problems with empty base layout algorithm

[Martin von Loewis]

> It seems strange to allow subobjects to be allocated in the middle
> of another subobject no?  It feels similar to the idea of
> interleaving field allocation.

Doing so is specifically allowed in the standard. 10/3 says

>> The order in which the base class subobjects are allocated in the
>> most derived object (1.8) is unspecified.

Also, relational operators for subobject addresses are unspecified -
unless they compare addresses of *member* subobjects, with no
intermediate access specifier. So you can compare base subobject
addresses only for equality.

Regards,
Martin