ojeq@find ubuntu 1303:21
meetingologyojeq: Error: "find" is not a valid command.03:21
pittiGood morning05:48
didrockshey larsu!07:51
larsuhi didrocks! ça va?07:51
didrocksça va bien, et toi ?07:52
larsubien aussi, merci07:53
* larsu dives into css again today07:53
didrocksgood luck!07:54
* didrocks dives into why machine-id isn't recreated on boot by systemd07:55
larsuit can't even do simple things like this? We're doomed. systemd will be the end of linux. And choice!07:58
didrocksahah :)07:59
didrockslarsu: seriously, I wonder why it fails, / is mounted rw, the perms on /etc is fine…07:59
didrocksand if you run a simple wrapper later on, it will create the file without any issue08:00
* didrocks takes the open() parameters one after another to see why fd < 008:00
larsuweird. well, poettering is this ---> way08:01
larsu(at least in my irc client :P)08:01
didrockslarsu: I want to debug a little bit more beforehand08:02
didrockslarsu: but if you can get him comment on pitti and I's thread on the upstream ML, that would be appreciated :)08:02
pittibonjour didrocks et larsu !08:02
larsuhi pitti!08:02
didrocksbonjour pitti!08:03
larsudidrocks: I can't do more than you, ping him :)08:03
larsudidrocks: he usually comes online later, though08:03
didrockslarsu: where is the German connection? :)08:03
pittisimple deal -- send one beer, get one answer :)08:03
pittididrocks: ah, having fun sprinkling lots of debug statements into this code and rebooting your VM?08:04
pittione of the few cases where there's *such* a good excuse for printf() debugging  :)08:05
didrockspitti: indeed, started yesterday night, so I know why it doesn't create it in the code, and the tmpfs behavior is expected08:05
larsuprintf debugging is the best, I don't know why people complain about it all the time08:05
didrockspitti: however, the all case is that open("/etc/machine-id", O_RDWR|O_CREAT|O_CLOEXEC|O_NOCTTY, 0444); return a negative fd08:06
pittididrocks: what's errno?08:06
didrockspitti: EROFS08:06
didrocksI opened /proc/mounts just before, / is rw08:06
pittioh, heh08:06
didrocks /etc is populated with the right perms08:07
didrocksand if you use the wrapper script later on, all is fine, the file is created as expected08:07
pittididrocks: but indeed that makes little sense -- it can apparently update the file if it exists as a 0-byte file08:07
didrocks(after the boot proceeded)08:07
pitti(hence our current workaround of truncating instead of rm)08:07
didrockspitti: right, there is a fallback08:07
didrocksthen, it tests if the file exists08:07
didrocksif the file exists and is empty, -> mount tmpfs08:07
didrocksand generate + write into it08:08
pittididrocks: tmpfs to where?08:08
didrocks(thanks to the flag writable = false; in the code)08:08
pittididrocks: oh, heh; but that's still wrong for our purpose indeed, as we actually want a permanent ID08:09
didrockspitti: I guess it's working actually, because, it's also writing it as the dbus-id08:10
didrocksand then, for generating, it's picking up the dbus-id if it's coherent :)08:10
didrocksso your workaround kind of works08:10
didrocksbut kind of ugly as well :p08:10
didrockspitti: in case you didn't notice, but most of the machine-id that are generated I guess is thanks to systemd postinst script which is using the wrapper (calling exactly the same code) to stamp /etc/machine-id the first time08:11
didrocksstill, I don't see why I'm getting this errno at boot, while all seems fine on the FS…08:12
pittididrocks: yes, I know that one; but that won't get re-applied after image build and cleaning /etc08:12
pittididrocks: it's not apparmor that early during boot yet, right? (and we don't confine init anyway)08:13
pittididrocks: so does the remount rw happen before? maybe that just takes a bit of time?08:13
didrocksI didn't look at that, I didn't see anything in dmesg when getting the errorno08:13
pittididrocks: did you try with putting an all-healing sleep(1) before the open()?08:13
didrockspitti: /proc/mounts should reflect the current state, right?08:13
pittididrocks: yeah, it should08:13
didrocksah, I did use printf, not went to the sleep-hack yet08:13
didrockswill try that08:14
pittirace condition debugger's best friend :)08:14
didrocksheh ;)08:14
* pitti bows to your new systemd hacker08:14
didrocksahah, the "printfer + sleeper" hacker :p08:14
jjohansendidrocks, pitti: yep we aren't applying apparmor policy that early or to init08:14
didrocksyeah, dmesg was too quiet to blame apparmor :)08:15
didrocksthanks for confirming08:15
pittiyeah, that wouldn't make sense either08:15
pittifor now my best suspicion is some kind of race between making r/w and open, although I would have assumed the remount to be synchronous08:15
didrockspitti: http://paste.ubuntu.com/9076436/08:15
didrockssee the second part, "without machine-id"08:16
didrocks(I stripped some parts)08:16
pitti[    2.739054] systemd[1]: Started Remount Root and Kernel File Systems.08:17
