[Bug c++/26757] [4.1/4.2 regression] C++ front-end producing two DECLs with the same UID

["jakub at gcc dot gnu dot org" <gcc-bugzilla@gcc.gnu.org>]

------- Comment #14 from jakub at gcc dot gnu dot org  2006-04-20 12:04 -------
The reversal of the DECL_UID (x) = DECL_UID (t); assignment in name-lookup.c
causes
FAIL: g++.dg/opt/inline2.C (test for excess errors)
as well as
// { dg-do link }
// { dg-options "-O1" }

static int g;

void f()
{
  extern int g;
  g++;
}

int main () {
  f ();
}
regression.
For some of the duplicate decls in different binding levels I'm pretty sure we
could just duplicate_decls instead of keeping two decls around, but as the
comment explains e.g. for
static int g (int x, int y = 8);

void f ()
{
  int g (int x, int y = 12);
  g (6);
}

int g (int x, int y)
{
  return x + y;
}

int main ()
{
  f ();
  g (6);
}
we need two decls, at least throughout the frontend.  On the other side, the
two
decls are really supposed to share all cgraph {,varpool} node properties,
except
the decl itself (and uid), so I wonder if we couldn't have indirect nodes in
cgraph_hash and cgraph_varpool_hash or something similar for this purpose.


-- 

jakub at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hubicka at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26757

------- You are receiving this mail because: -------
You reported the bug, or are watching the reporter.