[00:33] Laney, mihall119, Are you there? Can you help me answer this question? [00:33] https://sourceforge.net/p/xoscope/bugs/20/#2eef [00:33] mhall119 [00:45] Trevinho: does 1552537 make any sense to you? two people report it on wily and xenial :/ [00:52] sarnold: I can't reproduce that as well... [00:52] Trevinho: thanks for giving it a look. I've got a feeling that it's user-induced somehow -- it almost always is :) -- but these reports never have enough to go on :( [00:58] sarnold: sure, it's better to check. === JanC_ is now known as JanC === thumper is now known as thumper-afk [05:56] hi [07:15] Good morning === maclin1 is now known as maclin [09:03] morning all [09:04] morning [09:10] Laney: guten tag === Drac0 is now known as Guest90504 [09:34] hey davmor2 [09:34] wie gehts? [09:41] meh and it's too early in the morning for mehs === thumper-afk is now known as thumper [09:42] and too late in the week too [09:44] it's all going to be A OK! [10:08] * Sweet5hark1 lols at the npm javascript jenga breakdown. [10:08] oh cute, I'm seeing the new slideshow [10:08] is it me or should the gnome-software screenshot not have a bit of wallpaper on the rhs? [10:08] the others don't [10:19] Laney, doh! [10:19] you're right [10:20] * willcooke fires up gimp [10:20] heh [10:20] Laney, have you got that screenshot link handy? [10:20] I can probably grep for it [10:20] VERY! [10:20] http://people.canonical.com/~laney/weird-things/g-s.png [10:20] supe [10:20] r [10:21] I don't visit many other website with "weird" in the url [10:21] hah! [10:21] I tried "laney" in mine, but nothing [10:21] * Laney is a clickbait nonclicker [10:22] "Ten things that Laney made me click -- see if those links will make you click too!" [10:23] You won't believe ten things laney doesn't click on [10:23] number 8 made me spit out my coffee! [10:23] :D:D:D [10:24] "You have just inserted a medium" [10:24] #shitubuntusays [10:24] urgh [10:25] * Laney is going to use said medium to look at the hidpi ubiquity session [10:25] back shortly [10:26] realised that I don't have anywhere to install the thing to though which is a tad annoying [10:28] can you repartition your drive? [10:29] borrow 10GB from swap for a few hours? [10:29] Laney, speaking of hidpi, plymouth upstream git now has hidpi support (but no release as yet) do we want to cherry-pick those patches? [10:31] darkxst: I know a-nothing, talk to slangasek maybe [10:31] have you tried it? [10:31] Laney, I don't have a hidpi laptop, but ricotz says it works well [10:36] Laney, darkxst, jfyi a package is locate here https://launchpad.net/~ricotz/+archive/ubuntu/staging/+packages?field.name_filter=&field.status_filter=published&field.series_filter=xenial [10:42] Laney, slideshow fix: http://bazaar.launchpad.net/~willcooke/ubiquity-slideshow-ubuntu/ubiquity-slideshow-ubuntu/revision/735 [10:42] Laney, will MP it [10:45] thx [10:46] cyphermox, sorry (again). Last one. I'm fairly sure this time ;) https://code.launchpad.net/~willcooke/ubiquity-slideshow-ubuntu/ubiquity-slideshow-ubuntu/+merge/289880 [10:46] darkxst: ricotz: Suggest you speak to slangasek, IMO it would be nice but it's not my decision [11:00] good morning! [11:03] hi desrt [11:06] willcooke: one partially successful install from pending \o/ [11:06] willcooke: now to see if oem kills it [11:06] davmor2, what changed? [11:07] willcooke: I think infinity fixed the locale/glibc changes he made I could be wrong though [11:08] *I* fixed the locale problem and he fixed the vte one which was stopping ubiquity from launching [11:08] iz gud [11:10] what's up desrt [11:12] desrt, Laney - any thoughts on this one: https://bugs.launchpad.net/ubuntu/+source/libunity/+bug/1538471 [11:12] Launchpad bug 1538471 in libunity (Ubuntu) "scope-runner-dbus.py assert failure: *** Error in `/usr/bin/python3': free(): invalid next size (fast): 0x0000000001ca17b0 ***" [Medium,Confirmed] [11:13] jamesh thinks he's narrowed it down to gir === Drac0 is now known as Guest1023 [11:45] Laney: you're awesome sir remind me to slap you less ;) [12:14] hey all [12:18] hi andyrock [12:18] bad news we got a core file [12:18] but it's empty [12:19] ha [12:19] useful [12:19] https://www.irccloud.com/pastebin/P8dvdGBx/ [12:19] handy [12:20] but this error can be releated: Fatal IO error 11 (Resource temporarily unavailable) on X server :0. [12:21] but still not sure if it's the cause or a result of the bug [12:23] andyrock, whats next to try? [12:24] hikiko mentioned that there were some broken packages as well? [12:24] yes [12:24] i'll try to save the core file and restore the snapshot [12:25] likely we get a full bt [12:25] andyrock, I've dist upgraded