- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Sat, 25 Apr 2009 00:32:27 -0500
- To: Boris Zbarsky <bzbarsky@mit.edu>
- Cc: www-style list <www-style@w3.org>
On Sat, Apr 25, 2009 at 12:17 AM, Boris Zbarsky <bzbarsky@mit.edu> wrote: > Tab Atkins Jr. wrote: >> >> Whenever an element is clicked/activated/etc, it toggles between >> matching the :checked pseudoclass and not matching it. > > OK, what happens if an element that already matches that pseudoclass is > inserted into the document and is in a radio-group in which something else > is already checked? You mean creating an element which is pre-:checked? I don't think that's possible. Or do you mean moving a node from one document to another, where it matched :checked in the first document? I would be perfectly fine with saying that it's not :checked when moved to the new document. Are there any weirdnesses that I'm overlooking that would cause problems here? > Or has its radio-group property changed, with the same > effect? This is the interesting problem. I'm fine with taking the easy way out. Whenever radio-group changes, the element becomes un:checked. Some cursory thinking doesn't show this as causing any problems. I basically want :checked to change *only* as a result of direct user interaction, not due to some crazy conflict-resolution rules. (The closest thing to 'crazy confliction-resolution rules' I'd tolerate is, say, making an entire ancestor chain have the same radio-group value, then clicking the innermost child. The click event would bubble up the chain, making each parent :checked in turn and removing :checked from the child, leaving you with just the outermost element :checked. That's easy for authors to understand in terms of clicks, though.) ~TJ
Received on Saturday, 25 April 2009 05:33:08 UTC