/srv/irclogs.ubuntu.com/2011/12/17/#ubuntu-devel.txt

=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
=== dendrobates is now known as dendro-afk
barryslangasek: are you still around?00:57
slangasekbarry: yep!00:57
barryslangasek: i am seeing something very weird with dbus-python tests, which i think i've tracked down to a bug in oneiric's dbus.  i'm looking for a little sanity check if you have a few minutes00:58
slangasekbarry: ok, what are you seeing?00:58
barryslangasek: it starts with this: https://bugs.freedesktop.org/show_bug.cgi?id=4330300:58
ubottuFreedesktop bug 43303 in python "dbus-python test suite failures" [Normal,New: ]00:58
barryso the upshot is that this fails on oneiric but not wheezy00:59
barryslangasek: turning on DBUS_VERBOSE=1 gives some more information about what's happening00:59
slangasekinteresting01:00
barryslangasek: what i think is going on is that dbus is finding the dbus-python's .service file *and* correctly parsing it/adding it, but then the function in dbus which does this is returning a FALSE status unconditionally, even though it should return a TRUE status01:00
barryslangasek: now, grab oneiric's source for dbus01:00
slangasekdoes the problem occur or not with precise/01:00
barryslangasek: that's a great question.  i haven't been able to set up an environment to test that yet01:01
barryslangasek: but, by looking at the code, i think the bug is obvious ;)01:01
slangasekoh01:01
slangasekwell, let's just fix it then :)01:01
barry:)01:01
slangasekis it worth an SRU?  i.e., does it break things in practice besides the python-dbus test suite?01:02
barryslangasek: i'm still trolling through source package bugs, but i don't see a matcher yet.01:02
barryslangasek: but i think it only affects .service files in non-standard locations, so it's unlikely to affect normal operation01:02
barryslangasek: fwiw, the bug is in bus/activation.c, update_desktop_file_entry().  if you look at the bottom of the function, it turns FALSE unconditionally, but it should return retval (which would be set to TRUE in this case).01:03
barryslangasek: and in fact, on sid and precise, that's exactly what it does01:03
barryslangasek: i have verified that it works on sid, but i will definitely test precise next!01:04
slangasekok01:04
barryslangasek: i don't know if the bug was upstream but i don't see any debian/patches that would affect it01:04
* slangasek nods01:04
slangasekso provided that it's fixed now in precise, it probably makes more sense to just get you using precise and not bother with an SRU for this01:05
barryslangasek: agreed. unless there's a open bug on oneiric that could be caused by this01:06
slangaseknot one that I've heard of... I guess the desktop team might've heard more01:06
barryi guess i know one thing i'll be doing on xmas break (upgrading my main desktop to precise :)01:06
slangasek:-)01:06
barryslangasek: ok cool.  anyway, thanks for letting me talk this out.01:07
slangasekno problem!01:07
barryhave a good weekend!01:07
slangasekyou too01:08
barryslangasek: well, i don't think that's the whole story, but i'm giving up for now :)01:59
=== dendro-afk is now known as dendrobates
rubyplusplusAnyone ever use the twitter bootstrap modal?  I was using it successfully, now I'm not sure what changed, but now it's showing onload02:34
=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
dnewkirk#ubuntu-app-devel05:43
=== dendrobates is now known as dendro-afk
ManDayI guess you are a better audience for this question:10:42
ManDayStarting off from a debootstrap oneiric minimal chroot, I installed things like xfce4 and lightdm (and of course, all their dependencies) and also linux-generic and CASPER into that chroot and burned it onto an USB stick (not squashfs'ed). Now when I boot that USB stick, using the initrd of the linux-generic I get dropped into busybox with the prompt "(initramfs)". Why does it not boot normally? Is10:44
