=== robbiew is now known as robbiew_ === nigelb_ is now known as nigelb === chrisccoulson_ is now known as chrisccoulson === doko_ is now known as doko [16:00] * slangasek waves [16:00] o/ [16:00] howdy [16:00] hello [16:00] * stgraber waves [16:00] hey [16:01] hi [16:01] hi [16:01] #startmeeting [16:01] Meeting started at 10:01. The chair is slangasek. [16:01] Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] [16:01] o/ [16:02] [TOPIC] lightning round [16:02] New Topic: lightning round [16:02] $ echo $(shuf -e cjwatson barry doko csurbhi stgraber jhunt mvo ev vorlon bdmurray) [16:02] mvo bdmurray stgraber csurbhi cjwatson ev doko jhunt vorlon barry [16:02] mvo: tag [16:02] What I did last week (shorter week, half a day sick, friday on vacation): [16:02] Spec writing/workitem creation work [16:02] SRU for aptdaemon (#781874) and new oneiric version [16:02] Apt: Apt-ddtp update oneiric, review/merge donkults branch, work on lp:~mvo/+junk/apt-ordering-failure for SoC apt order [16:02] Apturl: work on a fix for overlong urls [16:02] Expenses uds budapest, travel preparing spring dublin [16:02] Review/merge lp:~hodgestar/software-properties/configurable-key-server [16:02] Software-center: add fallback to previous distro series if there are no reviews for the current one yet, upload new version, fix tests for oneiric, debug/fix broken apthistory.p caches, merged lp:~evfool/software-center/smallfixes, work on lp:~mvo/software-center/qml, review/merge lp:~evfool/software-center/qfix774201, review/merge lp:~gary-lasker/software-center/refactor, upload new 4.0.2 SRU with fixes in the reviews loader, work on lp:~mv [16:02] o/software-center/pyflakes branch, work on the lp:~mvo/software-center/refactor [16:02] Update-manager: debug/fix/upload UnitySupport integration problem, review/merge lp:~brendan-donegan/update-manager/bug699660-fix-settings-shortcut, update meta-release (now that new u-m is in -updates) [16:02] (this is for the last week from mon-fri, gtimelog seems to give me [16:02] this weekview whatever day I pick from the week, how do others get wed->wed only?) [16:02] (done) [16:02] I guess this comes down to "lots of software-center work" :) [16:03] :) [16:04] bdmurray: [16:04] created foundations-bugs team and mailing list, subscribed them to packages we care about [16:04] modification of ftbfs bugs for doko [16:04] mvo: i distill from the timelog.txt file [16:04] modification of bug squad standard replies, bug squad signature greasemonkey script [16:04] done [16:04] barry: so you don't use the "create weekly report" menu ? I should just use the file too I thnk :) [16:04] bdmurray: modification of ftbfs bugs is scripting something against bugs currently in LP? [16:05] mvo: exactly! super easy [16:05] slangasek: yes doko asked me to retag, target and milestone them [16:05] barry: thanks! [16:05] ah cool [16:05] stgraber: [16:05] I've spent a good part of Thursday and Friday working with upstream x2go. [16:05] I'm happy to do similar things for other people [16:05] Monday was a public holiday here. [16:05] Uploaded nxcomp and nxproxy merging changes from x2go and rebasing on a recent upstream. [16:05] Reviewed and uploaded python-x2go (in NEW) and started working on porting the software-center code to it. [16:05] These changes and new packages should also land in Debian soon (through pkg-devel-x2go). [16:05] Been doing a few merges too and otherwise worked on my IPv6 test environment and working on the NFS ipv6 MIR (bug 781516). [16:05] Launchpad bug 781516 in rpcbind (Ubuntu Oneiric) "[MIR] libtirpc, rpcbind" [High,New] https://launchpad.net/bugs/781516 [16:06] Started thinking of possible integration of containers to help the App Review Board. [16:06] Updated a few specs so they look good and are ready for review/approval. [16:06] (done) [16:07] stgraber: thanks for that MIR, I've kicked it over to the MIR team now :) [16:07] psurbhi: [16:07] *) Tested 2.6.7.x mdadm fixes for autoboot,partition naming etc again - uploaded a debdiff for the same. [16:07] *) visa documentation, appointment, ticket booking [16:07] *) Assessing the need for mountall interface - when root changes while booting (had initially thought that