- From: Mikko Rantalainen <mira@st.jyu.fi>
- Date: Wed, 09 Jul 2003 12:17:39 +0300
- To: "www-style@w3.org" <www-style@w3.org>
Ian Hickson wrote: > On Tue, 8 Jul 2003, Mikko Rantalainen wrote: > >>OK, it's the more general problem then. How about: >> >>span { position: relative; top: 0; } >>span:hover { top: 100%; } >> >>In that case, hovering the mouse moves the element and causes the >>element be no longer be hovered -- and the element moves back to it's >>original position and it is again hovered, right? > > Yes, but in that particular case there is a simple solution, which is to > say that :hover is only changed during a mouse-move event (this is what > UAs do, in fact). Then how is the :broken different? When the UA notices that the content is broken, it can send a single event that then causes styling backend to do the similar thing as with :hover. I think the issue you brought up with :broken is not that different from :hover if the UA reflows the page during :hover in case the styling requires that. If we have a clean solution for either problem, the other one can copy that. Am I missing something? Does it make the difference if the event comes from mouse movement instead of rendering engine? In both cases executing the actions required by the event can cause further events being triggered and AFAIK, it's what this issue is all about. >>Would it be somehow possible to modify rendered content without removing >>the :broken status? Perhaps adding some invisible content that would be >>regarded as :broken? > > > I don't really understand how such complexity is really that useful. Just throwing out some ideas. I was wondering if it were possible to prevent recursion without extra limits on UA part. I'm aware that the idea I wrote isn't perfect, but hopefully it guides thoughts of somebody else to the right track so he can come up with the solution. -- Mikko
Received on Wednesday, 9 July 2003 05:18:07 UTC