Empty base optimization, sharing vptrs

[Christophe de Dinechin]

From: "Matt Austern" <austern at isolde.engr.sgi.com>

> NONPOLYMORPHIC BASES
>
> We may allocate any number of empty base subobjects at the same  
location
> (the beginning of the complete object) except when, because of repeated 
> inheritance, we'd have multiple subobjects of the same type  
allocated in
> the same place.

You probably have noticed that I currently don't read the standard  
that way. But I also acknowledge that this is quite a reasonable way  
of implementing things. Jason was supposed to send a message to  
c++core regarding this issue. I did not receive this message, but  
this might be a problem with our mail setup. Was the message sent?

> (Oh, another point.  I don't have a proof, but my guess is that
> finding an optimal space-minimizing solution to the class layout
> problem is equivalent to the travelling salesman problem.  I'm not 
> going to try.

Ah ah, but that's the kind of optimization that could be useful to  
make people buy IA64-class machines :-)


Christophe