On Thu, May 10, 2012 at 9:58 PM, Christophe Fergeau <cfergeau@xxxxxxxxxx> wrote: > Hey, > > I've split this patch in 2 to make it more digestable. First one is > unchanged and only contains the gvir_config_object_set_node_content > changes > The 2nd one only has the gvir_config_object_attach_replace changes, > and it has some changes: I reordered the parameter to _attach_replace to > make the order similar to most other functions, and I changed the > GVIR_CONFIG_OBJECT() casts to (GVirConfigObject *) otherwise we'd get a > runtime check when passing a NULL pointer. Thats not true. I was also afraid of that and I wrote a test app to see if G_TYPE_CHECK_INSTANCE_CAST minds a null but it didn't. > gvir_config_domain_set_devices could try to unref a NULL pointer, I've > changed that, and gvir_config_domain_set_clock was still rejecting NULL > klock parameters through a g_return_val_if_fail. > > I've dropped the gvir_config_domain_source_pty_set_path change for now > since there are other places where _replace_child_with_attribute is used > and could handle NULL pointers as well. > > ACK on these 2 patches, but you should review the 2nd one for obvious > mistakes. Oh, I thought it was just that you divided them in 2. I'll have another look. -- Regards, Zeeshan Ali (Khattak) FSF member#5124 -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list