What does Generator mean in ES6? Introduction to Generator in ES6

The content of this article is about what does Generator mean in ES6? The introduction of Generator in ES6 has certain reference value. Friends in need can refer to it. I hope it will be helpful to you.

Generator in ES6

Generator is a very interesting feature of ES6, and it is also a feature that is not easy to understand. Different from the obvious purpose of let/const providing block-level scope, what is the purpose of this thing?

First of all, we need to clarify a problem. In JavaScript, as long as any function starts to execute, it cannot be stopped until the execution is completed (don’t mention debugging to me, have you ever seen any user who uses the product and has to turn it on? debug your code).

But, Generator provides this capability. Look at the following code:

function *g(){
    console.log('start');
    yield 1;
    console.log('middle');
    yield 2;
    console.log('end');
}
var g1 = g();
console.log(g1.next()); 
// start
// {value: 1, done: false}
console.log(g1.next());  
// middle
// {value: 2, done: false}
console.log(g1.next());  
// end
// {value: undefined, done: true}

According to the output results, we see that in function g, when the yield keyword is encountered, the running program will stop. Only by calling the next() method will the code in function g continue to be executed. So function g itself has a paused state.

At this point, we need to know:

• Generator is not a function, not a function, not a function;

• g() is not It will start execution immediately, but will pause as soon as it comes up and return an Iterator object;

• Every time g1.next() will break the pause state and execute until it encounters the next yield Or return

• When encountering yield, the expression after yield will be executed and the value after execution will be returned, and then it will enter the paused state again, at this time done: false.

• When a return is encountered, a value will be returned and the execution ends, that is, done: true

• The return value of g.next() each time It is always in the form of {value: ..., done: ...}

Generator and asynchronous

Since Generator can function stop Later, some people with bright minds thought about whether they could use Generator to process asynchronous programs.

Let’s first look at a traditional example:

    function asyn(fn) {
        return new Promise((resolve,reject)=>{
            setTimeout(()=>{
                fn();
                resolve(true);
            }, 1000);
        });
    }

    function main() {
        console.log('start');
        asyn(function(d) {
            console.log('async one');
            asyn(function(d) {
                console.log('async two');
                console.log('end');
            });
        });
    }

    main();

Then let’s look at an asynchronous program using Generator:

    function asyn(fn) {
        return new Promise((resolve,reject)=>{
            setTimeout(()=>{
                fn();
                resolve(true);
            }, 1000);
        });
    }

    co(function*() {
        console.log('start');
        yield asyn(function(d) {
            console.log('async one');
        });

        yield asyn(function(d) {
            console.log('async two');
        });
        console.log('end');
    });

    function co(fn) {

        return new Promise((resolve,reject)=>{
            let g = fn();

            onFullfilled();
            function onFullfilled() {
                let ret = null;
                ret = g.next();
                next(ret);
            }

            function next(ret) {
                if(ret.done) return resolve(ret.value);
                ret.value.then(onFullfilled);
            }
        } );
    }

The function does not use nested callbacks in the asynchronous program, but directly uses synchronization way to write it down. The principle is probably that there are two asynchronous programs, and Xiao Ming and Xiao Hong are used to refer to them. Xiaohong needs to wait for Xiaoming to finish executing before it can be executed. Then when we execute Xiaoming, we pause the running of the program (through yield), wait until Xiaoming returns the result, and then execute the following Xiaohong (next()).

In the above program, we added a co function, which is used to allow the Generator to execute automatically. To put it bluntly, when the first asynchronous function returns, the next() method is automatically called to run the following code.

Generator and Koa

Koa is a web application framework based on Node.js. In Koa, the asynchronous programs processed are mainly network requests (HTTP), file reading and data query. There are many asynchronous scenarios here. If you add program layering and use the traditional callback method, there will be too many callbacks.

app.on('get', function(){
    auth(function(){
        router(function(){
            find(function(){
                save(function(){
                    render(function(){
                        //......
                    })
                })
            })
        })
    })

})

This way of writing is extremely detrimental to program maintenance and is inconvenient at all. After having the generator, we can write the program in the same way as above. The original version of Koa used this method to convert intermediate processing programs into "yields" (middleware). Processing client requests in the form of middleware makes the development of App applications more flexible and not limited by the framework itself.

In the latest Koa2, Genetator has been abandoned and instead uses async/await.

But no matter which method is used, the essence is to use Promise.

The above is the detailed content of What does Generator mean in ES6? Introduction to Generator in ES6. For more information, please follow other related articles on the PHP Chinese website!