Empty base optimization, sharing vptrs

[Christophe de Dinechin]

From: "Matt Austern" <austern at isolde.engr.sgi.com>
> My take on the discussion is that everyone agrees there is a defect in 
> the standard, because it can reasonably be interpreted both ways,  
and that
> the standard needs to be clarified.  Several people expressed  
support for
> the interpretation under which this proposed layout rule is legal.  My 
> guess (perhaps Mike and Jason could comment, since they've both  
seen the
> discussion too) is that's how this issue will eventually be resolve. 

Even though I don't read the standard that way today, I still would  
favor this resolution, because it just seems to make sense.
Thank you for forwarding the messages, I will investigate why I  
don't receive them.

Christophe