[00:32] <cwayne> hmm, what actually calls system upstart on boot?
[00:35] <rsalveti> cwayne: initrd
[00:36] <cwayne> rsalveti: so if i needed to modify it (to look in /custom for jobs), what would I need to change?
[00:36] <cwayne> initramfs-tools-ubuntu-touch?
[00:38] <rsalveti> cwayne: what kind of jobs?
[00:38] <rsalveti> initrd only mounts everything up, including bind-mounts, and start upstart
[00:38] <cwayne> rsalveti: basically trying to do this: https://bugs.launchpad.net/savilerow/+bug/1315060
[00:40] <rsalveti> hm, didn't even know we wanted such things :-)
[00:40] <rsalveti> but makes sense
[00:41] <cwayne> rsalveti: yeah, also its good to have the option in case we need it :)
[00:42] <cwayne> but yeah, I'm not 100% clear on what changes we'd actually need to make to get it hooked up in the image
[00:42] <rsalveti> cwayne: initramfs-tools-ubuntu-touch
[00:42] <rsalveti> but in your case you actually want to change init
[00:44] <rsalveti> check the ubuntu-touch-generic-initrd deb, you'll have a init script in there
[00:44] <rsalveti> there's where the magic happen
[00:45] <rsalveti> trying to remember who provides that, maybe initramfs-tools
[00:45] <rsalveti> because basically it gets the touch script from initramfs-tools-ubuntu-touch and generates a standard initrd
[00:47] <rsalveti> yeah, initramfs-tools, /usr/share/initramfs-tools/init
[00:49] <rsalveti> but it's probably better to support such additional arguments by changing kernel cmdline
[00:49] <rsalveti> bbl, dinner
[00:49] <cwayne> hmm okay, i'll take a look and then probably bother you tomorrow with more questions :)
[01:35] <liuxg> when I am trying to chroot in sdk, I get the error like: E: Invalid Release signature (key id 40976EAF437D05B5). the detailed descript is at https://bugs.launchpad.net/ubuntu/+source/android/+bug/1343039. Does anyone have any clue on this?
[01:39] <sarnold> liuxg: I wonder if you need to run click chroot create ... by hand?
[01:50] <liuxg> sarnold, thanks! how can I do that? it is really troublesome.  I have been struggling it for a while.
[01:53] <sarnold> liuxg: I'd try something like: click chroot create -a armhf -f ubuntu-sdk-14.10
[01:54] <liuxg> sarnold, thanks. I will have a try very quickly
[01:55] <liuxg> sarnold,  click chroot create -a armhf -f ubuntu-sdk-14.10  AND click chroot create -a armhf -f click-ubuntu-sdk-14.10-armhf both give me errors.
[01:56] <sarnold> liuxg: can you pastebin them?
[01:58] <liuxg> sarnold, https://pastebin.canonical.com/113672/
[02:01] <sarnold> liuxg: aha. okay. looks like -a expects i386 or amd64
[02:03] <liuxg> sarnold, which one is supported so far? i386 seems the one.
[02:04] <liuxg> sarnold, https://pastebin.canonical.com/113673/ same error
[02:07] <sarnold> liuxg: darn, sorry. :( I don't know what to try next.
[02:08] <liuxg> sarnold, anyway, thank you for your reply.
[04:24] <lotuspsychje> morning
[07:41] <JoshStrobl> popey: just got my Ubuntu Pioneer t-shirt. Gonna have the wife take a pic.
[08:11] <JamesTait> Good morning all; happy Peach Ice Cream Day! :-D
[08:18] <popey> JoshStrobl: yay!
[08:24] <JoshStrobl> popey: uploaded it to my G+
[08:25] <JoshStrobl> https://plus.google.com/108419594942248098225/posts/c8r9QGLkTQZ
[08:28] <popey> \o/
[08:47] <mandel> seb128, morning! one question, where do I file a bug for ubuntu system settings (the upgrade page does not do the rotation correctly => http://youtu.be/Q0X3YJTp1iE)
[08:47] <mandel> popey, ^^
[08:48] <popey> https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+filebug
[08:48] <mandel> The credentials not found string is placed on top of the download progress bar
[08:48] <seb128> mandel, what popey said
[08:48] <seb128> we have a bug open about inconsistant rotation between panel, but none about that specific issue I think
[08:50] <mandel> seb128, I'll ping gatox about it, he should be the one taking care of it
[08:52] <seb128> mandel, thanks
[08:52] <popey> seb128: I'm getting bug 1337200 every morning, and Pat asked for a trace, do you have any specific instructions because the ones on the wiki seem to be more related to crashing processes
[08:53] <popey> ooh, we have strace in the image, that does seem better.
