[06:05] hi [07:54] good morning! [07:57] larsu, o/ [07:58] hi robert_ancell! How are you? [07:58] larsu, doing good [07:58] :) [08:05] Good morning [08:06] morgen pitti! [08:06] hey larsu, wie gehts? [08:06] gut gut, und dir? [08:06] sehr muede, aber okay [08:07] hehe :) [08:07] hello pitti larsu robert_ancell [08:07] hikiko, good morning [08:07] good morning [08:07] hey hikiko [08:08] hey robert_ancell, long time no see! How are you these days? [08:08] pitti, doing good. Was a nice day so took the kids to the pool and working tonight instead [08:08] * pitti tosses a snowball to robert_ancell [08:08] :P [08:09] did you ever get this error: "Failed opening dbus connection: org.freedesktop.DBus.Error.FileNotFound: Failed to connect to socket /sys/fs/cgroup/cmanager/sock: No such file or directory." ? [08:45] morning ganf [08:45] gang [08:47] morning willcooke! [08:52] morning all ! [08:53] morning larsu willcooke :)! [08:53] hi alexarnaud [09:00] lan3y, willcooke is anyone looking at the build failure of the desktop rootfs? [09:00] good morning :) === lan3y is now known as Laney === Laney is now known as Guest80303 [09:04] hi [09:04] might keep this nick [09:04] it's a good nick === Guest80303 is now known as Laney [09:04] nice feng shui [09:04] jibel: https://bugs.launchpad.net/ubuntu/+source/pyotherside/+bug/1534067 [09:04] Launchpad bug 1534067 in pyotherside (Ubuntu) "[MIR] pyotherside" [Undecided,Confirmed] [09:04] so "yes" [09:04] hi Laney! [09:05] hi larsu! [09:06] how are you? [09:06] Laney, thanks [09:07] Laney: good thanks. hacking an index into that local geonames database [09:09] nice [09:26] larsu: can you jhbuild the glib-2.46 branch? [09:26] it breaks for me [09:26] or is there some smart way of doing it that I don't know [09:27] let me try [09:27] how did you tell jhbuild to build that branch? [09:27] m4macros/glib-gettext.m4:39: error: m4_copy: won't overwrite defined macro: glib_DEFUN [09:27] buildone -n [09:27] after switching to it [09:28] are you sure that doesn't switch back? (I think you need `jhbuild make` for that) [09:28] unless it does it behind my back [09:29] anyway, cool command! but it fails in the same way :( [09:30] works fine for me. Did you run on a clean checkout? [09:30] git clean -dxf is your friend [09:30] yeah [09:30] * larsu figured, just making sure [09:33] Laney: ah got it now! [09:34] works on master [09:34] yeah [09:35] and if you cherry-pick 6b577196eed0754d2805fd48caa64f58f9bb8ee4 [09:36] yeah I just stumbled over that as well [09:36] maybe we should backport that? [09:36] but deprecating a marco on a relase branch is awkward [09:37] don't understand why it fixes it really [09:37] is it lazy or something? [09:37] the error is on an unchanged line [09:37] unchanged since 2002 :) [09:38] I meant unchanged by that patch, but unchanged itself for ages too(!) [09:41] another cool thing is if you run the autogen.sh invocation manually then it works [09:41] so I guess jhbuild is setting ACLOCAL_CFLAGS? [09:43] ACLOCAL_PATH=/home/laney/dev/gnome/install/share/aclocal:/usr/share/aclocal [09:43] intriguing [09:44] oh - maybe the macro is defined through an old glib installation in the jhbuild install dir? [09:45] ja [09:46] Laney: right, this works. Probably we had a newer version of glib installed [09:46] pruning the install dir fixes the issue for me [09:46] wonder how I never had this before [09:47] dunno [09:47] good morning Laney [09:47] I didn't becasue I never build older glibs from jhbuild [09:49] but glib_DEFUN should still have been there [09:49] clearly don't understand this properly [09:49] morning [09:49] maybe it's doing some deduplication [09:49] because the files were previously identical [09:49] hey pitti, hey andyrock [09:49] pitti: late night last night? [09:49] just back from the dentist. it was not a good start of the day [09:49] Laney: oh yes -- basketball, and then another 3 hours of p-m sprint [09:50] :( [09:52] Laney: hm indeed [09:55] * Laney tests this theory [09:58] mmm [09:59] so I guess it has some protection against scanning the same file twice [09:59] #pragma once [10:00] well that was a good learning experience [10:04] now I wonder why all keyboard shortcuts broke in terminator [10:06] hmmm, seems I can hardlock/crash my machine by building libreoffice-l10n in a pbuilder ... === lool- is now known as lool [10:30] uh oh [10:30] I have a suspicioooooooooooooooonnnnnnnnnnnnnn [10:33] it points to a person from canada who likes to go climbing [10:56] ruh roh [11:43] davmor2, can you make this happen? https://bugs.launchpad.net/unity/+bug/1536590 [11:43] Launchpad bug 1536590 in unity (Ubuntu) "Default text is inconsistent dash search box" [Medium,Confirmed] [11:43] no hurry [11:55] willcooke: Laney: if you land the unity-control-center merge, it might be a good idea to do the same change on gnome-control-center [11:56] cyphermox, ack, good plan [11:56] just waiting to see if seb128 has a strong feeling one way or the other [11:57] yeah [11:58] I don't mind either way. I don't especially like removing the switch from there, but in my defense it's also very poorly done [11:58] rfkill is all very ugly and painful [11:59] from my POV, there is almost always a hardware switch anyway [11:59] no [11:59] that's a very bad lie ;) [11:59] ?! [12:00] many switches are software-only [12:00] oooh [12:00] interesting [12:00] and a quite a few break in fun ways [12:00] What is wrong with using urfkill other than it is some work to do it? [12:00] * willcooke retracts his last assertion [12:00] ;) [12:00] Laney: not much [12:01] thought so [12:01] yeah nothing other than work [12:01] Laney: it should work just fine, it's just the incertainty of doing the work of testing it some more on the desktop [12:01] we already unbroke quite a few things when it was tested for the phone [12:02] (and I suspect it's pulled in if you were to install desktop-next or something) [12:02] willcooke: mine just always says search computer and online but let me update and see if it still says it [12:03] thanks davmor2 [12:04] cyphermox, Laney - thinking about this a little more, missing airplane mode from an LTS is probably a bad idea [12:05] the problem right now is that the switch lies, since it's not saved across reboots [12:05] I don't really like "it's a bit broken, so let's remove it" instead of "it's a bit broken, we should fix that" to be honest [12:06] but it's not going to be me doing the work so you shouldn't count my opinion very much [12:06] Laney: both options do fix the problem, they just take vastly different steps to do it [12:06] :D [12:07] all of the network panel is evil IMHO :) [12:10] willcooke: confirmed there are issues there [12:12] willcooke: when I open from the dash button it said search computer, after turning on then off the search online, However when I hit the meta key it then showed search computer and online still, then when I open from the dash it then says computer and online again [12:13] BUGGY CRAP! [12:14] willcooke: also xapian seems to be broken on upgrade I assume that is just a deb issue that might not of uploaded so I'll try that again after Lunch [12:14] fixed [12:15] davmor2, bug 1536093 [12:15] bug 1536093 in xapian1.3-core (Ubuntu) "package libxapian-1.3-5 (not installed) failed to install/upgrade: trying to overwrite '/usr/share/xapian-core/stopwords/dutch.list', which is also in package libxapian-1.3-4 1.3.3-0ubuntu2" [High,Fix released] https://launchpad.net/bugs/1536093 [12:16] jibel: ah thanks dude :) [12:17] thanks chaps [12:20] doko borked the fix and had to do it twice [12:20] good times [12:20] * Laney was irritating him about it yesterday [12:37] so, I had GNU make on xenial getting struck when building libreoffice-l10n twice, yet when I abort and restart it gets unstruck .... [12:37] (regression in make 4.1? fun!) [12:39] fun indeed [12:40] OTOH: various [ 8976.073014] "CPU3: Package temperature