function snafu() {
    blah;
}
var foobar = new domFunction(function() {
    snafu();
});

So I added a try...catch statement to the portion of domFunction that executes the desired code. Change lines 57-58 of the original script:

            //we can call the argument function and clear the timer
            if(!c) { f(); clearInterval(t); }

To:
            //we can call the argument function and clear the timer
            if(!c) {
                try {
                    f();
                } catch(err) {
                    alert('Error: ' + err.description);
                }
                clearInterval(t);
            }

What do you think of the change? I'm looking into making the alert message more informative, and possibly routing it to the Firefox console when available.

I like it :^) I've noticed the same problem, and it can be rather insane, as it loops through the timeout throwing the same error into the console over and over again!

But I hadn't thought of using a try{} catch{} construct .. that's probably the way to go, yeah :)

I get this error message in firefox in the demo page:

Error: uncaught exception: [Exception... "Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIDOMWindowInternal.focus]" nsresult: "0x80004005 (NS_ERROR_FAILURE)" location: "JS frame :: chrome://global/content/bindings/tabbrowser.xml :: setFocus :: line 753" data: no]

//if its value is "id" and the element with the given ID doesn't exist
//or its value is "tag" and the specified collection has no members
if
(
    (a[i] == 'id' && document.getElementById(i) == null)
    ||
    (a[i] == 'tag' && document.getElementsByTagName(i).length < 1)
    ||
    (a[i] == 'function' && window[i])
)

No need to check for all the functions you want to use, of course - just one from any .js file you rely on.

Anyway, just a thought (haven't actually tried it yet!). Thanks for the really useful bit of code :)

P.S. Looking at the error trapping mod above, it might be slightly better to re-throw like this:

//we can call the argument function and clear the timer
if(!c) {
    try {
        f();
    } catch(err) {
        throw err;
    } finally {
        clearInterval(t);
    }
}

//if its value is "id" and the element with the given ID doesn't exist
//or its value is "tag" and the specified collection has no members
//or its value is "function" and the specified function doesn't exist
if
(
(a[i] == 'id' && document.getElementById(i) == null)
||
(a[i] == 'tag' && document.getElementsByTagName(i).length < 1)
||
(a[i] == 'function' && !window[i])
)

Note the ! before window[i]

That's not true - JS files are loaded synchronously unless you use the [unreliable] "defer" attribute.

The problem here is the structure of your codebase; you're relying on dependencies that don't exist when you need them, and that's an architecture issue - something you can resolve manually by attention to dependency structures.

I'm guessing here that you're using a generalised framework like prototype? If so, that certainly won't make your job easier, because following dependency chains in external frameworks can be notoriously difficult (and is, in my view, a reason for not using them; but that's another argument for another time!)

I do appreciate however that a timed solution would fix this potential issue more swiftly in specific cases, but it's really a case of patchwork, of providing a temporary solution to a problem that you'd be much better advised to address at its root.

(a[i] == 'function' && !window[i])

should be like this:

(a[i] == 'function' && typeof window[i] == 'undefined')

But even then, it would of course only work for functions in the window scope.