17:30:38 <dougwig> #startmeeting networking_lib
17:30:39 <openstack> Meeting started Wed Apr  6 17:30:38 2016 UTC and is due to finish in 60 minutes.  The chair is dougwig. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:30:40 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:30:43 <openstack> The meeting name has been set to 'networking_lib'
17:30:48 <dougwig> #topic Open discussion
17:31:00 <dougwig> any interest in carving out some lib hacking time during the summit?
17:31:13 <HenryG> There will be a session for it
17:31:32 <HenryG> https://etherpad.openstack.org/p/newton-neutron-lib-next-steps
17:31:43 <HenryG> The etherpad ^^
17:31:49 <HenryG> Please submit ideas
17:32:48 <HenryG> I promised armax I would run the session
17:33:09 <dougwig> ok
17:33:22 <dougwig> you have the link to all the sessions handy?
17:33:52 <HenryG> We can also find some quiet space at other times to do some hacking on lib at the summit.
17:34:22 <HenryG> dougwig: no, I think you'll have to ask armax for that
17:35:48 <dougwig> from the last neutron meeting logs:
17:35:48 <dougwig> https://etherpad.openstack.org/p/newton-neutron-summit-ideas
17:37:45 <dougwig> i saw your note on splitting your patch.  eta on that?
17:37:54 <HenryG> today-ish
17:38:06 <HenryG> I have spent a lot of time trying to test it
17:38:35 <HenryG> I need to isolate the moving parts a little
17:39:03 <dougwig> have you done any test patches on the neutron side migrating the models to the new code?
17:39:11 <dougwig> migrating might have been a poor choice of words in this case.
17:39:15 <kevinbenton> :)
17:39:37 <HenryG> I have a test patch with lbaas
17:39:58 <kevinbenton> will we have a neutron migration migration script? ;)
17:40:16 <HenryG> please, my brain ;)
17:40:21 <HenryG> it hurts
17:40:51 <dougwig> it's neutron, your brain is used to pain.
17:41:39 <HenryG> Some patches to neutron core will need to land first, which switch to using the lib. Otherwise the models don't get imported right.
17:42:15 <HenryG> I have most of it lined up, will post patches soon.
17:42:45 <dougwig> yeah, i had that with exceptions too.  need common ancestors, or i bet sqla gets angry
17:43:48 <HenryG> I can't tell if it gets angry. It is quite passive aggressive about it though.
17:45:40 <HenryG> So anyway, I need to test quite a bit, because the patches to lib have to be published for core to use them, and if they are wrong then we have to re-publish to fix bugs, which feels wrong.
17:46:23 <dougwig> it's not wrong, it's just the annoyance of a library, and an incentive not to move the stuff that gets touched a lot.
17:47:34 <dougwig> do you have neutron reviews you need us looking at?
17:47:43 <HenryG> On that note, folks are continuously updating the files that I am trying to move to lib. I wonder how that's going to work in the future.
17:48:32 <kevinbenton> i'm not sure how we can motivate people to not do this
17:48:46 <kevinbenton> since neutron-lib doesn't really buy anything for core neutron
17:48:58 <HenryG> The changes are legitimate
17:49:41 <HenryG> But once the files are in lib, it will take a publish for core to get them.
17:50:36 <HenryG> I.e. merge to lib; publish lib; update g-r; wait for g-r bot
17:51:40 <HenryG> But I guess it's the same with oslo.* or any other lib
17:52:04 <HenryG> Folks will have to adjust
17:52:30 <dougwig> i made a post to the ML about this, but there isn't ever going to be a clean break where it's one side or the other.  it's always going to be a moving target.  which i don't consider a bad thing.  we don't want neutron development velocity defined by the lib.
17:52:55 <dougwig> that's just IMO.  i think if the lib becomes critical path for neutron, we might as well give up, it's just going to frustrate people.
17:53:59 <kevinbenton> have any other projects started importing the lib yet?
17:54:07 <HenryG> I am thinking about things like https://review.openstack.org/302329 and https://review.openstack.org/300017
17:55:32 <dougwig> kevinbenton: just neutron and lbaas at the moment, on purpose.
17:55:49 <HenryG> But those patches are the "feature" themselves. Not something that other features depend on. So I guess we are OK there.
17:57:02 <dougwig> if all of neutron.context moves, and something needs to go in there, folks will have to get used to deriving a neutron.context from the lib context and modifying that first.  this is all doable in regular python, without the lib being a blocker.
17:57:48 <HenryG> yup
17:58:58 <dougwig> i do wonder if that educational hurdle is too high, sometimes.
18:00:01 <HenryG> We should aggressively update the devref every time someone hits a problem and we help them.
18:01:47 <dougwig> my favorite pasttime.
18:02:11 <HenryG> ask the helpee to update the devref
18:02:55 <HenryG> -2 their patch until they have updated the devref ;)
18:03:13 <dougwig> oh, you want me to go full armax?
18:03:26 <dougwig> there's no coming back from full armax.
18:03:32 <HenryG> We learn from the best
18:04:12 <dougwig> alright, let's get out of here, and get some coding done.
18:04:20 <HenryG> ack
18:04:27 <dougwig> #endmeeting