11:07:30 <Stskeeps> #startmeeting Mer release management meeting 10/4/2012
11:07:30 <MerBot> Meeting started Tue Apr 10 11:07:30 2012 UTC.  The chair is Stskeeps. Information about MeetBot at http://wiki.merproject.org/wiki/Meetings.
11:07:30 <MerBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
11:07:37 <Sage> o/
11:08:06 <lbt> hi
11:09:07 <Stskeeps> so where are we at with next release: number of CVE fixes merged
11:09:13 <Stskeeps> nspr/nss upgrade
11:09:13 <Stskeeps> d
11:09:37 <Stskeeps> i'd like to ditch nss from glibc, like ubuntu and tizen does
11:09:56 <Stskeeps> libsolv has been merged, zypper upgrade around the corner
11:10:04 <Stskeeps> i'm testing sb2 2.3.48 in a seperate OBS project
11:10:20 <lbt> I should be able to test OBS with sb2 too
11:10:22 <lbt> Qt5 ?
11:10:26 <Sage> python 2.7
11:10:29 <Sage> ?
11:11:19 <Stskeeps> lbt: qt5 has gotten an update and i need to push an update from bostik's stuff
11:11:37 <Stskeeps> python 2.7 will go in after 2.3.48 as i need to be sure it's not caused by python
11:12:18 <lbt> so the sb2 change - does that cause SDK issues?
11:12:21 <Stskeeps> just been away from the computer this week
11:12:27 <Stskeeps> sb2 .. no, it should on the contrary improve
11:12:50 <lbt> OK - I need to look at improving the mapping flexibility there too
11:13:12 <lbt> it's a high-ish hurdle to getting started
11:14:34 <lbt> on my side I've been following-on the OBS packaging work from Islam with our OBS version
11:14:49 <lbt> Setup a virtualised network with distributed systems; documentation of setup and config so it's easy to do regression tests.
11:15:04 <lbt> Fixed some small bugs in api/webui too
11:15:33 <Stskeeps> :nod:
11:15:36 <lbt> the idea is to be able to do automated deploy of OBS and then build against MDS
11:15:43 <Stskeeps> as a test case?
11:15:50 <lbt> yes
11:15:52 <lbt> side effect is that it's also easier to deploy
11:16:09 <lbt> ie so when we use sb2 we can do test builds
11:16:25 <lbt> the tricky bit will be getting coverage
11:16:38 <Stskeeps> do a mer no-use-for-build test?
11:17:00 <lbt> I was thinking "build acl once" :)
11:17:07 <Stskeeps> you're not thinking big enough!
11:17:15 <lbt> hehe
11:17:26 <lbt> well, it'd be a useful thing to do on more HW
11:17:39 <Stskeeps> acl once, but for multiple architectures
11:17:43 <lbt> yes
11:18:27 <lbt> the objective is only obs install/regression initially
11:18:30 <Stskeeps> yep
11:18:43 <lbt> the problem is I can't do things like Xen workers
11:18:51 <lbt> or at least not easily
11:19:12 <lbt> and I can't do soak tests like the ones which kill cobs workers
11:19:34 <lbt> speaking of which https://build.pub.meego.com/monitor
11:19:45 <lbt> no dead workers - I may have fixed it :)
11:19:59 <Sage> \o/
11:20:22 <lbt> anyhow - that's been taking time just messing about setting up the infra for the virtualisation
11:20:41 <lbt> but it's looking good now and I've been making commits
11:21:24 <lbt> https://github.com/lbt/open-build-service/tree/lbt_master
11:22:40 <lbt> speaking of OBS/projects - should we be providing named versions of Mer core in the cobs?
11:22:49 <Stskeeps> named?
11:22:52 <Stskeeps> tagged, you mean
11:22:56 <lbt> yes
11:23:08 <Stskeeps> we (may) with mds2
11:23:17 <Stskeeps> depending on needs
11:23:24 <lbt> I'm kinda uncertain
11:23:44 <lbt> I would almost like to do it on request
11:23:44 <Stskeeps> source is easy
11:23:49 <Stskeeps> binary downloads, a little harder
11:24:07 <lbt> yes - but 'supporting' building against january releases?
11:24:54 <Stskeeps> i think it'd be up to policy
11:25:00 <lbt> or, come july ... supporting the march release :)
11:25:03 <Stskeeps> on a sidenote, we should consider to add Tizen targets to COBS
11:25:26 <lbt> yes, I saw that ... would be a good task
11:25:27 <Stskeeps> ie, a binary import
11:27:44 * lbt adds to his TODO
11:28:22 <Stskeeps> a prjconf is in the 'gbs' tools package
11:29:15 <lbt> I've also updated our release instructions over the past couple of weeks (including bz + ks now) so that should be useful. There's probably more to do on SDK, images and the like
11:29:32 <lbt> yes, I'll make that visible w/o a tunnel sometime
11:31:10 <lbt> so things for the next week(s) ?
