Re: content: url() is bad

On 12 Apr 2004 at 19:52, Boris Zbarsky wrote:

> 
> Ian Hickson wrote:
> > So the image in an <img> is not the element's contents?
> 
> No.  It's what the element _is_.  The contents of an element are 
> something the element is a box around.... bug again, maybe that's just me.
> 
> > Ah. I'm finding it difficult to understand why whether the content causes
> > height/width 'auto' to mean 'intrinsic' rather than 'fill' changes whether
> > it is content or not...
> 
> There are other differences between replaced and non-replaced elements 
> (for example, the handling of before/after may need to be different; see 
> previous discussions we've had on that).  So it's not just width/height 
> handling.
> 

I agree with Boris here. A possibly interesting case might be when the 
element is inline, but that might be more due to the inline model.

Eg: 
Should really 
span {border-style: solid; content: "Logo: " url(xyzcompany-logo)}
flow the border through the image of its content-property while it should 
not if the content property only held the image? 

Assuming that the above example should render a different height than 
span {border-style: solid; content: url(xyzcompany-logo)}
i think it means the interpretation of the span element is rather
different. 

CSS 2.1 firmly defines "replacedness" as a function of the markup element.
With the above definition the first case suggests the parts of the content 
value are similar to children in the document tree of the span node, while 
in the second the span "consists" of the content property. 

Personally i would prefer viewing an element with a valid content value 
as if it was replaced, but i am not sure such a distinction needs to 
be made explicit.

 /Staffan

Received on Tuesday, 13 April 2004 11:40:01 UTC