Re: Suggested text for 'column-span'

On 5/20/10 2:41 PM, David Hyatt wrote:
> The above model would be very easy to implement in WebKit.

Would it, though?  The existing block-inside-inline implementation in 
Gecko is pretty complex (especially in order to handle dynamic changes 
efficiently).  WebKit's approach to the dynamic change issue seems to be 
to just mishandle them in various cases [1].  There are also issues with 
generated content in WebKit [2].  And issues with containing blocks [3]. 
  So if the "easy to implement" claim is based on analogy with 
block-in-inline, I don't buy it, sorry.  My experience is that getting 
block-inside-inline right is a huge pain.

-Boris

All svn links below are for rev 1083; I don't know how to get 
test.csswg.org to link to a particular rev

[1] 
http://test.csswg.org/svn/contributors/mozilla/incoming/reftests/ib-split/remove-from-split-inline-2.xht 
for example

[2] 
http://test.csswg.org/svn/contributors/mozilla/incoming/reftests/ib-split/insert-into-split-inline-12.xht 
at least and the weird whitespace in this testcase:

   <head>
     <style>
       #i { display: inline; border: 2px solid; }
       #i::after { display: block; content: "Four"; }
     </style>
   </head>
   <body>
     <div id="i">One<div>Two</div><div>Three</div></div>
   </body>

[3] 
http://test.csswg.org/svn/contributors/bzbarsky/submitted/css2.1/visudet/anonymous-block-not-containing-block-001.xhtml

Received on Thursday, 20 May 2010 19:09:54 UTC