W3C home > Mailing lists > Public > www-style@w3.org > September 2009

RE: [css3-multicol] floats that overflow columns

From: Håkon Wium Lie <howcome@opera.com>
Date: Tue, 22 Sep 2009 14:53:18 +0200
Message-ID: <19128.51390.410860.770805@gargle.gargle.HOWL>
To: Alex Mogilevsky <alexmog@microsoft.com>
Cc: fantasai <fantasai.lists@inkedblade.net>, "www-style@w3.org" <www-style@w3.org>
Alex Mogilevsky wrote:

 >    RESOLVED: howcome to add an example of a float intruding into previous
 >              columns and wait for implementors to complain

This is was done:

  http://www.w3.org/TR/css3-multicol/#overflow-and-multicol-elements

The last call deadline is getting close, so it's a good time to discuss.

 > Here is what I propose for floats in multicol:
 > 
 > 1) In [multicol] spec, remove special treatment for overflow of
 > floats. Overflow floats should be clipped to the column exactly as
 > any other kind of overflow would. This way, content that was
 > initially designed for single-column layout has same behavior.
 > 
 > 2) Leave definition of floats that intrude across columns to GCPM
 > [2]. In fact, page floats as currently defined in GCPM are powerful
 > enough to work as shown in [1].

And Fantasai responed:

 > I think I would prefer if instead of 1) we let floats intrude into later
 > columns, but not into earlier columns. I.e. floats can intrude into other
 > columns, but they always overflow the end edge, never the start edge, of
 > their block. I'm not sure how hard it would be to handle that sort of
 > clipping behavior, but the layout part is straightforward.

My preliminary testing shows that no UA has implemented intrusion,
neither into previous nor later columns. Here's a simple table showing
current implementations:

                    SCALES IMAGE    CLIPS IMAGE  CLIPS TEXT BETWEN COLUMNS
gecko               yes             no           no
webkit              no (100% max)   no need      yes, at left gap edge
prince              no (100% max)   no need      no
antenna house       yes             no           yes, at right gap edge

My test case is here:

   http://people.opera.com/howcome/2009/multicol.html

This indicates that, perhaps, intrusion is too complex for now and
that we should:

  - remove it as a requirement
  - make sure it can be cleanly added later (in GCPM, perhaps) 

The easiest thing is probably to say that content should be scaled as
per the specified value, but clipped in the middle of the column. The
"right" answers would then be:

                    SCALES IMAGE    CLIPS IMAGE  CLIPS TEXT BETWEN COLUMNS
all                 yes             yes          yes, in middle of gap

Does this make sense?

-h&kon
              Håkon Wium Lie                          CTO °þe®ª
howcome@opera.com                  http://people.opera.com/howcome
Received on Tuesday, 22 September 2009 12:54:08 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:21 GMT