Re: Getting address of enclosing object from address of inner object

Hi and thanks again for answering.
You gents have convinced me is not the path to follow.
I'm proceeding in a different direction.
But by looking at Alexandre's example I thing you've misunderstood my
intentions.
And I'm confused by the statement that compilers might rearrange
members...
Let's suppose I have the following memory layout for objects A, B and
C.


A memory layout
________________________________________
0 - A::member1 <- address of objectA
1 - A::XXXX
2 - A::member2
________________________________________

In this case if I have an object A and the offset to XXXX (suppose 1 )
I can get
back to the address of A.

//

B: virtual A
________________________________________
0 - VPTR <- address of objectB
________________________________________
1 - A::member1 <- address of (A*)(&objectB)
2 - A::XXXX
3 - A::member2
________________________________________
4 - B::member1
________________________________________

In this case I still have the address of XXX and I want to get to the
address of
(A*)(&objectB). This means I want to get from 2 to 1, not from 2 to 0.
I understand that the compiler might put B members before or after A
members
but I don't think it can rearrange the members inside of A.

//

C: B, virtual A
________________________________________
0 - VPTR <- address objectC and (B*)(&objectC)
________________________________________
1 - A::member1 <- address of (A*)(&objectC)
2 - A::XXXX
3 - A::member2
________________________________________
4 - B::member1
________________________________________
5 - C::member1
6 - C::member2
________________________________________

Same thing as above if I have the address of XXXX I think I still can
get reliably
to the address of the enclosed A. That is from 2 to 1. I understand
that going from
2 to 0 would be unreliable. Since I cannot be sure that relative
position of
A members, B members and C members are in this way or another.

Thanks again,
Francesco

-- 
      [ See http://www.gotw.ca/resources/clcm.htm
for info about ]
      [ comp.lang.c++.moderated.    First time posters: Do this! ]