[cxx-abi-dev] Problem with vcall offsets

[Nathan Sidwell]

Mark Mitchell wrote:

> Section 2.5.3 says:
> 
>    If virtual base A has a primary virtual base class P sharing its
>    virtual table, P's vbase and vcall offsets come first in the
>    primary virtual table, in the same order they would appear if P
>    itself were the virtual base, and those from A that do not replicate
>    those from P precede them.
Every time I look at this stuff it makes my head spin. I think you're
falling into the same trap I do, because you've reached the same
(erroneous) conclusion. The wording about this is not good,
unfortunately I've never found a more obvious simpler version.

> 
> In other words, the vtable for A, when used as a virtual base is:
> 
>    A's vcall offset for A::a
>    A's vbase offset to P
>    P's vcall offset for P::p
>    P's vbase offset to V
>    V's vcall offset for V::v
>    Offset-to-top
>    Typeinfo
>    Entry for V::v
>    Entry for P::p
>    Entry for A::a
not quite. you've mislabeled things. Think of the vcall offset as belonging
to the derived clss. After all you only need the vcall offset when there's
a virtual base involved.
I.e. A's vtable is
	A's vbase offset to P
	A's vcall offset for P::p
	P's vbase offset to V
	P's vcall offset for V::v
	Offset to top
	typeinfo
vptr points here:	V::v
	P::p
	A::a

When A is a primary virtual (of say Z), the A-in-Z vtable is the same shape
as A's vtable. It is prepended by
	Z's vbase offset to A (which will be zero 'cos A is primary)
	Z's vcall offset to A::a (also zero 'cos of A's primality)


> This cannot possibly be right.  If you want to be able to generate
> code to cast from an A* to a P*, then the index of A's vbase offset
> for P must be constant; it cannot depend on whether or not A itself is
> a virtual base in something else. 
Huh? its always at aptr->vtable[-6]

> That implies that the vbase and
> vcall offsets cannot be interleaved; instead, the vbase offsets must
> always come nearest the address point.
Don't follow. When you have virtual polymorphic base first prepend a
vcall offset for all inherited virtual functions which have now
*just become* morally virtual, then prepend a vbase offset for the
virtually inherited base.

> I think that we need to always emit vcall offsets, even for bases that
> are not virtual.
No.

> Thoughts?
You are confused by a maze of twisty vcalls. I always find the output
of -fdump-class-hierarchy enlightening.

nathan

-- 
Dr Nathan Sidwell   ::   http://www.codesourcery.com   ::   CodeSourcery LLC
         'But that's a lie.' - 'Yes it is. What's your point?'
nathan at codesourcery.com : http://www.cs.bris.ac.uk/~nathan/ : nathan at acm.org