above threshold, cpu clock throttled (total events = 29018)" and now I even weirder error from make (possibly because of overheating?) [12:43] ETOOMANYPOSSIBLEROOTCAUSES, reducing build to two l10ns and disabling parallelism ... [12:52] when I initialise an array in C/C++ can I be sure that all the elements are set to 0? [12:53] willcooke: yes for global variables, no for stack (i. e. function local) vars [12:53] ah ha [12:53] thanks pitti [12:53] and most certainly not for malloc()/new'ed vars [12:54] willcooke: calloc() is quite convenient (in C) [12:54] both for safe array size allocation and for clearing [12:54] * willcooke reads [12:54] thx [12:57] g_new0! [12:58] * pitti ^5s Laney [12:58] willcooke: right, with more context (C or C++, glib or not, static or dynamically sized, etc.) we can become more specific [12:58] willcooke: if you are in C++(11) and use a std::array (which isnt a bad idea) the elements will be default constructed ... [12:59] Laney: glib: OOP like its 1989! [13:00] Sweet5hark: c'mon, for C is really quite nice OOP [13:00] pitti: dont feed the troll. ;) [13:02] and with attribute ((__cleanup__)) (when wrapped in some nice macros) it feels outright comfy :) [13:05] * Sweet5hark wises up on the fact that his "will be default constructed" statement is only true when providing an initializer. [13:05] It's my lunchtime learn some C session. I'm using Arduino's C++ which I understand is not "normal" but close enough to C [13:05] even though it's C++ [13:06] otherwise, C++ will stick to its principle of most surprise and do what the coder least expects. [13:07] (the latter got a little bit better with C++11 though. A little bit.) [13:07] Sweet5hark, hi, see pm [13:08] ricotz: I uploaded rc2, didnt copy to libreoffice ppa as -l10n ftbfs in interesting ways. [13:09] willcooke: C and C++ are really quite different; you shouldn't think of them as basically one language with a few differences [13:09] Sweet5hark, on local builds? the ppa build succeeded though [13:10] yeah, C++ is 'orrible :-( If you stick to a "safe subset" and modern C++, like with Qt, it's okay [13:10] ha [13:10] Sweet5hark, if you suspect a problem make 4.1 it wouldn't concern the backports [13:11] I really like to push those builds to get something recent [13:11] ricotz: arrrgh, it did see the old build still on my private ppa. [13:12] ricotz: thus local breakers are likely due to overheating only. [13:12] Sweet5hark, you should really give your builds some useful version strings! [13:12] suffixed with ~xenialX [13:12] to avoid those version clashes [13:13] Sweet5hark, hehe, don't blame make if your memory/cpu gets too hot ;) [13:16] ricotz: curiously its fine with the non-l10n stuff. for whatever weird reason, l10n stuff really heats (my) hardware ... [13:16] Sweet5hark, too clarify those https://launchpad.net/~bjoern-michaelsen/+archive/ubuntu/libreoffice-staging/+packages [13:17] ricotz: yes, those are the latest. [13:17] Sweet5hark, I guess I asked you several times, but please to ppa specific versions [13:20] pitti: the hard part is to know what the 'safe subset' of C++ is. And everyone is using a different 'safe subset' of C++. [13:22] pitti: I was quite enthusiastic about C++11 cleaning things up ... and then I read effective modern C++ and learn about the old pitfalls that remained and the new ones that were added. [13:24] ... still C++11 is an improvement by leaps and bounds over C++03, but it still has lots of weird corners. [13:25] * Sweet5hark makes note to self to really do a pet project in rust to see if it is as good as he hopes. [13:25] there is C++14 btw... [13:27] anpok: two quotes from https://bugs.documentfoundation.org/quips.cgi?action=show on that: "mst_: thorsten: the problem with C++14 is that it's still C++" and ... [13:27] anpok: "Well, except that: std::vector(3, 0) is doing something different from: std::vector{3, 0} just to make things more interesting. But hey, that's C++ for