[lldb] Use correct size when dumping DWARF64 DW_FORM_ref_addr

Not that we ever do that, because this is unused code, but if
someone was debugging lldb I guess they'd call this.

Was missed in llvm#145645

Relates to llvm#135208

Author: DavidSpickett

lldb/source/Plugins/SymbolFile/DWARF/DWARFFormValue.cpp

@@ -420,8 +420,7 @@ void DWARFFormValue::Dump(Stream &s) const {
       DumpAddress(s.AsRawOstream(), uvalue, sizeof(uint64_t) * 2);
     else
       DumpAddress(s.AsRawOstream(), uvalue,
-                  4 * 2); // 4 for DWARF32, 8 for DWARF64, but we don't
-                          // support DWARF64 yet
+                  m_unit->GetFormParams().getRefAddrByteSize());
     break;
   }
   case DW_FORM_ref1: