[00:47] Zzz === Drac0 is now known as Guest64403 === thumper is now known as thumper-afk [05:59] Good morning === thumper-afk is now known as thumper === hikiko_ is now known as hikiko [08:26] quiet desktop channel this morning! [08:26] * willcooke sings [08:27] oh, maybe that's why everyone left [08:27] hum, starts raining [08:27] :p [08:27] from Will's *singing*? wow [08:27] willcooke: ^ you can become a millionaire if you do that in California, the Sahara, or other deserts :) [08:28] :-) [08:28] hey pitti! wie gehts? [08:29] seb128: prima, danke! [08:29] :D [08:29] morning pitti [08:29] I had a bit of a "spring time fatigue" case in the last days, but since basketball last night I'm feeling great again [08:29] seb128: et toi, comment vas-tu ? [08:30] ça va bien ! [08:30] pitti: spring time fatigue meaning "enjoying to lie in the sun" [08:30] morning everyone! [08:30] hey larsu ! [08:30] a bit tired, went to bed later this week (still on the holidays sleep time schedule :p) [08:30] just went out to get some real coffee and read email&co at a coffee place [08:30] oh hi! [08:31] hey larsu Laney! [08:31] seb128: nice [08:31] the channel is waking up ;-) [08:31] hey Laney, how are you? [08:31] larsu: nah, more like "hayfever season starts" [08:31] uh oh :( [08:31] but doing exercise works wonders against that [08:32] which is a great motivation -- "work out or feel bad" :) [08:32] haha perfect! [08:32] * seb128 wants to go play tennis [08:32] btw, will anybody happen to be at the cloud sprint in May? [08:33] foundations now got invited there, as we moved team [08:33] you did? [08:33] you moved? [08:33] pitti: good thanks, you? [08:33] and foundations has its own sprint in June [08:33] we're going to work on upstreaming g-s changes today :-o [08:33] but it seems we never have joint sprints with the desktop team any more :( [08:33] Laney, \o/ [08:33] pitti, yeah, it's a shame :-( [08:33] pitti: you moved into cloud things? [08:33] seb128: not me personally, foundations got moved from UES to CDO [08:33] ah [08:34] all of 'core' or just you? [08:34] pitti, right, the you was = foundations, I somewhat missing the email about that (if there was any) [08:34] pitti, and no, I don't think we are going to join you :-/ [08:34] we have a team sprint in Prague at the end of april [08:34] but that's co-hosted with unity8/design teams [08:34] yeah, probably not public channel material, but horribly uninteresting/boring anyway for any non-Canonical folks :) [08:34] haha [08:35] still pondering whether I'll go to debconf this year [08:35] pitti: too far? [08:35] "great people and great conference" vs. "argh far away and already doing a lot of flights this year" [08:35] but you'd be in the cloud more! [08:35] a bit of the same here [08:35] (yeah, I just did that) [08:36] travelling 3 times in april [08:36] though I guess it's going to be a bit more quiet again then [08:36] both my CO2 foot print and my radiation intake must have reached absurd levels by now [08:37] shame for foundations moving out of UES [08:37] it means that if we get a UES sprint we are still not seeing each others [08:37] not entirely sure why this was done, but I entirely blame Rick :) [08:37] pitti, you are going to go to UbuCon Europe at least I guess? [08:37] though that's in nov. [08:38] maybe [08:39] seb128: could you look at https://code.launchpad.net/~nhaines/ubuntu-wallpapers/xenial-fcs/+merge/291209 perhaps? [08:39] Laney, sure [08:40] thanks! [08:40] Laney, what changes do you guys upstream? I though attente/robert_ancell were commiting everything that is upstream material directly to git and just stacking ubuntu specific backend/changes in another branch? did that de-rail? [08:41] or is upstream taking the apt backend in trunk now? [08:41] they're in branches, it's about merging as much as possible in trunk [08:42] cool [08:42] btw are we moving to get g-s sticking in bg as a service after all? [08:42] I saw attente reverted his patches to make it exit on close [08:43] seb128: yeah, i think we should if we want to think about checking for fwupd updates [08:43] you mean checking about those in bg? [08:44] yes [08:44] because I guess getting those updates from an active g-s would