ManDayit because of casper?10:44
ManDayJudging from the init in the initrd, it provides $REASON which claims "No init found".11:05
ManDayAnyone?11:08
ManDayI think I might be missing a kernel parameter, but I wouldn't know which!11:08
ManDayUnless... of course...11:09
ManDayYes, that must be it: If I don't casper-boot, I need to specify the root= explicity!11:09
ManDayYes, that works.11:19
=== bregma_ is now known as bregma
=== Guest97949 is now known as EvilJackyAlcine
=== EvilJackyAlcine is now known as Guest85587
=== Guest85587 is now known as JackyAlcine
=== dendro-afk is now known as dendrobates
=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
yogiI'm trying to load Precise Pangolin on a Dell 10 Mini (with the infamous Poulsbo chipset).  Both the alpha 1 and the daily are crashing on boot18:00
yogithe machine is locked stiff - there is a IRQ in the middle, native_smp_send_reschedule18:02
penguin42yogi: Report it then, I guess a screenshot of the oops is probably the most useful info from the crashed machine18:05
yogiUGH - attaching the screenshot crashed my Chrome with 25 tabs!18:15
yogihttps://help.ubuntu.com/community/ReportingBugs is not helpful18:17
penguin42yogi: File it against linux, and attach the screenshot, also if you can boot something older on it then run apport-collect   and the bug number just to collect info about the hardware18:20
yogiTimeout error18:22
yogiSorry, something just went wrong in Launchpad.18:22
yogipenguin42: I'm on https://bugs.launchpad.net/18:23
yogiwhat is the quickest way to submit a bug?18:23
yogiI understand the need to filter the incoming bugs somewhat18:23
yogibut there is not 'Submit a bug' button on the main page18:23
yogiI guess I should submit a ...   Nevermind!18:24
penguin42yogi: Just run ubuntu-bug linux     from something that boots18:24
yogiit asks me for admin password, on my other box18:25
yogiwhy wasn't I given the option to not attach all those log files?18:29
RiddellScottK: who care about boost now?  libboost-graph1.46-dev doesn't depend on libboost-graph1.46.118:34
* Riddell reports bug 90577218:39
ubottuLaunchpad bug 905772 in boost1.46 (Ubuntu) "libboost-graph1.46-dev does not depend on library" [Undecided,New] https://launchpad.net/bugs/90577218:39
debfxRiddell: http://bugs.debian.org/65133718:59
ubottuDebian bug 651337 in libboost-graph1.46-dev "libboost-graph1.46-dev: Dangling symlink" [Normal,Open]18:59
Riddellthat's the one19:00
debfxRiddell: does it cause a build failure?19:02
Riddelldebfx: it did cause rocs to fail in the ninjas PPA until I added it19:19
ScottKRiddell: No one, really.  ajmitch pretends he doesn't care, but deep down inside, he does.19:47
ajmitchScottK: trying to pin blame on me?21:39
* Frostbite pins a tail on ajmitch 21:40
ManDayCasper hangs for more than a minute after it says "Running init-bottom". It then says  "* Stopping configure virtual network devices      Waiting for network configuration...         Waiting up to 60 more seconds for network configuration"      I've got no idea where this is coming from, GREPing through the scripts returned nothing. Any idea what to do?21:55
broderManDay: check your /etc/network/interfaces file - it should only have a block for the lo interface21:56
penguin42ManDay: I could swear I saw a bug in the list for stuff to fix on PP along the lines of 'waiting to configure already configured network devices'21:56
ManDaybroder: Has   auto lo    iface lo inet loopback21:57
broderManDay: and nothing else? ok, i don't know then. but the waiting messages are coming from /etc/init/failsafe.conf21:58
ManDaypenguin42: I'm not certain it's actually a bug in Casper. Could be that I negelected to give it somethign that it wants21:58
ManDaybroder: Thanks, that's greatly helping me!21:58
broderManDay: it's not capser. by the time you get to "Stopping configure virtual network devices", you're past casper code21:58