[08:54] <popey> mvo_: thanks for the strace tip ☻
[08:56] <mvo_> popey: your welcome! I hope it helps finding the issue :)
[08:56] <seb128> popey, where did he ask for a trace?
[08:56] <seb128> on the bug he asked for a gdb backtrace
[08:57] <seb128> https://wiki.ubuntu.com/Backtrace#Already_running_programs
[08:57] <mvo_> seb128: I suggested that
[08:57] <mvo_> seb128: to get a first hint what is going on
[08:57] <seb128> mvo_, k
[08:57] <popey> as I said, the wiki is more detailing info about crashing programs
[08:57] <popey> this isnt crashing
[08:57] <seb128> popey, no, it's not, see that wikipage
[08:58] <popey> i did, and am
[08:58] <seb128> "If the program hangs but doesn't crash you can press ctrl+c in gdb while the program is frozen and then continue with the next step. "
[08:58] <seb128> that applies to normal run as well
[08:58] <seb128> you can ctrl-c, do a bt to see where it's stopped, what's doing
[08:58] <seb128> then "continue"
[08:58] <popey> it doesnt hang
[08:58] <seb128> then ctrl-C again
[08:58] <popey> or crash
[08:58] <seb128> well
[08:58] <seb128> it doesn't need to
[08:58] <popey> hence me dismissing that bit ☻
[08:58] <seb128> the bt tell you what it was doing when you pressed ctrl-C
[08:58] <popey> ok, will try that.
[08:59] <seb128> if you take a few bt you can see if it's always in the same code
[08:59] <popey> surprised I seem to be the only person seeing this.
[08:59] <popey> happens every day.
[09:07] <seb128> popey, you are probably the only person sitting on u-s-s open ;-)
[09:07] <popey> i guess
[09:20] <mandel> popey, what about this bug where when you tap the lock screen you get a bouncy dock => http://youtu.be/qT_1Ehq8Ssw
[09:21]  * mandel is trying to work but keeps findings little bugs that annoy him
[09:36] <popey> mandel: i thought that was intenional
[09:37] <JoshStrobl> popey: yea that seemed intention to me too
[09:37] <mandel> popey, really? 'cause I though that if you tap on the lock screen you can change the info it displays, messages, images etc..
[09:37] <popey> it looks to me like a hint
[09:37] <mandel> so if you tab to see diff info and that bounces..
[09:37] <popey> to let you know where the launcher is
[09:38] <popey> ---> design
[09:38] <mandel> popey, I can always double check with thm
[09:49] <mhr3> popey, ping? is there an official test plan for the clock app?
[09:49] <popey> mhr3: which clock app, the old one or the new one?
[09:49] <mhr3> popey, the old one is still the default, right?
[09:49] <popey> currently yes, but it will be replaced by the new one
[09:50] <popey> which is currently in the store as "Clock Reboot"
[09:52] <Transfusion> hi, i'm following the porting guide and two questions... isn't android 4.4.2 supposed to be based on CM11?? and I'm getting fatal: remote github not defined in ...manifest.xml. I tried adding a <remote /> there, but after a repo sync I don't see my vendor/ or device/[codename] tree anywhere
[09:53] <mhr3> popey, we just want to make sure that our change to u1db doesn't break everything
[09:53] <mhr3> popey, so if there's a test plan for it, would be useful
[09:54] <popey> i dont think we do have one, other than running the AP tests.
[09:58] <popey> jdstrand: latest click reviewers tools branch now broken because lxml is needed? installed python-lxml and it still complains.
[10:12] <Transfusion> where can I see a sample manifest and roomservice.xml ?
[10:12] <Transfusion> i must be doing something wrong if i can't get repo sync to clone the kernel repo into the root directory
[10:40] <Transfusion> is ubuntu touch based on AOSP or CM or a mix of both? why do I fetch the aosp code and then use the CM device tree
[10:45] <asac> alecu: hey. browser activity seems to have no site previews/surfaces
[10:45] <asac> alecu: known issue?
[10:46] <asac> the entries there are just grey boxes on #133
[10:46] <asac> davmor2: any idea?
[10:46] <asac> alecu: also, how am i supposed to add a bookmark? seems there is no star UI element etc.
[10:47] <asac> alecu: ok found the star feature inside activities... very hard to spot imo
[10:48] <asac> is that the final design we are shooting for here?
[11:08] <JoshStrobl> Transfusion: I imagine it is because CM is based on AOSP.
[11:09] <Transfusion> ahh, that somehow never occured to me.
[11:09] <Transfusion> i forgot ubout ubuntu touch not using most of the java portions of CM
[11:10] <JoshStrobl> Transfusion: We can all be thankful for that :D
[11:12]  * rickspencer3 updates phone
[11:13] <rickspencer3> messaging indicator is a bell now?
