On Thu, Jun 12, 2014 at 12:48 AM, Paolo Bonzini <[email protected]> wrote: > Il 11/06/2014 16:32, Peter Crosthwaite ha scritto: > >>> > In the end, I believe the conceptual cleanliness of destroy==unparent, >>> > and >>> > the removal of 200-odd memory_region_destroy calls, are worth the extra >>> > ugly >>> > [0]'s. >>> > >> >> I think you get to keep the uniqueness though under my latest >> proposal? No tree-wide collision resolution needed. When >> object_property_add_child_array sucessfully sets something to [0] (i >> == 0 iteration), just create an alias without [] at the same time. The >> naming system will behave identically to this patch as-is in all the >> collision cases (i >= 1). > > > Yes, but then you get duplicates, and some of them are wrong (the ones > without [0] if there is also a [1]). > > I think the solution could be to special case names ending with "[*]" > (perhaps straight in object_property_add_child!), and then audit manually at > all the 400+ init calls. > > For now I'd rather keep it simple. >
Fair enough. I guess we can add that alias later (and when aliases themselves are a little more baked). Regards, Peter > Paolo >
