Sent by Ingo Chao on 5 April 2005 19:07
Bruno Fassino schrieb:
....
> I haven't looked at the original page, but in the conditions that you
> describe it is indeed not clear what should happen.
> According to CSS2.1 10.3.7 that you mentioned, the width of the AP element
> should be computed with a shrink-to-fit algorithm, so it depends on the
> content.
> But the content has a percentage width, so (according to 10.2) we are
> perhaps in the following condition for such content div#featured-product :
> "If the containing block's width depends on this element's width, then the
> resulting layout is undefined" ?
>
> BTW, in a simple test case Mozilla shows a reasonable behaviour (div#content
> take all the available width, unless the content of div#featured-product is
> very small), while Opera (both 7 and 8b) insists in leaving non zero a right
> margin...
>
Bruno, thank you for the explanation, and for taking the time, much
appreciated.
And yes, I need a quick test case...
http://www.sinnverwandt.de/satzansatz/cssd/tmp/100307.html
The calculated width:auto shrinks to fit the content, and Opera and FF
respect this content (here: Floats and some text) in a different way,
depending on the viewport, zoom, text wrapping, and float width. too
bad, I like the idea of using percentages on a width:auto-a.p.container.
The first example (three floats w/o text) expands the width as if the
floats are positioned side-by-side, than take 50% and reflow them within
this length. Opera and Gecko are similar.
The second example divides Opera and Gecko in two interpretations of the
mentioned rules ... yes, to me it's "undefined" too (at least
unpredictable with more complex contents).
Ken's current page differs between Op7.54 (scrollbar), FF, and Op8b3
(narrow content).
Anyway. Have a nice day.
Ingo
______________________________________________________________________
css-discuss [EMAIL-REMOVED]]
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/