[11:13]  * rickspencer3 shrugs
[11:24] <popey> rickspencer3: yeah
[11:30] <Laney> it makes me think there's an alarm set
[11:30] <Laney> "huh, the menu is empty"
[11:34] <cwayne> and it turns green now instead of blue
[11:35] <popey> \o/
[11:37] <cwayne> but the LED flashes blue
[11:37] <cwayne> lol
[11:43] <rickspencer3> popey, any idea what the bell represents?
[11:43] <popey> notifications
[11:43] <rickspencer3> ah
[11:43] <rickspencer3> so, not just messages anymore
[11:43] <popey> "Ding a ling" your dinner is ready
[11:44] <rickspencer3> or "There's an Update" :)
[11:44] <rickspencer3> ?
[11:44] <popey> dunno, i only see sms in there
[11:46] <popey> rickspencer3: matches the bell in G+
[11:46]  * rickspencer3 nods
[11:46] <rickspencer3> interesting
[12:13] <dpm> ogra_, pitti, I was talking to a Korean translator yesterday, and it seems that while we're including Korean translations in the image, they cannot be displayed because of missing Korean fonts. Installing the 'fonts-nanum' package fixes it. I'm not sure if there is any other language dependency required, but I tested it myself and only that package was needed. Could we add it to the touch seeds?
[12:18] <ogra_> dpm, hmm, shouldnt essential fonts be part of the langpack deps ?
[12:19] <Laney> I think the seeds are better, langpacks shouldn't be enforcing choices like that
[12:19] <dpm> ogra_, I'm not sure, that's why I pinged pitti too. I know they used to be installed as part of language-support-* packages, but we dropped them a while ago in favour of language selector doing the installation IIRC
[12:20] <jdstrand> popey: you need python3-lxml
[12:20] <pitti> no, langpacks don't depend on fonts, as we want the fonts without the langpacks
[12:20] <pitti> so syncing the seeds with the desktop (i. e. include the missing ko font) is right
[12:20] <jdstrand> that should've been in debian/control
[12:20] <pitti> dpm: ^
[12:20] <popey> jdstrand: ah, doh! thanks
[12:21] <dpm> ok, thanks pitti. So ogra_, do you need me to file a bug or something to get the 'fonts-nanum' included in the seeds? ^
[12:21] <jdstrand> popey: so, we got so behind on hook checks that I implemented a check to warn if there is an unrecognized hook
[12:21] <ogra_> dpm, yeah, assign it to me so i dont forget
[12:21] <Laney> dpm: is korean the last entry in the u-s-s language list?
[12:21] <jdstrand> popey: that should help us stay up to date
[12:21] <Laney> I was just wondering what that one is - see #ubuntu-desktop ;)
[12:21] <popey> cool
[12:21] <dpm> Laney, indeed, it's _that_ unreadable one :)
[12:21] <Laney> grand
[12:22]  * jdstrand was shocked to see so many new hooks the other day
[12:22] <Laney> dpm: I'll just seed it, don't bother with a bug
[12:22] <dpm> ogra_, cool, which project should I file the bug against?
[12:22] <jdstrand> I guess they weren't that new, but without the check telling me they were, I never saw them :)
[12:22] <ogra_> ubuntu-touch-meta
[12:22] <Laney> but it would be ubuntu-touch-meta
[12:22]  * Laney peers at ogra_ 
[12:22] <ogra_> :)
[12:22] <Laney> are you giong to review achiang's branch?
[12:23] <ogra_> Laney, if i ever find the time ...
[12:23]  * ogra_ is fighting with system-settings 
[12:23] <Laney> ...
[12:23] <Laney> you what
[12:23] <ogra_> i try to implement the Develeoper Mode Ui
[12:23] <ogra_> http://people.canonical.com/~ogra/devmode/
[12:23] <Laney> interesting
[12:24] <Laney> is it designed?
[12:24] <ogra_> yep
[12:24] <ogra_> https://wiki.ubuntu.com/AboutThisDevice#developer-mode
[12:24] <Laney> k cool
[12:24] <ogra_> the QML is done ... just need toget it to talk to dbus ...
[12:25] <Laney> many examples of that in u-s-s already
[12:25] <ogra_> yeah, i try to get around having to write cpp though :)
[12:26] <Laney> good luck with that ...