pittididrocks: curious that this comes after the machine-id bits08:17
didrocksyeah, I wondered about that one08:17
pittisystemd[1]: rootfs / rootfs rw08:17
didrocksbut look at /proc/mounts reports before08:17
pittibut that comes before (but only 0.06 s)08:17
didrocksthis is just before the open()08:17
pittididrocks: oh wait, check line 4708:17
didrocksok, making sense08:18
didrocksclearly didn't spot that one yesterday night08:18
didrocks(when I added this debug)08:18
pittididrocks: keep that pastebin, makes an excellent artifact for reporting the bug08:18
didrocksso, there is another mount on / on top of the current one, which is ro?08:18
pittididrocks: maybe / is still the initramfs rootfs, I'm not sure08:19
didrocksoh, that early…08:19
pittialthough that would be weird, AFAIK it's the initramfs itself which does the pivot_root08:19
pittisystemd doesn't run in initramfs08:19
didrocksand systemd isn't in the initramfs yet08:19
didrocks(I heard there is some plan)08:19
seb128oh, that channel is busy this morning08:19
pittibonjour seb12808:20
pittiseb128: yeah, welcome to #ubuntu-init-system-hackers :)08:20
didrockspitti: ok, I think bug report time then? or should I try to look where this mounted / is coming from?08:20
pittididrocks: so the remounting is systemd-remount-fs.service, and thus happens out of process08:21
seb128pitti, wie gehts?08:21
pittididrocks: yeah, I guess so, then you can discuss how the code should get reordered to generate machine-id on a r/w root08:21
didrocksah, indeed08:21
pittiseb128: gut, danke! und Dir?08:21
didrockspitti: let me add a sleep, just to confirm it's racy08:21
pittididrocks: according to our current theory, no amount of sleep should help08:22
didrockspitti: you don't think this target was started before?08:22
pittididrocks: if it does help, we are back to "async remounting r/w"08:22
seb128pitti, auch gut, danke08:22
pittididrocks: nah, it first initializes, then starts units08:23
pittididrocks: but actually, the "Booting up is supported only when" indicates that it might not even be considered a bug08:23
didrocksyou're right, of course, [    2.326811] systemd[1]: Found ordering cycle on local-fs.target/start08:23
didrocksthe target is afterwards08:23
pittididrocks: i. e. you run into case 3), while our workaround is case 2)08:24
didrockspitti: yeah, I wonder however how this work with the empty /etc then08:24
pittididrocks: can't hurt to ask on the ML08:24
seb128pitti, didrocks, do you guys need systemd tester yet? or is that enough on the "known issue list" that you don't need more feedback yet?08:24
pittiseb128: both actually08:25
didrockspitti: starting a second thread today you mean? :)08:25
pittiseb128: we have a good deal of stuff to work on, but that by far doesn't mean that we know about every boot race or breakage with customizations08:25
seb128pitti, let me rephrase08:25
pittididrocks: the other was yesterday :)08:25
didrocksyeah, I didn't phrase it correctly :)08:26
didrocksok, let's do it08:26
seb128if I opt in, am I going to run into 10 known gotcha and waste half a day dealing with known issues?08:26
seb128e.g that wouldn't be useful work neither for me nor for you08:26
pittiseb128: ah -- no, it really should boot and work normally08:26
didrocksseb128: I guess you won't get blocked (hopefully) on anything08:26
seb128or is that supposed to be "everything works, if you have an issue it's interesting and we should debug it together"08:26
pittiseb128: unless you use things like root on NFS or server-ish packages which only have an upstart job08:27
didrocksseb128: you can even try one boot without turning the switch on definitively08:27
pittiseb128: if your desktop doesn't boot, or hardware doesn't work etc, please do tell me08:27
pittiseb128: as that's absolutely expected to work08:27
seb128do I need to install special packages or just systemd?08:27
pittiseb128: don't install anything; for the first try, just boot with init=/bin/systemd08:27
didrockshey willcooke08:28
pittiseb128: if you want to do that more permanently, you can install systemd-sysv, but it works just fine on the kernel cmd line08:28
pittihey willcooke08:28
seb128hey wik08:28
seb128hey willcooke08:28
larsumorning willcooke08:29
pittiseb128: so go ahead, do it, and be amazed of the Lennart mugshot on the framebuffer during boot!08:33
pittididrocks: ah, he thinks I was joking!08:34
seb128thanks, as much I would like to throw darts in my laptop's screen, I sort of need the screen :p08:34
pittioh, and no pets within 50 m!08:35
didrockspitti: "he"?08:35
seb128Dépaquetage de libfcitx-qt5-0:i386 (0.1.2-2ubuntu2) ...08:35
seb128dpkg: erreur de traitement de l'archive /var/cache/apt/archives/libfcitx-qt5-0_0.1.2-2ubuntu2_i386.deb (--unpack) :08:35
seb128 tentative de remplacement de « /usr/lib/i386-linux-gnu/libfcitx-qt5.so.0.1 », qui appartient aussi au paquet fcitx-libs-qt5:i386 0.1.2-2build108:35
