Indeed, removing these lines doesn't break (including FF) any tests, was added in 3c7f2af

Issue exist in FF 19 but it's no longer reproducible in modern browsers. Would you like to sends us a PR?

/cc @rwaldron

The code still acts as a defense against form aliasing: http://jsfiddle.net/Lz0k4bn4/

We can either drop support or add a unit test.

Is this change fixing a bug, or removing a feature? 😈 The length variable in that aliased test case is an HTMLInputElement so you couldn't use the form element as an arraylike object to iterate over its element collection.

The length variable in that aliased test case is an HTMLInputElement so you couldn't use the form element as an arraylike object to iterate over its element collection.

Nuh-uh: http://jsfiddle.net/Lz0k4bn4/1/ (check the console)

I must be missing something, here are some extra console logs: http://jsfiddle.net/Lz0k4bn4/2/

Never supported this usage, in fact, we have always closed issues like that, see the most recent one

I agree, we have been closing them not because we are happy with having this bug, but because the aliasing problem is too hard to solve.

I'm fine with removing it.

Working around form aliasing is hard, especially in the general case, but it is possible for length (and also nodeType and nodeName, for that matter). I'm fine if we don't, but it merits a wont-fix mention analogous to Object.prototype manipulation.