Re: [CSS21] Proposal for a replacement for section 17.2.1 (table anonymous objects)

2009/5/31 Boris Zbarsky <>:
> Brad Kemper wrote:
>>> 3) If a box T is not a 'table-column' box and has a parent P which is a
>>>  'table-column-group' box, discard T.
>> Does it go without saying that if the display of parent P ever changed to,
>> say, a table-body box or table-row box (via JavaScript, for instance), that
>> you may need to re-insert T? Or that if T changed its display to
>> 'table-column', that it should be re-inserted?
> I would think so; it ought to be a fundamental principle, if it's not
> already, that the box tree (and layout) only depends on the current computed
> values of properties for all the nodes, not on their history...  Hysteresis,
> while observed in some UAs in some cases, is highly undesirable, in my
> opinion.
> If desired, we could probably make display compute to none in cases like
> this instead of having the rule about throwing out the box.  It probably
> doesn't matter much either way; all that's affected is what the DOM style
> APIs return for the relevant nodes; the rendering is the same either way.

As a general rule, finding the computed value of a property does not
require laying out the page, so this point should change only the used
value. This is the reason we have specified, computed, used and actual
values, differently from CSS2.0.

> -Boris


Received on Sunday, 31 May 2009 12:26:38 UTC