I am trying to run a test suite for an object that returns a promise. I want to chain several actions together with short timeouts between them. I thought that a "then" call which returned a promise would wait for the promise to be fulfilled before firing the next chained then call.
I created a function
function promiseTimeout (time) {
return new Promise(function(resolve,reject){
setTimeout(function(){resolve(time);},time);
});
};
to try and wrap setTimeout in a Promise.
Then in my test suite, I am calling something like this ...
it('should restore state when browser back button is used',function(done){
r.domOK().then(function(){
xh.fire('akc-route-change','/user/4/profile/new');
}).then(promiseTimeout(2000)).then(function(t){
xu.fire('akc-route-change','/user/6');
}).then(promiseTimeout(10)).then(function(t){
expect(xu.params[0]).to.equal(6);
history.back();
}).then(promiseTimeout(10)).then(function(){
expect(xu.params[0]).to.equal(4);
done();
});
});
I can put a breakpoint on the first xh.fire call and a second one on the xu.fire call and would have expected a two second gap when a continues from the first breakpoint to the second.
Instead it reaches the second breakpoint immediately, and the value of t at that point is undefined.
What am I doing wrong?
TL;DR - you've wrapped setTimeout in a promise properly, the issue is you are using it improperly
.then(promiseTimeout(2000)).then
will not do what you expect. The "signature" for .then is then(functionResolved, functionRejected)
A promise’s then method accepts two arguments:
promise.then(onFulfilled, onRejected)
Both onFulfilled and onRejected are optional arguments:
- If onFulfilled is not a function, it must be ignored.
- If onRejected is not a function, it must be ignored.
source: https://promisesaplus.com/#point-21
You are not passing a function to then
Consider the way you are doing it:
Promise.resolve('hello')
.then(promiseTimeout(2000))
.then(console.log.bind(console))
vs how it should be done:
Promise.resolve('hello').then(function() {
return promiseTimeout(2000)
}).then(console.log.bind(console))
The first outputs 'hello' immediately
The second outputs 2000 after 2 seconds
Therefore, you should be doing:
it('should restore state when browser back button is used', function(done) {
r.domOK().then(function() {
xh.fire('akc-route-change', '/user/4/profile/new');
}).then(function() {
return promiseTimeout(2000);
}).then(function(t) {
xu.fire('akc-route-change', '/user/6');
}).then(function() {
return promiseTimeout(10);
}).then(function(t) {
expect(xu.params[0]).to.equal(6);
history.back();
}).then(function() {
return promiseTimeout(10);
}).then(function() {
expect(xu.params[0]).to.equal(4);
done();
});
});
Alternatively:
it('should restore state when browser back button is used', function(done) {
r.domOK().then(function() {
xh.fire('akc-route-change', '/user/4/profile/new');
}).then(promiseTimeout.bind(null, 2000)
).then(function(t) {
xu.fire('akc-route-change', '/user/6');
}).then(promiseTimeout.bind(null, 10)
).then(function(t) {
expect(xu.params[0]).to.equal(6);
history.back();
}).then(promiseTimeout.bind(null, 10)
).then(function() {
expect(xu.params[0]).to.equal(4);
done();
});
});
EDIT: March 2019
Over the years, things have changed a lot - arrow notation makes this even easier
Firstly, I would define promiseTimeout differently
const promiseTimeout = time => () => new Promise(resolve => setTimeout(resolve, time, time));
The above returns a function that can be called to create a "promise delay" and resolves to the time (length of delay). Thinking about this, I can't see why that would be very useful, rather I'd:
const promiseTimeout = time => result => new Promise(resolve => setTimeout(resolve, time, result));
The above would resolve to the result of the previous promise (far more useful)
But it's a function that returns a function, so the rest of the ORIGINAL code could be left unchanged. The thing about the original code, however, is that no values are needed to be passed down the .then chain, so, even simpler
const promiseTimeout = time => () => new Promise(resolve => setTimeout(resolve, time));
and the original code in the question's it block can now be used unchanged
it('should restore state when browser back button is used',function(done){
r.domOK().then(function(){
xh.fire('akc-route-change','/user/4/profile/new');
}).then(promiseTimeout(2000)).then(function(){
xu.fire('akc-route-change','/user/6');
}).then(promiseTimeout(10)).then(function(){
expect(xu.params[0]).to.equal(6);
history.back();
}).then(promiseTimeout(10)).then(function(){
expect(xu.params[0]).to.equal(4);
done();
});
});
To make a timeout which works as you want, write a function which takes a delay, and returns a function suitable for passing to then.
function timeout(ms) {
return () => new Promise(resolve => setTimeout(resolve, ms));
}
Use it like this:
Promise.resolve() . then(timeout(1000)) . then(() => console.log("got here"););
However, it is likely that you will want to access the resolved value of the promise leading into the timeout. In that case, arrange for the function created by timeout() to pass through the value:
function timeout(ms) {
return value => new Promise(resolve => setTimeout(() => resolve(value), ms));
}
Use it like this:
Promise.resolve(42) . then(timeout(1000)) . then(value => console.log(value));
await new Promise((resolve, reject)=>{
// wait for 50 ms.
setTimeout(function(){resolve()}, 50);
});
console.log("This will appear after waiting for 50 ms");
This can be used in an async function and the execution will wait till given interval.
This as already been answered above, but I feel this could be done easily with:
const setTimeoutPromise = ms => new Promise(resolve => setTimeout(resolve, ms))
setTimeoutProise function accept wait time in ms and passes it down to the setTimeout function. Once the wait time is over, the resolve method passed down to the promise is executed.
Which could be used like this:
setTimeoutPromise(3000).then(doSomething)
Another approach for adding delays to Promises without having to predefine or import a helper function (that I personally like best) is to extend the property of the Promise constructor:
Promise.prototype.delay = function (ms) {
return new Promise(resolve => {
window.setTimeout(this.then.bind(this, resolve), ms);
});
}
I'm leaving out the reject callback since this is meant to always resolve.
DEMO
Promise.prototype.delay = function(ms) {
console.log(`[log] Fetching data in ${ms / 1000} second(s)...`);
return new Promise(resolve => {
window.setTimeout(this.then.bind(this, resolve), ms);
});
}
document.getElementById('fetch').onclick = function() {
const duration = 1000 * document.querySelector('#duration[type="number"]').value;
// Promise.resolve() returns a Promise
// and this simply simulates some long-running background processes
// so we can add some delays on it
Promise
.resolve('Some data from server.')
.delay(duration)
.then(console.log);
}
<div>
<input id="duration" type="number" value="3" />
<button id="fetch">Fetch data from server</button>
</div>
Or if you need it to also be .catch()-able, here is when you need the second argument (reject). Note that the catch() handling will also occur after the delay:
Promise.prototype.delay = function(ms) {
console.log(`[log] Fetching data in ${ms / 1000} second(s)...`);
return new Promise((resolve, reject) => {
window.setTimeout(() => {
this.then(resolve).catch(reject);
}, ms);
});
}
document.getElementById('fetch').onclick = function() {
const duration = 1000 * document.querySelector('#duration[type="number"]').value;
Promise
.reject('Some data from server.')
.delay(duration)
.then(console.log)
.catch(console.log); // Promise rejection or failures will always end up here
}
<div>
<input id="duration" type="number" value="3" />
<button id="fetch">Fetch data from server</button>
</div>
