I was reading through the source for the _.isFunction() function and saw this line:

if (typeof (/./) !== 'function') {

and I don't understand why it's there. /./ is a regex that always seem to have the type object. Why wouldn't _.isFunction be redefined if /./ type was a function?

I was reading through the source for the _.isFunction() function and saw this line:

if (typeof (/./) !== 'function') {

and I don't understand why it's there. /./ is a regex that always seem to have the type object. Why wouldn't _.isFunction be redefined if /./ type was a function?

• javascript
• underscore.js

• 4 Answered here: stackoverflow./questions/5054352/… – Dogbert Commented Jul 11, 2013 at 17:49
• Ah, so the check is to avoid accidentally having regex considered function? Nice. Excellent question. – ZenMaster Commented Jul 11, 2013 at 17:52
• 3 "Why wouldn't _.isFunction be defined" - it's being redefined. The original toString() definition is two blocks up. – Rup Commented Jul 11, 2013 at 17:52

1 Answer 1

Some versions of various JavaScript engines have allowed for calling RegExp as another way of using .exec():

var pattern = /./;

pattern('abc');
pattern.exec('abc');

And, since they were Callable, typeof considered them functions:

Type of val: Object (native or host and does implement [[Call]])
Result: "function"

To my knowledge, though, current versions don't exhibit this behavior and will throw a TypeError. But, if you're concerned with backwards patibility, as Underscore is, you may need to check for it.