Paolo Carlini wrote: However... Looking at gcc/NEWS of 3.4.0, it seems the *wrong* one, that is, that of 3.3.x! Is this a known issue? Disregard the above: I see that you have already fixed it a few days ago. Thanks, Paolo.