still work even if it was not in service mode? [08:46] seb128: i guess. but i'm also not sure what happens if gs is terminated while its checking for updates [08:47] seb128: there's also some other ui crashers that i think are being caused because we allow the ui to be destroyed [08:48] attente, hum, k [08:48] is there already a release date planned for 16.04 ? [08:48] xclaesse, https://wiki.ubuntu.com/XenialXerus/ReleaseSchedule [08:48] seb128, thanks [08:49] attente, is there any side issue due the fact that we have 3 components doing index updates now? apt has a systemd timer unit (I think), update-notifier and now gnome-software [08:49] attente, also does gnome-software start with the session? or does the user need to try to use it once to have it activated? (in which case if you don't use it you still never get notified about firmware updates) [08:50] seb128: doesn't start with a unity session afaict, but does with gnome-shell iirc [08:51] seb128: not sure about side issues, but if gs is doing an update, it prevents using apt directly from the command line [08:52] right, it locks the db [08:52] but it still means we do that update more often that needed [08:52] which might also means hitting the servers 3 times more [08:52] which might not be low-cost on the infra [08:53] like if you have 10M users which have a daily trigger, and g-s does the same it means we have another 10M server hit daily [08:53] seems suboptimal [08:54] stop update-manager doing the update itself then [08:54] err, yes please [08:54] but u-m isn't supposed to do taht automatically, and I don't think it does [08:54] u-n gets *triggered* by new apt indexes being available [08:54] but it doesn't request them by itself [08:55] Laney, update-manager is not a service and not doing any index update in bg afaik [08:56] what pitti said [08:56] apt has that cron, now system timer unit [08:56] I don't know what g-s does exactly [08:56] does it trigger refreshes by itself in a periodic way? [08:56] attente, ^ do you know? [08:58] seb128: yeah, the default behaviour is it checks once a day [08:58] well, if we have it acting as a service imho it should [08:58] - start with the session [08:58] - don't dup that apt is already doing [08:59] +work [08:59] why would gnome-software start with the session? [08:59] that seems rather expensive [08:59] because they use it in service mode to do what update-notifier does for us [08:59] and index updates are even more expensive of course [08:59] except that it does it for more than debs [09:00] ah, this will replace u-n? [09:00] e.g for firmware updates as well [09:00] I guess it should [09:00] but it's a bit late in the cycle [09:00] apt's trigger at least has some protection to not run on battery, and when the box is offline, etc. [09:00] well, we ship firmware through normal packages [09:01] unsure [09:01] that's the recent fwupd stack superm1 worked on [09:01] https://bugs.launchpad.net/ubuntu/+source/gnome-software/+bug/1544376 [09:02] Launchpad bug 1544376 in gnome-software (Ubuntu) "[ffe] Enable firmware support" [High,Fix released] [09:02] that does efi bios updates === Drac0 is now known as Guest35512 [09:37] tjaalton, pitti, thanks for figuring that xorg/kbd issue out [09:37] well, need to decide if we still want that caching or not [09:38] Trevinho, andyrock, just as a follow up, https://launchpad.net/ubuntu/+source/xorg-server/2:1.18.3-1ubuntu2 [09:38] tjaalton, would still be good to have imho [09:48] hey seb128 [09:48] hey andyrock ;-) how are you? [09:49] i just wake up :D [09:49] *o [09:53] you are on Trevinho's time? ;-) [10:25] seb128, pitti: I got a private mail from sabdfl about the call for testing message: "Thank you, nicely done." So we probably did the right thing. :) [10:25] GunnarHj, hey, good ;-) [10:26] seems the update went well, no negative feedback/bugs that I can see [10:30] jibel, willcooke, cyphermox: no issues on all amd box going from fglrx to amdgpu [10:30] davmor2, sorry? [10:30] No issues? [10:30] * willcooke rubs eyes [10:30] willcooke: no issues upgrading an all amd box [10:30] wooooohooo [10:30] great work [10:30] thanks davmor2 [10:31] willcooke: 14.04 to 16.04 going from fglrx to the new gfx stack amdgpu [10:32] Laney, seb128 - thanks for sorting the wallpapers [10:32] seb128: Just saw a negative Chrome report: [10:32] https://bugs.launchpad.net/ubuntu/+source/language-selector/+bug/1468027/comments/113 [10:32] Launchpad bug 1468027 in kubuntu-meta (Ubuntu) "change default CJK fonts to Noto CJK" [Medium,In progress] [10:32] Don't know yet if it's due to the fc upgrade or Noto Sans CJK in general. [10:32] k === hikiko is now known as hikiko|ln [11:24] davmor2: cool, thanks [11:35] there is no autoremove of old kernels running time to times? [11:35] on 14.04 at least [11:35] is that something fixed in 16.04 ? [11:35] Realized that my mother's computer has 10G (!) of useless kernels [11:36] * ricotz notes removing 8GB worth of kernels was fun [11:36] yeah, going to take a looong time [11:36] basically 2 years of kernel updates [11:41] I really hope it's fixed in 14.04, it's like the most critic bug I've had with ubuntu [11:41] *16.04 [11:41] I remember I already had years ago /boot full because of that [11:41] in the middle of a kernel upgrade => system won't boot anymore [11:42] xclaesse, what to be fixed? [11:42] seb128, removing 10G of kernel updates [11:42] yeah, that's fixed [11:42] not in 14.04 at least [11:42] they are marked for autoremoval/cleared [11:42] right [11:42] doing the upgrade to 16.04 atm [11:42] it was fixed properly in 15.10 I think [11:42] seems to be working for me at least [11:43] will it do the autoremove itself, or do I still have to run sudo apt autoremove myself? [11:43] Trevinho, Afternoon. [11:43] Trevinho, https://code.launchpad.net/~ubuntu-mate-dev/compiz/fix-1567354/+merge/291235 [11:45] xclaesse, update-manager is doing it for you [11:46] seb128, cool, thanks :) [11:46] yw! [11:55] Trevinho, Sorry. My earlier merge proposal was stacked incorrectly. [11:56] I've corrected it here - https://code.launchpad.net/~ubuntu-mate-dev/compiz/fix-1567354/+merge/291239 === hikiko|ln is now known as hikiko === alan_g is now known as alan_g|lunch [12:15] seb128, Trevinho do we need a FFe for the hud change? [12:15] andyrock, yes [12:16] seb128: k ta [12:16] i'll take care of it [12:16] thanks [12:18] seb128: a technical question: in bamf we need dbusmenu-glib/gtk now [12:18] i made the dep optional in configure.ac [12:18] but i'm wondering what i should do in debian/rules [12:19] explicit is better than implicit [12:19] so if there is a --enable/disable just set what is appropriate [12:19] there is a --enable [12:19] and it's on by default [12:20] stilll explicit is better [12:20] but e.g. elementary guys can turn it off [12:20] so please add it to the rules [12:20] don't forget the build-depends in debian/control as well ;-) [12:20] k [12:37] Laney: do you know off the top of your head the magic string to search for in https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-xenial/xenial/s390x/g/glib2.0/20160406_154338@/log.gz [12:37] Laney: i. e. which test failed? [12:37] * pitti tried "fail:", "error:", and "not ok" so far [12:38] and also through the hundreds of matches of "fail" (which are mostly test names which contain that word) [12:40] pitti: gnome-software does set up a monitor for network status changes with a GNetworkMonitor and all the plugins then have code on what to do based on network status [12:40] (re that discussion that i was mentioned above) [12:43] i don't believe there is any check for whether the system is on AC vs battery today, but I doubt there would be opposition to that being added [12:44] well, as long as it's only testing for BIOS updates and perhaps snaps, that at least doesn't duplicate the apt cronjob [12:44] pitti, "Test glib/gdbus-proxy.test failed" I guess is hte issue? [12:44] even if the metadata offers a FW update for some hardware on the system, individual fwupd plugins have controls (and policy) whether to offer firmware updates to fwupdmgr (CLI) or gnome-software (gui) [12:44] (not sure if snappy already has some cron job/timer for checking for updates) [12:44] well that's actually something i was going to mention [12:44] maybe it's worthwhile having a similar cron job to apt for snaps and fwupd appstream data [12:45] if not using gnome-software's refresh service [12:45] GLib-GIO:ERROR:/build/glib2.0-E9pInk/glib2.0-2.48.0/./gio/tests/gdbus-proxy.c:826:fail_test: code should not be reached [12:45] seb128: ah, thanks [12:45] pitti, yw [12:45] ok, that's nothing new then, this test has failed fairly often recently [12:46] desrt ^ is that a known issue? [12:49] superm1, pitti, willcooke, I didn't look at when g-s does it index refresh exactly but we should keep a good eye on it, if it does it everytime you connect/get online it might DoS the servers a bit [12:49] andyrock: Yes, use enable and off by default [12:49] eww, yes [12:49] *every* time? [12:49] *if* [12:49] it also DoSes your 3G connection [12:49] as said I didn't look at the code [12:49] attente or Laney probably know better [12:50] Laney, attente, ^ can you figure that out? when does g-s refresh its indexes exactly? does it has a notion of "don't do it on mobile"? [12:50] it does monitor for metered connections i know === alan_g|lunch is now known as alan_g [13:06] huh, now empathy is misbehaving again :-/ [13:12] dobey, :-( [13:12] do you get any warning/errors? [13:13] no [13:14] well, nothing visible [13:14] when did that start? empathy didn't change in ages [13:14] neither did the telepathy stack [13:14] and .xsession-errors doesn't show anything [13:15] well i've been using pidgin forever. i just started trying to use empathy again a few days ago [13:15] dobey, logs are in .cache/upstart [13:15] .xession-errors get nothing nowadays [13:15] just the upstart errors [13:15] since that's what xsession starts [13:16] then upstart manages the session so it gets the log [13:16] under unity7 or dbus logs, depending how you start the thing [13:16] hmm, ok [13:16] i'll see if there's anything in there [13:17] k [13:17] brb [13:30] seb128, holidays this week, sorry [13:32] My network just bounced a bit and APT's download description went a little wonky. I really admire the optimism of whoever put petabyte units in there. [13:48] seb128: hughsie said that it does the right thing by default for metered connections (doesn't do updates), but allows the user to override through the ui [13:49] attente, where in the ui? [13:49] attente, also if you are on eth or wifi, how often does it refresh/hit the server? every time you connect to a ap or cable? [13:49] or once a day? [13:49] or...? [13:49] src/gs-shell-updates.c [13:49] daily [13:49] can we just disable that? [13:50] if we have a systemd timer job [13:50] if->since [13:50] seb128: pops up a dialog [13:51] saying that? [13:51] "Checking for updates while using mobile broadband could cause you to incur charges." [13:51] hum [13:52] so that dialog pops daily out of nowhere if you are on 3g? [13:52] seems to be so [13:52] :-/ [13:52] oh well, that's a detail/bug we can fix later [13:53] the important bit is to make sure that both apt and g-s don't hit the server to refresh the index daily [13:53] or we double the number of requests to the server [13:53] who says it does that? [13:53] shrug [13:53] the dialog thing [13:54] attente just did? [13:54] i'm just looking through the code [13:54] src/gs-shell-updates.c +858 [13:55] actually it seems like this only happens when the user explicitly clicks the refresh button though... [13:55] yes [13:55] seems like an non issue then [13:55] good :-) [14:04] qengho: i've seen the PB/s before, without /my/ network going wonky [14:08] seb128: it already ignores the background updates [14:08] »···if ((flags & GS_PLUGIN_REFRESH_FLAGS_UPDATES) == 0) [14:08] »···»···return TRUE; [14:09] Laney, meaning? it doesn't do autorefresh of the index by itself? [14:09] just if you click the button? [14:09] correct [14:09] k [14:09] so what's the point of having it always active as a service? [14:09] I might be overlooking something it does by itself [14:09] if it only reacts to UI actions it can as well only be active when the ui is open right? [14:12] firmware [14:13] other plugins than apt in general [14:13] is that frequent enough to justify a service active all the time? [14:13] talk to superm1 [14:14] superm1, ^ [14:14] do you think we need a service for that? [14:14] or it's going enough to let user know about those updates when they use the store? [14:16] why is it such a big problem? [14:19] Laney, it's not the end of the world, but it's the 3rd memory user process on my session with 45M of RAM used and it's one more process active/creating wakeups/... [14:19] shared RAM [14:20] we always tries to limit bloat, just part of that [14:20] but we can't get around then it is what it is [14:20] just a bit sad to see us drift [14:20] seb128: if it's running as a service it has an update monitor that will run and update metadata once a day not only when you click the refresh button [14:20] to me this actually replaces update-notifier [14:21] I'm +1 for that [14:21] or at least it should, maybe it's not banged on hard enough for that now [14:21] but then we actually need to check if we can disable update-notifier [14:21] or does it still do other things? [14:21] but when the whole concept was initially coming together that's what i was expecting was going to happen [14:21] right [14:21] that sounds good [14:21] but it's a bit late in the cycle to swap out update-notifier [14:22] at least if we do we need a ffe and somebody actually making sure we don't throw away anything useful [14:22] yeah [14:23] but Laney and attente were with hughsie all week, hopefully they understand the approach that's supposed to be used with this well now [14:23] right [14:25] I think that part is not the issue [14:25] it's just making sure we don't have custom parts doing the same things [14:25] it's not replacing update notifier yet [14:25] should do in future though, and update-manager [14:25] what is update-notifier still doing? [14:25] * seb128 looks to refresh his mind [14:26] gnome-software isn't a full apt upgrader yet [14:26] well, update-notifier doesn't do much nowadays iirc [14:26] the apt index refresh is a systemd timer [14:26] it's not even telling you about all updates [14:26] the apport prompting used to come from it but it's an upstart job now iirc [14:27] u-n is what auto-open update-manager though right? [14:27] also it handles upgrade medias, e.g if you put a dvd with a new version of Ubuntu [14:28] superm1, I guess the question is to know how often/useful firmware upgrades notifications are to users, e.g is it worth having a process eating 45M RAM all year long to prompt users about those once a year [14:29] seb128: well at least for my company updates are issued approximately every 2 months for the systems we're supporting it on [14:29] fwupd is going to be supporting all sorts of random firmware upgrades at some point though too [14:30] and i'm sure other companies will come on board to use it for EFI eventually [14:30] yeah, don't get me wrong, I see that it's useful [14:30] it's just a cost/benefit thing [14:31] so if it's not ready to replace update-notifier, is there a better way to pull down metadata updates (cron?) and open gnome-software when relevant? [14:31] not one ready [14:32] we would need to work on a solution [14:32] seb128: http://people.canonical.com/~bjoern/xenial/5.1.2/libreoffice-l10n_5.1.2-0ubuntu1_source.changes http://people.canonical.com/~bjoern/xenial/5.1.2/libreoffice_5.1.2-0ubuntu1_source.changes [14:32] attente, Laney, superm1, in any case I think that needs a ffe [14:32] seb128: ^^ upstream has released 5.1.2, kindly asking you to consider sponsoring. [14:32] if we decide we want those firmware update notifications we need to turn back g-s to a proper service and have it to start on login [14:32] Sweet5hark1, thanks, looking in a bit [14:33] seb128: well i guess leadership in your team needs to get all this stuff documented somewhere and decide on the approach to take then [14:33] willcooke, ^ that would be you? ;-) :p [14:33] i suspect if it's going to be a cron updates and open gnome-software some other way it will have to land after release as an SRU of sorts [14:34] and if it's before, leave gnome-software as a service it's an FFe (maybe turn that bug that firmware updates weren't working into an FFe) [14:34] right [14:42] xnox: hey, is there a way to get a trace of the execution of upstart jobs? Like a log of what has been launched, when and in which order [14:43] Trevinho, yes. is this for desktop or pid 1? [14:43] xnox: session [14:44] modify things in xsession that launch upstart, and add --debug there [14:44] and then in your ~/.xsession.log or what not will have all the glory details [14:44] xnox: ok, thanks [14:44] seb128, sorry was tied up [14:44] read the backlog [14:44] * Trevinho wonders who launches upstart now :) [14:45] so I think the issue is. If g-s doesn't run as a service we don't get FW update notifications. If we turn off u-n and rely purely on g-s then we don't get "technical items" notifications [14:46] We're not going to turn off u-n now. [14:46] xnox: also ~/.config/upstart/ANY_SERVICE.conf will always override the ANY_SERVICE system one, right? [14:46] and it sounds like we need g-s running as a service anyway to fix other issues [14:46] so I think we just have to eat it for now [14:46] and run both [14:47] Trevinho, yes, the manpage lists priority of locations. [14:47] xnox: sure, just to be sure :) [14:53] superm1: just uploaded glib to jessie-backports [14:54] willcooke, k, I sort of got to the same conclusion [14:54] xnox: --debug added but got nothing in .xsession.log :o [14:54] Laney, attente, k, let's turn g-s to a service and start it a login then ... I think that should get a ffe but up to Laney to decide if he agrees on not, I'm not going to argue if you think it's fine to change without one [14:55] Trevinho, .xession-errors? [14:55] thanks seb128 willcooke [14:55] seb128: thanks [14:55] nope [14:55] Trevinho, :-/ [14:56] Laney: awesome thanks. [14:56] yw [14:56] are you going to backport fw* and gnome-software and everything? [14:57] cheers guys [14:58] just seen someone's gnome-shell lock screen [14:58] they show notifications on there [14:58] this is nice ;_; [14:58] I've -marco 7545 1.3 0.3 46852 5096 ? Ss 16:56 0:01 /sbin/upstart --user --debug --verbose but no logs [14:59] Laney, did you learn about their secret $companey_plan through the email summaries? ;-) [14:59] Laney: hopefully the whole stack, but have to see what other push back I hit. If nothing else at least command line options [15:07] seb128: just that he got highlighted on IRC :P [15:07] in #bring-down-canonical [15:07] !!!! [15:08] lol [15:09] hmmm, it hails outside and I could be on gran canaria instead (libreoffice espania hackfest) [15:18] seb128: so, it seems the log goes to /var/log/lightdm/x-0-greeter.log, xnox possible? [15:18] I guess [15:18] or is that the greeter session log? [15:18] Trevinho, that would be the upstart session of the greeter, it runs an upstart too. [15:19] ah [15:19] Trevinho, you probably after ~/.xsession* something log, and modify the Xsession scripts [15:20] xnox: no, there's nothing related... Maybe there's some option in xorg to disable this by default [15:20] hm [15:20] I would just manually launch that by hand, but... I neeeded some user feedback, and thus nothing to be hardly modified [15:21] Trevinho, can you change the command to "> /tmp/upstart.log" or something? [15:22] seb128: I tried, but it didn nothing too [15:22] it's in /etc/X11/Xsession.d/99upstart and the command is exported as a variable... which I don't know who runs [15:22] Trevinho, cat ~/.xsession-errors [15:23] ? [15:23] should have the user desktop init messages [15:23] ah, by another script [15:23] * Trevinho didn't grep :) [15:23] xnox: I get errors there, but no stdout data [15:24] Trevinho, the Xsession.d/99upstart sounds like the right thing to adjust, or there should be 01upstart too [15:24] and it should be in ~/.xsession-errors... =( [15:25] xnox: ok I got it [15:25] hi xnox! [15:25] /etc/X11/Xsession.d/99x11-common_start -> exec $STARTUP &> /tmp/xlog.log [15:25] Trevinho, there are loads of logs in ~/.cache/upstart/* too [15:25] from individual logs [15:25] xnox: yeah, scanned and sorted. none was there [15:27] =( [15:45] Laney, do you know why debian/copyright in ubuntu-wallpapers only list the copyright of the current release set of images? the older ones are still distributed so technically we should add the new one but remove the old ones (asking because it seems it was already done like that in previous cycles) [15:51] seb128: you are probably interested in https://bazaar.launchpad.net/~ubuntu-art-pkg/ubuntu-wallpapers/ubuntu/revision/155#debian/copyright and