Daniel Doesburg wrote:
> http://www.export-lead-platform.nl/2206.0.html?&L=1
>
> If you narrow the screen the 2 columns in the big left part
> "unfloat". Why?
> But, worser, if you make the screen wider again, the columns
> refuse the float again.
I've not looked at the details (you have at least fourteen nested divs :-),
but what happens is more or less this:
You have two floats with width:45% _and_ some padding. If the window is
narrow enough, the two floats do to fit anymore one at the side of the
other, so the second one drops down (because 45% plus the padding becomes
greater than 100%.)
The funny thing in FF is, as you noticed, that making the window wider
again, the two floats remain one below the other (until you refresh the
page.)
But I don't think that FF is to blame here. The above description miss an
important thing: the container of your floats is floated itself and has
width:auto, so its width should computed with a shrink to fit algorithm.
But this means that this width depends on the content, so we are probably on
a condition that the spec marks as undefined [1] ("... If the containing
block's width depends on this element's width, then the resulting layout is
undefined in CSS 2.1...")
Workaround? Set a width for the container ".content_2blok". I see that you
feed it width 100%, but for IE only, I guess you have probably some
complications that led to this, I haven't investigated.
For the interested ones, I have reproduced the Gecko behaviour in a minimal
test case [2], but I don't think this is even a bug.
hth,
Bruno
[1] http://www.w3.org/TR/CSS21/visudet.html#the-width-property
[2] http://www.brunildo.org/test/test/flinflpercw.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/