seb128happyaron, known issue?08:35
pittididrocks: read the two lines above08:36
didrocksahah :)08:36
happyaronnope, I "though" it was handled, but could be wrong08:36
seb128happyaron, just happened with today's vivid updates08:37
seb128brb, restarting08:37
* didrocks grrrr at thunderbird freezing again08:40
happyaronhaven't received that update (by not using main archive server08:41
didrockspitti: sent, let's see how it goes, I hope my explanations are clear enough, feel free to annotate08:45
pittididrocks: ack, thanks08:45
pittididrocks: sounds clear enough to me, thanks!08:46
didrocksgreat ;)08:46
=== alf is now known as alf_
seb128didrocks, pitti, if I've a process "/bin/systemd splash" running it's that systemd is in use?08:55
seb128well in use as pid108:55
pittiseb128: yes; you can call "systemctl" to see if it works, too08:55
seb128pitti, initctl works, but I guess that does magic to redirect to systemd jobs?08:56
didrocksseb128: then, can you try systemctl status as well and report the "State:" line?08:56
seb128    State: degraded08:56
pittiseb128: no, it doesn't; if "sudo initctl" works, then you have upstart running08:56
pittiseb128: but I believe you are looking at the session upstart08:56
seb128pitti, indeed08:56
seb128k, systemd is in use then08:57
pittiseb128: systemctl --failed ?08:57
didrocksok, systemctl should have red (failed) lines, what are they?08:57
seb128why is "Degraded"§?08:57
pittiseb128: (without the ?)08:57
seb128● ifup@eth1.service   loaded failed failed ifup for eth108:57
seb128● ifup@virbr0.service loaded failed failed ifup for virbr008:57
seb128● ifup@wlan0.service  loaded failed failed ifup for wlan008:57
didrocksahah :)08:57
didrockspitti: see, I'm not alone! :)08:57
pittisudo systemctl show ifup@eth1.service08:57
seb128is that fixed with the update from this morning?08:57
pittididrocks: oh, you figure it's that same bug?08:57
didrocksI'm pretty sure08:58
didrocksseb128: no /run/network directory?08:58
pittiseb128: most likely, yes08:58
seb128pitti, http://paste.ubuntu.com/9091049/08:58
pittiseb128: do you have ls -l /etc/rc*.d/*networking*08:58
seb128didrocks, no such dir08:58
pittiseb128: sorry, my bad -- I mean sudo systemctl status ifup@eth1.service08:58
pittiseb128: ok, same bug then -- I figure you don't have any sysvinit links to networking08:58
seb128pitti, http://paste.ubuntu.com/9091050/08:59
pittithus /etc/init.d/networking doesn't get started08:59
seb128lrwxrwxrwx 1 root root 20 août  11 10:19 /etc/rc0.d/K07networking -> ../init.d/networking08:59
pittiright, as I expected08:59
didrocksseb128:   [ Didier Roche ]08:59
didrocks  * debian/ifup@.service: add a ConditionPath on /run/network, to avoid08:59
didrocks    failing the unit if /etc/init.d/networking is disabled. (Closes: #769528)08:59
pittiseb128: so, that failure is fixed with -6ubuntu108:59
didrocksyw ;)08:59
seb128didrocks, <seb128> is that fixed with the update from this morning?08:59
pittiseb128: the more interesting question is: what disabled networking in the first place?08:59
seb128yeah, I read the changelog on -changes08:59
didrocksso yeah ;)08:59
seb128pitti, not me08:59
pittiseb128: you don't care much, as you use NM09:00
pittibut on a server that would be wrong09:00
didrockspitti: yeah, same for me, I didn't remember to have done that09:00
pittiI suspect that this is some kind of upgrade failure09:00
pittimaybe ifupdown or initscripts do something funky09:00
didrocksor it's a deskop-team cabale :)09:00
pittilrwxrwxrwx 1 root root 20 Okt  5 17:59 /etc/rc0.d/K06networking -> ../init.d/networking09:00
pittilrwxrwxrwx 1 root root 20 Okt  5 17:59 /etc/rc6.d/K06networking -> ../init.d/networking09:00
pittilrwxrwxrwx 1 root root 20 Nov 17 11:00 /etc/rcS.d/S09networking -> ../init.d/networking09:00
pittiI have this09:00
pittii. e. rcS.d.09:01
seb128I don't have the rcS.d one09:01
pittiseb128: ok, so that's interesting, thanks for pointing out; I was writing off didrocks' as local breakage09:01
pittiseb128: anyway, it's entirely harmless for now09:01
pittiseb128: (the "degraded")09:01
seb128well, I'm going to wait for the update to be available and reboot09:01
pittiseb128: desktop, networking, bluetooth etc. all as it should be?09:02
seb128pitti, yeah, I didn't notice a visible difference so far09:02
seb128oh, no09:02
seb128no bluetooth!09:02
seb128no indicator and the settings list no adapter found09:03
pittiseb128: you probably disabled it in sysetm-settings09:03
seb128disabled what? the indicator?09:03
pittiseb128: upstart scripts have a bug to never actually remember the status across reboots, although they try09:03
didrocksseb128: bluetooth entirely09:03
pittiseb128: yeah, indicator bug -- it doesn't appear if you have BT soft-killed09:03
seb128no, unity-control-center sees no hardware09:03
didrocksif you enable it in g-c-c, you should get the indicator09:03
seb128well I didn't kill it09:03
seb128g-c-c thinks there is no hardware09:04
