[00:35] slangasek, i'm pretty sure its /run . [00:35] which is blocked on the mounting of / [00:35] but somehow on intel this works. [00:38] ie, the change that caused this failure was [00:38] /etc/init/cloud-init-local.conf [00:38] - start on mounted MOUNTPOINT=/ [00:38] +start on mounted MOUNTPOINT=/ and mounted MOUNTPOINT=/run [02:12] smoser: mmm. Remind me the intended semantics of /etc/init/cloud-init-local.conf? You don't /want/ to block the filesystem events from returning, right, you just want cloud-init-local to run only after / and /run are mounted? [02:12] smoser: did you ever try 'start on mounted MOUNTPOINT=/ and virtual-filesystems'? === iulian is now known as Guest21730 === henrix_ is now known as henrix [07:22] doko: done (subscriptions) === tjaalton_ is now known as tjaalton [07:48] hi -- https://wiki.ubuntu.com/UtopicUnicorn/ReleaseSchedule suggests Utopic is to be released today. Is that still on schedule? [07:52] wouter: yes! [07:54] okay, thanks [09:07] argh is the archive broken? [09:07] my upgrade crashed out [09:07] i've been running aptitude full-upgrade, and still have 1200+ packages to install [09:07] at least it's progress ¬_¬" [09:16] Is it too late to request some sync requests from Debian? [09:17] There are some upstream fixes that landed in Debian a few hours ago for MATE. [09:18] is that in main or universe? [09:19] i think you can still request syncs for universe packages even after final freeze [09:19] hyperair, universe [09:19] at least, it won't be impacting the iso images === alexbligh1 is now known as alexbligh === alexbligh is now known as alexbligh2 === alexbligh2 is now known as abligh === abligh is now known as alexbligh1 [12:21] cjwatson: argh, because I can't type; new one coming [12:25] nemo: you can use apport-unpack on the .crash file (see manpage) [12:28] pitti: ah. thanks [12:28] pitti: fortunately it turned out to be really really really easy to reproduce so they didn't need the bt [12:28] but I'll remember that [12:29] pitti: hm. where *is* the .crash ? [12:29] nemo: aside from that, the .crash is just a plain text file, so it's quite human readable [12:29] nemo: /var/crash/ [12:33] pitti: oh nice. yeah. I can just copy and paste this [12:33] pitti: shame I couldn't just ctrl-c from the tool, but that might just be a limitation of the gtk widget [12:34] nemo: there's a nice close button in the window title bar :) [12:36] pitti: oh, I just mean if I ever run into an ubuntu user who hits a crash in our stuff, would be nice if I could just tell them to ctrl-c [12:36] but. nbd for me, that /var/crash thing is pretty much perfect [12:37] ctrl-c is spelled "Cancel" in GTK [12:37] hm. [12:37] that's how you copy in a fair number of gtk apps [12:38] in fact, some, like gedit/pluma don't even let you use the X copy ಠ_ಠ [12:38] pitti: oh. btw, if you're curious about the crash, it is kinda funny. http://www.fnordware.com/j2k/jp2samples.html save the jp2 link to a MATE 1.8 or 1.10 desktop and you'll get an infinite series of crashes and respawns of caja [12:39] pitti: assertion in jp2 decoder - looks like they decided to move thumbnail generation into caja instead of keeping it in a separate process [12:39] presumably for performance [12:39] I mentioned it in #MATE so hopefully someone will fix it soon-ish [12:39] only way to make it stop infinitely crashing and respawning is to move the image off the desktop [12:41] I guess they could change the assertions to just aborting the thumbnail creation, but given how nutty image formats can be, thinking keeping this in a dedicated process might be safer in the long run. ah well. [12:42] nemo, I am on the MATE team. Whats up. [12:43] flexiondotorg: oh. are you in #MATE ? I mentioned it last night there and has been quiet since [12:43] flexiondotorg: but. what I mentioned above. http://www.fnordware.com/j2k/jp2samples.html [12:43] save the JP2 link to the Desktop results in infinite series of crashes of caja due to an assertion in jp2_dec [12:43] Yep. [12:44] * flexiondotorg goes testing in a VM... [12:44] flexiondotorg: I nattered on a bit in there about how screwed up image/video (and audio) decoding can be, and how this is probably best kept out of important user stuff [12:44] flexiondotorg: thumbnailer could be a dedicated process and such [12:45] nemo, Confirmed. [12:47] If you save it anywhere and view that folder it will crash Caja. Obvoiusly on the Desktop that