What does the exclamation mark do before the function?

JavaScript syntax 101. Here is a function declaration:

function foo() {}

Note that there's no semicolon: this is just a function declaration. You would need an invocation, foo(), to actually run the function.

Now, when we add the seemingly innocuous exclamation mark: !function foo() {} it turns it into an expression. It is now a function expression.

The ! alone doesn't invoke the function, of course, but we can now put () at the end: !function foo() {}() which has higher precedence than ! and instantly calls the function.

So what the author is doing is saving a byte per function expression; a more readable way of writing it would be this:

(function(){})();

Lastly, ! makes the expression return true. This is because by default all immediately invoked function expressions (IIFE) return undefined, which leaves us with !undefined which is true. Not particularly useful.

The function:

function () {}

returns nothing (or undefined).

Sometimes we want to call a function right as we create it. You might be tempted to try this:

function () {}()

but it results in a SyntaxError.

Using the ! operator before the function causes it to be treated as an expression, so we can call it:

!function () {}()

This will also return the boolean opposite of the return value of the function, in this case true, because !undefined is true. If you want the actual return value to be the result of the call, then try doing it this way:

(function () {})()

There is a good point for using ! for function invocation marked on airbnb JavaScript guide

Generally idea for using this technique on separate files (aka modules) which later get concatenated. The caveat here is that files supposed to be concatenated by tools which put the new file at the new line (which is anyway common behavior for most of concat tools). In that case, using ! will help to avoid error in if previously concatenated module missed trailing semicolon, and yet that will give the flexibility to put them in any order with no worry.

!function abc(){}();!function bca(){}();

Will work the same as

!function abc(){}();(function bca(){})();

but saves one character and arbitrary looks better.

And by the way any of +,-,~,void operators have the same effect, in terms of invoking the function, for sure if you have to use something to return from that function they would act differently.

abcval = !function abc(){return true;}() // abcval equals falsebcaval = +function bca(){return true;}() // bcaval equals 1zyxval = -function zyx(){return true;}() // zyxval equals -1xyzval = ~function xyz(){return true;}() // your guess?

but if you using IIFE patterns for one file one module code separation and using concat tool for optimization (which makes one line one file job), then construction

!function abc(/*no returns*/) {}()+function bca() {/*no returns*/}()

Will do safe code execution, same as a very first code sample.

This one will throw error cause JavaScript ASI will not be able to do its work.

!function abc(/*no returns*/) {}()(function bca() {/*no returns*/})()

One note regarding unary operators, they would do similar work, but only in case, they used not in the first module. So they are not so safe if you do not have total control over the concatenation order.

This works:

!function abc(/*no returns*/) {}()^function bca() {/*no returns*/}()

This not:

^function abc(/*no returns*/) {}()!function bca() {/*no returns*/}()