- From: Peter Moulder <peter.moulder@monash.edu>
- Date: Sun, 09 Jan 2011 19:12:59 +0000
- To: www-style@w3.org
In the following example: <span style="display:table-row;"> <span style="display:table-cell;">Cell1</span> <span>Inline <span style="display:block; width:50%; border:1px dashed blue;"> Block </span> </span> <span style="display:table-cell;">Cell3</span> </span> I believe we want the width:50% to be calculated with respect to the anonymous table cell even though it's anonymous (and is thus not the "closest non-anonymous ancestor box"). So I believe we only want to skip past the anonymous block boxes created by 9.2.1.1 rather than all anonymous boxes. Another problem with the way this provision was quickly added to the specification is that it results in contradictions with each of the places whose behaviour it was intended to modify: if propdef-top says that "percentages refer to height of containing block", and the containing block of a certain box is (according to §10.1) defined by an anonymous block box, then that contradicts the claim in §9.2.1.1 that the anonymous block box is ignored. Avoiding these contradictions requires actually changing propdef-top and so on (e.g. introduce a new concept "containing height" that propdef-top etc. can reference), or changing the definition of "containing block". pjrm.
Received on Sunday, 9 January 2011 19:13:30 UTC