[00:18] Laney, did you drop the filtering_out_users patch from accountsservice on purpose? [03:30] * Logan_ pokes siretart [03:30] can we sync libav? [03:32] our only delta can be dropped === wgrant changed the topic of #ubuntu-devel to: Launchpad offline 06:00 - 06:30 UTC | Trusty Final released! | Archive: Open | Devel of Ubuntu (not support or app devel) | build failures -> http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of lucid -> trusty | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots: [04:10] Good morning [04:11] mapreri: well, it's always office hours somewhere on the world :) (I'm in Germany) [04:11] Howdy. [04:11] hey Unit193 [04:21] xnox: oh BTW, yesterday I figured out the trouble with the py3 AP porting; so once the next autopilot lands in utopic, I'll update shotwell, autopilot-gtk, and ubuntu-system-settings-online-accounts to python3 [04:21] xnox: I have the patches already more or less, they just need the autopilot3-sandbox-run [04:30] hey pitti :) [04:30] hey Logan_, how are you? [04:30] doing well, you? [04:31] quite fine, thanks [04:50] apw: I did the same for linux-exynos5 and also restricted the test to armhf (it will always fail on x86); files are again on http://people.canonical.com/~pitti/tmp/linux-autopkgtest/ [06:07] pitti, ping === wgrant changed the topic of #ubuntu-devel to: Trusty Final released! | Archive: Open | Devel of Ubuntu (not support or app devel) | build failures -> http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of lucid -> trusty | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots: [06:47] doko: FYI, you filed a bunch of dh-autoreconf bugs in Debian twice [06:47] I'm merging them as I come across them [07:03] tvoss: sorry, missed your ping; hello [07:03] pitti, hey there :) sorry, forgot what I meant to ask you :) [07:07] tvoss: heh - contentless pings FTL :/ [07:07] pitti, fair [07:30] dholbach: Hi, re bug #1320612: it looks like your german locale slipped into the test suite. Based on the failure log I guess it could be fixed by setting the locale to "en_US" (like the previous test) but is it worth an additional Ubuntu delta? [07:30] bug 1320612 in python-babel (Ubuntu) "Merge python-babel 1.3+dfsg.1-3" [Undecided,Triaged] https://launchpad.net/bugs/1320612 [07:31] geser, hum... wouldn't this affect users of a non-English Debian session as well? [07:32] only when running the test-suite I guess [07:32] of course I would forward this to Debian or upstream [07:33] geser, sure... I meant users of a non-English Debian session who are trying to run a build [07:33] geser, if you want, I can try to build it again locally with LC_ALL=C and see if that works [07:34] but it sounds like a bug to me to expect en_US in the testsuite [07:35] yes, I'll try to fix it and forward it to Debian. Should this block the merge for now? [07:35] geser, no, I'll do another test build now [07:35] thanks [07:38] geser, hum [07:39] geser, 2ubuntu1 builds fine, 3ubuntu1 won't build, even not with LC_ALL=C [07:39] ah of course - "Runs the unit test on build (Closes: #727616)" :) [07:42] dholbach: hmm, why does my pbuilder have no problem in building this package? [07:44] I have no idea [07:46] I'll try to reproduce this myself and fix the locale issue in the test suite [07:50] geser, it seems to be the LANG and LANGUAGE environment variables [08:11] dholbach: it's already fixed upstream: https://github.com/mitsuhiko/babel/commit/2eb475f835375c1b3b7fefcb7d8aad7047d7fda7 [08:24] geser, great! === _salem is now known as salem_ [09:05] is there a way to get PPA builders spit out all the changed symbols instead of stopping at the first library? [09:07] Mirv: i think that got implemented in Debian, not sure if we have that merged yet. [09:08] That wouldn't be a property of PPA builders; it's dh_shlibdeps or thereabouts [09:08] xnox: ok, well good to know [09:09] yeah, not related to PPA [09:09] I'll check dh_shlibdeps' current status [09:09] Worst case you could pass some permissive -c option to dpkg-gensymbols and then apply logic to the output [09:09] It's in dh_makeshlibs [09:10] Er yeah that [09:10] Looks like 9.20140227 [09:10] Which we should have in trusty and utopic, going by the versions [09:11] Laney: only utopic, no? trusty is so 2013 [09:11] oh yes [09:11] ha, I just saw the 0227 at the end [09:11] s/0// [09:14] hmm, so it should actually already work but doesn't for me [09:14] unsurprisingly it was the pkg-kde's Qt developer asking for the feature === salem_ is now known as _salem === _salem is now known as salem_ === salem_ is now known as _salem [10:52] lool: looks like that worked for fakeroot, btw [10:52] err, echan, but whatever [11:25] xnox: Could you merge php5? php-geoip is waiting for it, and probably anything else that uses dh-php5. [11:26] ack. === MacSlow is now known as MacSlow|lunch [11:27] ta === jono is now known as Guest41797 === vrruiz_ is now known as rvr === _salem is now known as salem_ [12:18] Elv1313: yo, still need me to look at sflphone? [12:27] Elv1313: I've uploaded to trusty and utopic, please add a test case to clearly describe how to test for the fix === sil2100_ is now known as sil2100 === MacSlow|lunch is now known as MacSlow [13:25] cjwatson: ping, can we get a pkcon local_uninstall? I'm writing a debug helper script for the SDK that is going to take a click package, install it , run a hook from it, and uninstall it to leave a clean device [13:26] cjwatson: but i have no root priviledges at that point [13:26] zbenjamin: https://lists.launchpad.net/ubuntu-appstore-developers/msg00553.html [13:27] zbenjamin: "pkcon remove" [13:27] cjwatson: awesomeeee thanke :) [13:27] Just note the bit at the end of that mail about the particular form of the argument you need to pass to pkcon remove [13:29] cjwatson: ok so thats packagename;packageversion; ? whats the all? the target arch? [13:30] yes [13:31] cjwatson: so if i have a fat package, how would that look like? only armhf for a armhf device? [13:31] IIRC it's not checked so you can just always shove "all" in that slot, I think [13:31] ah good i'll try that [13:32] I think for a fat package it would strictly be "multi", but whatever [13:33] yeah i can put that in once we are there [13:34] strictly it would match the filename, so com.ubuntu.test_1.3_all.click => all [13:35] (well, even more strictly it would match what "dpkg-deb -f foo.click Architecture" says) [13:43] vila, ev: writing tests for the debci worker is fun :) (creating 100 requests, 30 workers, killing two thirds of the workers, and making sure that all requests still arrive) [13:43] \o/ [13:44] pitti: next step: test that chaos in production [13:44] http://anonscm.debian.org/gitweb/?p=users/mpitt/debci.git;a=commitdiff;h=3c03eec [13:44] who can i beg/bribe to accept some SRUs into their respective -proposed pockets? [13:44] whoa, ruby? [13:44] ev: well, not quite the next step (still need to implement the swift up/download and the health monitor), but getting closer [13:45] heh, indeed [13:45] ev: Antonio wrote some parts of it in ruby (most things are in sh), he's the ruby maintainer [13:45] ah [13:45] ev: not that happy, but as long as it's mostly test helpers and RSS generators I can live with it [13:45] that's a shame :-P [13:47] ev: but the implementation of the worker is absurdly simple, I really like that AMQP stuff (http://anonscm.debian.org/gitweb/?p=users/mpitt/debci.git;a=blob;f=bin/debci-worker) === pete-woods is now known as pete-woods-lunch [13:48] the actual core logic is like 3 lines, the rest is just prettyness [13:48] nice [13:53] Riddell: Hi, thanks. what kind of test do you want? It is an integration problem (in Akonadi), so it is not really possible to create an unit test without re-creating the contact topology that cause the freeze [13:53] Maybe I can ask the 2 users that had it to test the new package? [13:53] * Laney wows at the number of tests python-defaults has triggered [13:54] Laney, did you drop the filtering_out_users patch from accountsservice on purpose? [13:57] mterry: it's upstream now, no? [14:00] Elv1313: a manual test is needed "install sflphone and akonadi, start sflphone, check e-mail, notice freeze, repeat