[12:27] <ogra_> well, QDbusActionGroup looks promising :)
[12:27] <dpm> ah, thanks Laney, I hadn't seen your comment re: the seeds update for Korean. Ok, then I won't bother with the bug and I'll leave it in your hands
[12:28] <Laney> ogra_: only if you're trying to poke a menu model
[12:28] <ogra_> oh, i thought i could get states too :(
[12:29]  * ogra_ just needs to query a boolean from dbus
[12:29] <Laney> not in general, states of GMenuModel actions exposed over dbus yes
[12:29] <Laney> e.g. the indicators do this
[12:29] <sergiusens> ogra_: just create the properties as I showed you and fill the value with the query to the bus
[12:30] <ogra_> sergiusens, yeah, thats what i will do ... i was just hoping there was a way to do that directly from QML without having to hack up cpp backends
[12:31] <Laney> the about panel already has a backend so you won't have to start from scratch in any event
[12:39] <dpm> Elleo, are you the best person to ask questions about the OSK? A Korean translator was asking how to get a Korean keyboard layout in Ubuntu, and I'm wondering if there is any documentation about adding new keyboard layouts to ubuntu-keyboard
[12:41] <pngo_> why location detection and GPS features activate by itself? how can I disable it permanently?
[12:42] <Elleo> dpm: iirc there's not really any documentation at the moment; but the process is relatively simple in that the keyboard layouts are just QML
[12:43] <dpm> Elleo, ok, thanks. Do you have perhaps a MP I can point him to, that shows how a layout was added in the past, and which files are required?
[12:43] <Elleo> dpm: basically they'd just need to copy one of the existing keyboards from plugins/$langid and change the characters, then add it to the list in language menu and keyboardcontainer
[12:43] <ogra_> pngo_, known bug,, you cant yet ...
[12:44] <Elleo> dpm: plus some change to let system settings know about it (I'm not familiar with that side of things though)
[12:44] <Elleo> dpm: but if they can put together the layouts I'm happy to handle any of the other stuff involved
[12:44] <pngo_> ogra_: thnx
[12:45] <Elleo> dpm: the layouts are basically like this: http://bazaar.launchpad.net/~phablet-team/ubuntu-keyboard/trunk/view/head:/plugins/en/qml/Keyboard_en.qml
[12:45] <Elleo> dpm: oh, unless Korean requires something special like pinyin, where characters are created out of multiple key presses, then things get a lot more complicated
[12:46] <dpm> Elleo, it will need an input method too, yes
[12:46] <alecu> asac: pong. You asked me a few hours ago about the browser activity, but I've never worked on that. Perhaps you were thinking of somebody else?
[12:46] <Elleo> dpm: pinyin is are only example of that sort of thing at the moment, so that'd probably be the best starting point: http://bazaar.launchpad.net/~phablet-team/ubuntu-keyboard/trunk/files/head:/plugins/pinyin/
[12:46] <dpm> but I was trying to find out if first the basic layout could be created, as I know input methods are rather complex
[12:49] <Elleo> yeah, well for just the basic layout if they make a Korean version of all the QML files here: http://bazaar.launchpad.net/~phablet-team/ubuntu-keyboard/trunk/view/head:/plugins/en/qml/ I can put everything else together for them and at least get them a skeleton for producing predictions (they'll need to implement the actual korean side though, or point to a lib that can do it for us like libpinyin)
[13:01] <rickspencer3> loving the in call indicator
[13:03] <nik90> dednick: ping
[13:07] <dednick> nik90: yo
[13:07] <nik90> dednick: hey I was trying out the dbus connect() to listen for property changes in dbus, here is what I did http://paste.ubuntu.com/7808936/
[13:08] <nik90> dednick: I don't see my onSettingsChanged() function called even after I changed the property using dconf-editorp
[13:09] <dednick> nik90: i don't see any PropertiesChanged signal on the Alarmproperties interface
[13:09] <dednick> nik90: there is one on the org.freedesktop.DBus.Properties interface though.
[13:10] <nik90> dednick: it is in the org.freedesktop.DBus.Properties interface which is in the AlarmProperties
[13:10] <nik90> dednick: ah so I should use that interface
[13:10] <dednick> connection.connect(service, path, interface)
[13:10] <dednick> nik90: ^
[13:10] <nik90> dednick: but is my declaration of the onSettingsChanged(QVariantMap) correspond correctly to the signal definition.
[13:11] <nik90> dednick: I got that from https://bazaar.launchpad.net/~phablet-team/telephony-service/trunk/view/head:/libtelephonyservice/callmanager.cpp
[13:11] <dednick> nik90: no
[13:12] <dednick> nik90: onPropertiesChanged(QString, QVariantMap, QStringList)
[13:13] <nik90> dednick: the signal definition doesn't tell me what these arguments correspond to..
[13:13] <dednick> nik90: onPropertiesChanged(const QString& interface, const QVariantMap& changed, const QStringList& invalid)
[13:14] <dednick> nik90: dfeet does. PropertiesChanged(String, Dict of {String, Variant }, Array of [String])
[13:15] <nik90> dednick: yeah I meant like does String in the dfeet point at the PropertyName?