pittiseb128: output of "rfkill list"?09:04
seb1281: dell-bluetooth: Bluetooth09:04
seb128Soft blocked: yes09:04
seb128Hard blocked: no09:04
pittithere you go09:04
pittithat's the thing that the inidicator should look at09:04
pittiseb128: you can enable it in g-c-c09:04
seb128how did I do that?09:04
seb128no I can't09:04
pittihm, I can09:04
seb128g-c-c says I've no hardware09:04
seb128hum, it's acting weird09:05
pittiseb128: "rfkill unblock 1"09:05
seb128if I toggle the switch it just goes back to disabled09:05
seb128that worked09:05
didrocksseb128: does it display the indicator then?09:05
didrocksseb128: there is a bug I noticed between the indicator and g-c-c09:05
seb128didrocks, no, but I disabled that, I don't need the indicator09:05
pittiseb128: right, that's what g-c-c is supposed to do; it does that here, even if it goes back to "disabled"09:06
didrocksif you turn it on in g-c-c, it goes back like 80% to disable09:06
Laneygood day systemd fans09:06
didrocksif you use the indicator, it updates its status09:06
didrockshey Laney!09:06
pittiyeah, the indicator grays out if you disabled BT in the current session, but it doesn't appear at all if you boot with BT hard or soft killed09:06
seb128didrocks, pitti, is any of you looking to the "go back to disable"?09:06
didrocksseb128: I don't think it's systemd, I got it with upstart09:06
seb128is that bug filed?09:06
seb128didrocks, I didn't say it was, but you seem to know about it09:06
didrocksI was thinking it was blueman + indicator at the time09:06
seb128didrocks, so I was curious if it was filed/being debugged09:06
pittino, the difference with systemd is merely that saving state across reboot actually works09:06
seb128pitti, ^09:07
didrocksand I just killed blueman to see if that was the cause, it wasn't09:07
pittiwhereas with upstart it doesn't -- it always comes back enabled (at least here)09:07
seb128no everything has to do with systemd :p09:07
seb128we should still fix bugs09:07
pittiseb128: yes, absolutely09:07
pittiseb128: I filed a related bug many months ago, hang on09:07
didrocksseb128: I can open the bug, I have just vague reproducer steps though and didn't look at it in detail, just mentionned it in the bluez session09:07
seb128didrocks, seems like pitti already did ;-)09:08
pittiah, bug 112610809:08
didrocksseb128: I bet it's an indicator + g-c-c kind of thing09:08
ubot5bug 1126108 in indicator-bluetooth (Ubuntu) "Indicator disappears entirely when disabling / turning off bluetooth" [Medium,Fix released] https://launchpad.net/bugs/112610809:08
seb128didrocks, oh, I remember that one, it's a different issue09:09
pittistill the very same description/symptom, apparently this either regressed, or wasn't fixed completly09:09
seb128the one I was speaking about is that the g-c-c control goes back to disabled when you try to enabled bt09:09
didrocksseb128: you disabled the indicator and you still have it?09:09
seb128pitti, the issue was discussed a bit on robert_ancell's mp iirc09:09
didrocks(it being the bug)09:10
pittiseb128: ah; no, I didn't file one for that09:10
pittiseb128: it feels like it doesn't wait long enough for BT to appear or something such09:10
seb128didrocks, no, I disabled showing the icon, the indicator is still running09:10
didrocksseb128: I just chmod -x the indicator and killed it09:10
didrocksthe bug is still there09:10
seb128didrocks, you know you can "stop indicator-bluetooth"? ;-)09:11
seb128upstart management ftw09:11
pittiseb128: didrocks likes the blunt and heavy tools :)09:11
didrocksseb128: I'm using old school trick learnt from an ex-french guy :)09:11
didrocksso anyway, it's not the indicator which is resetting the status under g-c-c feets09:11
seb128indeed not09:12
didrocksseems it's shooting itself in its feet09:12
pittispeaking of ex-french -- seb128, hoe goed spreek je Nederlands door nu?09:12
seb128pitti, ik ben Sebastien09:12
pittiIk hou van je!09:13
seb128that's about what I learnt :p09:13
* didrocks smells some google translate under this :)09:13
* pitti looks innocent -- what is this Gogle trans-late?09:14
didrocksit's an early thingy09:15
Laneyhoe gaat het09:15
pittiLaney: goed, dank je!09:16
didrocksseb128: seems to be quite an old bug: bug #105694709:18
ubot5bug 1056947 in gnome-control-center (Ubuntu) "Cannot turn on bluetooth after turning it off from the indicator" [Low,Confirmed] https://launchpad.net/bugs/105694709:18
* mlankhorst tries reading afrikaans and giggles09:18
seb128didrocks, indeed, thanks for checking09:19
* Laney reboots with systemd too09:23
* Laney doesn't want to be left out09:23
didrocksLaney: tell us about your system state, if you are in degraded mode or not09:24
Laneyand actually I use /e/n/i on my desktop :-)09:24
didrocksold school guy!09:24
didrocksseb128: what do you think about desktoppers not using n-m? :)09:25
Laneyneeded a bridge interface for lxc09:25
seb128didrocks, be nice to old people09:26
