I have seen anonymous functions inside for loops to induce new scope on the web in one or two places and would like to know if it makes sense.
for example:
var attr, colors = ['green','blue','red'];
for ( attr = 0; attr < colors.length; attr++) {
(function() {
var colorAttr = colors[attr];
// do something with colorAttr
})();
}
I understand it has something to do with keeping the scope inside the for loop clean, but in what situations would this be necessary? Would it be good practice to do this everywhere you need to declare a new var inside the for loop?
When you have inner functions that are not executed immediately, as part of the loop.
var i, colors = ['green', 'blue', 'red'];
for (i = 0; i < colors.length; i++) {
var color = colors[i];
setTimeout(function() {
alert(color);
}, i * 1000);
}
// red
// red
// red
Even though var color
is inside the loop, loops have no scope. You actually only have one variable that every loop iteration uses. So when the timeouts fire, they all use the same value, the last value set by the loop.
var i, colors = ['green', 'blue', 'red'];
for (i = 0; i < colors.length; i++) {
(function(color) {
setTimeout(function() {
alert(color);
}, i * 1000);
})(colors[i]);
}
// green
// blue
// red
This one captures the value at each iteration into an argument to a function, which does create a scope. Now each function gets it's own version of a color
variable which won't change when functions created within that loop are later executed.
You're almost there. It makes only sense in your snippet, if you pass in the attr
value into your self invoking function as an argument. That way, it can store that variable within its very own scope object
(function( attr ) {
var colorAttr = colors[attr];
// do something with colorAttr
})( attr );
Now, the activation object respectively lexical environment record (which are ES3 and ES5 scope objects) will have an entry for whatever value is behind attr
and therefore, its closured.