13:00:22 <bstinson> #startmeeting CBS/Infra 13:00:22 <centbot> Meeting started Mon Oct 20 13:00:22 2014 UTC. The chair is bstinson. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:00:22 <centbot> Useful Commands: #action #agreed #help #info #idea #link #topic. 13:00:57 <imcleod> .hellomynameis imcleod 13:01:00 <bstinson> #chairs kbsingh alphacc Arrfab MerlinTHP 13:01:13 <bstinson> #chair kbsingh alphacc Arrfab MerlinTHP 13:01:13 <centbot> Current chairs: Arrfab MerlinTHP alphacc bstinson kbsingh 13:01:19 <bstinson> hi imcleod 13:01:22 <bstinson> and kbsingh 13:01:27 <imcleod> Hey bstinson and all 13:01:28 <MerlinTHP> Oooh 13:01:30 <MerlinTHP> It's that time 13:01:31 <MerlinTHP> Hi 13:01:39 <bstinson> should be a short meeting 13:01:45 <bstinson> #topic Agenda 13:01:50 <bstinson> #info Topic: Status Updates 13:01:55 <bstinson> #info Topic: Open Floor 13:01:55 * MerlinTHP "copy run start"s so he doesn't forget 13:02:20 <bstinson> #topic Status Updates 13:03:17 <bstinson> #info next action for Centpkg is to get with MerlinTHP about starting on the lookaside script 13:03:47 <MerlinTHP> OK, we can talk about that this week :) 13:05:17 <bstinson> MerlinTHP: i think we both have access to the cbs-tools repo on git.c.o 13:05:33 <MerlinTHP> Regarding my actions from last week... I've started a requirements document, but it's still at the semi-random bullet point stage. 13:05:37 <MerlinTHP> Yep, I do :) 13:05:43 <Arrfab> bstinson: normally yes, I've given access to both of you to that git repo 13:06:06 <MerlinTHP> I still need to talk to kbsingh about ACL granularity 13:06:29 <MerlinTHP> I'll try and get the lookaside stuff sorted this week. 13:07:37 <MerlinTHP> That's my actions from last week, basically :) 13:08:29 <Arrfab> MerlinTHP: yeah, I was supposed to also work with on that requirements document, but poodle and 6.6 got in the way 13:08:59 <MerlinTHP> Mm :) 13:09:19 <MerlinTHP> I was going to look at it more last week, but there was unwatched TV on my DVR, sorry ;) 13:09:35 <bstinson> i should have a light week at work, ping me if you want a hacking partner 13:10:03 <MerlinTHP> Sure :) 13:10:38 <bstinson> kbsingh: can you look at http://bugs.centos.org/view.php?id=7649 and http://bugs.centos.org/view.php?id=7648 when you get a chance? 13:10:51 <bstinson> i should have one more repo request coming in for centos-packager 13:11:33 <kbsingh> bstinson: wilo tomorrow AM - am going to also try the gitblit upgrade then 13:11:43 <kbsingh> bstinson: the blocker there is really working out where the repos need to end up at 13:11:54 <kbsingh> they dont really fit into a SIG model 13:12:05 <kbsingh> so maybe right at the top level with centos-git-common ? 13:13:20 <bstinson> that would probably work, i was thinking for each project we keep 1 repo for the actual source and 1 in dist-git format to build through koji 13:13:39 <bstinson> other workflows are welcome 13:14:23 <kbsingh> how do you mean actual source and dist-git ? 13:14:49 <MerlinTHP> Think of it as dist-git vs upstream 13:14:56 <bstinson> taking centpkg for example, there is already https://git.centos.org/summary/?r=centpkg.git 13:15:11 <MerlinTHP> dist-git is spec + tarball, "actual source" is where the tarball comes from 13:15:27 <bstinson> indeed 13:15:33 <kbsingh> hum 13:15:48 <MerlinTHP> It just happens that in this call, it's all us 13:15:55 <kbsingh> this is something that came up the other day as well, i dont think we've actually got a plan or a procss for people who want to do dev work against git.centos.org 13:16:40 <bstinson> what trouble were they running into? 13:17:00 <kbsingh> lack of process :) 13:17:13 <bstinson> heh :) 13:17:15 <kbsingh> git.centos.org is setup to host rpms, thats about it 13:17:55 <kbsingh> for content that isnt in the rpm format, the sig's can setup their own repos, eg for tests or docs etc 13:18:08 <kbsingh> i guess in this case, we can do something similar 13:18:29 <kbsingh> but its worth having a wider conversation on the centos-devel list a bit, so we make sure that a chunk of people are aware of this 13:19:13 <bstinson> is the plan to have separate projects for them to put those in? 13:21:31 <kbsingh> bstinson: i think thats just it - there is no plan. 13:21:55 <kbsingh> bstinson: what i said up there, was mostly meant to be for tests / docs - i dont think we actually hve a plan for a development git repo 13:22:16 <kbsingh> one challenge will be the branch naming, even if its a dev git repo, it will still limit you to the signame branch 13:22:41 <kbsingh> if we need to have repo's that dont restrict people to this ( which would make sense fro dev git repos ) - then i will need to go do some work at the backend 13:24:07 <bstinson> another option would be to spin off dev/tests/docs someplace else and do rpms-only in git.c.o 13:24:12 <bstinson> of course that would be a lot of extra work 13:28:17 <bstinson> kbsingh: would it help if I sent out an informal poll to the list, asking about preferred workflow? 13:28:41 <kbsingh> where would the someplace else be ? 