pittiLaney: the default one doesn't do? I use LXC all the time09:26
Laneypitti: I wanted it to get real IPs on my LAN (+ avahi, etc), does it do that now?09:26
pittiLaney: ah no, the automatic virbr0 is just a
pittiLaney: anyway, it should work just fine, did it?09:27
Laneydist-upgrading first09:27
pittiLaney: err, I mean lxcbr0, not virbr09:29
pittiit doesn't even add a global IPv6 address, sheesh!09:29
Laneyoh noes, Errors were encountered while processing:09:30
* Laney looks at this first09:30
GunnarHjHi pitti!09:55
=== ara_ is now known as ara
Laneyah, good to see Lennart's face10:19
Laneyand now we're booted10:19
Laneydidrocks: so what is degraded mode?10:19
didrocksLaney: systemctl status | grep State10:19
didrockssweet, no failure then :)10:20
Laneyit took a while, might have been fscking or something10:20
Laneyno plymouth to tell me10:20
didrocksyeah, this is known10:21
didrocksotherwise, just try basic system like bluetooth and so on, and ensure it's working :)10:21
didrocksbut not having any units failing is already something10:21
Laneydon't have any fancy hardware there10:22
Laneyexcept the webcam?10:22
Laneyholy shit, cheese is segfaulting10:22
* Laney runs10:23
didrocksit doesn't here, let me upgrade gstreamer :)10:23
Laneyseems gtk-3.14-ish10:23
LaneyI get a ton of criticals10:23
Laneyseb128: confirm/deny?10:24
didrocksok, I'm on 3.1210:24
seb128Laney, xfm10:25
seb128wfm even10:25
seb128no critical10:25
Laneywith 3.14?10:27
Laneyyou capture my thoughts precisely10:29
Laneyhttp://paste.ubuntu.com/9092839/ that's G_DEBUG=fatal-criticals10:30
larsuuh oh icon view10:36
Laneyyou have clue?10:38
larsuno, but icon view is notoriously bad to find issus in10:39
larsuthis only started happening with systemd?10:40
LaneyI only tried it just now10:40
Laneycould it really be that?10:40
larsujust asking after glancing at the scrollback10:41
seb128I'm using systemd as well :p10:41
Laneywe're all so futuristic10:41
seb128Laney, does it happen if you unset gtk_modules?10:41
seb128do you have o-s enabled?10:41
Laneycriticals yes, segfault no10:41
seb128just asking, because who knows10:42
larsuLaney: the bt you linked is a critical as well. Where's the crash?10:42
Laneyit doesn't happen when I gdb10:43
Laneylemme see if apport got a dump10:43
larsuneat - seems unrelated to the cell_area one...10:44
Laneythis trace is not entirely helpful10:44
larsuhm, indeed not10:48
Laney3.14 fixes the criticals but not the crash11:08
mitya57Laney: oh, I did add Breaks, but to the wrong binary package. Thanks for fixing it!11:39
Laneymitya57: I didn't delete the old ones, you might want to do that. Wasn't going to fully understand the updates in order to work that out. :)11:40
mitya57I see, will delete in next upload.11:41
larsuI'm not an expert, but I don't think I need 114 instances of indicator-14:11
seb128larsu, lol14:26
seb128kenvandine, hey14:26
seb128kenvandine, trying to backport most of trunk changes back to rtm? do you need some help there?14:26
kenvandineseb128, done already :)14:27
kenvandinejust getting them ready so we can work on them one at a time as the bugs get approved14:27
seb128kenvandine, done with everything, we can go home? ;-)14:27
kenvandinefor ota-114:27
kenvandineyeah, there are no more bugs, ever :)14:27
kenvandineseb128, did you enjoy the MP spam? :-p14:28
seb128kenvandine, did! I bounced some back your way14:30
seb128kenvandine, we need to figure out a strategy for string changes14:30
kenvandineyes.. we do14:30
kenvandineseb128, we've let some string changes in vivid, and at some point vivid is going to rtm :)14:31
kenvandineroll baby roll!14:31
seb128kenvandine, that's fine since by that time vivid is going to have updated langpacks14:33
seb128kenvandine, which is not the case of rtm if we land one of your settings changes14:33
kenvandine2 of them i think14:33
kenvandinebut, they said we are going to update the langpacks in rtm14:34
kenvandinewe just need to coordinate that with the string changes14:34
seb128well, how do you coordinate?14:34
seb128you need the new template to be in place before generating the langpacks14:34
seb128otherwise they are not going to include the correct translations14:34
kenvandinei think olli decided the string changes would be for ota-1, we just need to make sure the langpacks get generated after landing14:34
kenvandineyeah, generate them after landing and before images14:35
seb128well, as long as it happens14:35
seb128we should probably flag all string changes14:35
kenvandinei suggested we plan all the packages that have string changes to land at the same time14:35
seb128and have a special day where we land all those14:35
seb128then do langpack exports14:35
kenvandinethat's what i suggested14:35
* seb128 ^5 kenvandine14:35
seb128didrocks, pitti, where/how do we define display-manager.service to be lightdm and what would be the way for a sysadmin to change that to e.g gdm?14:51
