So with Gecko, will borders with widths that equate to odd pixel values
I believe the rendering of such is somewhat undefined (and I'm not quite sure how we handle them, nor whether we handle them consistently).
After reviewing the CSS 2 and 2.1 spec, I'm still a little skeptical as to if rendering the tables collapsed border over the top of it's containing blocks edge is the intended way of doing it.
It's what the spec says. As to whether it's intended by the spec authors, that's a good question. The issue has been raised with the CSS WG several times now... They've confirmed that this is what the spec says and that it's suboptimal. Nothing past that yet.
Based on this last statement, if borders which spill into margins should be taken into account when determining overflows, shouldn't they also be taken into account when rendering?
No; the rendering just depends on the outer edge of the margin box, which is not affected by the borders (they spill into the margin, but don't affect the margin dimensions).
Notice that when you scroll to the right and bottom there is a 1px white gap.
That sounds like a bug to me.
Further on the positioning of block elements in normal flow, section 9.4.1 states 'In a block formatting context, each box's left outer edge touches the left edge of the containing block.' My interpretation of this is that the table border should not overlap the containing div's border, it should appear just inside of it.
Please look up the definition of "left outer edge" and note that in the presence of negative left margin the left outer edge can be sorta in the middle of the box...
-Boris _______________________________________________ mozilla-layout mailing list [EMAIL PROTECTED] http://mail.mozilla.org/listinfo/mozilla-layout