[13:15] <nik90> dednick: I compared it to the GetAll() function and there it makes sense
[13:15] <dednick> nik90: right. well i just told you them above :)
[13:15] <nik90> dednick: yeah :)
[13:15] <dednick> dbus interface, changed properties, and invalid properties
[13:16] <dednick> nik90: you should get a interface="com.canonical.datetime.AlarmProperties" when the DefaultVolume, etc properties change.
[13:17] <dednick> need to test incase there are other properties with the same name on other interfaces
[13:17] <nik90> ok
[13:17] <cwayne> rsalveti: heya, so how would I go about changing the kernel cmdline?
[13:18] <rsalveti> cwayne: was just thinking here that we might not necessarily be able to change kernel cmdline for a few devices
[13:18] <rsalveti> I think manta is one that basically ignores what we add in the build
[13:19] <rsalveti> ogra_: cwayne needs to add an option to upstart so it can look for a different directory for system related jobs
[13:19] <rsalveti> and that option needs to come from the initrd if I understood correctly
[13:20] <rsalveti> so we might need to change the init script logic in the initrd, which is fine as this change can be generic enough to work on any ubuntu image
[13:21] <cwayne> that'd be preferable i think
[13:21] <rsalveti> ogra_: I thought initially that adding the additional flags via kernel cmdline would be enough, but I just remembered that some devices ignore that
[13:27] <ogra_> rsalveti, how would you do that ? upstart resets the env when switching to /sbin/init ... and the system upstart doesnt allow settings vars via initctl
[13:27] <ogra_> *setting
[13:27] <cwayne> we need to change the actual call to sbin/init
[13:27] <ogra_> ??
[13:28] <ogra_> ou would have to patch /sbin/init to allow this ... the call wont help
[13:28] <cwayne> https://bugs.launchpad.net/savilerow/+bug/1315060
[13:28] <cwayne> already done
[13:28] <rsalveti> the patch is already there
[13:28] <rsalveti> would basically call init with special arguments
[13:28] <rsalveti> --configdir or something like that
[13:29] <ogra_> hmm, not sure that will work with run-init ... but you can try indeed :)
[13:41] <nik90> dednick: hey it works :) ..I need a quick advice, when I tell it to output what's changed it returns QMap(("DefaultVolume", QVariant(int, 80) ) ). How do I check which property has changed programmatically and assign it to my local variable?
[13:42] <nik90> dednick: I did http://paste.ubuntu.com/7809087/ and that fails because the unchanged properties is assumed 0. So it changes one correctly and the other is given an incorrect value of 0.
[13:42] <nik90> dednick: sry, I am terrible at this
[13:42] <dednick> nik90: if (properties.contains("DefaultVolume"))
[13:43] <nik90> dednick: ah..perfect. thnx
[13:44] <cwayne> ogra_: rsalveti: so I'd need to change run-init in ubuntu-touch-generic-initrd?
[13:44] <ogra_> cwayne, well ...
[13:45] <ogra_> cwayne, we explicitly do use the upstream /init script in initramfs tools ...
[13:45] <ogra_> you would have to hack up initramfs-tools itself
[13:46] <ogra_> (run-init is the last call in /init)
[13:46] <ogra_> (this is pretty intrusive)
[13:49] <ogra_> jodh, hey ... so with the fix of bug 1315060 ... how do you imagine that this option gets appended to the /sbin/init execution ? we would have to hack up the run-init call in initramfs-tools, no ?
[14:04] <jodh> ogra_: we implemented what was requested. I would have thought the simplest option might be to tweak /usr/share/initramfs-tools/init since that already passes a custom option to upstart for recovery mode?
[14:05] <ogra_> ah, so the same thing but with the new options ?
[14:06] <jodh> ogra_: yes, minimally, you could now add '--prepend-confdir=/foo/oem' to the last line of that script next to the '${recovery:+--startup-event=recovery}' bit.
[14:06] <ogra_> yeah
[14:07] <jodh> ogra_: which would cause upstart to read configuration from /foo/oem (and any subdirectories) *before* it read jobs from /etc/init/.
[14:07] <ogra_> right
[14:07] <ogra_> alternatively --append-confdir ...
[14:08] <jodh> I guess you might also want to modify /usr/bin/ubuntu-touch-session along similar lines for OEM Session Init customisations.
[14:08] <ogra_> yep
[14:08] <jodh> ogra_: well there is discussion on both those options on the MP - depends if you want the OEM customisations to take priority over the default.