13:29:35 <bstinson> that would be another machine somewhere, analogous to fedorahosted.org 13:30:12 <kbsingh> bstinson: thats an interesting though 13:30:45 <kbsingh> somedays, i can type - today is not one of those days 13:30:53 <mikem> centoshosted++ 13:31:11 <kbsingh> bstinson: i think we can solve this by just having /rpms/ as a project be the dist-git stuff, and let everything else be in the other projects 13:31:34 <Evolution> the downside with doing a 'centoshosted' is that to some extent fedora's had failure with that 13:31:40 <kbsingh> mikem: I'm thinking that its going to be another piece of infra to maintain 13:31:43 <Evolution> fedorahosted isn't exactly winning. 13:32:05 <bstinson> kbsingh: that makes sense too 13:32:20 <mikem> kbsingh, yeah, otoh I think we discussed this before and folks suggested that github could just be used instead 13:32:30 <kbsingh> mikem: yeah 13:32:47 <MerlinTHP> as a project, centos seems a lot less allergic to github than fedora is 13:32:49 <kbsingh> i beleve we have a github.com/centosproject/ as well - we might be able to reuse that for adhoc upstreams 13:33:27 <mikem> kbsingh, doesn't have to be a different server. could be a different topdir/project on git.c.o 13:34:35 <mikem> but, while gitblit is pretty nice, github is probably nicer for folks working on projects 13:35:21 <kbsingh> yeah, the biggest win with github is someone else gets to run the signup and merge-req process 13:35:34 <MerlinTHP> :) 13:35:41 <kbsingh> bstinson: i think we can thrash out a solution, fairly quickly, but i dont think we should do that here. 13:35:57 <bstinson> sure, to the list 13:35:58 <kbsingh> its a pretty big deal and i think it would be best if we moved this specific item to the centos-devel list and work it there 13:37:14 <bstinson> kbsingh: did you want me to write something up, or were you going to send that out? 13:39:11 <kbsingh> sure, why dont you go ahead and send something in 13:40:14 <bstinson> #action bstinson will write up a discussion item on centos-devel about dev work on ad-hoc upstreams 13:40:24 <bstinson> #topic Open Floor 13:40:39 <bstinson> this has all been pretty much open floor :) but what else? 13:40:46 <MerlinTHP> Nothing else from me 13:41:17 <kbsingh> there was an issue on building from git.centos.org - did that get looked into ? 13:41:37 <kbsingh> i think gwd sent off ( requested! ) a build from rpms/xen.git - but it failed to show up in cbs 13:45:25 <bstinson> was there a bug filed? 13:45:39 <MerlinTHP> ISTR the build failed with a "can't find spec file" error 13:45:48 <MerlinTHP> Not sure of alphacc got a chance to look at it 13:45:52 <MerlinTHP> *if 13:48:04 <Evolution> MerlinTHP: would you be offended if I asked about auth? 13:48:16 <MerlinTHP> Not at all 13:48:24 <Evolution> in that case... auth? 13:48:28 <MerlinTHP> I was watching tv last week, though ;) 13:48:37 <MerlinTHP> Sorry, I was being slack last week. 13:48:44 <imcleod> I have a minor open floor request 13:48:49 <MerlinTHP> shoot 13:49:01 <imcleod> Could we get Colin Walters' CBS access sorted out: http://bugs.centos.org/view.php?id=7699 13:49:10 <imcleod> For Atomic SIG work. 13:49:11 <imcleod> :-) 13:50:38 <kbsingh> imcleod: given that the git.c.o -> cbs integration stuff isnt working quite yet, is there much value there ? 13:50:45 <kbsingh> imcleod: i guess he could hit scratch builds 13:51:28 <imcleod> kbsingh: My understanding is that cbs will currently allow tagged non-scratch builds from directly uploaded SRPMS. Appreciate this is not ideal but it at least starts getting us Atomic content built in a public context. 13:51:50 <imcleod> kbsingh: I have my account and was going to start trying to create a stream of RPM builds. Colin has offered to participate as well. 13:52:16 <imcleod> kbsingh: We will, of course, move to git.c.o ASAP. 13:53:04 <kbsingh> ok, works for me 13:53:12 <imcleod> kbsingh: Cool. 13:53:15 <bstinson> this goes to alphacc? 13:53:28 <Evolution> iirc MerlinTHP should have rights as well. 13:53:32 <kbsingh> yea, I'll sync with him once he is back on the right side of the pond 13:53:40 <MerlinTHP> what eh banana, not yet 13:53:50 <MerlinTHP> On the todo list for this week, I believe. 13:53:55 <Evolution> ah, okay 13:53:56 <Arrfab> Evolution: not that I'm aware if 13:54:39 <bstinson> what else for open floor? 13:54:52 <kbsingh> imcleod: ETA is going to be a few days on that req. 13:55:03 <imcleod> kbsingh: OK. 13:55:23 <imcleod> Me again: http://bugs.centos.org/view.php?id=7685 - Enabling koji image building features 13:55:52 <imcleod> Depending on the precise version of koji in use in the CBS, this may be as simple as adding the Image Factory packages to the lone koji builder and firing up libvirtd. 13:56:59 <alphacc> imcleod: we didn't have time to look at this one. sorry. I'll try to sort that out this week. 13:57:24 <imcleod> alphacc: No worries, just wanted to reaffirm my interest in it. Also, I can assist with details, as I helped implement the feature in koji. 13:57:34 <imcleod> alphacc: Thanks! 13:58:06 <alphacc> imcleod: I'll drop a msg when I start 13:58:21 <imcleod> alphacc: Cheers. 13:59:11 <bstinson> great! if there's nothing else, i'll be closing the meeting in 1 minute 14:00:20 <bstinson> #endmeeting