manually [12:25] <- out: dentist [12:25] and I see many problems here and there [12:25] I believe that if we fix broken packages/dependencies etc [12:25] the session wont crash [12:26] I'll note down the errors [12:26] and send them to cyphermox [12:26] he might be able to make more sense [12:26] andyrock, I got no logs from the gui upgrade either [12:26] mmm [12:27] because the filesystem was corrupted after the crash [12:27] it's very improbable [12:27] that compiz causes this problem especially when the bug is "can't find libc" [12:27] https://bugs.launchpad.net/ubuntu/+source/ubuntu-release-upgrader/+bug/1555237/comments/10 [12:27] Launchpad bug 1555237 in ubuntu-release-upgrader (Ubuntu) "Upgrade from 14.04.4→ 16.04 dies midway taking out the session." [Critical,In progress] [12:28] here he's able to make the session crash just upgrading few packages [12:32] could be helpful :) [12:34] andyrock, [12:34] tbh I don't understand how [12:34] it points out a problem in compiz? [12:35] how did he end up with these packages? [12:36] insserv: warning: script 'cron' missing LSB tags and overrides [12:36] I got this one too for another package [12:36] for the record: It may indeed turn out to not be a problem in Compiz (quite likely in fact, but lets see) - however, the point here is not to rule out compiz but to narrow it down to what *is* causing the problem so that we can get it fixed. [12:36] (willcooke's comment) [12:36] that looks much more relevant [12:36] to the error [12:37] I believe that we have to fix the packages anyway [12:37] what I would do [12:37] is to dist-upgrade manually [12:37] and fix every broken package [12:37] every time it stops with an error [12:38] and then try the visual upgrade [12:38] the problem is [12:38] I don't know how to do the fixes at all :/ [12:38] if someone could help on that [12:39] we could make sure that there aren't any upgrade blockers [12:39] when I perform the visual upgrade something is so broken that causes a kernel panic [12:40] and the error is that there's no root fs [12:40] without a proper fs [12:40] it's impossible to retrieve any other info [12:40] plus we ll have to fix the packages anyway [12:40] because we would like the user to be able to complete the dist-upgrase [12:40] upgrade* [12:46] well the kernel panic can be a result of a unity crash [12:46] because the unity crash can cause an X crash [12:47] so if you leave the upgrade in a bad moment you can get a kernel crash [12:48] hikiko: things do *not* stop with an error. As the dist-upgrade gets done, compiz crashes and the session gets taken down before you can finish the upgrade, and it would continue if only there was still X to display prompts [12:48] hikiko: you may use an older image to test that, or a newer one than 20160322 provided that glibc was fixed [12:49] ie. glibc is a different issue that is already being fixed [12:49] s/being// [12:49] on the latest one anyway [12:49] cyphermox, which images you suggest I use? [12:49] when I dist upgrade, compiz doesnt crash [12:49] 20160323 should be good, I think [12:50] hikiko: no, it's not *just* upgrading compiz [12:50] you also need to wait for the screensaver to kick in [12:50] so cyphermox what's the theory? [12:50] that the screensaver crashes the visual update? [12:51] and leaves the system in a weird condition? [12:51] I have no idea. compiz crashes, I upgraded manually until there was just the few desktop packages left I wrote down in my comment, because I couldn't reduce it any more than that [12:51] and in between I'd let the screensaver activate to try and get the system to crash again [12:51] I don't know enough about X or GL or compiz to know why the screensaver would change anything for compiz [12:52] is it possible [12:52] that's why we asked for help [12:52] that the screensaver [12:52] uses a library that is upgraded by one of those packages [12:52] and that's why it crashes? [12:53] that's still X or desktop. [12:53] have you tried upgrading things at all? [12:53] cyphermox, [12:54] cyphermox: X crashes as well here [12:54] if you disable the screensaver [12:54] you still have the issue [12:54] x crashes too [12:55] hikiko: sure. if you prefer you may well look into that crash? [12:55] cyphermox, I'm looking at it but it seems to me that the crash is a consequency of the upgrade crash [12:56] the upgrade does not appear to crash, what makes you think that? [12:57] I think pitti had the same prob with willcooke and me and saw a similar message for insserv and cron [12:57] cyphermox, [12:57] after the crash I rebooted [12:57] do you have a stack trace? [12:57] and got a kernel panic [12:57] no [12:57] the fs