * 将 inputPromise 的状态由未完成变成 rejected */ defer.reject('inputpromise rejected');// 控制台打印 fail: inputpromise rejected /** * 将 inputPromise 的状态由未完成变成 fulfilled */ //defer.resolve('inputpromise fulfilled');
可以使用 progress(function (progress)) 来专门针对进度信息进行处理,而不是使用 then(function (success) { }, function (error) { }, function (progress) { })
var Q = require('q'); var defer = Q.defer(); /** * 获取初始 promise * @private */ function getInitialPromise() { return defer.promise; } /** * 为 promise 设置 progress 信息处理函数 */ var outputPromise = getInitialPromise().then(function (success) { }).progress(function (progress) { console.log(progress); }); defer.notify(1); defer.notify(2); // 控制台打印 1,2
promise 链
promise 链提供了一种让函数顺序执行的方法
函数顺序执行是很重要的一个功能。比如知道用户名,需要根据用户名从数据库中找到相应的用户,然后将用户信息传给下一个函数进行处理
var Q = require('q'); var defer = Q.defer(); // 一个模拟数据库 var users = [{ 'name': 'andrew', 'passwd': 'password' }]; function getUsername() { return defer.promise; } function getUser(username) { var user; users.forEach(function (element) { if (element.name === username) { user = element; } }); return user; } // promise 链 getUsername().then(function (username) { return getUser(username); }).then(function (user) { console.log(user); }); defer.resolve('andrew');
我们通过两个 then 达到让函数顺序执行的目的。
then 的数量其实是没有限制的。当然,then 的数量过多,要手动把他们链接起来是很麻烦的。比如
foo(initialVal).then(bar).then(baz).then(qux)
这时我们需要用代码来动态制造 promise 链
var funcs = [foo, bar, baz, qux] var result = Q(initialVal) funcs.forEach(function (func) { result = result.then(func) }) return result
当然,我们可以再简洁一点
var funcs = [foo, bar, baz, qux] funcs.reduce(function (pre, current),Q(initialVal){ return pre.then(current) })
看一个具体的例子
function foo(result) { console.log(result); return result + result; } // 手动链接 Q('hello').then(foo).then(foo).then(foo); // 控制台输出: hello // hellohello // hellohellohello // 动态链接 var funcs = [foo, foo, foo]; var result = Q('hello'); funcs.forEach(function (func) { result = result.then(func); }); // 精简后的动态链接 funcs.reduce(function (prev, current) { return prev.then(current); }, Q('hello'));
对于 promise 链,最重要的是需要理解为什么这个链能够顺序执行。如果能够理解这点,那么以后自己写 promise 链可以说是轻车熟路啊
promise 组合
回到我们一开始读取文件内容的例子。如果现在让我们把它改写成 promise 链,是不是很简单呢?
var Q = require('q'), fs = require('fs'); function printFileContent(fileName) { return function () { var defer = Q.defer(); fs.readFile(fileName, 'utf8', function (err, data) { if (!err && data) { console.log(data); defer.resolve(); } }) return defer.promise; } } // 手动链接 printFileContent('sample01.txt')() .then(printFileContent('sample02.txt')) .then(printFileContent('sample03.txt')) .then(printFileContent('sample04.txt')); // 控制台顺序打印 sample01 到 sample04 的内容
很有成就感是不是。然而如果仔细分析,我们会发现为什么要他们顺序执行呢,如果他们能够并行执行不是更好吗? 我们只需要在他们都执行完成之后,得到他们的执行结果就可以了
我们可以通过 Q.all([promise1,promise2...]) 将多个 promise 组合成一个 promise 返回。 注意:
1. 当 all 里面所有的 promise 都 fulfil 时,Q.all 返回的 promise 状态变成 fulfil
2. 当任意一个 promise 被 reject 时,Q.all 返回的 promise 状态立即变成 reject
我们来把上面读取文件内容的例子改成并行执行吧
var Q = require('q'); var fs = require('fs'); /** *读取文件内容 *@private */ function printFileContent(fileName) { // Todo: 这段代码不够简洁。可以使用 Q.denodeify 来简化 var defer = Q.defer(); fs.readFile(fileName, 'utf8', function (err, data) { if (!err && data) { console.log(data); defer.resolve(fileName + ' success '); } else { defer.reject(fileName + ' fail '); } }) return defer.promise; } Q.all([printFileContent('sample01.txt'), printFileContent('sample02.txt'), printFileContent('sample03.txt'), printFileContent('sample04.txt')]) .then(function (success) { console.log(success); }); // 控制台打印各个文件内容 顺序不一定
现在知道 Q.all 会在任意一个 promise 进入 reject 状态后立即进入 reject 状态。如果我们需要等到所有的 promise 都发生状态后(有的 fulfil, 有的 reject),再转换 Q.all 的状态, 这时我们可以使用 Q.allSettled
var Q = require('q'), fs = require('fs'); /** *读取文件内容 *@private */ function printFileContent(fileName) { // Todo: 这段代码不够简洁。可以使用Q.denodeify来简化 var defer = Q.defer(); fs.readFile(fileName, 'utf8', function (err, data) { if (!err && data) { console.log(data); defer.resolve(fileName + ' success '); } else { defer.reject(fileName + ' fail '); } }) return defer.promise; } Q.allSettled([printFileContent('nosuchfile.txt'), printFileContent('sample02.txt'), printFileContent('sample03.txt'), printFileContent('sample04.txt')]) .then(function (results) { results.forEach( function (result) { console.log(result.state); } ); });
结束 promise 链
通常,对于一个 promise 链,有两种结束的方式。第一种方式是返回最后一个 promise
如 return foo().then(bar);
第二种方式就是通过 done 来结束 promise 链
如 foo().then(bar).done()
为什么需要通过 done 来结束一个 promise 链呢? 如果在我们的链中有错误没有被处理,那么在一个正确结束的 promise 链中,这个没被处理的错误会通过异常抛出
var Q = require('q'); function getPromise(msg, timeout, opt) { var defer = Q.defer(); setTimeout(function () { console.log(msg); if (opt) defer.reject(msg); else defer.resolve(msg); }, timeout); return defer.promise; } /** * 没有用 done() 结束的 promise 链 * 由于 getPromse('2',2000,'opt') 返回 rejected, getPromise('3',1000) 就没有执行 * 然后这个异常并没有任何提醒,是一个潜在的 bug */ getPromise('1', 3000) .then(function () { return getPromise('2', 2000, 'opt') }) .then(function () { return getPromise('3', 1000) }); /** * 用 done() 结束的 promise 链 * 有异常抛出 */ getPromise('1', 3000) .then(function () { return getPromise('2', 2000, 'opt') }) .then(function () { return getPromise('3', 1000) }) .done();
附录:一个 Promise 简单的应用(Node.js笔记(5)promise)
附:Promises/A+ 规范
promise 代表一个异步操作的最终结果。主要通过 promise 的 then 方法订阅其最终结果的处理回调函数,和订阅因某原因无法成功获取最终结果的处理回调函数。
更对详细见:Promises/A+
A 与 A+ 的不同点
A+ 规范通过术语 thenable 来区分 promise 对象
A+ 定义 onFulfilled/onRejectd 必须是作为函数来调用,而且调用过程必须是异步的
A+ 严格定义了 then 方法链式调用时,onFulfilled/onRejectd 的调用顺序
相信看了本文案例你已经掌握了方法,更多精彩请关注php中文网其它相关文章!
推荐阅读:
The above is the detailed content of How to replace the callback function in the code with promise. For more information, please follow other related articles on the PHP Chinese website!