W3C home > Mailing lists > Public > www-style@w3.org > August 2012

Re: Region overflow proposal

From: Florian Rivoal <florianr@opera.com>
Date: Tue, 07 Aug 2012 11:17:49 +0200
To: www-style@w3.org
Message-ID: <op.winyrzww4p7avi@localhost.localdomain>
> I've written a draft further fleshing out this idea.  This draft is
> available at:
>
>   http://lists.w3.org/Archives/Public/www-archive/2012Aug/att-0005/Overview.html
>
> I'm hoping we can find some time to discuss this at the face-to-face
> meeting next week.

Thanks for drafting it. I wanted to do it at some point, but I am glad I
didn't find the time, since you did this better than I could.

> I think such a draft may also
> be a sensible place to put the paginated overflow ideas that are
> currently in http://dev.w3.org/csswg/css3-gcpm/#paged-presentations .
> The draft currently has a placeholder section for this (and the
> introduction and abstract mention it as well).

I definitely agree. Paged overflow could use a more detailed specification
than it has now, and given the similarities, this is definitely the right
place to do it.

> Styling an ‘::nth-region()’ pseudo-element with the ‘overflow’ property  
> has no effect;

I think it should be allowed. If you set overflow to something else than  
region on the 3rd region, no 4th region will be generated. But I find it  
reasonable to be able to express "when overflowing, clone this box up to 5  
instances, then hide whatever's left".

#foo { overflow:region; }
#foo::nth-region(5) { overflow:hidden; }


> Every region box is a fragmentation container, and for each region box  
> which ends with a fragmentation break, (which could happen because  
> breakable content overflows in the block dimension or because of a  
> forced break), there must be another region box created [...]

This is the right idea, but I am wondering if it covers something I would  
consider being an important use case:

.artcile { max-height: 30vh; column-count:2; overflow:regions; }

Would that work? I think it should, but I am not sure it would with your  
definition, because multicol's overflow isn't really in the block  
dimension.

If we can't fix the definition to accommodate columns' somewhat unusual  
overflowing behavior, There are more verbose ways to address the same use  
case:

.artcile { max-height: 30vh; overflow: regions; }
.article::nth-region(n) { column-count: 2; }

However, I like the first syntax better, and it can't really mean anything  
else that this or a no-op, so I hope we can make it work.

  - Florian
Received on Tuesday, 7 August 2012 09:18:23 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:58 GMT