Re: [RFA 2/4] dwarf2_physname

[Daniel Jacobowitz <dan@codesourcery.com>]

On Fri, Feb 05, 2010 at 12:23:48PM -0800, Keith Seitz wrote:
> It might be a gcc-ism, but we actually get a DW_TAG_lexical_block
> between DW_TAG_subprogram and DW_TAG_variable. I presume this is to
> mark the prologue. It is a very nice way to know that a variable is
> not global. O:-)

It's a GCC-ism :-)

> After revisiting that, I think it clearer to understand if I invert
> the original test:
> 
>     case DW_TAG_variable:
>       {
> 	struct attribute *attr;
> 
> 	/* We only need to prefix "globally" visible variables.  These include
> 	   any variable marked with DW_AT_external or any variable that
> 	   lives in a namespace.  [Variables in anonymous namespaces
> 	   require prefixing, but they are not DW_AT_external.]  */
> 
> 	if (dwarf2_attr (die, DW_AT_specification, cu))
> 	  {
> 	    struct dwarf2_cu *spec_cu = cu;
> 	    return die_needs_namespace (die_specification (die, &spec_cu),
> 					spec_cu);
> 	  }

On this part we are in complete agreement.

> 	attr = dwarf2_attr (die, DW_AT_external, cu);
> 	if (attr || die->parent->tag == DW_TAG_namespace)
> 	  return 1;
> 
> 	return 0;
>       }
> 
> How does that look?

I'm not sure, but I think it's close enough for now.

(Things I'm wondering about that we can sort out later: namespace X {
int f() { extern int y; } }, does y go in a namespace?  What
namespace?  It certainly doesn't go in X::f()::y but I don't know what
GCC emits.  namespace X { class Y { static int x; } }, is that
DW_AT_extern?)

-- 
Daniel Jacobowitz
CodeSourcery