Looking at my pull_leaf() routine the translation nodes are properly extracted but there is a problem with restoring the primitive back to (0,0,0). looks like there is a bug in line with the ft_xform routine, and I need some clarifications with with the pull_comb() routine, the 4th argument of the rt_db_get_internal() does it represent the matrix transformation of the database object? With these clarifications i'll be able to debug the routine removing the NAN being pulled from the mged objects.