Emptyness vs. virtualness

[Jim Dehnert]

Mark Mitchell wrote:
> 
> >>>>> "Jim" == Jim Dehnert <dehnert at transmeta.com> writes:
> 
>     Jim> Well, I think it's well-defined now, and can be clarified if
>     Jim> necessary.  Further, I think the alternative will be a
>     Jim> messier rule.  So would it be beneficial enough to be worth
>     Jim> changing it?
> 
> I'm sorry; I'm not sure I understand what you think is the status
> quo.  That only non-virtual parts are considered for overlap, or that
> virtual parts are?

I thought that all parts were considered, but that the virtual base A
couldn't be placed at offset zero because the non-virtual one was
already there.  Did I misread the example?

On looking at the ABI again, though (my copy), I do see a typo that isn't
helping.  The first paragraph of 2.4 (III) says "II-2 (if empty) or II-3
(if non-empty)" but should swap the parenthesized comments.

Jim
-- 
-	    Jim Dehnert		dehnert at transmeta.com
	    (408)919-6984	dehnertj at acm.org