was in bad state [12:57] sure, but that's normal if the system freezes while you upgrade [12:58] my theory is that: [12:58] our goal is for the system not to freeze and the X session not to die while upgrading [12:58] either a post installation script is doing something or something is missing and makes x crash [12:58] or [12:59] a visual program crashes and stops the upgrade [12:59] but I think 1 is more possible [12:59] because I had dist-upgrade problems [12:59] plus [12:59] compiz seems to crash because it cant find libc [12:59] hikiko: we asked for help because it would be very nice if you could look into the compiz crash itself or the X crash or whatever, and see if that can be fixed -- if a postinst script tries to do things and causes a crash, it is overwhelmingly still likely to be a bug [13:00] if libc is missing [13:00] how could i fix that in compiz? [13:01] hikiko: we'll always be replacing libraries from under things. that's what upgrades do, and it wouldn't be different from what upgrading from wily to xenial would do, or vivid to wily [13:01] something is triggering that crash, and perhaps there is a way to avoid it === alan_g is now known as alan_g|lunch [13:03] as mentioned earlier, even if you disable the screensaver the whole session eventually freezes, that's verybad and if we'll need help to figure that out too [13:03] I agree [13:03] wait [13:03] you mean that [13:03] before the crash [13:04] you disabled the screensaver during the upgrade [13:04] and got a crash? [13:04] I tried both things [13:04] sorry I didn't understand that [13:04] with the screensaver, as soon as the screensaver starts, compiz crashes and I get thrown back to lightdm, and then I can't login [13:04] that's during the upgrade? [13:05] what are the steps to reproduce? [13:05] you start the update manager in compiz [13:05] then? [13:05] without the screensaver (with the screensaver disabled), the upgrade seems to continue slightly longer but eventually the session *freezes*, I can't move the mouse anymore [13:05] you just run 'sudo update-manager -d' in a unity session [13:05] that's what I do [13:05] but in my case [13:06] the screensaver doesn't seem to do anything [13:06] just [13:06] at some point [13:06] with or without screensaver [13:06] everything blacks out [13:06] and when I ssh [13:06] I see no logs and X or compiz running [13:06] and if I reboot [13:06] I get the kernel panic [13:07] and the message is that the fs cant be mounted [13:07] we explicitly can't care about what happens after you reboot [13:08] in this case, the upgrade has not completed so the system may be in any state [13:14] davmor2: want to confirm bug #1560973 / dupe it if you already know? [13:14] bug 1560973 in Ubuntu "EFI booting Ubuntu image to grub removes existing boot entries" [Undecided,New] https://launchpad.net/bugs/1560973 [13:28] cyphermox, question: [13:28] i'll try to run the upgrader without compiz [13:28] on top of metacity [13:29] the new lts has cron or uses the systemd cron? [13:29] +1 andyrock [13:31] Laney: I'll take a look in a second for you just need to finish off test the bit's I'm doing currently [13:32] ja [13:32] hikiko: I have no idea about cron [13:33] either way the usage should be the same [13:33] cron stops [13:33] before the crash [13:33] I don't know if it's relevant or not [13:34] but I got the same message as willcooke and I think pitti got it too [13:35] I don't know if it's worth investigating [13:37] muktupavels: hey, do you know if metacity is supposed to be all good with hidpi? [13:37] trying to understand https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1560162 (sorry for the inflammatory title) [13:37] Launchpad bug 1560162 in ubiquity (Ubuntu) "Xenial: scaling is horribly out on xps13 install session" [Undecided,New] [13:37] Laney: probably not... [13:38] seems like window placement / decoration drawing is not right [13:42] Laney: can you ack https://requests.ci-train.ubuntu.com/#/ticket/1157 ? [13:42] Laney: also... Hey! :) === alan_g|lunch is now known as alan_g [13:43] ok [13:43] hi Trevinho [13:43] how goes? [13:43] quite good, you? [13:43] lots of weird stuff going on [13:43] weird? [13:44] just bugs [13:44] Laney: metacity most likly need something like this - https://git.gnome.org/browse/mutter/commit/?id=57c1078ee742d9f01e80e7fe05c88adfec4b4ac3 [13:44] I have not done that because I have no HiDPI display :( [13:45] muktupavels: do they have something for window placement too? [13:45] ubiquity has GTK_WIN_POS_CENTER_ALWAYS [13:46] but you can see it's being positioned in the bottom right [13:46] you can set some keys / environment variables to use 2× scaling [13:46] * Laney will try on the installed system anyways, but after lunch [13:47] probably have to look at glib/s390x too, which still hangs [13:47] Laney: don't remember [13:50] ok [13:55] Laney: do you know what environment variables I should use? [14:09] Laney: so oddly it removed the efi entry for the first ubuntu install however it added it correctly from grub2 so now I guess I really need to test against something other than ubuntu so I'll dig out an image from somewhere [14:11] Laney: I assume the issue is that the files and path are the same [14:13] Laney: feel free to change the title of that bug to something more manageable for yourself, I just write what I see and hope it makes sense :D [14:49] Laney: hi, sorry for the trouble with ubuntu-snappy, please just seed ubuntu-snappy-cli instead of ubuntu-snappy [14:49] re from the dentist (still drugged though) === JoseeAntonioR is now known as jose [15:04] mvo: ok [15:05] davmor2: what do you mean with 'added'? [15:06] Laney: I just uploaded a new ubuntu-snappy that ensures its all working with ubuntu-snappy-cli [15:07] davmor2: note I don't have anything other than xenial installed [15:07] ricotz: abi-fixed bamf should be in proposed [15:07] Laney: so the EFI folder path and file names for Ubuntu installs are the same so it looks to me like they overwrite with the newer version installed. However there is a new entry in Grub2 for the older install which I think is expected, I'm installing fedora now and I'll do a side by side with that as then the paths should be different [15:08] Trevinho, great :) [15:08] davmor2: why is it overwriting anything if I just look at the menu and then turn the system off? [15:08] cyphermox: https://bugs.launchpad.net/ubuntu/+bug/1560973 -> grub2, or something else? [15:08] Launchpad bug 1560973 in Ubuntu "EFI booting Ubuntu image to grub removes existing boot entries and makes system unbootable" [Undecided,New] [15:08] Laney: I don't know, one for the foundations team [15:09] davmor2: yeah, just throwing it out there :P [15:09] ricotz: it should also fix your issue with new starting bamf app created with no desktop file. Although as you said you should fix plank to properly show them [15:09] you don't need to investigate that much [15:09] or launchers around [15:10] Laney: yeah, grub possibly, though I can't see why it would change anything on the hardware [15:10] Laney: no I'm just doing a side by side with another os to see if the efi entries are kept then I am assuming they will be but it is always good to know that too [15:10] Laney: either it's that or it's some freak BIOS bug [15:10] cyphermox, Laney: grub2. Grub doesn't support efi [15:11] it was more 'grub2 or some other component completely?' :) [15:11] ^ indeed, but nobody uses grub-not-2. [15:11] cyphermox: yeah, I don't have any other efi installs to check against though I'm afraid [15:11] * Laney has grub2-ified it [15:11] Laney: I should be able to reproduce on my hardware. [15:11] 'k [15:11] if not, it's a BIOS issue [15:12] good job we're friends with superm1 [15:12] haha yeah ;) [15:12] oh, wait [15:12] Laney: is that a new install? [15:12] what image did you use? [15:12] 20160323 [15:12] fun. [15:13] just looking at the grub-efi menu thing [15:13] could it be the install itself that confused your BIOS? [15:13] you know the one that appears from the install [15:13] s/install/image/ [15:13] wait [15:13] before you even get to ubiquity or anything [15:13] I mean, the system you used to write the image, was it a newly-installed system? [15:13] erm [15:13] no [15:13] ok [15:13] I just dd-ed it from my regular install [15:13] ok [15:14] there's a strong smell of BIOS, in any case [15:14] cyphermox: confirmed the initial issue if you install ubuntu along side ubuntu I'm just trying fedora23 with ubuntu alongside now to see if that behaves differently I'm hoping it does :) [15:15] davmor2: what initial issue? [15:15] cyphermox: the efi one [15:16] bug #? or are you talking about the same thing as Laney? [15:16] cyphermox: same one as Laney [15:17] man fedora takes an age to install [15:17] well, it's not the same thing if you're installing two things on the same disk side by side [15:38] andyrock, and cyphermox I check the post inst scripts that have warnings or errors to see if i find something relevant [15:38] i just managed to reproduce the crash with metacity [15:39] i told you it cant be compiz [15:39] the other things i want to check is: [15:39] 1 libc is multi arch did we changed the dirs? [15:39] cyphermox hikiko i always get the crash after udev restarting [15:40] me too [15:40] and the gvfs-common [15:40] is the last package i