Richard Garbutt wrote:
The problem is with the navigation on the right. The elements with red
borders are an unordererd list (#menulist) inside an absolutely
positioned div (#menu). The unordered list has width set to 9em and
margin: auto. The page is meant to behave like it does in Firefox, and
how it does before the hover event is triggered in IE.
<http://www.westcornwallweb.co.uk/dev/blog/index.htm>
Before thinking of fixing IE, we should fix the difference in the
interpretation of your page in Fx and Opera8.
Fx shows a wide, Opera a narrow #menu (red border)
When #rap gets a position:relative, Opera 8 and Fx show nearly the same.
But, is it what you wanted to achieve?
You container have no width, and percentage margins, and the menu
container is positioned absolute, without a width, but with an
percentage offset.
position:relative states that #rap shall be the containing block for
this a.p. container, and the right:6% is computed relative to its origin.
----
IE seems to suffer from similar relational problem: no width, %-margins,
positioning: the calculation is done with respect to widths not fully
known while rendering, and when you hover over a link, a redraw will
rearrange these elements with respect to other lengths that are not
necessarily right.
I've tried to stabilize IE6 with a
#quirksfix {_height:1%;}
as an additional wrapper:
<div id="quirksfix">
<!-- begin sidebar -->
<div id="menu">
...
</div>
<!-- end sidebar -->
</div>
well, it's not equal to Fx and Opera, but not jumping anymore.
--
In conclusion, I'm a little bit clueless what lenght relates to what
element, and the three browsers seem to feel the same. Maybe, to prevent
the quirky percentages, you might switch to em-based margins and
paddings, and help the browser to calculate with some width?
Ingo
--
http://www.satzansatz.de/css.html
______________________________________________________________________
css-discuss [EMAIL PROTECTED]
http://www.css-discuss.org/mailman/listinfo/css-d
List wiki/FAQ -- http://css-discuss.incutio.com/
Supported by evolt.org -- http://www.evolt.org/help_support_evolt/