https://bazaar.launchpad.net/~ubuntu-art-pkg/ubuntu-wallpapers/ubuntu/revision/166#debian/copyright [15:53] Laney, seb128: (hopefully) quick question: `debuild -S` calls dpkg-buildpackage with "-us -uc", but I want the package to be signed... [15:53] do you know where it pulls those options from? [15:53] larsu, it does sign by default [15:53] I called all the tools manually before, which worked [15:53] what's in ~/.devscripts? [15:53] seb128: maybe I set something somewhere? I can't find anything :( [15:53] do you have a private key matching the email in the changelog? [15:53] I usually sign it manually though [15:53] Laney: *big hug* [15:53] you can debuild -S -k otherwise to force a key to sign [15:54] I just didn't know I put that there - probably years ago [15:54] seb128: ah that's handy as well [15:54] thanks! [15:54] yw! [15:54] probably because you wanted it to not sign automatically in the past [15:54] it's annoying default behaviour if you build a lot of packages [15:55] right [15:55] why do you sign manually? [15:55] only need to do it when actually uploading [15:56] ah, right [15:56] thank you! [16:03] Laney, thanks, in fact it gets me confused even more :-/ [16:04] I guess going to need to ping Nathan / Daniel [16:04] bah and dholbach *just* left IRC [16:06] looks like he made it a list instead of a generic thing but didn't do the old ones [16:07] well he removed the old ones from the first section [16:07] which is CC-BY-SA 2.0 [16:07] and the * is for 3.0 [16:07] but I'm unsure why we have every 3.0 but the new ones 2.0 [16:07] or if that's wanted [16:07] that doesn't make any sense to me [16:10] I asked Nathan but he doesn't seem around [16:10] let's wait to see if he replies [16:19] Laney, https://git.gnome.org/browse/gnome-software/commit/?h=wip/rancell/apt&id=48db971d839823dad9eeb49b43175d07bd978f5a fixes https://bugs.launchpad.net/ubuntu/+source/gnome-software/+bug/1554164 right? [16:19] Launchpad bug 1554164 in gnome-software (Ubuntu) "Invalid read in get_changelog()" [High,New] [16:19] Laney, I'm going to assume so and change the bug to be assigned to you/fix commited :p [16:23] umm don't know [16:23] probably [16:25] the trace looks similar [16:29] Sweet5hark1, libreoffice sponsored, thanks! [16:30] Sweet5hark1, btw did you see https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1566050 ? it's seems to hit quite some users [16:30] Launchpad bug 1566050 in libreoffice (Ubuntu) "soffice.bin crashed with SIGSEGV in MenuItemData::~MenuItemData()" [High,Confirmed] === alan_g is now known as alan_g|EOD [17:14] attente: Laney since it sounds the user service for gnome-software will come back it's worth mentioning that you should probably also make sure this doesn't start when on a live disk (probably something needed in the session script to detect casper or a casper script to modify the livefs while booted ) [17:16] actually applying an update from a live disk is technically possible for some FW types live, but others you would have to stage the system properly before trying (eg mount an existing ESP to the right place first or create one) [17:18] does update-notifier/update-manager do that? [17:19] a patch would be nice :) [17:19] happy to do a casper patch, need to see how you will normally start the service though [17:20] looks to me that casper does turn off update-notifier [17:20] with 31disable_update_notifier [17:23] it'll be an xdg autostart file [17:23] look at the current one, fixing OnlyShowIn there [17:23] /etc/xdg/autostart/gnome-software-service.desktop [17:30] Laney: okay added in casper 1.371 [17:46] superm1: merci, did/will you upload? [18:08] mvo: do you know where "apt show" gets APT-Sources from? [18:08] I can't follow the code [18:09] Laney: it looks at the current indexfile that matches the record and ask where it comes from, its in private-show.cc line 176 [18:10] Laney: is it not working? or why do you ask? [18:10] mvo: It is working, I want to steal the logic for gnome-software [18:10] you want to see something scary? [18:11] mvo: https://git.gnome.org/browse/gnome-software/tree/src/plugins/gs-plugin-apt.c?h=wip/rancell/apt#n387 [18:12] Laney: uff [18:12] Laney: there is no great way right now to check this but that looks