[Dwarf-discuss] class members and DW_AT_specification

[Daniel Jacobowitz]

On Tue, Feb 27, 2007 at 07:13:29AM -0500, Ron Brender wrote:
> My speculation is that the g++ practice you describe reflects some kind 
> of normalization that occurs within the g++ implementation. Harmless but 
> unnecessary. At least I hope unnecessary because I would find it 
> unreasonable for gdb to depend on this practice for its operation. YMMV...

As far as I know it is internal to g++ and GDB does not depend on it.
In fact, for many years GDB mishandled it; a certain amount of legwork
is required to get namespace scope correct when the specification is
in the correct namespace but the definition is not.

-- 
Daniel Jacobowitz
CodeSourcery