Re: [css3-gcpm] Multi-column floats overflowing the column

Also sprach Brady Duga:
 > 
 > 
 > On May 4, 2007, at 11:21 AM, Alex Mogilevsky wrote:
 > 
 > >
 > > There is no contradiction here. GCPM defines new kinds of float  
 > > behavior, specifically with the intent of overriding clipping  
 > > behavior of columns. Therefore definition of how these new floats  
 > > interact with other objects (including columns) belongs in GCPM.
 > 
 > That's fine - I'm not saying it shouldn't be allowed to flow out of  
 > the columns, but I do see a contradiction (or at least potential  
 > confusion) in the current wording. The draft says it extends the  
 > functionality of the multi-column spec, not that it replaces it.  
 > Since the multi-column spec says things get clipped and this draft  
 > never explicitly modifies that constraint, it appears that it should  
 > still be in place.

I think you're right in asking for this to be clear. Perhaps we can
find better wordings -- both specs are still Working Drafts.

Let's look at the current wordings. The multicol draft says:

  Content that extend outside the column box (e.g., long words and
  floats that are higher and/or wider than the column box) is clipped.
  This only applies to content for which the column box is the
  containing block (i.e., as if the column box had 'overflow:hidden').

  http://www.w3.org/TR/2005/WD-css3-multicol-20051215/#the-multi-column

So, when the float leaves the column box, the rules about clipping in
that box no longer applies. However, GCPM says:

  The column box of the column the element flows to becomes the
  containing block of the element.

  http://www.w3.org/TR/2007/WD-css3-gcpm-20070504/#advanced

>From this, you could argue that the float should be clipped because it
now belongs to a new containing block with its own clipping regime. So,
arguably, there is an inconsitency.

 > I would like to see some text explaining when (for  
 > paged media) it is allowed for floated content to flow into  
 > neighboring columns. Always? Only when using one of the new float  
 > values? Only when float-offset is specified?

The simplest way to solve this would be to drop clipping altogether
and allow floats and other content to extend into other columns.
However, implementors like clipping.

Or, we could say that clipping should not be performed on floats. 

Or, as you suggest, the new float values or 'float-offset' could be
triggers.

What's your preferred solution?

And, how to you like the rest of the GCPM draft?

-h&kon
              Håkon Wium Lie                          CTO °þe®ª
howcome@opera.com                  http://people.opera.com/howcome

Received on Friday, 4 May 2007 19:08:33 UTC