I have a class, ChatRoom
, that can only render after it receives a long-running HTTP request (it could take 1 second or 30 seconds). So I need to delay rendering un
Consider this:
(function wait() {
if ( chatroom.json ) {
chatroom.render();
} else {
setTimeout( wait, 500 );
}
})();
This will check every half second.
Live demo: http://jsfiddle.net/kBgTx/
You could also achieve this using lodash's debouncer with recursion.
import _debounce from 'lodash/debounce';
const wait = (() => {
if ( chatroom.json ) {
chatroom.render();
} else {
_debounce(wait, 500)();
}
})();
The answer I came up with is like this:
var count = 0;
// Number of functions that need to run. This can be dynamically generated
// In this case I call check(data, cb) a total of 3 times
var functionNum = 3;
function toCallAfter(){
console.log('I am a delayed function');
}
I had this for a check function that ran once regularly and twice in a loop:
check(data, function(err){ // check is my asynchronous function to check data integrity
if (err){
return cb(null, { // cb() is the return function for the containing function
errCode: 'MISSINGINFO',
statusCode: 403,
message : 'All mandatory fields must be filled in'
});
} // This part here is an implicit else
count++; // Increment count each time required functions complete to
// keep track of how many function have completed
if (count === functionNum) {
return anon();
}
return;
});
// Run twice more in a loop
for(var i = 0; i < 2; i++) {
check(data, function(err) { // calls check again in a loop
if (err){
return cb(null, {
errCode: 'MISSINGINFO',
statusCode: 403,
message : 'All mandatory fields must be filled in'
});
}
count++;
if (count === functionNum) {
return toCallAfter();
}
return;
});
}
Lastly, I'd like to point out a significant performance fault in the alternate (and extraordinarily common) answer:
(function wait() {
if ( chatroom.json ) {
chatroom.render();
} else {
setTimeout( wait, 500 );
}
})();
In this case, you are essentially holding the browser or server (If using node.js) hostage for 500 milliseconds for every check which is an incredibly long time for a computer. Meaning a huge performance hit. My solution of directly keeping track of the required completed functions is free from time restraints and will run instantly as soon as all functions complete.