you." === bigon_ is now known as bigon [13:28] yeah different code means different things.. just get over it. [13:34] anpok: the whole '{}' initializer thing was introduced to clear things up, but it didnt really in the end. For 99% of the cases ctor(2,0) and ctor{2,0} do the same, but in some corner cases they do different things. That is bound to cause subtile and hard to triage bugs. [13:35] anyway: true to the principle of most surprise. [13:37] It is backward compatible.. so the possible improvements were very limited. [13:40] c++ developers that enjoy the language are a flock of people enjoying complex things.. [13:43] anpok: there are only two kinds of programming languages: those that suck, and those that nobody uses. [13:44] and Python [13:44] ;p [13:44] willcooke: [13:44] willcooke: ^^^ does that code run for you? its entirely whitespace [13:44] is it friday yet? [13:44] soon! [13:45] * Sweet5hark tries to hold his guns then. [14:00] hi team!! [14:07] hey desrt! [14:08] * desrt hugs pitti [14:08] good morning :) [14:08] * pitti hugs desrt back, how are you? [14:09] looking forward to FOSDEM, I hope I see some of you there! [14:09] you will see me. [14:09] with Ubucon I guess didrocks and seb128 won't come? [14:09] desrt: ah, nice! [14:09] assuming my travel documents are in order [14:09] Laney: you as well? [14:09] Sweet5hark: and you? [14:11] pitti, they're still going :) [14:12] willcooke: great [14:12] pitti, correction, didrocks is going. Seb not [14:15] * desrt has an ingress date with didrocks [14:16] :D [14:28] pitti: yup [14:29] pitti: what are your plans? [14:29] ahoy desrt [14:29] Laney: I'll go to Leuven next Wednesday, to visit Michael again [14:29] Laney: and on Sat/Sun we come to FOSDEM in the morning; I'll stay at his place Sat night, though [14:29] Laney: and I take the 18:20 ICE back to Augsburg Sun evening [14:30] Laney: I might pop up at the systemd hackfest on Friday (was just announced yesterday) === ChrisTownsend1 is now known as ChrisTownsend [14:35] michael? [14:35] oo. systemd hackfest. [14:35] i never miss a chance to hug lennart [14:36] pitti: ah nice, see you around then! [14:50] Laney, desrt: hey, curious on your opinion here. I have some users that plan to upgrade to xenial and they use evolution. webkit/main-ness aside, what are your thoughts on evolution in terms of stability, usability, etc these days? [14:51] Laney, desrt: I'm going to use it (under appamor, /me shakes fist at webkit) until the release. I have to say-- it so far seems to be maintained quite well === jdstrand_ is now known as jdstrand [14:52] jdstrand: i switched to webmail ages ago :( [14:55] hi jdstrand [14:55] I don't worry too much when I have to file bugs on evolution because it has a nice responsive upstream [14:55] I think it pretty much works well although it's always been a bit of a big beast [14:56] some operations aren't the quickest [14:56] sort of reminds me of Outlook last time I used that but I guess that they aim for similar things [14:56] yeah [14:56] so that's not surprising [14:56] probably going to punt it to universe soon [14:56] I used to use it all the time, then switched to tbird personally when Canonical switched to supporting it [14:56] because webkit1 [14:57] I always liked evolution [14:57] it is still in main? [14:57] yeah [14:57] oh, probably for support [14:57] (UA) [14:57] guess so [14:57] the security team has a note on not providing updates to webkit [14:58] just read that webkitgtk is going to start publishing CVEs [14:58] I'd love for someone to get excited about oxide and create gtk3 bindings for it [14:58] that's webkit2gtk for us [14:59] they tried that before iirc (being responsive to CVEs). the problem was resources-- they couldn't backport easily so they needed new versions and they couldn't maintain that for the duration of one of our stable releases (even non-LTS) [15:00] classic webkit support problem [15:00] anyhoo, I hope they figure out how