On Thu, Jun 12, 2014 at 12:03 AM, Paolo Bonzini <pbonz...@redhat.com> wrote: > Il 11/06/2014 15:42, Peter Crosthwaite ha scritto: > >>> > >>> > - object_deinit(obj, ti); >>> > object_property_del_all(obj); >>> > + object_deinit(obj, ti); >> >> Does this means you lose access to your own properties before calling >> your own finalisation? I guess its a non-issue if all clients just use >> private fields directly, but it seems weird you cant use your own >> getter at the end because it's dissappeared on you already. > > > It is just as weird that a property's finalizer cannot do its work because > it needs something that instance_finalize has already freed. :) > > This is important with the previous patch, that calls arbitrary code of > children objects at finalization time. >
Ok fair enough. Acked-by: Peter Crosthwaite <peter.crosthwa...@xilinx.com> > Paolo >
