Re: [cssom] Supporting JS properties whose names are CSS property names on CSSDeclaration objects

On Wed, Jun 5, 2013 at 5:02 AM, Boris Zbarsky <bzbarsky@mit.edu> wrote:

> On 6/4/13 3:56 PM, Glenn Adams wrote:
>
>> Just because it is implemented by some UA is no cause to put in spec
>>
>
> Again, we need to either get the UAs that implement it to drop it or we
> need to spec it.  People are writing content depending on this I didn't
> bring it up for the fun of it, but rather because of pages that were broken
> in Gecko because they assumed that all browsers are WebKit and have this
> behavior....


And the way to do this (dropping it) is to not spec this behavior but
anti-spec this behavior, i.e., explicitly rule out in the spec. If we want
to get this behavior dropped, then I think we need to add language that
translates to:

"A CSS property name SHALL NOT be used as an <identifier-name-string> when
accessing a named property using the MemberExpression dot notation or
bracket notation on a CSSStyleDeclaration object unless that CSS property
name is an exact, case sensitive match of the named property's member name."

"Note: This rules out using expressions like style.float and
style['font-size']. In these cases, the correct expressions are
style.cssFloat and style['fontSize']."