to do it this time :) [15:00] Laney: thanks for the feedback. with it still being in main, that gives me a little more to think about [15:01] not for long [15:01] larsu fixed a bug in the theme that was blocking the switch [15:01] so it's more or less ready to go after I test my port of software center a bit more [15:02] Laney: to go to universe? [15:03] sorry I don't know what the other half of that question is [15:03] Laney: you said it's "more or less ready to go", to go what? [15:03] Laney: to get punted to universe? [15:03] ah yes [15:04] Laney: you might want to talk to the UA folks [15:04] swap for the new one which removes the webkit api [15:04] webkit1 [15:04] sure I will but it's not really an option to stay on the ancient version any more [15:04] so it's going to happen one way or another [15:04] oh, I think I didn't understand what you were talking about [15:05] if that's keep two webkits in main then you get to get angry [15:05] Laney: so there is an evolution that is webkit1 and one that is webkit2? [15:05] no [15:05] there is only webkit1 [15:05] tpopela at redhat is working on it [15:05] one for webkit1 [15:05] indeed [15:06] I see [15:06] they are adding API to webkit itself to support evolution's usecases [15:07] wysiwyg editing stuff [15:29] that's pretty cool [15:34] :| [15:34] time to play whack a mole on more failing tests [15:35] pitti: can you add testing of reverse test-deps to your sprint? :) [15:36] Laney: won't be able to do that today, but it's on my WI list for Jan/Feb indeed [15:36] needs some prerequisite work in dpkg [15:36] pitti: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-xenial/xenial/amd64/a/autopilot-gtk/20160121_135055@/log.gz like this [15:36] python-psutil broke autopilot yet migrated [15:37] actually wouldn't that have been a runtime dep even? [15:37] yes [15:38] but no recorded test for this upload [15:40] oh I see, autopilot-gtk -> python{,3}-autopilot -> python-psutil [15:40] so this is a recursive case [15:41] ah, wow, it's a feature not a bug. sort of. if one accidentally enables High Contrast mode in lightdm, it starts looking ugly, not really high contrast and loses user backgrounds. [16:36] FFS [16:43] mitya57: I don't suppose you know about https://launchpadlibrarian.net/232311658/buildlog_ubuntu-xenial-s390x.autopilot-legacy_1.4.1+15.10.20150911-0ubuntu1_BUILDING.txt.gz do you? [16:43] it happens on all arches, just s390x tried to build later on [17:23] now if I forget it again in a year or two, I may find my own questio and answer at http://askubuntu.com/questions/723902/lightdm-doesnt-show-wallpaper-background-and-colors-are-wrong/723903 :) [18:00] Laney, looks like a bug in Sphinx :) Unfortunately I don't have time to look at it today, will do tomorrow. [18:00] okay! [18:10] night night [18:25] kenvandine, hey [18:26] kenvandine, "with new LE devices we never know the type before they are connected." ... does it mean we can't display device icons anymore? seems like an user regression, it seems weird that new protocols have such limitations [18:31] hey seb128! [18:31] How goes? [18:31] hey willcooke! [18:31] seb128, sorry for the spam :) [18:31] good, watching Mark's keynote atm [18:32] no worry [18:32] I didn't read the airplane mode thing yet [18:32] but GNOME got that working I think, even the hotkey [18:32] we should maybe look at backporting what they did? [18:36] night all [18:39] bah video streams not working for ubucon [18:39] * willcooke hits mhall119 and runs [19:54] seb128, i'm not sure, morphis could explain better [19:54] he convinced me it would be better to remove the isSupportedType all together [19:54] i think pat and bill are convinced too [19:55] seb128, he said even if a device isn't something we claim to support, pairing would usually succeed [19:55] the issue is we don't have an app that utilizes them all [19:55] well, fine, but get the design updated then ;-) [19:56] i'll bring it up in tomorrow's meeting