Re: [css3-multicol] Applies to block containers

Le 20/02/2013 23:55, Håkon Wium Lie a écrit :
> Also sprach Simon Sapin:
>
>   > The columns, column-width and column-count properties all have the same
>   > "Applies to" line:
>   >
>   > > non-replaced block-level elements (except table elements), table
>   > > cells, and inline-block elements
>   >
>   > This should really be "block container" instead of an inclusive list of
>   > the block containers that exist in CSS 2.1.
>
> It would be good to use a term rather than a list. But CSS 2.1 isn't
> crystal-clear on what a "block container" is. For example, the
> property index [1] sometimes calls it "block containers" (note plural)
> and sometimes "block container elements". Is there an updated
> definition in CSS3 somewhere?
>
>   [1] http://www.w3.org/TR/CSS21/propidx.html

Yes, CSS 2.1 can be a bit sloppy at times.

I use "block container" as short for "block container box", defined here:
http://www.w3.org/TR/CSS21/visuren.html#block-container-box

AFAIK "block container element" is undefined. I’d much rather use "block 
container box" everywhere instead, if possible. If not, I suggest this 
definition:

> A block container element is an element that
> <a href="http://www.w3.org/TR/CSS21/visuren.html#box-gen">generates</a>
> a block container box.

Not sure where it should go, though.


>   > Except that elements that multicol are not block containers anymore.
>
> You mean, since the column boxes act as containing blocks, the
> multicol elements are no longer block containers? I'm not sure we need
> to answer that question.

Block containers contain either only block-level boxes or only line 
boxes. This is not the case of multicol boxes that contain column boxes. 
(The column boxes *are* block containers, however.)

In my mental model, CSS21/visuren.html#box-gen would first generate a 
block container box. css3-multicol would then be considered, and can 
later replace it by a multicol box.


>   > So I guess that the properties really apply to "elements that would
>   > otherwise have been block containers." But that’s awkward. "Applies
>   > to: block containers" is probably a good enough approximation.
>
> Yes.


-- 
Simon Sapin

Received on Wednesday, 20 February 2013 23:18:16 UTC