[14:08] <ogra_> thats up to cwayne
[14:09] <jodh> ogra_: one thing to look out for - whatever directory or directories you explicitly specify with --prepend-confdir / --append-confdir must exist before Upstart starts; if they get created later, too bad as they will be ignored (inotify limitation).
[14:10] <rickspencer3> ogra_,  ?? http://www.der-postillon.com/2014/07/barack-obama-schenkt-angela-merkel-zum.html
[14:10] <ogra_> jodh, yeah, i guess we'll pre-create the dirs at image build time for touch
[14:10] <rickspencer3> (ftr, I get that it's a joke )
[14:11] <jodh> ogra_: perfect.
[14:11] <ogra_> rickspencer3, the best german satire page :)
[14:11] <ogra_> rickspencer3, ooooh !
[14:11] <ogra_> thats cool !
[14:12] <rickspencer3> the penny drops :)
[14:12] <alecu> Hi all, I need some pointers for this: the click scope needs to run a binary when the image is being built, to create a database of departments for the preinstalled apps. We have the binary ready, and we'd like to know how to run it in the image build process, and who to ping about this.
[14:12] <ogra_> rickspencer3, yeah :D
[14:12] <ogra_> haha
[14:29] <jgdx> brendand, hey, can you give me a quick introduction to using phonesim?
[14:30] <jgdx> brendand, e.g. pastebin the magic you displayed in the hangout? :)
[14:30] <brendand> jgdx, in a meeting right now - glad to later on
[14:30] <jgdx> brendand, ack, thanks
[14:56] <jgdx> Laney, kenvandine, could you take a look at https://code.launchpad.net/~jonas-drange/ubuntu-system-settings/1297418-apply-new-designs-to-background-panel/+merge/223571 in seb's absence?
[14:56]  * jgdx quickly checks for conflicts vs trunk
[14:57] <Laney> he should be around
[15:01] <kenvandine> jgdx, i'd rather seb128 since he's been reviewing it already, but if he can't I will
[15:02] <jgdx> kenvandine, sure
[15:03] <jgdx> Laney, ack
[15:04] <seb128> jgdx, kenvandine: that one looks good to land to me
[15:05] <seb128> kenvandine, want to do the landing ?
[15:05] <jgdx> seb128, thanks
[15:06] <achiang> ogra_: if you don't mind sooner rather than later... the seed branch is highly contended so the longer it waits, the more likely it is i'll get Yet Another Merge Conflict
[15:08] <cwayne> jodh: ogra_: so if the /custom/init or whatever doesn't exist but it's added to --append-confdir, will upstart die? or just not load stuff from there
[15:08] <kenvandine> seb128, let me try to finish up my call forwarding branch and we can land them together, i'm close :)
[15:08] <ogra_> cwayne, it will ignore it
[15:08] <seb128> kenvandine, great
[15:08] <cwayne> ogra_: perfect
[15:08] <ogra_> cwayne, but we can make sure it always exists on touch images
[15:09] <ogra_> by creating it at build time
[15:09] <cwayne> we shouldn't do that thought
[15:09] <cwayne> because we want only the custom tarball to touch /custom
[15:09] <ogra_> well you want it writable, dont you ?
[15:09] <cwayne> no
[15:09] <ogra_> ah, k
[15:09] <cwayne> it'll be populated with the custom tarball, then be r/o
[15:09] <ogra_> (for that it would have had to exist)
[15:10] <ogra_> achiang, im super busy finishing developer mode this week
[15:19] <mhall119> kenvandine: ping
[15:19] <achiang> ogra_: sad panda
[15:20] <kenvandine> mhall119, pong
[15:20] <ogra_> achiang, i'll try my best, but no promises ... any core-dev can merge that though
[15:20] <mhall119> kenvandine: is there a package for the C++ api docs for OnlineAccounts?
[15:20] <zeebok> this has probably been asked a dozen times, but is there any future plans of integrating Online Account IMing into messaging-app?
[15:21] <achiang> ogra_: no one seems to be jumping up and down to do it though. :)
[15:21] <awe_> cyphermox_, do you have a pull request for the code referenced in: https://bugs.launchpad.net/ubuntu/+source/urfkill/+bug/1339794/comments/3?
[15:21] <achiang> ogra_: i understand your priorities, thanks
[15:21] <awe_> cyphermox_, if so, could you add the pull request to the bug?
[15:21] <cyphermox_> I did not, I'd like it to be tested before
[15:22] <kenvandine> mhall119,  you probably want libaccounts-qt-doc
[15:22] <mhall119> is that qt specific, or can it be used by scopes?
