Re: [css3-multicol] new editor's draft

On Mon, Oct 25, 2010 at 1:40 PM, Håkon Wium Lie <howcome@opera.com> wrote:
> Also sprach Tab Atkins Jr.:
>
>  > Given this markup:
>  >
>  > <div break-after:column>foo</div>
>  > <div break-after:column>foo</div>
>  > <div break-after:column>foo</div>
>  > <p>bar</p>
>  > <h1 column-span:all text-align:center>baz</h1>
>  > <p>qux</p>
>  >
>  > Where would "bar" show up? (Given a column-count:3 on the containing element.)
>  >
>  > Would it be like any of the following?
>  >
>  > |-- container---|
>  >
>  > (A)
>  > foo | foo | foo | bar
>  >                   qux
>  >       baz
>  >
>  > (B)
>  > foo | foo | foo | bar
>  >       baz
>  >                 | qux
>  >
>  > (C)
>  > foo | foo | foo
>  > bar |
>  >       baz
>  > qux |
>
> It's a good example. Here's my solution:
>
> (D)
> foo | foo | foo | bar
>      baz
> qux
>
> because:
>
>  - "bar" is in a column of its own as the foo element before it has
>    'break-after:column'
>
>  - baz has 'colspan: all' and thereby a line of its own. It is
>    centered wrt the multicol box.
>
>  - qux is flows into the columns after baz
>
>  > >  > It seems like perhaps this feature (column spanners in overflow
>  > >  > columns still showing up in the main area) should instead be a
>  > >  > function of the column-overflow mode.  With 'column-overflow:inline',
>  > >  > overflow column spanners don't span at all.  With
>  > >  > 'column-overflow:block', they do span, in the way you describe.
>  > >
>  > > Yes, 'column-overflow' when/if defined, could provide control over
>  > > this. But we need to define what the rendering should be even without
>  > > 'column-overflow'.
>  >
>  > Oh, I agree.  I was suggesting that we may want to kick the particular
>  > behavior you're suggesting to later, and for now stick with the
>  > behavior that Alex was suggesting.
>
> That's also an option.

Actually, given the rendering you and Alex pointed to, I'm happy with
specifying this behavior as the default.  It seems reasonable.

(It doesn't solve the column-height-bigger-than-viewport problem,
though it may make it somewhat less important, if there are regular
headings or other spanning elements.)

~TJ

Received on Monday, 25 October 2010 20:45:13 UTC