with new version, freeze will not occur" that sort of thing [14:00] ok [14:01] cjwatson: works perfectly [14:02] pitti: nice [14:02] Laney, parts of it, but not the login.defs support [14:03] pitti: I'm not sure how you assert how many requests are processed when many are killed though [14:04] vila: if I submit N requests I should get N results, as long as at least one worker is still running [14:04] pitti: oh, that's the one or two [14:04] vila: i. e. if a worker dies in the middle of a test, the AMQP request should go back to the queue as it won't get ack'ed [14:05] pitti: so worse case you have duplicates ? [14:05] vila: no, the 1 or 2 is how many result directories you get (as the dir gets created when the test starts) [14:06] vila: right, worst case is you get two results, when it crashes after it wrote the "exitcode" file but before ack'ing the request [14:06] vila: I can probably refine that a bit to remove an incomplete result directory (when there's no exit code), still on my list [14:06] Riddell: done [14:07] zbenjamin: great [14:08] pitti: ack [14:09] ev: that confirms that we either need to stop using prefetch_count=1 or have the worker commit suicide to fix bug #1320205 [14:09] bug 1320205 in Ubuntu CI Engine "nacking a rabbit message is not enough to have another consumer get it" [High,Triaged] https://launchpad.net/bugs/1320205 [14:09] Elv1313: lovely, now we just need a friendly archive admin to review and accept the trusty SRU, try batting your eyelids at ScottK [14:09] thanks [14:09] mterry: that wasn't part of the patch, this is new upstream behaviour [14:09] is it causing a problem? [14:10] Laney, no we used to support login.defs in our patch, but upstream doesn't. So dropping the patch means dropping support for login.defs [14:11] Laney, (to define a MIN_UID for considering which users are human) [14:11] or UID_MIN, can't remember which way :) [14:12] sec, just got a phone call [14:12] cjwatson, how is click chroot create supposed to be run? [14:13] cjwatson, click chroot create -a i386 gives option errors [14:13] cjwatson, not particularly helpful ones [14:14] click chroot -f ubuntu-sdk-14.04 -a i386 create [14:14] order matters (perhaps unhelpfully) [14:14] that is, the subcommand name needs to come after the options to the chroot command [14:15] (oh and the whole thing needs to run under sudo) [14:17] mterry: okay, the patch added login.defs and went upstream [14:17] mterry: we had that in the previous AS in ubuntu [14:17] and now in this release they've dropped that support [14:17] argh, python multiprocessing is basically the worst [14:18] Laney, ah interesting... do they say why? [14:18] so yes indeed the minimum UID is specified in configure [14:18] but I think it should be okay [14:18] agonisingly confusing errors [14:18] mterry: fragile and broken or wording to that effect [14:18] if necessary we could bring it back as a configure option I guess [14:18] Laney, why should it be OK? The touch images use users at 1000 and 1001 and we were relying on that support to not show them in the greeter [14:19] what were we doing? [14:19] wait [14:19] that has a different login.defs? === ValicekB_ is now known as ValicekB [14:21] Laney, on Ubuntu Touch images, we set UID_MIN in login.defs to 1002 [14:21] Laney, there are radio and system users at 1000 and 1001 used by android [14:21] SRUs are such a pain to do :( [14:22] bah [14:22] mterry: okay I guess we can revert this then :/ [14:22] do you want to upload that? [14:22] ba13b59c [14:23] Laney, sure I can do that [14:23] I'll see about making it configurable [14:24] Also a UID_MAX of 60000 is funny, especially if it's compiled into the binary and not overridable. === salem_ is now known as _salem [14:25] the new behaviour doesn't consider a maximum [14:27] Yay. === _salem is now known as salem_ [14:28] and under the old one login.defs was able to override it [14:28] ho hum === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [14:46] vila: can you add that to the bug report? That pitti has confirmed this [14:51] sorry, what did I confirm? [14:52] pitti: that killing a worker that is processing a message puts that message back onto the queue so another worker picks it up [14:52] vila: yes, as long as it doesn't get ack'ed yet; isn't that the whole point of it? [14:53] pitti: yes, but we thought that *nacking* a message will have the same effect but it doesn't [14:53] pitti: the worker has to die [14:53] vila: ah; it sounds like it should, but I haven't played around with that yet [14:57] pitti: all your workers are on the same host right ? [14:57] vila: for the test suite, yes (not in production, of course) [14:57] pitti: yup, that was the question indeed [14:58] ah, I figured out how to run rabbitmq-server in a test suite as my user [14:58] vila: oh, does that matter? [14:58] pitti: not sure, but I can imagine that the behavior may be different when different hosts are involved [14:59] vila: I hope not.. [15:00] pitti: I mean, at least when thinking about failures, a host dieing with a single worker is not the same as a host dieing with multiple workers :) [15:00] vila: ah, of course; I meant, it's hopefully transparent wrt. queue ack/nack handling [15:01] pitti: yes, me too === pete-woods-lunch is now known as pete-woods === ted is now known as tedg [15:29] slangasek: pitti: what is "fastboot" mode in the boot initscripts?! was that ever supported with mountall? [15:34] xnox: I'm afraid I never heard of it [15:34] xnox: I thought "fast boot" was "upstartification in lucid" [15:35] xnox: I don't know anything about fastboot as regards initscripts [15:35] ack. [15:36] pitti: slangasek: should the re-introduced scripts actually just be lsb-headers, source init-functions and otherwise empty? [15:38] xnox: I don't think it hurts that much to just take the debian ones, to make merges easier; there's no other waste than a few kB of hard disk, is there? [15:39] * pitti needs to run out now, cu tomorrow [15:47] pitti: slangasek: pushed updates to sysvinit merge proposal. And responded to a bunch of comments. === roadmr is now known as roadmr_afk [16:33] xnox: I much prefer the re-introduced scripts be as close as possible to the Debian ones, for future merging [16:35] slangasek: ok. === salem_ is now known as _salem === sz0 is now known as sz0` === sz0` is now known as sz0 === roadmr_afk is now known as roadmr [17:50] is launchpad being sad, or is it me? (code hosting / branch scanning?!) [18:05] it was a bit slow about an hour ago, took about 10 minutes to produce a diff [18:06] mvo: bug 1320683 is surely because of phased updates [18:06] bug 1320683 in update-manager (Ubuntu) "Update manager does not show all available updates" [Undecided,Confirmed] https://launchpad.net/bugs/1320683 [18:07] mvo: you can see the status of packages phasing at http://people.canonical.com/~ubuntu-archive/phased-updates.html [18:09] bdmurray: ah, thanks. I'd forgotten about that entirely. [18:10] sarnold: no problem === _salem is now known as salem_ === sz0 is now known as sz0` === salem_ is now known as _salem === sz0` is now known as sz0 [19:05] Riddell: you around by chance? [19:05] or xnox [19:06] lamont: hi! === sz0 is now known as sz0` [19:07] I have sflphone questions, and your name is on it === sz0` is now known as sz0 [19:08] lamont: only marginally =) go on =) [19:11] it believes in binding to the IP address of the default gateway, rather than the IP address associated with the route to the server... I wish to fix that, and it's a rather large codebase... hoping for a pointer [19:11] s/of/associated with/ [19:12] xnox: ^^ [19:12] or I an pester msp later [19:14] lamont: wow. definitely beyond what i've done with sflphone. [19:17] xnox: no worries === sz0 is now known as sz0` [19:32] pitti, do you know much about how the volume up/down events are handled in Ubuntu Touch? === cyphermox_ is now known as cyphermox [19:45] xnox: and figured out how to do it, though there is a bug === sz0` is now known as sz0 [19:58] xnox: gnome/src/config/accountconfigdialog.c line 1048... help me understand "CONFIG_LOCAL_INTERFACE" there? [19:59] which is more of a dbus question thananything else [20:03] lamont: that does look wrong. [20:04] xnox: can you approve things to go into -proposed for SRUs? [20:05] dobey: no. you want ~ubuntu-sru [20:05] (loop up their team on launchpad to see who is in it) [20:06] lamont: what is account_lookup? [20:06] * xnox presume calling into gnome account settings api. [20:06] infinity: can you accept some things to -proposed for me? [20:06] lamont: and i guess you want to make sure CONFIG_PUBLISHED_SAMEAS_LOCAL is false. [20:07] stgraber, slangasek: ^^ or one of you can accept stuff to -proposed for these SRUs for me? [20:08] dobey: it's typically useful to mention which package you are after, and for which of the stable series.... [20:09] there 68 unapproved requests at the moment [20:09] xnox: yeah it sure looks like it wants a bug [20:09] ubuntuone-client and ubuntuone-storage-protocol for saucy and precise [20:09] anyway, I'm off till Monday and on a pretty sucky internet so I won't do any queue review until I'm back some place with better connectivity (that may very well be on Monday in Malta) [20:10] stgraber: ok, thanks. didn't know you were away. enjoy [20:11] lamont: so the full line is all withing sflphone configuration of the account (and it's properties). [20:11] ok [20:11] lamont: that's not really stock dbus api or anything, just a bunch of functions that sfphone wraps. [20:12] dobey: I'll have a look. [20:12] infinity: thanks [20:12] xnox: what I want is for it to pick the IP address of the interface I override, and not the default-gateway interface [20:14] lamont: right, that dbus call is defined in ./daemon/src/client/dbus/configurationmanager-introspec.xml [20:14] lamont: which takes you to getAddrFromInterfaceName C++ method [20:15] which is defined as [20:15] return SipTransport::getInterfaceAddrFromName(interface); [20:16] line 107 in ./daemon/src/sip/siptransport.cpp [20:16] lamont: i think you'll need to fix that function ^ [20:17] cause that's where it computes the value it uses, in a cryptic way... [20:17] lamont: does that help at all?! [20:21] dobey: Why the en_US.UTF-8 requirement for tests on saucy? [20:24] infinity: becasue some of the cert files have non-ascii names, and the default locale is that weird ANSI thing that only does ascii, which breaks python encode/decode stuff [20:25] dobey: Ahh. C.UTF-8 would probably do the trick without making you need a langpack build-dep. [20:25] dobey: But I this works too. *shrug* [20:25] dobey: Was just curious. [20:26] maybe. we were already doing the en_US.UTF-8 bit elsewhere though, so i just copied that [20:26] TÜBİTAK_UEKAE_Kök_Sertifika_Hizmet_Sağlayıcısı_-_Sürüm_3.pem [20:26] Wow. [20:26] yeah [20:26] Kay, explanation accepted. :) [20:27] dobey: The only followup, then, is why not the same patch for precise, in case ca-certs gets updated there with similar consequences? [20:27] i thought i did do the en_US thing on precise too [20:27] xnox: that may help. C++ may not. [20:27] Not in this upload. If you did, it was a previous one. [20:28] * infinity looks. [20:28] lamont: that function, is pure C by the looks of it thought. [20:28] huh. well, i did it like a month ago, so i don't remember exactly [20:28] if i didn't do it on precise, it was because the tests didn't fail when i did the test build, most likely [20:29] Oh. precise might not run a testsuite at all. [20:29] oh or that [20:29] It doesn't have the override with ./tun-tests [20:29] precise is quite old, yeah [20:29] So, yeah. [20:29] Alrighty. Acceptiferating. [20:29] thanks [20:30] dobey: Also, this isn't carte blanche to be annoying, but when you have deadlines like the U1 shutdown in 10 days, bugging people earlier wouldn't be a bad thing. [20:30] infinity: i have been bugging :( [20:30] dobey: Oh. You should have bugged me. And maybe included the "hey, there's a deadline" bit. Oh well. [20:31] dobey: As it stands now, I think I'm going to ask you guys to verify the crap out of those u1-client uploads as best you can, so we can phase them to 100% when we release, and hope they get out to most people ASAP. [20:31] xnox: wondering if maybe