ManDaybroder: It works without casper (meaning unsquashed, boot=local) though.21:59
ManDaybroder: Weird script... It just has sleeps in it...22:00
broderManDay: it's designed to delay the boot until manually configured interfaces in /etc/network/interfaces finish initializing22:01
ManDayI spot "initctl" in there. I did NOT follow the instructions of "CustomizedLiveCDFromScratch" which said something about replacing initctl22:01
ManDayThese here:22:01
ManDayhttps://help.ubuntu.com/community/LiveCDCustomizationFromScratch#Make_the_ChRoot_Environment22:02
ManDayA bit further down22:02
ManDayCould it be related?22:02
broderno. you only need to move initctl out of the way while you're building the chroot; you should undo it before you squash the fs22:03
ManDayI wasn't planning on becoming an Upstart expert :-/22:04
ManDayI guess I'll just delete those sleeps and let it fall right through22:04
ManDayI have no clue what I'm actually doing, though22:04
ManDayI did not install any of casper-lupin, discover, laptop-detect, os-prober either, as the CustomizeLiveCDFromScratch Tutorial suggests. I guess that's not a problem, either?!22:06
broderthose shouldn't matter. when you say you didn't follow the instructions about initctl, does that mean you didn't do the dpkg-divert step, or you didn't undo the dpkg-divert in the "Cleanup the ChRoot Environment" section?22:08
ManDayNeither, broder22:08
=== EyesIsServer is now known as EyesIsAsleep
ManDaySince everything seemed to work and it wasn't really clear what those steps were good for (and the tutorial is a little aged, anyway) I didn't deem them necessary22:09
broderthey are necessary to prevent random programs from the chroot getting left behind on your system22:09
broderbut they shouldn't affect the output22:09
ManDayon MY system!? outside of the chroot?22:09
broderyes22:09
ManDay**** sake22:09
ManDayyou mean running or actually persistent on disk?!22:10
broderhttp://upstart.ubuntu.com/cookbook/#run-upstart-in-a-chroot-environment22:10
broderrunning22:10
ManDayok, not a problem them22:10
ManDayfor a second I was dreaded ubuntu had shat into my perfect gentoo setup22:10
ManDay:P22:10
broderoh, if you're not using upstart outside of your chroot then yeah, it probably wouldn't matter22:11
ManDaynvm anyway, i just rememberd that i only did that on the first few tries. later on I always did that from a livecd enviroment22:11
broderand your live environment comes up once the wait in failsafe times out22:14
broder?22:14
ManDaybroder: Yeah. Goes straight into console with the "ubuntu" user that I set up, as casper should22:15
ManDayFrom thereon I can startx and everything and it works without an issue22:16
broderis the lo interface up when you get to the console?22:16
ManDayany more questions?22:16
ManDaycos I'll have to boot back into it to check22:16
ManDayso I'd rather check several things at once, if you require more info22:17
ManDayor wait,22:17
broderi guess...what's the status of the lo interface22:17
ManDayno, dont wait. For a second I thought I could use another computer, but that would complicate things22:17
broderand what does "sudo ifquery --list --allow auto" print out22:17
ManDay"ifquery" - never heard of22:17
ManDayok, ill check that22:17
ManDaysee you in a bit22:18
=== jibel_ is now known as jibel
ManDaybroder: ping22:25
=== doko_ is now known as doko
broder?22:25
dokodamn libubuntuone22:25
ManDayThat command returns   lo  eth0 eth1 eth2 ath0 wlan0    - suprisingly though, I only got ath0 and wlan0 - I don't have ethernet adapters.22:26
ManDaybroder: ^22:26
broderwell, that would be the problem. for some reason it's deciding that you need to be bringing up all of those interfaces before your networking is initialized22:26
broderyou're really sure that your livecd's /etc/network/interfaces doesn't list any of those?22:27
