On 2008-10-02, dorayme <do************@optusnet.com.auwrote:
In article <sl*********************@bowser.marioworld>,
Ben C <sp******@spam.eggswrote:
>Firefox doesn't allow table cells to establish containing blocks for
positioned descendents.
The *only* browser of mine that does establish a containing block is the
now defunct one of MacIE5! They don't make menschy browsers like that no
more!
<http://dorayme.890m.com/alt/relativelyAbsolute.html>
compared with
<http://dorayme.890m.com/alt/relativelyAbsolute_table.html>
It works in Konqueror 3.5.9, but not in Opera 9.5.1.
Anyway the spec is arguably ambiguous on this as it says in 9.3.1:
The effect of 'position:relative' on table-row-group,
table-header-group, table-footer-group, table-row,
table-column-group, table-column, table-cell, and table-caption
elements is undefined.
but it really it's talking offsetting, not the side-effect of
establishing containing blocks.
The section about containing blocks doesn't mention any exceptions for
table-cells. It just says "the nearest ancestor"-- it's even supposed to
work for inlines (which can create containing blocks with negative
widths, whatever that's supposed to mean).