create a burning fireball of death. [12:47] flexiondotorg: yeah. I figured anywhere would work too [12:47] flexiondotorg: I'd initially tried gdb --args caja /tmp [12:47] but, was too much of a pain to keep my main caja from taking over [12:48] that's why I dropped by here to find out how to get a trace from apport instead [12:48] nemo, apport trpped it in my VM. [12:48] yep === beisner-afk is now known as beisner [12:50] slangasek, wrt semantics of cloud-init-local. i want to block. [12:50] and i think virtual-filesystems would not block. [12:51] smoser: right; and it doesn't work so well to block both events when one might be dependent on the other [12:51] flexiondotorg: by getting a trace from apport, I mean, I couldn't figure out how to copy a trace out of apport-gtk - that's why I came to #ubuntu-devel and was told how to find the text in /var/crashes [12:51] er. /var/crash [12:52] nemo, What thumbnailer is used on your system? [12:53] flexiondotorg: ummmmm. well, this is caja 1.8.1 so I'm gonna guess that the crash means the thumbnailer is caja [12:53] flexiondotorg: but in older versions and in gnome2, it was a separate process [12:53] gnome thumbnailer or something [12:53] slangasek, but why are they dependenton another ? [12:53] other than /run not existing (which it does). [12:53] flexiondotorg: for a large directory of images there was noticeable lag, probably due to repeated process respawn [12:53] flexiondotorg: I'm guessing that's why they decided to move it in-process [12:54] flexiondotorg: 'course that makes handling bad code a bit more interesting. [12:54] flexiondotorg: a single dedicated process that could be passed urls might be a compromise [12:54] slangasek, this would make sense to me, and i wouldn't have done the thing if it didnt work in intel. [12:55] flexiondotorg: perhaps easier than ensuring all image parsing is rigorous and safe - over the years it seems image parsing is constantly a source of insanity, esp for obscure formats, or effed up ones like TIFF [12:56] flexiondotorg: and, yeah, my recollection from messing around w/ AMP back in the day is that MP3 is just as much of a pain (the headers that is). HTML is bad too - basically any format w/ a ton of implementers and a bit too broad a spec [13:05] lool: this is kind of old, but LP: #1276347 [13:05] Launchpad bug 1276347 in ubuntu-download-manager (Ubuntu) "Some files are never downloaded (while under test)" [Undecided,New] https://launchpad.net/bugs/1276347 [13:07] cjwatson: hey! I see ldlinux.c32 is now in the utopic netboot directory in archive.ubuntu.com, but according to my tests, 2 more files are also needed libcom32.c32 and libutil.c32, is it possible to also put them there? [13:08] roadmr: the Debian bug indicated that that was unnecessary because they are now found by a path search [13:08] we should dig into that if it is not in fact true [13:08] roadmr: but I cannot deal with this right now, I'm working on releasing 14.10 [13:09] please hold until later [13:10] cjwatson: oh ok, I'll test that, and no problem, I can wait. Release yay, thanks! [13:17] Any chance anyone can comment one way or the other on whether the fix for https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1274779 will ever appear in 14.04 or 14.10? [13:17] Ubuntu bug 1274779 in xserver-xorg-video-intel (Ubuntu) "[ivb] hang on pageflip (IPEHR: 0x0a000001 or 0x0a080001 depending on pipe)" [High,Fix committed] [13:36] exarkun: what fix? upstream bug says it should be reopened [13:37] ah, fixed in 3.17.. you'll have to wait for that [13:37] unless the fixes are marked for stable@ [13:46] tjaalton: It sounds like that means it won't be backported. Is my understanding correct? [13:46] tjaalton: The fix will appear in whatever release of Ubuntu packages Linux 3.17, and not before? [13:48] has anyone have issues with cryptsetup in 14.04? [13:48] uhm. 14.10 [13:54] exarkun: otherwise someone needs to backport it to 3.13/3.16 [14:04] tjaalton: Thanks. [14:54] pitti: hey, were you the one complaining about rtm for emulator? [14:54] it seems to be broken [14:55] unity-system-compositor is a defunct process here [15:02] rsalveti: yeah, we sat next to each other when I noticed :) [15:02] rsalveti: oh sorry, actually not -- that was sergiusens [15:02] pitti: yeah :-) [15:02] pitti: I think I got what is wrong [15:02] rsalveti: so I used ubuntu utoppic, that one still works (older unity or so) [15:02] rsalveti: in rtm, unity failed with some EGL_BLAH error [15:02] pitti: yeah [15:02] pitti: https://launchpadlibrarian.net/186134877/glibc_2.19-10ubuntu1_2.19-10ubuntu2.diff.gz [15:02] this is the reason [15:02] EGL_BAD_MATCH [15:03] I'm copying that over to rtm [15:03] 'zactly [15:03] rsalveti: oh, awesome === roadmr is now known as roadmr_afk === roadmr_afk is now known as roadmr === charles_ is now known as charles [15:48] slangasek: so, reinstalled and didn't check the "download updates" box, and it still downloaded a whole bunch of updates during install :P [15:48] pretty curious [15:49] hmmm [15:49] just FYI === roadmr is now known as roadmr_afk [15:49] I'll file a bug === roadmr_afk is now known as roadmr [16:26] smoser: so what is the root filesystem device as listed in /etc/fstab, for your success vs. failure cases? [16:27] smoser: I see root=/dev/vda on the kernel commandline for both, but maybe /etc/fstab differs [16:27] smoser: in particular, if ppc64el lists something UUID= ish, mountall doesn't resolve that until udev announces the device's presence, which won't happen until virtual-filesystems is emitted [16:28] slangasek, both success and failure [16:29] $ cat /etc/fstab [16:29] LABEL=cloudimg-rootfs / ext4 defaults 0 0 [16:30] seemingly no change when i 'sed s,LABEL=cloudimg-rootfs,/dev/vda, /etc/fstab' [16:31] jamespage, can the neutron-plugin-nuage neutron-plugin-opencontrail neutron-plugin-sriov-agent ceilometer-agent-ipmi be demoted? if not please seed [16:32] doko, they can indeed [16:32] be demoted that is [16:32] ok [16:32] wait. strike that comment. [16:33] well. i had failed to actualy make the change. but the change doesn't seem to have affect. === roadmr is now known as roadmr_afk === roadmr_afk is now known as roadmr [17:46] Has anybody used travis-ci.org to build debian packages ? [17:47] zerick: paulproteus has some scripts to use pbuilder at https://github.com/paulproteus/travis-debcheck [17:47] haven't played with them myself. I hear they're incredibly fragile but also work way better than you'd expect [17:48] the use case is automated testing, not actually distributing the packages to anyone. [17:53] I see [18:03] smoser: hmm ok [18:05] maybe its a rathole. [18:05] but i'm stuck on why this would be different per arch. [18:05] it could just be race condition and luck on intel i guess. [18:18] smoser: is udev running at the point it gets stuck? [18:24] slangasek, you have a suggestion on how to know that ? [18:26] smoser: hack the system to open a root shell somewhere that you can get access to it (VT?) in early boot [18:28] tedg: ping [18:29] slangasek, not terribly easily. i'm just going to add a start on started job. and sleep N, then start printing debug info [18:30] zbenjamin, Howdy === cjwatson changed the topic of #ubuntu-devel to: Archive: closed | Devel of Ubuntu (not support or app devel) | build failures -> http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of lucid -> utopic | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots: === seb128_ is now known as seb128 [18:52] smoser: so on the x86 machine you're passing 'ro' on the commandline, but not on the power one? [18:53] same on both. [18:54] passing 'rw' "fixes" [18:55] slangasek, http://paste.ubuntu.com/8644088/ [18:55] mvo_: I filed bug 1384864 about what we talked about this morning [18:55] bug 1384864 in command-not-found (Ubuntu Vivid) "Please blacklist pg_ctl" [Undecided,Triaged] https://launchpad.net/bugs/1384864 [18:55] that is with 'my-debug.conf' providing the debug output there (http://paste.ubuntu.com/8644091/) [18:59] pitti: thanks, added and I am running the extraction again [19:04] mvo_: danke! [19:05] infinity: \o/ === _salem is now known as salem_ === salem_ is now known as _salem === ara is now known as Guest78810 === Daryl is now known as Guest41002 === roadmr is now known as roadmr_afk === roadmr_afk is now known as roadmr [22:03] tedg: already solved, we talked in the coffee break about the UbuntuAppLaunch python typelib [22:21] doko: were you interested in gdb bugs as well? my vague memory served me poorly, but the "gcc" bug I mentioned was actually gdb bug #1332484. still there in utopic though, so there's nothing to backport. [22:21] bug 1332484 in gdb (Ubuntu) "GDB Demangler crashes in 14.10 and 14.04" [Undecided,New] https://launchpad.net/bugs/1332484 === roadmr is now known as roadmr_afk === roadmr_afk is now known as roadmr