- From: Orion Adrian <orion.adrian@gmail.com>
- Date: Wed, 26 Oct 2005 15:15:43 -0400
- To: www-style@w3.org
On 10/26/05, Matthew Raymond <mattraymond@earthlink.net> wrote:
>
> Bert Bos wrote:
> > On Wednesday 26 October 2005 09:39, David Woolley wrote:
> >>>I haven't thought about it for more than 5 minutes, so I'm sure
> >>>there are lots of problems. But at least in 5 minutes I discovered
> >>>fewer
> >>
> >>You get the normal back-reference type problems, including:
> >>
> >> INPUT:hover -> LABEL {display: none}
> >>
> >>for the common nested case.
> >
> > Isn't that the same problem as 'INPUT:hover {display: none}' ?
> >
> > I think it is well-defined, just not very useful in practice.
>
> He's talking about this:
>
> | <label>Control: <input></label>
>
> When you hover over the <input>, the parent is set to "display:
> none", so the <input> is hidden as well, but that stops the hover, which
> causes the element to reappear...
>
> Same thing can occur when changing the size of the <label>, so
> ideally there would need to be a limitation on what properties you could
> use with this selector. Outline, background and other would be
> unaffected, but anything that can change the bounding box size is a no-no.
Given the sheer number of problems that occurs when you can specify
display based on a dynamic case, has anyone considered removing that
capability?
--
Orion Adrian
Received on Wednesday, 26 October 2005 19:15:48 UTC