Re: [RFA] stabsread.c: Fix for `ptype class Static' stabs FAIL

[Elena Zannoni <ezannoni at redhat dot com>]

Peter.Schauer writes:
 > The patch below fixes the `ptype class Static' stabs FAIL from classes.exp.
 > 
 > Ok to commit ?
 > 

If it only affects c++ (and it looks like it does) and if DanielJ says
Ok, it's fine with me.

Elena


 > 	* stabsread.c (read_member_functions): Fix is_stub test for
 > 	static member functions, improve comment.
 > 
 > --- ./stabsread.c.orig	Sat Feb 23 12:17:26 2002
 > +++ ./stabsread.c	Wed Feb 27 20:08:37 2002
 > @@ -3287,13 +3287,27 @@ read_member_functions (struct field_info
 >  	      }
 >  	    case '?':
 >  	      /* static member function.  */
 > -	      new_sublist->fn_field.voffset = VOFFSET_STATIC;
 > -	      if (strncmp (new_sublist->fn_field.physname,
 > -			   main_fn_name, strlen (main_fn_name)))
 > -		{
 > -		  new_sublist->fn_field.is_stub = 1;
 > -		}
 > -	      break;
 > +	      {
 > +		int slen = strlen (main_fn_name);
 > +
 > +		new_sublist->fn_field.voffset = VOFFSET_STATIC;
 > +
 > +		/* For static member functions, we can't tell if they
 > +		   are stubbed, as they are put out as functions, and not as
 > +		   methods.
 > +		   GCC emits the fully mangled name if
 > +		   dbxout.c:flag_minimal_debug is not set, so we have to
 > +		   detect a fully mangled physname here and set is_stub
 > +		   accordingly.  */
 > +		if (!(strncmp (new_sublist->fn_field.physname,
 > +			       main_fn_name, slen) == 0
 > +		      && new_sublist->fn_field.physname[slen] == '_'
 > +		      && new_sublist->fn_field.physname[slen + 1] == '_'))
 > +		  {
 > +		    new_sublist->fn_field.is_stub = 1;
 > +		  }
 > +		break;
 > +	      }
 >  
 >  	    default:
 >  	      /* error */
 > 
 > -- 
 > Peter Schauer			pes@regent.e-technik.tu-muenchen.de