this was needed) [16:07] *) ext4 deadlock bug - re-worked on the patches that I sent out earlier. [16:07] (done) [16:08] slangasek: ok, I guess it can be approved with the current content but we should wait for a new libtirpc in Debian or drop the build-dep ourself and then sync whenever Debian does the same thing [16:08] stgraber: yep [16:09] cjwatson: [16:09] This week: [16:09] * more work on live-build, now properly specced/wikied (foundations-o-live-build) - x86 mostly sorted, working on ARM requirements now [16:09] * wrote up foundations-o-great-cd-debate, hopefully clearly [16:09] * started initial oneiric CD builds [16:09] * moved transition tracker onto DC infrastructure [16:09] Next week: [16:09] * make sure the installer is working properly for alpha-1 [16:09] psurbhi: and that mdadm diff is ready for sponsorship, right? It's on my todo list [16:09] * if possible, cut over to live-build, at least for x86 [16:09] slangasek, yes [16:10] just that its quiet a big one.. :-/ [16:10] yeah :) [16:11] cjwatson: I was puzzled that foundations-o-great-cd-debate only has one work item [16:11] I thought somebody would say that :) [16:11] surely I should have a work item there to investigate the 703MB CD question further :) [16:11] ah, yes, that would be a good plan [16:11] I didn't want to write a vague work item that said "maybe increase the CD limit" [16:12] and a lot of the items are in reality in other specs (e.g. desktop-o-cdspace) [16:12] * slangasek nods [16:12] I'll lob a work item in for you [16:12] ta [16:12] ev: [16:12] (the size benefit from live-build was a pleasant surprise) [16:12] Done: [16:12] - Investigated Mozilla Input (http://input.mozilla.com and [16:12] http://feedback.mozilla.com) and the old kampyle.com-based feedback system [16:12] used in the Funnel Cake project (recording why people were quitting the [16:12] Firefox install for a day). Set up a local Django instance of Mozilla Input [16:12] and started poking around the code. [16:12] - Taught Wubi to boot straight into the second stage when in Windows Vista or [16:12] later. Working out an approach for Windows XP. [16:12] - Implemented code to present the list of keyboard variants in pertinent to the [16:12] selected language in ubiquity in the panel keyboard indicator, using [16:12] python-xklavier, which I'll have to MIR and seed. [16:13] oh? there was a size savings with live-build? [16:13] - Started trying to untangle kvm-autotest for automated Wubi testing. [16:13] - Had a quick play with kexec, enough to confirm that it works. Still need to [16:13] integrate into ubiquity. [16:13] - Lots of discussion around the crash database. [16:13] - Tried to address a tech board member's concerns around the community not [16:13] having the full data available for installation success/failure measurement. [16:13] - Created an Ubuntu Metrics team to keep the momentum from mpt's UDS plenary. [16:13] Sent out a kickoff email to get things started. [16:13] - Had a call with Elliot about his team massively helping plan and resource [16:13] bigger items from the Ubuntu Metrics work. [16:13] - Wrote a brief for the Web Team for work on the Spread Ubuntu project. Sadly [16:13] looks like it might be deferred again :-/. [16:13] TODO: [16:13] - Catch up on the discussion around the crash database and investigate what [16:13] Mozilla is doing with Socorro, per Elliot's advice. [16:13] - Respond to the Technical Board's concerns around the installation [16:13] success/failure measurement. [16:13] - Finish drafting results from "Whats wrong with UDS?" session and engage with [16:13] Jono to come up with firmer plans. [16:13] - Mock up a network-manager 0.9 D-Bus API for unit testing and to use while I [16:13] wait for the real thing to be packaged. Create the wireless page in the [16:13] installer. [16:13] - Investigate the libcheese pygi bindings and see how much work is required [16:13] for the picture page. [16:13] (done) apologies for the poor formatting on that. [16:14] slangasek: yeah, about 5 MiB - probably a bit of a memory tradeoff, I think it mostly comes from doing a better job at cleaning up *-old files in /var/lib/dpkg/ and that kind of thing [16:14] (since