[Bug debug/51668] New: class DW_AT_name does not match method's linkage name prefix (char)1 vs. '\001'

[Bug debug/51668] New: class DW_AT_name does not match method's linkage name prefix (char)1 vs. '\001'

[jan.kratochvil at redhat dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51668

             Bug #: 51668
           Summary: class DW_AT_name does not match method's linkage name
                    prefix (char)1 vs. '\001'
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: debug
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: jan.kratochvil@redhat.com
                CC: dodji@gcc.gnu.org
            Target: x86_64-unknown-linux-gnu


PASS: gcc (GCC) 4.6.3 20111222 (prerelease)
FAIL: gcc (GCC) 4.7.0 20111222 (experimental)

>From the GDB point of view it is a regression because gcc-4.7 now provides the
non-matching DW_AT_linkage_name for ctors/dtors which GDB prefers over its
physname computed from class's DW_AT_name.

template<char c>
class C {
public:
  void m () {}
  ~C() {}
};
typedef C<1> D;
D d;
int main () { d.m (); }

nm:
0000000000400604 W _ZN1CILc1EED1Ev
0000000000400604 W _ZN1CILc1EED2Ev
nm -C:
0000000000400604 W C<(char)1>::~C()
0000000000400604 W C<(char)1>::~C()
readelf -wi:
 <1><3a>: Abbrev Number: 3 (DW_TAG_class_type)
    <3b>   DW_AT_name        : (indirect string, offset: 0xac): C<'\001'>       

The gcc-4.6 -> gcc-4.7 regression:

-PASS: gdb.cp/templates.exp: print destructor of template typedef
+FAIL: gdb.cp/templates.exp: print destructor of template typedef

(gdb) p D::~C
$1 = {void (C<(char)'\001'> * const)} 0x40056e <C<(char)'\001'>::~C()>
->
There is no field named ~C