Re: [CSS21] Contradiction in the spec with auto-offset fixed-position elements

Ian Hickson wrote:

[snipped reference to part of spec I really should have noticed]
> So you work out the position as it would have been relative to the top of 
> the initial containing block (in this particular case, a box at the canvas 
> origin, with the size of the viewport) and then you place the box relative 
> to the viewport, using that offset.

I'm not sure that's sufficient...  consider the following marginally 
more complicated testcase:

<root>
   <scrollable style="overflow: scroll">
     Long content
     <fixed style="position: fixed">Text</fixed>
   </scrollable>
<root>

Now scrolling the <scrollable> will affect the static position of 
<fixed> with respect to the initial containing block, which means it 
should really move when we scroll <scrollable>.  Again, I don't think 
that's intended or desirable.

> That's a fancy way of saying that the static position of a fixed-
> positioned box is its initial position before scrolling, and that 
> subsequent scrolling doesn't affect it.

True for scrolling of the viewport itself, but not for scrolling in 
general.  :(

-Boris

Received on Sunday, 7 November 2004 05:21:42 UTC