11:31:39 <lbt> I'm hoping to get IMG to do builds as part of CI/DBC
11:31:48 <Stskeeps> :nod:
11:32:09 <lbt> we had a nasty problem with BOSS (ruote) on c.obs last week too
11:32:12 <lbt> FYI
11:32:24 <lbt> ran out of disk and corrupted the DB
11:32:27 <Stskeeps> eep
11:32:33 <Stskeeps> [ ] backup for BOSS?
11:32:34 <Stskeeps> :P
11:32:38 <lbt> indeed
11:32:47 <lbt> [ ] all backups...
11:32:56 <lbt> head -> sand
11:33:05 <lbt> managed to do some recovery but it was a bit scary
11:33:05 <Stskeeps> have you found phaeron's copy-project stuff yet?
11:33:22 <lbt> not looked - I saw him pop up last week and expected him to be about
11:33:28 <Stskeeps> ok
11:33:46 <Stskeeps> btw, what was the ultra-show-webui problem we had?
11:34:00 <lbt> on the mer-cobs?
11:34:02 <Stskeeps> yes
11:34:10 <lbt> not resolved
11:34:18 <Stskeeps> ok
11:34:23 <lbt> to do with internal name != external name
11:34:31 <lbt> and reverse proxies and ssl
11:34:56 <lbt> I was planning to erase the setup and redo it using the new packages
11:35:15 <lbt> which would also clean the setup (which is now a hacky mess)
11:35:16 <Stskeeps> makes sense
11:35:40 <lbt> we did find many undocumented things to tweak
11:37:13 <Stskeeps> there's another COBS meeting this afternoon/evening if you remember
11:37:30 <lbt> I forgot but saw a mention earlier
11:37:32 <lbt> when?
11:37:37 <Stskeeps> 16 utc i think
11:38:09 <lbt> ok
11:40:07 <lbt> so: OBS, then IMG/DBC as my main focus. I'll try and continue the git/obs work too.
11:40:15 <Stskeeps> ok
11:40:51 <lbt> let me know if you need to try anything on OBS with sb2
11:40:56 <Sage> is there any high priority package updates btw that should be handled?
11:41:03 <Sage> for next release
11:41:12 <lbt> I'd like to have a way to test experimental OBS branches
11:41:17 <Sage> that are not in review that is
11:42:39 <lbt> libxcb ?
11:42:49 <lbt> https://bugs.merproject.org/show_bug.cgi?id=226
11:43:22 <lbt> udev? #230
11:43:24 <Sage> I though someone was already doing that thus didn't touch
11:44:02 <lbt> *nod* ... just browsing high prio bugs :)
11:44:04 <Sage> I'm a bit confused with udev thingy probably should update the x86 adaptation and recheck after that.
11:44:23 <lbt> seen the udev/systemd news?
11:44:41 <Sage> I noted that udev is in systemd git now but didn't note any news
11:44:44 <lbt> https://lwn.net/Articles/490413
11:44:59 <Stskeeps> oh yeah
11:45:01 <Stskeeps> i forgot to mention that
11:45:26 * lbt waits for dbus to go in there too
11:46:08 <Sage> One can read along the lines that "systemd is the one and only sysinit for the future ;)"
11:46:34 <Stskeeps> one thing for discussion: SMACK support has begun to be entered into meego 1.2
11:46:41 <lbt> I wonder when systemd will incorporate lua
11:46:50 <Stskeeps> and it looks to be of a kind that we might be able to utilize in Mer too
11:49:57 <lbt> I'm not aware of the implications for this kind of decision
11:50:13 <lbt> eg the degree to which this limits vendor choice
11:50:22 <Stskeeps> :nod:
11:50:40 <lbt> vs the degree to which we say "Mer uses SMACK, deal with it" :)
11:50:54 <lbt> (cf systemd)
12:00:11 <lbt> Stskeeps: did I mention an issue on the DBC process to you last week ?
12:00:14 <Stskeeps> hmm?
12:00:20 <Stskeeps> if you did, i forgot
12:03:19 <lbt> the DBC is run in submit order which != accept order and doesn't account for A being submitted, checked w/o B and then B being accepted before A is accepted
12:03:40 <Stskeeps> DBC is being run against "what's merged right now"
12:03:44 <lbt> yes
12:03:55 <Stskeeps> that's the thing about topic branches
12:04:00 <Stskeeps> and yes, it's a risk point
12:04:17 <lbt> yep - that's all ... FYI
12:05:24 <lbt> OK, are we done?
12:05:37 <Stskeeps> think so
12:05:39 <Stskeeps> thank you for coming
12:05:41 <Stskeeps> #endmeeting