seb128is that an alternative still? where is the magic? ;-)14:51
pittiseb128: it hasn't changed really -- debconf and /etc/X11/default-display-manager14:52
pittiit can't change either (it's reeealy complex), as we need to stay compatible to sysvinit and upstart14:52
seb128pitti, I was just wondering what handles /etc/systemd/system/display-manager.service14:53
seb128to make it point to the right service14:53
pittiseb128: that's in {lightdm,gdm,...}.postinst14:54
pittiseb128: they check debconf and make this link point to your selection14:54
seb128pitti, ok, because I just installed xdm14:54
didrockspitti: speaking of which, again something quite incompatible with the "default distro not in /etc"14:54
seb128and selected in debconf to make it default14:54
seb128and /etc/systemd/system/display-manager.service still points to lightdm.service14:54
seb128didrocks, ^14:55
didrocksseb128: probably not patched, mind opening a bug, I can look at this?14:55
pittiseb128: I think I only tried that with gdm3 and lightdm (kdm is also patched now); right, probably it needs to learn about this, too14:55
seb128didrocks, debian has https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=75900514:56
ubot5Debian bug 759005 in xdm "xdm: Missing xdm.service, can't use with systemd" [Wishlist,Open]14:56
seb128though it's slightly different14:56
seb128https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=748668 as well14:56
ubot5Debian bug 748668 in slim "slim: Under systemd, randomly hijacks default-x-display-manager ignoring default selection" [Important,Open]14:56
didrocksseb128: yeah, a little bit, I'll look at both14:56
seb128didrocks, thanks14:56
didrocksyw, see you found desktop-work on systemd :)14:57
seb128not sure we should spend much time making random dms work14:57
seb128but I guess it's something we should fix if it's not too much work14:57
didrocksyeah, shouldn't be that much14:58
didrocksand I'll use that as a way to think about the empty /etc case14:58
didrocks(even if not applying for now)14:58
=== m_conley_away is now known as m_conley
maxbI have a weird issue with CUPS (Printers added with the lpadmin command line produce strange errors on system-config-printer). Is this a sensible channel to look for help in filing a useful bug, or is there a better one?15:40
seb128tkamppeter, ^15:42
larsumaxb: best place is probably #openprinting on this server. What are the errors you're getting?15:42
larsuseb128: beat me :P15:42
maxbAttempting to access the properties page for a device added with lpadmin results in "There was a problem connecting to the CUPS server.". Running system-config-printer with --debug doesn't seem to show anything that would indicate the problem's source15:43
larsumaxb: is CUPS running?15:44
maxbYes; devices added using the system-config-printer UI can have their properties pages viewed fine15:45
maxbStrangely, using "Duplicate" on a problem device in the UI produces a copy which does not experience the same problem. I've attempted to deduce what the difference is by studying the contents of /etc/cups/ but have been unable to locate an issue15:45
larsumaxb: very weird. Please try to talk to tkamppeter or twaugh (here or in #openprinting)15:46
tkamppetermaxb, if you are able to print on your lpadmin-generated print queue and the only problem of it is to access its "Properties" window on system-config-printer, then report a bug on Launchpad, for the system-config-printer package. Tim Waugh will see the report and ask you further questions for debugging it.15:54
maxbtkamppeter: printing doesn't work either, I was just using the properties page as something easy to reproduce15:57
tkamppetermaxb, what error message do you get when you print something from the command line, using the "lp" or "lpr" command?15:59
maxbHmm, that's interesting, printing from "echo foo | lpr" worked, whilst printing from Chrome just waited indefinitely with nothing happening16:00
tkamppetermaxb, in which stage does Chrome's printing hang? When opening the print dialog or when clicking "Print" inside the print dialog?16:06
maxbAt the point of clicking the final "Print" button. It doesn't hang as such in that it can still be closed, but nothing happens even when waiting for quite some time16:07
tkamppeterDoes the print dialog close when clicking on "Print" and the job does not come out or does the dialog stay open looking like CUPS is not answering to the click on "Print"?16:08
maxbPrinting from gedit works, if that helps narrow anything down16:08
tkamppetermaxb ^^16:09
maxbThe second - dialog is mostly greyed out after clicking print, apart from the "Cancel" button16:09
maxbThat was with Chrome's custom application print dialog. I just tried the same using "Print using system print dialog" - the dialog closed after clicking Print, and the job appears to have never been submitted16:11
tkamppetermaxb, is it only Chrome which does not print or are there other applications?16:11