see here [15:40] so we must probably check its scripts [15:40] but I have to go [15:41] I ll be back later to continue investigating [15:41] I just wanted to let you know what I am looking at [15:41] maybe you get some better ideas [15:41] andyrock: crash with metacity> the session dies and you get back to the login screen? [15:42] i don't get back to loginscreen [15:42] but x died [15:42] libc looks very suspicious to me that's what I am going to check first when i come back [15:42] good enough I suppose [15:42] https://www.irccloud.com/pastebin/AVhCBoaU/ [15:42] cyphermox: ^^^ [15:42] hikiko: please show what makes you think it's libc; if that's the issue, we can deal with it by letting infinity look at it hard again [15:43] andyrock: dmesg? [15:43] because compiz couldnt find the libc [15:43] i think [15:43] at the point of the crash [15:43] the new libc [15:43] is installed [15:43] right, but what was compiz doing at the time? [15:43] BUT [15:43] do you have a backtrace? [15:43] the dynamic loader [15:44] still looks at the old directories [15:44] i am going to check if it works [15:44] in i386 [15:44] please show logs for this [15:44] when i come back [15:44] andyrock, [15:44] do you have the log [15:44] cyphermox, it was andy's [15:44] he pasted it yesterday [15:44] if it's an issue with libc's pt_chown or something, it might already have been fixed in the new image [15:45] cyphermox, sorry I haave to go I ll look at it tonight and ping you [15:45] sorry [15:45] +it's just a theory [15:45] I am not sure [15:45] bbl [15:46] mmkay [15:47] cyphermox i don't have it, i already restarted the vm [15:47] i'll try to see if I can get a bt of Xorg [15:47] because it's Xorg that's crashing with or without unity/compiz [15:48] well, I suppose that's good news [15:59] mvo: Guess this wants to wait until after beta now [15:59] but I've seeded it anyway [16:04] Laney: thanks [16:04] desrt: you here? [16:08] https://www.irccloud.com/pastebin/Wl0AcISq/ [16:08] cyphermox ^^^ [16:08] do you think it's the libc bug that has already been solved? [16:08] pitti: any chance I/you/we can reboot 'aupkg01' to get it on the -15 kernel? [16:09] Laney: sure [16:09] I'm abusing it a little bit to try to reproduce this blasted glib2.0 hang [16:09] Trevinho: at lunch. bbiab. [16:09] pitti: JFDI or something more? [16:09] stop workers I guess [16:09] Laney: if you want to be kind, pkill worker and let the current tests finish (or at least settle down a bit) [16:09] * Laney nod [16:10] Laney: actually, there's just two pyzmq tests running for 4/7 hours which look stuck [16:10] Laney: so yes, stop workers, dist-upgrade, and reboot [16:10] andyrock: no idea tbh [16:10] desrt: ok, enjoy... Let me know when you're back [16:11] cyphermox: is the bug fixed in the last image? [16:11] andyrock: it doesn't look that way, and it seems to me like writes should just work too [16:11] andyrock: yeah, the bug I was thinking of should laready be fixed in the last image (20160323) [16:31] infinity: Hi Adam, did you notice bug #1560577? If you think my latest idea (minor script modifications) is ok, I can make a real patch. Suppose that change could be accepted in Debian as well. [16:31] bug 1560577 in glibc (Ubuntu) "Confusing new locale-gen behavior" [High,New] https://launchpad.net/bugs/1560577 [16:35] GunnarHj: Try again with -0ubuntu2, that behaviour changed. [16:36] infinity: Aha, didn't know. Will do. [16:52] Trevinho: back! [16:53] desrt: so, i've slightly updated the MP for hiding window buttons... [16:53] desrt: it's at https://code.launchpad.net/~3v1n0/gtk/unity-maximized-headerbar-buttons-hide/+merge/288552, basically I'm checking whether the headerbar is also the toplevel's titlebar before hiding stuff... [16:53] desrt: but, we can't avoid to do that if we don't want a double-title/buttons [16:54] in the case where the headerbar contains controls maybe we could hide the title label itself? [16:54] I don't think that would cause any breakage since these values are still shown. [16:54] cyphermox, Laney: so after an hour and 3 minute installing fedora turns out they lvm by default thankfully I haz t'interwebz to help me modify it all. I now have ubuntu and fedora cohabiting and both entries show in uefi menu [16:55] ok... [16:55] what does this teach us? [16:55] desrt: which kind of controls you mean?= [16:55] fedora's path is EFI/fedora/shim.... and ubuntu's is EFI/ubuntu/shim... [16:55] like all the stuff that apps like gedit shove into the headerbar [16:55] we don't want to hide this stuff [16:55] desrt: in fact we don't. [16:56] desrt: only the title label is hidden and the window controls (close, maximize, hide) [16:56]