ManDayanother strange thing is that when I boot the FS with casper WICD (XFCE wlan client) only finds 2 networks - as opposed to a dozen when I boot "normally"22:27
ManDaythat's all a little strange with casper22:27
ManDayit better should not mess so much with my system!22:27
ManDaybroder: hold on I check again22:28
ManDaybroder: positive22:29
ManDayit says here, verbatim22:29
ManDay# interfaces(5) file used by ifup(8) and ifdown(8)22:29
ManDayauto lo22:29
ManDayiface lo inet loopback22:29
ManDay(in the squashfs' etc/network/interfaces)22:29
ManDaylet me check the initrd, too22:30
broderinitrd shouldn't matter22:30
ManDaythought so. I'll check nonetheless22:30
ManDayyep. It's not even there in the initrd22:31
broderare /run and /var/run in the squashfs empty?22:32
ManDaylatter has a symlink to former, and former is by far not empty22:32
ManDaydbus  do-not-hibernate  init.upgraded  lock  motd  network  utmp  wicd22:33
broderwhat's in /run/network?22:35
ManDayjust let me understand what it's happing: It's after the point where Casper has invoked the native init, right? That init is Upstart which walks the /etc/init and brings up services as configured.22:35
ManDayRight?22:35
broderyes22:35
ManDayempty.22:35
broderhmm22:36
ManDaySo and ONE of the services causes that delay. Which service is it?22:36
broder/etc/init/failsafe.conf is blocking the boot because it thinks you still have more network interfaces that need to be brought up22:36
ManDay"Failsafe" is the name of the service?!22:36
broderwell, it's the name of the job22:37
broderit's intended to solve a problem on servers where the system was booting fast enough that old-style /etc/init.d/ scripts would get run before the network devices were initialized -  bug #58031922:38
ubottuLaunchpad bug 580319 in upstart (Ubuntu Natty) "init.d controlled services launch before all interfaces are up, thus failing to start" [Medium,Triaged] https://launchpad.net/bugs/58031922:38
broderand you're triggering this logic because ifquery is concluding that it eth0, eth1, eth2, ath0, and wlan0 are all statically configured interfaces22:38
ManDayI'm looking at that job's config and it seems rather unconditionally waiting  - there is no IF check whatsoever. It just has those couple of sleeps in there and the config of being started upon filesystem and "net-device-up" events...22:38
broderyes, but there's the "stop on static-network-up" at the top22:39
ManDayI'm wondering how things would have to work out so that Upstart does *not* stumble across those sleep's22:39
ManDaybroder: Ah!22:39
broderwhen static-network-up gets emitted, upstart kills the job22:39
ManDayUnderstood22:39
ManDayso the problem is actually in static-network-up which does not get fired!22:40
broderright. static-network-up is fired off by /etc/network/if-up.d/upstart22:40
ManDayunhuh... I'm grep'ing like crazy ;)22:40
broder/etc/network/if-up.d/upstart is run by ifup, which is run by either /etc/init/network-interface.conf or /etc/init/networking.conf22:41
ManDayBut the real question is why Casper affects this...22:41
broderit absolutely should not22:42
brodercasper should be totally out of play at this point22:42
SpamapSugh22:42
SpamapSI think I' about 10 minutes late from saving you guys .. well.. 10 minutes. ;)22:42
broderoh look! SpamapS! :)22:42
SpamapSYou have, in fact, discovered the secret of static-network-up. Well done. ;)22:43
broderSpamapS: i knew that. my current question is why ifquery --list --allow auto is printing out 5 interfaces that aren't in /etc/network/interfaces22:43
ManDaybroder: I'll unsquash again just to be absolutly sure I'm not telling you nonsense22:43
SpamapSthe key is really not what /etc/network/interfaces says, but what ifquery says.22:43
SpamapSbroder: something is likely messing with the ifstate file22:43
broderSpamapS: i was deliberately delaying fetching up the ifupdown source for as long as i could get away with :)22:44