maxbJust tried Thunderbird - that worked16:12
maxbHmm, a second try of Chrome-using-system-dialog seems to have worked, so maybe the problems are confined to apps attempting to query cups about printers in some way16:15
maxbOh great, the behaviour of my system appears to have changed.16:18
maxbNow I get "Unable to get queue details. Treating queue as raw." when opening the properties window for BOTH the UI-configured device and the CLI-configured device16:18
tkamppetermaxb, please report two bugs on Launchpad, one on system-config-printer for your problem with the "Properties" dialog, a second one on cups for the problem with Chrome (add chrome-browser as a second task). To both add an appropriate error_log as described on https://wiki.ubuntu.com/DebuggingPrintingProblems, section "CUPS error_log". For creating the one about the print dialog of Chrome, follow also the instructions of "R16:18
tkamppetereporting Bugs".16:18
* maxb will give it a go. Currently CUPS is failed to save settings after I tick the "Save debugging information for troubleshooting" box.16:21
maxbPerhaps a reboot would be wise at this point16:22
maxbThe good(?) news is that after a reboot the bug is back to the original presentation and I am able to turn on the debug options16:29
maxbErm, the excitement continues, apparently I'm out of inodes because my /tmp is full of millions of symlinks to nonexistent PPD files16:49
maxbtkamppeter: Apparently "There was a problem connecting to the CUPS server." means "I can't open the PPD file as user maxb because it's mode 750 root:lp"16:58
maxbAnd this is the issue for Chrome too16:59
desrthi hi desktopers17:00
seb128hey desrt, how are you? had a good few days off work?17:04
seb128didrocks, pitti, that systemd update fixed the failed ifup units issue, confirmed after an upgrade/reboot, thanks17:04
didrocksseb128: nice!17:08
didrockshey desrt :)17:08
mitya57bregma: are you taking care of lp:~albertsmuktupavels/compiz/support-libmetacity-private-3-14?17:09
didrockspitti: ok, I have another proposal (upstream) for the alternatives in systemd. I'll wait to hopefully get some answers on the current one maybe first17:09
bregmamitya57, yes indeed, it's awaiting a ci-train silo assignment17:09
mitya57ok, thanks17:09
didrockspitti: I'm happy to expose it to you though :)17:09
Sweet5harkclang plugins are awsome.17:24
Sweet5hark... if you are a bit morbid, it finds you bool ? true : false statements containing >60 individual statements including other conditional operators. Or assignments that contain other assignments and also are containing more than 50 individual statements. yay for sideeffects ....17:28
didrockspitti: actually, my proposal is implemented in systemd looking at the code, so I guess it makes sense! I will just propose a way we change how we handle this to be more declarative17:40
pittididrocks: which one now? the want.d/ or machine-id, or something else?17:45
didrockspitti: how we do handle alternatives, like display manager, I need to go out soon, but I'm happy to talk to you about it tomorrow morning (finishing up some end of day cleanup tasks)17:48
seb128oh, just a fyi I've a day off tomorrow17:49
seb128in case somebody wonders why I'm not around17:49
seb128I'm back on friday17:49
pittididrocks: yeah, me too; just came back quickly upon an SMS :)17:49
pittiseb128: enjoy!17:50
seb128pitti, thanks!17:50
didrockssee you on Friday seb128! :)17:51
seb128didrocks, have a good evening!17:51
Sweet5harkseb128: have fun! (and when you are back, consider uploading the libreoffice pkg for vivid) ;)17:51
Sweet5harks/consider/please consider/17:52
seb128Sweet5hark, hey, you didn't reply to my ping from yesterday about the version bump, did you?17:52
seb128Sweet5hark, upload is blocked on that to happen on your side17:52
seb128Sweet5hark, or did you update since and I missed your ping about it?17:52
Sweet5harkseb128: hmm, must have missed that. whats missing?17:53
seb128Sweet5hark, the version you put up for sponsoring is already in utopic-updates and it can't exist with different builds in different series17:54
seb128Sweet5hark, we can't copy the binaries because of the libpoppler abi change in vivid17:54
seb128Sweet5hark, so you need at least to bump the version to 0ubuntu217:54
Sweet5harkseb128: ah, ok. willdo.17:54
Sweet5harkseb128: have fun on your day off!17:55
seb128Sweet5hark, thank you!17:55
=== alan_g is now known as alan_g|EOD
* willcooke -> EOD18:11
willcooke*mic drop*18:11
=== m_conley is now known as m_conley_away
RAOFdesrt: Hey, dude. Where's the place to go to propose a new desktop item field?23:11
desrtxdg list23:11
desrtwhat did you have in mind?23:11
desrti might be able to save you some time by telling you why you won't get what you want :)23:11
RAOFI want a Needs-GPU: field.23:15
RAOFWith... some contents.23:15
desrtyou're not going to get that :)23:15
desrtwhy do you want it?23:16
