[11:25] <Laney> cjwatson: I just saw you've been working on Built-Using support... is that close to being ready at all? The new proposed-migration is now enforcing this but it's a little bit odd to do so without LP actually keeping the sources around. Steve grumbled about it a bit and forced qemu in over a complaint, and I'm wondering what we should do here really. I think in principle enforcing BU is a good idea...
[11:37] <cjwatson> Laney: Making progress but it'll probably be a little while yet
[11:38] <cjwatson> Maybe a month or so if there are no more serious roadblocks?  Wild guess though
[11:38] <Laney> nod
[11:38] <Laney> I guess we should turn the policy off until then
[11:41] <cjwatson> The basic idea is that if sources referenced by Built-Using aren't published in their own right then they'll stick around (or if necessary be copied in) with a new "Referenced" publishing status for as long as necessary
[11:42] <cjwatson> But it's a new publishing status, significant dominator work, guards around copying and deletion, etc. - not straightforward.  And the last time we tried to turn part of the work in progress on, we discovered a significant roadblock with kernel workflow that forced a rethink
[11:43] <Laney> Yeah, I've not thought through all of the implications but I can imagine there are many
[11:45] <Laney> I think it only really matters for the release team when it entangles things in -proposed
[11:54] <cjwatson> The interaction with -proposed is a bit complex in itself.  I think it will likely involve publications being temporarily copied into -proposed in a Referenced status in order to satisfy Built-Using (this is less confusing than trying to do cross-pocket domination IMO).  Hopefully proposed-migration will just ignore those because they'll be duplicates of publications in the release pocket