[15:23] <robotfuel> Wellark: ping I have this crash bug in indicator-network can you triage? https://bugs.launchpad.net/ubuntu/+source/indicator-network/+bug/1343341
[15:24] <kenvandine> mhall119, not sure if that can be used by scopes
[15:24] <kenvandine> there's also the glib api, libaccounts-glib-doc
[15:26] <mhall119> cwayne: mhr3: ^^ which docs should I publish to the Scopes APIs section of developer.u.c?
[15:26] <mhall119> libaccounts-qt-doc or libaccounts-glib-doc?
[15:27] <mhr3> mhall119, ehm... no idea
[15:27] <mhr3> doesn't have much to do with scopes, does it?
[15:27] <mhall119> -qt is C++, but also uses things like QString
[15:27] <mhall119> -glib is C
[15:28] <cwayne> ssweeny: kyleN ^
[15:28] <mhall119> and uses things like gchar
[15:29] <mhr3> mhall119, you're basically asking me what's the recommended library for online accounts
[15:29] <ssweeny> i would imagine -qt for scopes right?
[15:30] <mhall119> mhr3: from scopes code, yes
[15:30] <ssweeny> i've used the -qt api from a scope before. i can't imagine trying to use glib
[15:30] <mhall119> I already know and have the recommended library for accessing it via apps
[15:30] <mhall119> ssweeny: ok, thanks
[15:33] <mhr3> mhall119, qt is pain to use from a scope, and c is pain to use from c++... your pick :)
[15:34] <nik90> charles: hey, So I implemented listening to PropertiesChanged() signal so that clock app dynamically updates its state when changes are made to the dbus property. I have a MP at https://code.launchpad.net/~nik90/ubuntu-clock-app/20-alarm-settings/+merge/227194, it is quite small (<80 lines of code). Can you check to make sure I have implemented all the
[15:34] <nik90> necessary precautions for it work properly.
[15:34] <mhall119> mhr3: if ssweeny says the -qt API is usable and has been used already, I'll go with that
[15:49] <Chipaca> i'm just going to leave this here: http://www.der-postillon.com/2014/07/barack-obama-schenkt-angela-merkel-zum.html
[15:53] <mterry> slangasek, morning!  Did you get a chance to noodle about passwd -d and all that jazz?
[16:03] <pmcgowan> jhodapp, got an audio issue, can we discuss?
[16:04] <jhodapp> pmcgowan: yeah
[16:04] <pmcgowan> jhodapp, playback of ringtones in system settings stopped working...
[16:04] <pmcgowan> so I made a simple app that just makes a QML Audio object and plays a file
[16:05] <pmcgowan> works on desktop not on phone
[16:05] <pmcgowan> https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1342321
[16:05] <jhodapp> pmcgowan: can you check syslog for apparmor DENIED messages?
[16:05] <pmcgowan> I checked no DEN there
[16:06] <jhodapp> pmcgowan: is there a log that we can get from system-settings?
[16:06] <pmcgowan> everything else seems to work, music, ringer, camera shutter etc
[16:06] <pmcgowan> jhodapp, that bug has the only weird output I see
[16:06] <jhodapp> ok let me take a look at the bug
[16:06] <pmcgowan> jhodapp, about video texture
[16:07] <jhodapp> pmcgowan: that's normal
[16:07] <pmcgowan> jhodapp, why would setting an audio url cause a video texture warning?
[16:08] <pmcgowan> if you say so
[16:08] <rickspencer3> phone really turned Green
[16:08] <nik90> jhodapp, pmcgowan: Just a note, this issue has been there at least since #120. It was first detected by the music app dev.
[16:08] <jhodapp> pmcgowan: it's just the way qtubuntu-media works right now
[16:08] <pmcgowan> nik90, say more as music works
[16:08] <pmcgowan> 120 sounds about right
[16:09] <jhodapp> pmcgowan: I can reproduce, let me see if media-hub-server is seeing it request to play that
[16:10] <jhodapp> pmcgowan: I see the problem
[16:10] <jhodapp> pmcgowan: invalid appamor security context...something with apparmor changed
[16:10] <jdstrand> ?
[16:11] <jdstrand> oh, the libapparmor api is returning something different?
[16:11] <jhodapp> jdstrand: media-hub-server is giving me this message: "Client denied access since it's an invalid apparmor security context"
[16:11] <jdstrand> jhodapp: ^
[16:11] <jhodapp> jdstrand: yeah I'm assuming so
[16:11] <jdstrand> tyhicks: can you assist jhodapp with that ^
[16:12] <jhodapp> jdstrand: it's bad that many clients try to play sounds but don't check if there was an error in doing so
[16:12] <jdstrand> tyhicks: is there a new rule we have to add to the media-hub policy to allow it to query the connecting process' label or something?
[16:12] <tyhicks> jhodapp: can you strace the media-hub-server process and then trigger the error?
