Re: [RC4] active-selector-002

On 12/28/2010 01:25 PM, Tab Atkins Jr. wrote:
>
> (I don't accept your conclusion in that thread, though - by common
> agreement between all browsers, clicking is an activation behavior,
> and all elements respond to clicks by becoming :active.  This isn't a
> matter of bugwards compat or not, it's just a matter of your personal
> opinion differing from the interoperable opinion of all browsers.  We
> should indeed be testing and requiring this behavior.  If you believe
> we need a disclaimer about the it, so be it, but that shouldn't
> prevent us from testing interoperable behavior.)

Sorry, but this is not a matter of my personal opinion. The HTML4 spec
gives no support to the idea of <span> being activateable, while it
does so for links and form elements; and the HTML5 spec explicitly
excludes it from its list of activateable elements. Given both these
facts and the current text of the CSS spec, there is absolutely no
reason why the CSS test suite should expect <span> to be activateable.

>> True, but activating the <a> should not cause its child <span> to be
>> :active.
>
> Clicking on the text of the link means clicking on the <span>.  By
> virtue of the <a> being the <span>'s container, it's also clicked
> on/activated.  There is no grounds for assuming that only the <a>
> should be :active and not the <span>, though.

You are absolutely correct, under the assumption that the <span>
is activateable. However, that assumption is not valid under the
HTML specs. If the <span> is not activateable, and it is not in
the ancestor chain of an activateable element, then per CSS2.1 it
cannot match :active.

~fantasai

Received on Tuesday, 28 December 2010 18:44:11 UTC