setInterval
Async
Modern version of setInterval for promises and async functions available in Node.js and browsers.
setIntervalAsync
works both on Node.js and in the browser, providing the same
familiar interface as setInterval
for asynchronous functions, while preventing
multiple executions from overlapping in time.
Getting Started
Node.js
First, install setIntervalAsync
using npm or yarn:
# Using npm:
npm install set-interval-async
# Or using yarn:
yarn add set-interval-async
Now, you can require setIntervalAsync
in CommonJS:
const { setIntervalAsync, clearIntervalAsync } = require('set-interval-async');
Or else, you can use ES6 modules syntax:
import { setIntervalAsync, clearIntervalAsync } from 'set-interval-async';
Browser
In the browser, you can add a script tag in your HTML:
<!-- Load from unpkg.com -->
<script src="https://unpkg.com/set-interval-async"></script>
<!-- Or from jsDelivr -->
<script src="https://cdn.jsdelivr.net/npm/set-interval-async"></script>
After the script is loaded, a variable called SetIntervalAsync
will be defined in the global context. From there you can retrieve the setIntervalAsync
and clearIntervalAsync
functions.
var setIntervalAsync = SetIntervalAsync.setIntervalAsync;
var clearIntervalAsync = SetIntervalAsync.clearIntervalAsync;
Usage
In the most basic scenario, you can use setIntervalAsync
the same way you would use vanilla setInterval
. For example, the following code will print 'Hello' to the console once every second.
const { setIntervalAsync, clearIntervalAsync } = require('set-interval-async');
setIntervalAsync(() => {
console.log('Hello')
}, 1000);
However, you can also provide an async function (or a function returning a promise), which has the added nicety that now you can wait until the cycle is fully stopped before moving on by using clearIntervalAsync
.
const { setIntervalAsync, clearIntervalAsync } = require('set-interval-async');
const timer = setIntervalAsync(async () => {
console.log('Hello')
await doSomeWork()
console.log('Bye')
}, 1000);
// Or equivalently:
const timer = setIntervalAsync(() => {
console.log('Hello')
return doSomeWork().then(
() => console.log('Bye')
)
}, 1000);
// Later:
await clearIntervalAsync(timer);
// At this point, all timers have been cleared, and the last
// execution is guaranteed to have finished as well.
This is particularly useful when - for example at the end of a unit test - you want to make sure that no asynchronous code continues running by the time your test manager moves on to the next one.
it('should test something', async () => {
const timer = setIntervalAsync(/* ... */);
// Do some assertions.
await clearIntervalAsync(timer);
// Interval is now fully stopped.
});
When Should I Use setIntervalAsync
?
Where setIntervalAsync
really shines is in those situations where the given asynchronous function might take longer to compute than the configured interval and, at the same time, is not safe to execute more than once at a time. Using vanilla setInterval
will break your code in this scenario, whereas setIntervalAsync
guarantees that the function will never execute more than once at the same time.
For example, consider the following code:
async function processQueue (queue) {
if (queue.length === 0) {
return;
}
let head = queue[0];
await doSomeWork(head);
queue.shift(); // Removes the first element.
}
The function above should never get called again before the previous execution is completed. Otherwise, the queue's first element will get processed twice, and the second element will be skipped.
However, with setIntervalAsync
, the following code is perfectly safe:
setIntervalAsync(processQueue, 1000, queue)
since setIntervalAsync
will guarantee that the function is never executed more than once at any given moment.
You can choose whether you wish to use the Dynamic
or Fixed
strategies, which will either launch every execution as soon as possible or set a fixed delay between the end of one execution and the start of the next one. See Dynamic and Fixed setIntervalAsync
for more details.
Dynamic and Fixed setIntervalAsync
setIntervalAsync
provides two strategies which can be used to prevent a recurring function from executing more than once at any given moment:
-
Dynamic: If possible, the given function is called once every
interval
milliseconds. If any execution takes longer than the desired interval, the next execution is delayed until the previous one has finished, and called immediately after this condition is reached. -
Fixed: The given function is called repeatedly, guaranteeing a fixed delay of
interval
milliseconds between the end of one execution and the start of the following one.
You can choose whichever strategy works best for your application. When in doubt, the Dynamic
strategy will likely suffice for most use cases, keeping the interval as close as possible to the desired one. The default strategy is Dynamic
.
In Node.js
You can require a specific strategy for setIntervalAsync
using CommonJS with the following snippets:
// Dynamic strategy.
const { setIntervalAsync, clearIntervalAsync } = require('set-interval-async/dynamic');
// Fixed strategy.
const { setIntervalAsync, clearIntervalAsync } = require('set-interval-async/fixed');
Or else, you can use ES6 modules syntax:
// Dynamic strategy.
import { setIntervalAsync, clearIntervalAsync } from 'set-interval-async/dynamic';
// Fixed strategy.
import { setIntervalAsync, clearIntervalAsync } from 'set-interval-async/fixed';
In the Browser
After the library has been loaded to the page, a variable called SetIntervalAsync
will be defined in the global context. From there you can retrieve the setIntervalAsync
from your desired strategy and clearIntervalAsync
functions.
