On May 14, 2013, at 12:21 PM, Andrew Pinski <pins...@gmail.com> wrote: > I think gdb should support __null anyways. Can you see if a newer > version of gdb supports it and if it does not, please file a bug?
__null is a private implementation detail of g++. While gdb could support it, I don't think they have to. Certainly, a consumer of dwarf isn't required to support it by the dwarf standard. I think all handling of __null should be in the compiler. Now, g++ could emit a -g3 style #define to 0 for it, that would also address the issue. I think I'd rather have Jason weigh in on what he feels the right design is, he is our debug expert. :-)
