Mangling: Special names

[Jim Dehnert]

> From loewis at informatik.hu-berlin.de  Mon Mar 27 06:05:59 2000
> 
> No, that won't work. Is this the in-charge, or not-in-charge
> constructor? Is it allocating or not?
> 
> Reconsidering this, I think it is right to have the special names as
> unqualified names in a nested name. Just as you'd have
> 
> _ZN3FoongE
> 
> to denote Foo::operator!(); you also have C1,... as *names* of the
> constructors. Since we need more than one name per constructor, we
> need to make them special

Yes, I think you're right.  I've changed it.

> > > Likewise, _ZN3Foo3BarTVE would be the Foo::Bar virtual table.
> > 
> > No, it goes:
> > 
> >     <encoding>	-> _Z <name>	     <type>
> > 		-> _Z <special-name> <class-enum-type>
> > 		-> _Z TV	     <nested-name>
> > 		-> _Z TV	     N 3Foo 3Bar E
> 
> In my view, the Foo::Bar virtual table is a static member of Foo::Bar,
> which would be written down as 'Foo::Bar::__vtable'. Just as this
> specific member would be encoded as
> 
>   _ZN3Foo3Bar8__vtableE
> 
> I would assume that the vtable is encoded with TV as the name of the
> static member. Your approach would also work, but I feel this is more
> consistent that way.

Hmmm, I hadn't thought of it that way.  Your method would work, too,
though sometimes it would result in extra N...E delimiters, as in the
vtable for a simple class Foo, which would become _ZN3FooTVE instead of
_ZTV3Foo.  Not a huge difference -- does anyone else have an opinion?

Jim

-	    Jim Dehnert		dehnert at sgi.com
				(650)933-4272