that's about the only file-list difference that's left) [16:15] ev: I improved the second-stage boot menu a bit, BTW, by bringing lupin back into sync with GRUB [16:15] huh, neat [16:15] cjwatson: yay, thanks [16:15] cjwatson: is it to the point where we don't show the windows option in grub when under wubi? [16:15] ev: is the TB concern that raw data gathered about successful installs, etc. should be available to the community also? [16:15] ev: woah, impressive list! [16:15] ev: one thing that occurred to me in the process though is that there's no way to do certain things in the Wubi environment, because we don't get to write to the GRUB environment block [16:15] mvo: mostly noise, really [16:15] :) [16:16] ev: no, I haven't changed that - that's quite awkward to do right now [16:16] slangasek: that was this particular member's concern, yes [16:16] which is a hard balancing act against "we don't do headcounts" [16:16] ev: do I have a work item for that? [16:16] clever objection though... if Canonical can be trusted with it, why shouldn't the community :) [16:16] cjwatson: no, do you want one? It was something I noticed when doing a Wubi install and I find it confusing [16:17] ev: mm, it's technically accurate and occasionally helpful but I can see how it might be confusing [16:17] slangasek: it was phrased in terms of "If I'm a user, I want to be able to look at a database and roughly see my anonymous information in there, knowing the time, etc" [16:17] ev: if it bothers you and you want rid of it, give me a work item, yes :) [16:17] cjwatson: will do :) [16:17] (it means showing two bootloader menus, ick) [16:18] doko: [16:18] ev: (taking to /msg) [16:18] (actually #u-i) [16:18] - filed MIR's, reviewed and promoted packages for component mismatches [16:18] - built OpenJDK b22 for armel, started jtreg tests on armel [16:18] - GCC and binutils updates, GCC bug triage [16:18] - Linaro toolchain WG call [16:20] anything interesting coming out of the WG call? [16:20] no, business as usual, addressed the powerpc build failure, but nothing else [16:21] * slangasek nods [16:21] jhunt: [16:21] Post-holiday inbox boxin'. Added workitems to all blueprints bar [16:21] foundations-o-upstart-for-admins. Provisional planning for Upstart [16:21] "helpers" (abstract jobs/events). Merging fun (ongoing). Investigating [16:21] nih+upstart build failures (ongoing) - looks like a toolchain issue on [16:21] armel. [16:21] EOF [16:22] * mvo read "post-holiday inbox bombing" [16:22] heh [16:22] sounds like a holiday in itself [16:22] love it [16:22] heh [16:23] jhunt: eta on work items for foundations-o-upstart-for-admins? [16:23] (slangasek: https://lists.ubuntu.com/archives/technical-board/2011-May/000873.html - he's made his concerns public now) [16:23] doko: thanks for the c-m work, it's been greatly appreciated [16:23] slangasek: should be done by cob today. [16:23] ok, great :) [16:23] what's the deadline for the specs btw? [16:24] last Friday ;) [16:24] ah... [16:24] slangasek: speaking of which, I'm approver of several undrafted specs where you're the drafter ;-) [16:24] * slangasek ducks and runs [16:24] yep, will have those done today [16:25] * cjwatson cracks the whip [16:26] my turn [16:26] * blueprint review/approval [16:26] * followed up on wrong memory requirements information for natty [16:26] * documenting our UDS output in the proceedings: https://wiki.ubuntu.com/UDSProceedings/O [16:26] * looking into the possibility of contracting a Windows programmer to help with Wubi [16:26] * trying to figure out a schedule for an upstart sprint this cycle [16:26] ... you guys need to plan your vacations together so I have something to work with ;P [16:27] EOF [16:27] oh, I intend on waiting until the last minute and taking 26 straight days off :-P [16:27] the "bye, see you next year" option is always fun [16:27] heh [16:27] barry: [16:27] bug 787822 and debian bug 625784 (virtualenv -p python3 broken), did my first upload to fix this in oneiric!; bug 784662 (winpdb; fix in ppa, maybe backports?); started looking at python3-defaults to drop py3.1; blueprinting; ubuntu packaging guide (copy last bits of wiki to rest and fulfill jorge's mandate