I haven't found practical use of it other than that but here is one scenario in which James Padolsey nicely uses this technique for IE detection in a while loop:
var ie = (function(){
var undef,
v = 3,
div = document.createElement('div'),
all = div.getElementsByTagName('i');
while ( // <-- notice no while body here
div.innerHTML = '<!--[if gt IE ' + (++v) + ']><i></i><![endif]-->',
all[0]
);
return v > 4 ? v : undef;
}());
The comma operator allows you to put multiple expressions in a place where one expression is expected. The resulting value of multiple expressions separate by a comma will be the value of the last comma separated expression.
I don't personally use it very often because there aren't that many situations where more than one expression is expected and there isn't a less confusing way to write the code than using the comma operator. One interesting possibility is at the end of a for loop when you want more than one variable to be incremented:
// j is initialized to some other value
// as the for loop executes both i and j are incremented
// because the comma operator allows two statements to be put in place of one
for (var i = 0; i < items.len; i++, j++) {
// loop code here that operates on items[i]
// and sometimes uses j to access a different array
}
Here you see that i++, j++ can be put in a place where one expression is allowed. In this particular case, the multiple expressions are used for side affects so it does not matter that the compound expressions takes on the value of the last one, but there are other cases where that might actually matter.
Comma operator is not specific to JavaScript, it is available in other languages like C and C++. As a binary operator this is useful when the first operand, which is generally an expression, has desired side effect required by second operand. One example from wikipedia:
i = a += 2, a + b;
Obviously you can write two different lines of codes, but using comma is another option and sometimes more readable.
The following is probably not very useful as you don't write it yourself, but a minifier can shrink code using the comma operator. For example:
if(x){foo();return bar()}else{return 1}
would become:
return x?(foo(),bar()):1
The ? : operator can be used now, since the comma operator (to a certain extent) allows for two statements to be written as one statement.
This is useful in that it allows for some neat compression (39 -> 24 bytes here).
I'd like to stress the fact that the comma in var a, b is not the comma operator because it doesn't exist within an expression. The comma has a special meaning in varstatements. a, b in an expression would be referring to the two variables and evaluate to b, which is not the case for var a, b.
I'd disagree with Flanagan, and say, that comma is really useful and allows to write more readable and elegant code, especially when you know what you're doing:
One typical case I end up using it is during optional argument parsing. I think it makes it both more readable and more concise so that the argument parsing doesn't dominate the function body.
/**
* @param {string} [str]
* @param {object} [obj]
* @param {Date} [date]
*/
function f(str, obj, date) {
// handle optional arguments
if (typeof str !== "string") date = obj, obj = str, str = "default";
if (obj instanceof Date) date = obj, obj = {};
if (!(date instanceof Date)) date = new Date();
// ...
}
This was a fairly complex, but real-world scenario. Bear with me while I explain what is happening, and in the process make the case for the Comma Operator.
Take apart all of the options passed to this function using pairs
which will turn { a: 1, b: 2} into [['a', 1], ['b', 2]]
This array of property pairs is filtered by which ones are deemed to be 'actions' in the system.
Then the second index in the array is replaced with a function that returns a promise representing that action (using map)
Finally the call to reduce will merge each "property array" (['a', 1]) back into a final object.
The end result is a transformed version of the options argument, which contains only the appropriate keys and whose values are consumable by the calling function.
You can see the reduce function starts with an empty state object, state, and for each pair representing a key and value, the function returns the same state object after adding a property to the object corresponding to the key/value pair. Because of ECMAScript 2015's arrow function syntax, the function body is an expression, and as a result, the Comma Operator allows a concise and useful "iteratee" function.
Personally I have come across numerous cases while writing Javascript in a more functional style with ECMAScript 2015 + Arrow Functions. Having said that, before encountering arrow functions (such as at the time of the writing of the question), I'd never used the comma operator in any deliberate way.
Another use for the comma operator is to hide results you don't care about in the repl or console, purely as a convenience.
For example, if you evaluate myVariable = aWholeLotOfText in the repl or console, it will print all the data you just assigned. This might be pages and pages, and if you'd prefer not to see it, you can instead evaluate myVariable = aWholeLotOfText, 'done', and the repl/console will just print 'done'.
Oriel correctly points out† that customized toString() or get() functions might even make this useful.
(function() {
"use strict";
var global = (function () { return this || (1,eval)("this"); })();
console.log('Global === window should be true: ', global === window);
var not_global = (function () { return this })();
console.log('not_global === window should be false: ', not_global === window);
}());
You can get access to the global variable because eval works differently when called directly vs called indirectly.
You could replace the comma with either an || or &&, but then you'd need to know what the function returns.
More important than that, the comma separator communicates intent -- the code doesn't care what the left-operand evaluates to, whereas the alternatives may have another reason for being there. This in turn makes it easier to understand and refactor. If the function return type ever changes, the code above would not be affected.
Naturally you can achieve the same thing in other ways, but not as succinctly. If || and && found a place in common usage, so too can the comma operator.
When you push onto that array, you are rarely interested in push's return value, namely the new length of the array, but rather the array itself:
arr.push('foo') // ['foo'] seems more interesting than 1
Using the comma operator, we can push onto the array, specify the array as the last operand to comma, and then use the result -- the array itself -- for a subsequent array method call, a sort of chaining:
Also, Hack-style pipelines are already feasible without introducing new syntax today:
let $; // Hack-style topic variable
let result = (
$= books,
$= filter($, _ => _.title = "..."),
$= map($, _ => _.author),
$);
The use of comma expressions here can kind of fake the pipeline operator that isn't in the language yet.
Eliminating the space between $= simulates the feeling of a proper pipe token, |>. Note that the "topic" variable, $, can be anything here and that it's just shorthand for repeatedly overwriting the variable. So something more akin to ...
// blocking inside an IIFE
let result = (() => {
let $;
$ = books;
$ = filter($, _ => _.title = "..."),
$ = map($, _ => _.author),
return $;
})()
The "comma" version successfully cuts out some of the noise, getting you closer to what the proposal would be:
let result = books
|> filter($, _ => _.title = "..."
|> map($, _ => _.author)
Here's another example using it to compose functions:
The comma operator (,) evaluates each of its operands (from left to right) and returns the value of the last operand. This lets you create a compound expression in which multiple expressions are evaluated, with the compound expression's final value being the value of the rightmost of its member expressions. This is commonly used to provide multiple parameters to a for loop.
