JavaScript Order of operations plus advanced thoughts


Example

Without a try catch block, undefined functions will throw errors and stop execution:

undefinedFunction("This will not get executed");
console.log("I will never run because of the uncaught error!");

Will throw an error and not run the second line:

// Uncaught ReferenceError: undefinedFunction is not defined

You need a try catch block, similar to other languages, to ensure you catch that error so code can continue to execute:

try {
    undefinedFunction("This will not get executed");
} catch(error) {
    console.log("An error occured!", error);
} finally {
    console.log("The code-block has finished");
}
console.log("I will run because we caught the error!");

Now, we've caught the error and can be sure that our code is going to execute

// An error occured! ReferenceError: undefinedFunction is not defined(…)
// The code-block has finished
// I will run because we caught the error!

What if an error occurs in our catch block!?

try {
    undefinedFunction("This will not get executed");
} catch(error) {
    otherUndefinedFunction("Uh oh... ");
    console.log("An error occured!", error);
} finally {
    console.log("The code-block has finished");
}
console.log("I won't run because of the uncaught error in the catch block!");

We won't process the rest of our catch block, and execution will halt except for the finally block.

// The code-block has finished
// Uncaught ReferenceError: otherUndefinedFunction is not defined(…)

You could always nest your try catch blocks.. but you shouldn't because that will get extremely messy..

try {
    undefinedFunction("This will not get executed");
} catch(error) {
    try {
        otherUndefinedFunction("Uh oh... ");
    } catch(error2) {
        console.log("Too much nesting is bad for my heart and soul...");
    }
    console.log("An error occured!", error);
} finally {
    console.log("The code-block has finished");
}
console.log("I will run because we caught the error!");

Will catch all errors from the previous example and log the following:

//Too much nesting is bad for my heart and soul...
//An error occured! ReferenceError: undefinedFunction is not defined(…)
//The code-block has finished
//I will run because we caught the error!

So, how can we catch all errors!? For undefined variables and functions: you can't.

Also, you shouldn't wrap every variable and function in a try/catch block, because these are simple examples that will only ever occur once until you fix them. However, for objects, functions and other variables that you know exist, but you don't know whether their properties or sub-processes or side-effects will exist, or you expect some error states in some circumstances, you should abstract your error handling in some sort of manner. Here is a very basic example and implementation.

Without a protected way to call untrusted or exception throwing methods:

function foo(a, b, c) {
    console.log(a, b, c);
    throw new Error("custom error!");
}
try {
    foo(1, 2, 3);
} catch(e) { 
    try {
        foo(4, 5, 6); 
    } catch(e2) {
        console.log("We had to nest because there's currently no other way...");
    }
    console.log(e);
}
// 1 2 3
// 4 5 6
// We had to nest because there's currently no other way...
// Error: custom error!(…)

And with protection:

function foo(a, b, c) {
    console.log(a, b, c);
    throw new Error("custom error!");
}
function protectedFunction(fn, ...args) {
    try {
        fn.apply(this, args);
    } catch (e) {
        console.log("caught error: " + e.name + " -> " + e.message);
    }
}

protectedFunction(foo, 1, 2, 3);
protectedFunction(foo, 4, 5, 6);

// 1 2 3
// caught error: Error -> custom error!
// 4 5 6
// caught error: Error -> custom error!

We catch errors and still process all the expected code, though with a somewhat different syntax. Either way will work, but as you build more advanced applications you will want to start thinking about ways to abstract your error handling.