by deleting many wiki pages); blog about python plans for oneiric; upgraded dev vm to oneiric; released python 2.6.7rc2; done. [16:28] Launchpad bug 787822 in python-virtualenv (Ubuntu) "'virtualenv -p python3 /tmp/xx' should work but doesn't" [Medium,Fix released] https://launchpad.net/bugs/787822 [16:28] Debian bug 625784 in python-virtualenv "python-virtualenv: 'virtualenv -p python3 xx' should work but doesn't" [Important,Open] http://bugs.debian.org/625784 [16:28] Launchpad bug 784662 in winpdb (Ubuntu) "winpdb doesn't work with python2.7" [Undecided,Won't fix] https://launchpad.net/bugs/784662 [16:28] slangasek: cool, that proceedings page makes us look good. :) [16:28] being the only team to ACTUALLY fill it out does that :) [16:29] :-D [16:29] slangasek, ev: the ultimate - http://dilbert.com/strips/comic/1991-10-31/ [16:30] ooo, I have to try that [16:30] heh [16:30] [16:30] barry: congrats on the upload! :) [16:30] slangasek: thanks! [16:30] barry: was it automatix? [16:31] ev: automatix? [16:31] so with the ubuntu packaging guide moved out of the wiki, are there stubs left somewhere that direct to the new guide? [16:31] ev: harsh [16:31] * barry didn't sleep a wink last night [16:31] barry: automatix, because every package management system needs a four-speed blender [16:31] barry: http://mjg59.livejournal.com/77440.html [16:32] killall -9 dpkg! yay [16:32] [TOPIC] AOB [16:32] New Topic: AOB [16:32] yes [16:32] https://lists.ubuntu.com/archives/technical-board/2011-May/000873.html [16:32] [LINK] https://lists.ubuntu.com/archives/technical-board/2011-May/000873.html [16:32] LINK received: https://lists.ubuntu.com/archives/technical-board/2011-May/000873.html [16:33] I'm interested to know if anyone has thoughts on how I can remove myself from between the rock and hard place that is giving developers access to the raw data and not giving people data to infer head counts with [16:33] people always infer head counts anyway, don't they? [16:33] or how I can play the tech board against itself for massive win [16:33] heh [16:34] ev: perhaps assign people a random uid and offer to email that to them so they can track their own data but no one elses? [16:34] oh wait [16:34] Scott has an interesting point about the data being genuinely useful to third parties as well [16:34] I think I'd prefer to discuss this in a TB meeting though ... [16:35] infer head counts> one option would be expiring the data [16:36] right - we can share whatever data we have, but just not keep interesting data [16:36] if what we want is success rate over time, only keep that [16:36] sorry, not to imply that's not interesting data :) [16:37] expire the real data, but keep the percentages forever? [16:37] yeah [16:37] that works for me [16:37] but sure, we can discuss this in the TB meeting [16:37] so have the server store the data for the day, at the end of the day generate the stats and remove the data? [16:38] or something like that - possible compromise anyway [16:38] yep [16:38] you can't aggregate those numbers over time because you don't have the scale [16:38] so it does limit the utility to some degree [16:38] food for thought, at least : [16:38] ) [16:38] anything else? [16:39] yeah, and we'll still have "interesting data" like how many installs do we get on release day... ;) [16:39] I created a metrics team [16:39] people should join it and participate [16:39] ++ [16:39] if they're interested in this sort of stuff, of course [16:39] http://launchpad.net/~ubuntu-metrics [16:39] LINK received: http://launchpad.net/~ubuntu-metrics [16:39] * slangasek creates ~ubuntu-imperials in response [16:40] ;-) [16:40] lol [16:42] #endmeeting [16:42] Meeting finished at 10:42. [16:42] thanks, all :) [16:42] thanks! [16:43] * jhunt is on holiday - please leave a message after the prompt [16:43] Oops - sorry! :) [16:43] bye [16:43] jhunt is coughing... :D [16:43] in the good news section (that we did not have): I played with qml and its really really nice for UI work [16:43] thanks [16:43] thanks [16:43] mvo: nice [16:43] thanks [16:43] mvo: oh cool, it's on my todolist of stuff to look at (since UDS-O) ;) will try to spend some time on it this weekend [16:44] qml> ick [16:44] "people are moving to the web, and they're using javascript as part of that. I know, we'll stop the tide by putting javascript on the desktop! This cannot fail!" [16:44] barry, stgraber: llp:~mvo/software-center/qml for the curious [16:44] ev: gnome-shell! [16:45] yep, equally awful [16:45] Also, Qt bad, very bad. [16:45] ev: out of curiosity, what do you dislike about the approach? just the layout stuff in qml is vastly more powerful than gtk [16:45] ev: ecmapython [16:46] I am very weary of us moving to Qt in general for a number of reasons [16:46] not least of which, it doesn't have an owner [16:46] but more importantly, it's just reliving the battle of the 1990s [16:46] we don't need another WIMP model to compete with Windows 95 [16:47] we need something new and revolutionary, that, if we're not going to develop upstream software ourselves, leverages the hard work being done in our open source industry [16:47] so, webkit + other things [16:47] mvo: cool, I'll have a look. Off for lunch now, see you. [16:47] if I have to pick between Google creating an open source layout engine and a company that laid off most of its employees on Qt, I'm going to go with Google. [16:48] (equally, I don't think this is the kind of thing we can hope for some PhD student to knock out as part of their thesis) [16:48] javascript on the desktop: I think I confirmed at UDS that I wasn't the only one who was reminded of Active Desktop [16:48] right, fair enough. the future is very uncertain for qt, that is unfortunately true [16:50] not that i have a horse in the race, but qt is free software and has a big fan base right? it's not going anywhere, though it may not progress as quickly. heck, even tcl is still around :) [16:50] I just want us to be forward thinking. We're not going to compete with the big boys by playing with yesterday's technology, or even today's. [16:51] that i agree with [16:52] right, and yet we need to have solution today that are compelling [16:59] sure, but this seems to be positioned as the future [16:59] and too many transitions will kill our developers and users [16:59] so I think it's wiser to invest heavily in developing future technology and smoothing out the rough bits of what we already have [17:00] that is something I very much agree with, we need to keep the transitions as low as possible [17:04] /topic [17:05] QA meeting? [17:06] hggdh ^ [17:07] darn it! [17:07] OK [17:07] was over about 25 minutes ago [17:08] hggdh, I can help gather the troops [17:08] patrickmw: isn't it at 1200? [17:08] I see 1700 UTC on the fridge [17:09] QA is next hour about 50 minutes away [17:09] yes [17:10] patrickmw: 'date -u' return 16:09 ere [17:10] grrr Foiled again by google calendar [17:10] here [17:10] patrickmw: welcome! Been there so many times... [17:10] cp1: thank you [17:10] :-) [17:10] pffuiiii, not guilty this time [17:11] hggdh, sorry for the hearth attack :) [17:11] :-) [17:11] my heart seems to be still strong, of sort of ;-) [18:00] almost time.... [18:00] no, it *is* time [18:00] ~o~ [18:00] OK [18:00] #startmeeting [18:00] Meeting started at 12:00. The chair is hggdh. [18:00] Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] [18:00] good morning/afternoon/evening/night to all [18:01] we will go thru the following agenda: [18:01] good * [18:01] Previous Actions (all) [18:01] Community Efforts/Testing [18:01] Automated/Systems Testing [18:01] Engineering Team Bug Status [18:01] Other Topics [18:01] Chair Selection [18:01] [TOPIC] Previous Actions [18:01] New Topic: Previous Actions [18:01] I cannot see any pending actions. Anybody has a better view than mine (known to be flaky)? [18:02] There was an action to move the meeting on #u-meeting at 1700UTC and update the calendar [18:02] here we are! [18:02] done [18:02] :-) [18:02] so [18:02] [TOPIC] Community Efforts/Testing [18:02] New Topic: Community Efforts/Testing [18:03] we are starting a new cycle, and today the first ISOs for the desktop were put available. I understand they are quite unstable, though [18:03] * charlie-tca waves [18:03] ooh yes :-) [18:03] but, generically, help testing the new Oneiric images is extremely welcome [18:04] Anything