not ideal [18:13] Laney: is this linking to libapt, i.e. could libapt be used here? [18:13] it's using aptdaemon in other places but not libapt [18:14] Laney: if not, you will need to look from the {,In}Release file to the packages file (or rather from the Packages file back to the release file) and check the origin label etc. [18:14] I can probably eliminate most of the worst parts [18:14] like looking at Origin [18:14] I wanted the APT-Sources thing to get the section and release [18:15] yeah okay, so starting from Release is probably a good idea [18:15] Laney: indeed, it might be simpler to use libapt, but maybe not, it seems there is a lot of reimplementation already so maybe its not worth the effort [18:15] robert_ancell wrote this, I'm coming to it a bit blind [18:15] Laney: so yeah, use release, walk down to the matching Packages file (you could use the hashes as a cheap way to find the matching ones) [18:15] Laney: I know :) [18:16] bleh [18:16] Laney: not blaming you [18:16] how do I even know which Release to look at? [18:16] * Laney is reimplementing apt [18:16] Laney: you will ned to look at them all and then see if the packages file matches [18:16] Laney: yes you are [18:17] Laney: a cheap way of finding the right packages file for the release file might be to use use the hashes, but its a bit slow, or you need to construct filenames accorindly and escape in the right way etc [18:17] (or you use libapt ;) [18:18] this is sounding more sane [18:18] this API is easy to use right? :) [18:19] I just need to iterate through all the available packages and get some fields like origin/section/component/release [18:24] Laney: *cough* easy *cough* - the python bindings are, but the c++ is medium easy, let me think a little bit if I can help [18:36] mvo: that would be nice if you have time, even a demo program would be great or a pointer to a similar example [18:56] Laney: yes uploaded [19:03] seb128: hmmm, that crash is odd ... [19:04] seb128: because the crashing thread includes #6 0x00007fe082149088 in __run_exit_handlers (status=0, listp=0x7fe0824d35f8 <__exit_funcs>, run_list_atexit=run_list_atexit@entry=true) at exit.c:82 in the stacktrace [19:06] seb128: so it was already exiting when it crashed (the latter possibly due to some race with the dbus stuff). but the question is: why was this exiting in the first place if the user didnt intend that? [19:08] seb128: oh, and thanks for the sponsoring! [19:19] hmmm, so for bug 1566050 -- I guess the stacktrace is misleading: something is causing LO to exit and it is crashing when doing that -- the real question though is what triggered the exit (and the stacktraces dont help with that -- it could even be many very different things, each an odd cornercase in itself) [19:19] bug 1566050 in libreoffice (Ubuntu) "soffice.bin crashed with SIGSEGV in MenuItemData::~MenuItemData()" [High,Confirmed] https://launchpad.net/bugs/1566050 [19:26] ok, ~90% of the cmdlines opened a csv -- so the stacktrace is meaningful and sc/source/ui/dbgui/csvgrid.cxx:88 will soon regret its lifecycle management choices ... [19:27] Sweet5hark1: the Dependencies.txt shows two modified files in the librreoffice-common package.. [19:28] sarnold: thats even ... woot? [19:30] Sweet5hark1: dunno. it might be relevant it may not be.. [19:30] sarnold: ok, thats horrible. I have a guess what is responsible for that though and its unlikely the cause of the crash ... [19:34] quittin time. night all [19:34] holiday tomrrow and Monday, travelling all next week [19:34] email or telegram if you need anything [19:34] l8r [19:37] Sweet5hark1: interesting, two dups too, I just made the second one public [19:41] sarnold: my bet would be someone played with the lifecycle management in the dialog in unhealthy ways. My current best bet is https://github.com/LibreOffice/core/commit/2660d24a07866e083c5135ea263030f3e3a2e729#diff-0438608831b6633414398d7cb4b44950R1360 ... [19:42] Sweet5hark1: yikes,that certainly seems like the scope of problem that could do that.. [19:45] I need help with gtk+ , how add shortcut to window? [20:37] === fredp is now known as Guest14494 === Guest14494 is now known as fredp === fredp is now known as Guest68109