// Dynamic strategy.
var setIntervalAsync = SetIntervalAsync.dynamic.setIntervalAsync;
var clearIntervalAsync = SetIntervalAsync.clearIntervalAsync;
// Fixed strategy.
var setIntervalAsync = SetIntervalAsync.fixed.setIntervalAsync;
var clearIntervalAsync = SetIntervalAsync.clearIntervalAsync;
API
Function setIntervalAsync
Executes the given handler
every intervalMs
milliseconds, while preventing multiple concurrent executions. The handler will never be executed concurrently more than once in any given moment.
See Dynamic and Fixed setIntervalAsync
for more details on which strategies can be used to determine the effective interval between executions when the handler takes longer than the target interval to complete. The default strategy is Dynamic
.
function setIntervalAsync<HandlerArgs extends unknown[]>(
handler: SetIntervalAsyncHandler<HandlerArgs>,
intervalMs: number,
...handlerArgs: HandlerArgs
): SetIntervalAsyncTimer<HandlerArgs>;
Note: when intervalMs
is less than 1, it will be set to 1. When intervalMs
is greater than 2147483647, it will be set to 2147483647. Non-integer values are truncated to an integer.
Function clearIntervalAsync
Stops an execution cycle started by setIntervalAsync
. Any ongoing function executions will run until completion, but all future ones will be cancelled.
async function clearIntervalAsync<HandlerArgs extends unknown[]>(
timer: SetIntervalAsyncTimer<HandlerArgs>
): Promise<void>;
The promise returned will resolve once the timer has been stopped and the ongoing execution (if available) has been completed. If the last execution ends in a promise rejection, the promise returned by clearIntervalAsync
will reject with the same value.
Type SetIntervalAsyncHandler
Synchronous or asynchronous function that can be passed in as a handler to setIntervalAsync
.
type SetIntervalAsyncHandler<HandlerArgs extends unknown[]> = (
...handlerArgs: HandlerArgs
) => void | Promise<void>;
Type SetIntervalAsyncTimer
Return type of setIntervalAsync
. Does not have any observable properties, but must be passed in as an argument to clearIntervalAsync
to stop execution.
type SetIntervalAsyncTimer<HandlerArgs extends unknown[]>;
Avoiding Deadlock When Clearing an Interval
While calling clearIntervalAsync
to stop an interval is perfectly safe in any circumstance, please note that awaiting its result within the async handler itself will lead to undesirable results.
For example, the code below leads to a cyclical promise chain that will never be resolved (the console.log
statement is unreachable).
const timer = setIntervalAsync(async () => {
// ...
if (shouldStop) {
await clearIntervalAsync(timer);
console.log('Stopped!');
}
}, interval);
This is the case because:
await clearIntervalAsync(timer)
will not resolve until the last execution has finished, and- the last execution will not finish until
await clearIntervalAsync(timer)
has been resolved.
To prevent this cycle, always allow the async handler to complete without awaiting for the interval to be cleared. For example, by removing the await
keyword entirely or by using an immediately-invoked function expression:
if (shouldStop) {
(async () => {
await clearIntervalAsync(timer);
console.log('Stopped!');
})();
}
Motivation
If you've ever had to deal with weird, subtle bugs as a consequence of using setInterval
[1] on asynchronous functions, or had to manually reimplement setInterval
using setTimeout
[2] to prevent multiple executions of the same asynchronous function from overlapping, then this library is a drop-in replacement that will solve your issues.
setInterval
runs a given function repeateadly, once every fixed number of milliseconds. This may cause problems whenever the function takes longer to execute than the given interval, since it will be called again before the first execution has finished. This is often a problem for non-reentrant functions; ie. functions that are not designed to allow multiple executions at the same time.
setIntervalAsync
is a drop-in replacement of setInterval
which shares the same API but is safe to use with non-reentrant, asynchronous functions.
[1] https://developer.mozilla.org/en-US/docs/Web/API/WindowOrWorkerGlobalScope/setTimeout
[2] https://developer.mozilla.org/en-US/docs/Web/API/WindowOrWorkerGlobalScope/setInterval
Contributing
In order to contribute to this project, you will need to first clone the repository:
git clone https://github.com/ealmansi/set-interval-async.git
Make sure that Yarn is installed globally on your system, install all project dependencies, and build the project:
yarn
yarn build
Now, you can run the tests and make sure that everything is up and running correctly:
yarn test
If the previous step succeeds, you're ready to start developing on this project.
Pull requests are welcome!
You can verify that your code follows the project's style conventions the with the following command:
yarn lint