Problem occurs with python and its get_current_arch () as it selects
selected_frame and current_frame while still inferior_ptid is valid for the
original parent.  But since this place it is already attached and later
unwinders try to access it, breaking:
  -PASS: gdb.threads/watchpoint-fork.exp: child: singlethreaded: breakpoint after the first fork
  -PASS: gdb.threads/watchpoint-fork.exp: child: singlethreaded: watchpoint after the first fork
  -PASS: gdb.threads/watchpoint-fork.exp: child: singlethreaded: breakpoint after the second fork
  -PASS: gdb.threads/watchpoint-fork.exp: child: singlethreaded: watchpoint after the second fork
  -PASS: gdb.threads/watchpoint-fork.exp: child: singlethreaded: finish
  +FAIL: gdb.threads/watchpoint-fork.exp: child: singlethreaded: breakpoint after the first fork
  +FAIL: gdb.threads/watchpoint-fork.exp: child: singlethreaded: watchpoint after the first fork
  +FAIL: gdb.threads/watchpoint-fork.exp: child: singlethreaded: breakpoint after the second fork
  +FAIL: gdb.threads/watchpoint-fork.exp: child: singlethreaded: watchpoint after the second fork
  +FAIL: gdb.threads/watchpoint-fork.exp: child: singlethreaded: finish

--- ./gdb/infrun.c	2009-12-21 20:26:30.000000000 +0100
+++ ./gdb/infrun.c	2009-12-21 20:26:11.000000000 +0100
@@ -375,6 +375,9 @@ follow_fork (void)
 	  }
 	else
 	  {
+	    /* Possibly referenced PARENT is no longer valid.  */
+	    reinit_frame_cache ();
+
 	    /* This pending follow fork event is now handled, one way
 	       or another.  The previous selected thread may be gone
 	       from the lists by now, but if it is still around, need