Re: Value returned for missing DOMStringList properties

On 8/3/2009 5:55 PM, Martin Honnen wrote:
> The ECMAScript binding
> (http://www.w3.org/TR/DOM-Level-3-Core/ecma-script-binding.html) clearly
> says:
>
> "Objects that implement the DOMStringList interface:
>
> Properties of objects that implement the DOMStringList interface:
>
> length
> This read-only property is a Number.
>
> Functions of objects that implement the DOMStringList interface:
>
> item(index)
> This function returns a String.
> The index parameter is a Number.
> Note: This object can also be dereferenced using square bracket notation
> (e.g. obj[1]). Dereferencing with an integer index is equivalent to
> invoking the item function with that index."
>
> so with JavaScript whether you use someDOMStringList.item(100) or
> someDOMStringList[100] should not matter, the result should be the same.

Thanks, I forgot to look at that part. That leaves no ambiguity.

Apparently, other interfaces such as NodeList or NamedNodeMap have the 
same behavior. So I would expect something like:

document.body.childNodes.item(100);
	-> null
document.body.childNodes[100];
	-> null
document.body.attributes.item(100);
	-> null
document.body.attributes[100];
	-> null

However, Gecko/WebKit/Opera return:

null
undefined
null
undefined

IE does:

(throws an exception)
undefined
null
null

So I'm not sure where to go from there regarding DOMStringList: make it 
conformant but non consistent with other DOM interfaces, or 
non-conformant but consistent? Is there an ongoing effort to standardize 
what browsers are currently doing?


Sylvain