SpamapSManDay: whats in /run/network/ifstate ?22:44
SpamapSshould be nothing but lo=lo22:44
ManDaySpamapS: Second. I'm currently trying this on another machine to see whether the results differ22:45
ManDaySame results there.22:47
ManDaySpamapS: You mean in the squashfs, or after I booted?22:47
SpamapSManDay: at the point where ifquery --list --allow auto is printing 5 interefaces22:47
ManDaySpamapS: That would be after boot. I'll have to leave you again to check this.22:48
ManDayBe back in a second, I'll go online on another machine22:48
ManDayWhich file did you want to know again?!22:51
SpamapSManDay: /run/network/ifstate22:52
ManDayHas lo=lo wlan0=wlan022:53
ManDayifconfig only has lo and lwan0, too, if that matters22:54
SpamapSbut does ifquery --list --allow auto show more?22:55
ManDayyes, it shows those 5 I mentioned22:55
ManDayeth0 to eth2, wlan0, lo, and ath022:55
ManDaywlan0 and ath0 being two physically existance devices (wlan and bluetooth)22:55
SpamapSok, can you do 'strace -e trace=open,stat ifquery --list --allow auto' and pastebin it?22:55
ManDayi don't have strace. I would have to install that which may take a while,22:56
SpamapSits pretty small22:56
ManDayin fact, it should take so long that i'd rather do it tomorrow22:56
SpamapSand its worth it to figure out where ifquery is getting its faulty information22:56
ManDayyes but i need to integrate it into the squashfs or something22:56
SpamapSOh you're not able to write?22:56
ManDaySpamapS: Like what?22:57
ManDaywell, i could get strace from my host system22:57
SpamapSsudo apt-get install strace22:57
SpamapSno?22:57
ManDaySpamapS: no, somehow, when I boot with casper my wireless doesn't pick up my home network22:57
SpamapSAH22:57
ManDay(it picks up other networks, but only 2)22:57
SpamapScan you hang on a second I'll try and dig through the code and see what else it might be consulting than ifstate and /etc/network/interfaces22:58
ManDaylet me see whether the local strace of my gentoo host is compativle22:58
SpamapSwhat version is this btw?22:58
ManDayoneiric23:00
SpamapSManDay: can you look at /etc/NetworkManager/nm-system-settings.conf ? Are those other 3 interfaces listed there?23:00
ManDaystrace worked23:00
ManDayit opens:23:00
broderi think that's spelled /etc/NetworkManager/NetworkManager.conf these days, unless ifupdown is specifically looking at the old path23:00
ManDay/etc/ld.so.cache   /lib/x86_64-linux-gnu/libc.so.6   /etc/network/interfaces   and /var/run/network/ifstate23:01
ManDaySpamapS: I don't have NetworkManager23:02
SpamapSchar *nm_system_settings = "/etc/NetworkManager/nm-system-settings.conf";23:02
SpamapSbroder: from ifupdown. :p23:02
brodereww23:02
SpamapSManDay: weird, so thats very confusing.23:02
SpamapSanyway, I have family stuff to attend to23:03
ManDaySpamapS: well, at least we have seen where ifquery got the faulty info from23:03
SpamapSbut this may be a helpful clue as to why there are a few people out there with inexplicable waiting.23:03
ManDay/etc/network/interfaces23:03
SpamapSManDay: are all 5 interfaces listed there?23:03
ManDayyes23:03
SpamapSOH23:03
SpamapSthats the problem23:03
SpamapShow did they get there?!23:03
ManDaycasper23:03
ManDaywhat else :-/23:03
ManDaythey are not in the squashfs23:04
ManDayor...23:04
ManDayill just double and triple check23:04
SpamapSthat does make sense I suppose23:04
ManDaydont want to pass out wrong info23:04
SpamapSbut yes, that is the problem23:04
SpamapSI don't know anything about casper23:04
SpamapSbut if it is generating an interfaces file that brings up *all* interfaces, then it will result in a long wait unless they all are up instantly23:05
SpamapSI'm starting to think that we need a 'nowait' group to put interfaces in for systems that don't use NetworkManager23:05
SpamapSso they can be brought up, but not waited on23:05
SpamapSanyway, I have to run23:05
