[01:57] <sadmac> Keybuk: did you look at this? https://bugzilla.redhat.com/show_bug.cgi?id=432384
[02:00] <Keybuk> sadmac: I hadn't seen that
[02:00] <Keybuk> for some reason, I'm not getting mails from your bugzilla
[02:01] <sadmac> hm. weird.
[02:02] <sadmac> Keybuk: do you have an account? (I don't think it filters on that, but maybe).
[02:02] <Keybuk> I do, yes
[02:03] <Keybuk> and I set myself to be cc'd on the meta bug
[02:03] <sadmac> hmm.
[02:03] <sadmac> spam filter?
[02:04] <Keybuk> I've checked that 
[02:04] <Keybuk> my spam filter only grey lists on spam
[02:05] <ion_> ma164709 < ion_> keybuk: Please take a look at the refcount branch now. (nih_alloc_reparent, nih_realloc still have  to be taken care of.)
[02:05] <ion_> ma164728 < ion_> keybuk: Does that look okay?
[02:05] <sadmac> Keybuk: I'll hit devel-list about it.
[02:06] <sadmac> Keybuk: or you could (might be nice to introduce yourself to the people :)
[02:06] <Keybuk> ion_: it's on my TODO to look at :)
[02:11] <Keybuk> ion_: did you push the branch to LP?
[02:13] <Keybuk> cause I can't remember the URL :)
[02:13] <ion_> keybuk: Hm, i probably didn’t make LP mirror it. Let me see...
[02:14] <Keybuk> s'ok, just the original URL would do too :p
[02:14] <ion_> Yeah, just grepping for it. :-)
[02:15] <ion_> http://johan.kiviniemi.name/software/bzr/libnih.refcount/
[02:21] <Keybuk> ion_: why uint32_t for refcount?
[02:21] <Keybuk> unsigned int seems the proper type?
[02:22] <ion_> Oh, i forgot to change that. We already discussed this. :-)
[02:23] <ion_> What i changed now is just the most recent commit.
[02:26] <Keybuk> looks good though
[02:26] <Keybuk> can't see any obvious problems
[02:29] <Keybuk> I'm wondering whether you need the restriction that you can't ref/unref a non-parent object
[02:29] <Keybuk> it should work if you can
[02:29] <Keybuk> since the parent holds one ref, and the other object another
[02:29] <Keybuk> if the parent is freed, the object will stay around because it's still ref'd
[02:29] <Keybuk> that other object would have to take care of unref'ing in its destructor though
[02:32] <ion_> OTOH, it would be somewhat confusing if an object is in multiple “containers”, only one of which is the parent. I was thinking that a “root” object (parentless) and all its children would be a single refcounted entity.
[02:35] <ion_> keybuk: Ok, i fixed refcount’s type. Pushed.
[02:36] <Keybuk> it might be interesting if each of the objects could have an n-way relationship with each other, rather than strict parent/child
[02:36] <Keybuk> that way unref'ing would be automatic
[02:36] <ion_> True...
[02:36] <Keybuk> maybe I'm overcomplicating things as usual there though
[03:25] <sadmac> Keybuk: nobody else is reporting missing emails. Could it be something in your bugzilla account settings?