from the community relating to test? [18:04] o/ [18:04] (BTW -- there are no server ISOs available) [18:04] jibel: please go ahead [18:05] only alternate and desktop on i386 are available. [18:05] so, we need help with sru testing and especially 3 bugs: [18:05] bug 665185 sysvinit (Lucid) High [18:05] Launchpad bug 665185 in sysvinit (Ubuntu Lucid) "/etc/init.d/sendsigs fails to kill some processes" [High,Fix committed] https://launchpad.net/bugs/665185 [18:06] it is targeted to 10.04.3 and needs verification asap [18:06] bug 695290 grub2 (Lucid, Maverick) High [18:06] Launchpad bug 695290 in grub2 (Ubuntu Maverick) "10_lupin case problem with ntfs UUIDs" [High,Fix committed] https://launchpad.net/bugs/695290 [18:06] Requires dual RAID controllers [18:06] if anyone has a dual raid controller please give it a try [18:06] bug 778026 lirc (Natty) Critical [18:06] Launchpad bug 778026 in lirc (Ubuntu Natty) "lirc-modules-source breaks in-kernel lirc support" [Critical,Fix committed] https://launchpad.net/bugs/778026 [18:06] Requires an Infrared Remote Control [18:07] if anyone has an Infrared Remote Control and is running natty please give it a try [18:07] for the record, Lucid 10.04.3 is looming... we need to check these SRU bugs really, really, ASAP [18:07] .. [18:07] OK. Moving on [18:07] [TOPIC] Automated/Systems Testing [18:07] New Topic: Automated/Systems Testing [18:07] Hello all [18:07] We are currently running the individual test suites that get executed for Kernel SRU to collect data: [18:07] What type of tests get run [18:07] What kind of results are generated [18:07] How do they perform in a CI environment [18:08] We have been making steady progress and have been able to run LTP lite to completion. After reviewing the other test suites (DBench, TBench, TioBench, SignalTest) we will need to transform the results to a standard format. Most of these test suites started development before test runners like unittest were widely used. [18:08] Ubiquity testing: [18:08] There is currently a piloted set of test scripts for automating the Ubiquity installer. Those scripts will be developed into a full-fledged framework so we can continue to add tests [18:08] Not much to report until after the Automation Sprint [18:08] Package continuous integration: [18:08] Engineering teams have been asked to submit lists of critical packages to Ubuntu. Those packages are being added to the CI process and will be built, tested and reported upon on a more frequent basis whether it be daily, hourly, or upon code push. [18:08] Questions? [18:08] o/ [18:09] * patrickmw is all ears [18:09] we are currently running kernel SRUs -- what should we change, if anything? [18:10] nothing right now. We are still in discovery mode. As soon as we can get the SRU tests in the hands of Jenkins we will let you know :) [18:10] ok [18:11] anything else? [18:11] not from me [18:11] moving on [18:11] [TOPIC] Engineering Team Bug Status [18:11] New Topic: Engineering Team Bug Status [18:11] pedro_: ? [18:12] (or am I jumping the gun?) [18:12] as you all know the desktop on Oneiric is a 'bit' broken [18:12] and probably more breakage is coming since the transition to Gnome 3 [18:13] so if you encounter a bug that i'm not aware of (commented on) or anybody from the desktop team, please raise it to me so i can assign it to the right person and get it fixed [18:13] (pedro_ can usually be found on the #ubuntu-bugs channel) [18:14] yes on that and at #ubuntu-desktop if i'm not around [18:14] and also! on the desktop page, we're having a bug day for the amazing Compiz [18:14] which is tomorrow https://wiki.ubuntu.com/UbuntuBugDay/20110526 [18:14] * hggdh forgot -desktop, shame, oh shame [18:14] so if you have some time and wanna learn a bit more about it please join us :-) [18:14] * pedro_ hugs hggdh [18:14] * hggdh hugs pedro_ back [18:15] done? [18:15] guess that's all from here [18:15] yeah [18:15] folks, when you are done typing, please enter a line with just two dots -- like [18:15] .. [18:15] (just an example) [18:15] o/ [18:15] charlie-tca: please go ahead [18:16] Are we going to need oneiric to work those compiz bugs, or will testing in natty be good? [18:16]