RAOFHybrid graphics23:16
desrtyou might be interested in TryExec=23:16
RAOFNo, that's not going to help.23:16
desrtit's the name of a utility program to run to decide if the desktop file in question should be displayed or not23:16
RAOFThe shell wants a way to know whether or not to start this application on the discrete chip or whether the integrated chip is ok.23:16
desrtthat's a bit more interesting.23:17
RAOFeg: Games always want the fastest chip, totem won't care.23:17
desrtdoes it affect the libGL that gets linked or something?23:17
RAOFImplementation dependent.23:17
RAOFBut on the free stack it's basically setting the DRI_PRIME environment variable.23:17
desrtmy answer to you would be to figure it out at runtime23:17
RAOFHow can the shell figure this out at runtime?23:18
desrtnot the shell -- the just-started app23:18
RAOFHow does the just-started app respect your powersaving prefernces?23:18
desrtpresumably we would have some library component in there responsible for the decision making23:18
RAOFThis sounds super-complicated vs Wants-GPU: Fastest / Wants-GPU: don't care23:19
desrtconsidering that the general trend right now is moving over to using dbus activation to launch apps via systemd, the shell isn't going to be able to control the environment that this happens under anyway23:20
desrtalso: it wouldn't only be the shell in any case... imagine an app that gets started on account of a file association from nautilus (or firefox, or anything else...)23:20
RAOFIncidentally, why is the general trend towards starting apps via dbus activation over systemd?23:21
desrtfor a lot of reasons23:22
desrtmostly because apps are increasingly being seen as services, only one feature of which is "please launch now"23:24
desrtother features might be "please do something appropriate in response to this push notification" or "please reply to the shell's search query"23:24
desrtanother reason is because it means that apps are always started in a clean environment, as a direct child of the session manager, able to be monitored and in their own cgroup23:24
desrtinstead of as a fork() off of whatever random app wanted to open text/plain without really caring who got called23:24
RAOFThis has a certain charm.23:24
RAOFBut what I'm hearing here is that I want to patch systemd :)23:24
desrti think what really needs to happen is that the app needs to tell the system on startup (ie: at the time of initialising GL) "i want the big one"23:25
desrtalthough i understand that the our diverse landscape of drivers from not-entirely-cooperative vendors might make that difficult to do23:25
RAOFI always love boil-the-oceans solutions :)23:25
desrtwell, eventually the ocean will boil23:26
desrtthese things can sometimes happen pretty quickly23:26
desrtmeanwhile, maybe use a wrapper script or something23:26
desrtExec=hardcore-graphics /path/to/original/exe23:26
RAOFThat doesn't really satisfy my goal of “this information should be upstream”23:27
desrtwhere that hardcore-graphics thing consults [whatever] and sets environment variables as needed23:27
desrtya... you gotta boil the ocean for that :)23:27
desrtbut seriously, though.... how do we handle this?23:27
RAOFWe don't, at all.23:27
RAOFHybrid graphics remains the preserve of the nerdy.23:28
desrtyou have one thing called libGL23:28
desrtand it is what it is23:28
RAOFActually... I'm not sure that's true. I think the *proprietary* drivers actually have some semblance of reasonable user interface. But I've not used them.23:28
desrti wonder if something like libepoxy could help23:30
RAOFWell, the “we have one thing called libGL” (a) isn't a problem for the free stack, and (b) is getting fixed by a new Linux GL ABI.23:30
desrtthat's really more of a consumer-side helper, though23:30
desrtbut this sort of pointer trickery is what i have in mind23:31
desrtGL people love that crap23:31
desrtyou can switch out the entire guts of the implementation on them at runtime :)23:31
desrt...at least it would help to substantially reduce the amount of manual indirection you are responsible for maintaining...23:33
RAOFSo, design goals: The app should be able to say “give me the awesome one”, the desktop should be able to determine which one the app gets.23:34
desrti think the way the app should do this is with a call to some function early in main() before GL is initialised23:34
desrtthis call might well be setenv()23:34
desrtit's certainly not -too- difficult to imagine a fake libGL that connects you with a different real implementation under it depending on some random things determined at runtime23:35
desrtyou could even make it based on ifuncs...23:36
desrt(although those are known to have issues with environment varaibles and they might be 'too early' to be useful to you)23:37
RAOFI guess it's much easier if you drop the “shell should have control” requirement.23:45

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!