that's getting called with DEFAULT_INTERFACE [20:32] lamont: hm. not sure. if you file a bug CC / subscribe me on it. for me to have a look, when i'm bored on a connection flight. [20:33] lamont: cause i'm deep in async upstart code at the moment and hope to get this done before EOW. [20:35] dobey: And poke me personally when it's verified, I think the client one is worthy of an early release if you've tested it well. [20:35] xnox: sure [20:36] lamont: the argument that should be called, should be whatever was stored in the account key, no?! [20:36] infinity: sure [20:36] dobey: Oh, except https://bugs.launchpad.net/ubuntu/+source/ubuntuone-client/+bug/1306225/comments/3 isn't very promising... [20:37] Launchpad bug 1306225 in Ubuntu One Client "The user should be informed about the u1 file service shutdown" [Critical,In progress] [20:37] infinity: thanks much [20:37] dobey: Is there no way to get that working at runtime instead of shutdown? [20:38] infinity: yeah, though at this point i'm thinking it's worth it to just release the update anyway, given how close june 1 is. the "not connecting" bit is going to be more important by the time we could get a new upload to fix the notification on precise (i have no idea why the notification is behaving that way on precise. it might be a much larger problem) [20:38] trying to test on saucy right now. just waiting for the archive to get updated so apt will work [20:38] ioctl(13, SIOCGIFADDR, {ifr_name="tun0", ifr_addr={AF_INET, inet_addr("10.172.126.2")}}) = 0 [20:38] hrmpf === pbn_ is now known as pbn [20:39] lamont: cheat =) debugging c++ with strace =))))) [20:39] xnox: I was doing thatbefore I looked at the source [20:39] because that's how I roll [20:39] =))))) [20:39] s/roll/troll/ ?! [20:39] lamont: Elv1313 is the sflphone man, try tracking him down if you have questions [20:40] mvo: can you confirm there's nothing to do with apt in bug 1016643? [20:40] bug 1016643 in apt (Ubuntu Quantal) "add-apt-repository downloads gpg key in an insecure fashion" [High,Triaged] https://launchpad.net/bugs/1016643 [20:40] (/me is closing ancient bugs...) [20:43] xnox: and now I can't getit to fail [20:44] Riddell: thanks. here or deban? [20:44] debian [20:47] lamont: upstream [20:50] amusingly, I can't actually reproduce the bug [20:50] things do, in fact, seem to work as expected. [20:51] turns out that setting your clock 2 weeks forward really screws up some things [20:52] bah [21:08] what a bother :( [21:14] infinity: https://bugs.launchpad.net/ubuntuone-client/+bug/1306225/comments/5 doesn't look good :( [21:14] Launchpad bug 1306225 in Ubuntu One Client "The user should be informed about the u1 file service shutdown" [Critical,In progress] [21:15] dobey: Is that something you can turn around a quick(ish) fix for? I'm happy to be responsive and process that one quickly if you can. [21:16] infinity: well, i haven't been able to do so in the ~15 minutes i've been poking at it (trying to get it to exit cleanly anyway). maybe i can if i just try to make it not ever connect, and leave the process running and acting normally like a disconnected client [21:41] infinity: so i managed to make a patch that makes it not connect, but keep the process alive and working otherwise normally, for 13.10, but 12.04 code is a bit different so i'll have to make a different patch for it [21:42] dobey: Alright. Keep me in the loop. I'm around most of the evening if you're working late, or tomorrow if not. [21:43] (Need to run out soon to visit a post office and remind myself how to send paper through the air, though. Weird backward-ass lawyer people) [21:44] infinity: yeah. i need to go now. i'll probably need to finish this up in the morning, so ping me when you hop on-line tomorrow === michagogo_ is now known as michagogo === ubott2 is now known as ubottu === sraue_ is now known as sraue === mnepton is now known as mneptok === Trevinho_ is now known as Trevinho === dkessel_ is now known as dkessel === Sarvatt_ is now known as Sarvatt === xymox_ is now known as xymox === TheMuso` is now known as TheMuso