broderSpamapS: might be able to abuse the "allow" syntax for that23:05
Bert_2Hi, I would like to ask some questions concerning possible incorrect dependencies of a package, is this the right channel ?23:05
SpamapSManDay: definitely need to stop casper from making them 'auto', or the system will wait for them to come up, by design.23:05
SpamapSbroder: I intend to actually23:06
ManDayaccording to broder casper is harmless...23:06
ManDay"wouldn't do such a thing"23:06
ManDayeh, broder ;P23:06
SpamapSIt makes since that casper would do that, if it wasn't going to use NetworkManager23:06
broder.......oh god23:06
ManDayi can confirm that the file does not look like that on the squashfs23:06
broderwhere did /usr/share/initramfs-tools/init-bottom/23networking come from23:06
SpamapSBut really all we've done is reinstate the old behavior of ifup in Debian past where it would wait until 'ifup -a' finished23:06
SpamapSanyway, I have to go23:07
SpamapSgood luck!23:07
ManDaygood bye SpamapS23:07
ManDaysee you around23:07
broderManDay: http://paste.ubuntu.com/773836/23:07
broderfrom /usr/share/initramfs-tools/init-bottom/23networking23:07
ManDaybroder: Is that in the initrd?23:08
broderyeah23:08
broderi've never noticed that code before, but it is absolutely most definitely responsible for your problem23:08
broderand it's probably a bug that it's there23:08
Bert_2anyone ?23:08
ManDaybroder: Looks like that's it.23:10
ManDayCurious how that has worked in the first place...23:10
ManDayDoesn't seem to happen with the "normal" LiveCD23:10
broderManDay: installing NM is the easy solution. ln -s /bin/true /usr/sbin/NetworkManager might work as a substitute but i can't make any promises23:10
broderbecause the normal livecd has NM installed23:10
ManDayMeaning what? Does NM undo that?23:10
ManDay(Remove the IFs again)23:11
broder"[ ! -x /root/usr/sbin/NetworkManager ]"23:11
broderthat code only runs if NM isn't there23:11
ManDayOh, sorry, didn't look23:11
ManDayclosely enough23:11
ManDayOkay. Want me to file a bug on it?23:11
broderprobably. looks like that code dates back to dapper23:12
Bert_2Hi, I would like to ask some questions concerning possible incorrect dependencies of a package, is this the right channel ?23:13
broder!ask | Bert_223:14
ubottuBert_2: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) See also !patience23:14
Bert_2Alright, so I get constant question about the Belgian ID application at the local computerclub23:14
ManDayJeez broder , I'm sorry but I guess you will have to do this. Launchpad sends me arround in circles.23:15
Bert_2the BeID package that people install from softwarecenter doesn't install the requered smartcard-deamon and the driver for the smartcard-reader23:15
broderManDay: there's a link at the bottom of the wiki page it bounces you to23:15
Bert_2therefor they can plug in their readers, install the app and sit there with the app open, claiming there is not smartcard-reader connected23:15
Bert_2that's why I think that beid-gui should have dependencies on pcscd and libacr38u and a few other packages23:16
Bert_2now is this a valid argument that I should file a bug about or is this just against the "rules" of dependencies ?23:16
ManDaybroder: Got it23:16
ManDaybroder: initramfs-scripts is an integral part of casper, is it not?23:17
broderManDay: yes, but the bug here is in casper23:17
ManDayUbuntu's pastebin is a pain in the ass, just by the way...23:18
ManDayDon't use it, please23:19
ManDaybug #90582823:23
ubottuLaunchpad bug 905828 in casper (Ubuntu) "init-bottom/23networking causes 2 minute delay" [Undecided,New] https://launchpad.net/bugs/90582823:23
ManDayThank you broder23:24
Bert_2people, can someone please answer my question cause I'm going to bed soon...23:32
Bert_2oh well, I'll try again tomorrow23:40
* doko demoted python-central, one more to go for python-support ...23:52

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