[16:13] <tyhicks> jdstrand: no
[16:13] <jhodapp> tyhicks: sure, something I can grep for?
[16:13] <jdstrand> ok (I didn't remember seeing that)
[16:13] <jhodapp> tyhicks: it'll spit out a ton of lines
[16:13] <tyhicks> jhodapp: yeah, just give me all of those lines in a pastebin
[16:14] <jhodapp> tyhicks: I'll have to gzip and email you...won't fit in a pastebin
[16:14] <tyhicks> jhodapp: oh wait... is media-hub-server getting the apparmor context from dbus or from libapparmor?
[16:14] <jhodapp> tyhicks: should be from libapparmor, let me double check that
[16:15] <popey> jdstrand: click-reviewers-tools barfs on the calculator app.. http://paste.ubuntu.com/7809755/
[16:15] <tyhicks> jhodapp: a pointer to the media-hub-server code that generates the error would be helpful, too
[16:15] <pmcgowan> jhodapp, I dont think there was an error at QML level
[16:15] <jhodapp> tyhicks: actually no, it's from dbus
[16:16] <jhodapp> pmcgowan: yeah, we need to change that to make our clients more robust
[16:16] <pmcgowan> jhodapp, I'l look to see if there is a signal to watch
[16:16] <tyhicks> jhodapp: ok, so an strace wouldn't be of any help
[16:17] <jdstrand> popey: can you paste the .desktop file?
[16:17] <jhodapp> tyhicks: http://bazaar.launchpad.net/~phablet-team/media-hub/trunk/view/head:/src/core/media/player_skeleton.cpp#L135
[16:17] <jhodapp> pmcgowan: ok
[16:18] <jhodapp> tyhicks: jdstrand: oh it might be the naive check on line 167/168
[16:18] <jhodapp> jdstrand: remember we thought this might come back to bite us
[16:18] <jdstrand> oh
[16:18] <jdstrand> this is system-settings?
[16:18] <jhodapp> yes
[16:19] <jdstrand> I know what the problem is
[16:19] <ogra_> nik90, are you sure about 120 or could it be 130 :) http://people.canonical.com/~ogra/touch-image-stats/130.changes
[16:19] <jdstrand> yes, the check is naive, but there was a change to system-settings
[16:19] <jdstrand> we should revert that
[16:19] <ogra_> pmcgowan, check if system-settings didnt use libcanberra ... that was dropped in 130 http://people.canonical.com/~ogra/touch-image-stats/130.changes
[16:19] <jhodapp> jdstrand: was the change in it's apparmor policy file?
[16:19] <jdstrand> tyhicks: nm, I know the problem
[16:19] <pmcgowan> ogra_, no it wouldnt its all qml
[16:20] <ogra_> pmcgowan, if it did use that it needs to be ported to use plain gstreamer
[16:20] <nik90> ogra_: no pretty sure this issue was there for some time now...130 was like 2 days ago..I would remember :P
[16:20] <pmcgowan> indicator had that issue but it was not qt
[16:20] <tyhicks> jdstrand: whew... I'm glad bc I couldn't think of any recent changes that would cause that
[16:20] <ogra_> pmcgowan, QML doesnt provide audio playback, there must be some backend
[16:20] <popey> jdstrand: http://paste.ubuntu.com/7809782/ is com.ubuntu.calculator_calculator.desktop
[16:20] <jdstrand> jhodapp: the change was http://bazaar.launchpad.net/~system-settings-touch/ubuntu-system-settings/trunk/revision/767
[16:20] <pmcgowan> ogra_, qtmm to mediahub
[16:20] <ogra_> ah
[16:20] <jdstrand> jhodapp: ie, it ships an apparmor profile
[16:20] <jhodapp> ah ok
[16:20] <jdstrand> jhodapp: and that profile name doesn't have a '_'
[16:21] <jhodapp> jdstrand: there we go :)
[16:21] <jdstrand> but we don't need that profile now
[16:21] <jhodapp> jdstrand: can we make it have one?
[16:21] <jdstrand> seb128: hey, can we revert http://bazaar.launchpad.net/~system-settings-touch/ubuntu-system-settings/trunk/revision/767?
[16:21] <popey> jdstrand: so, arabic in line two
[16:21] <jdstrand> seb128: we aren't going to be implementing bug #1296415 after all
[16:22] <jhodapp> pmcgowan: so there we go, that's the issue
[16:22] <jdstrand> seb128: and there is a side-effect that it is preventing system-settings from playing sound files
[16:22] <jhodapp> pmcgowan: I'll mark the bug as invalid
[16:22] <jdstrand> jhodapp: can you add an ubuntu-system-settings task?
[16:23] <jhodapp> jdstrand: task?
[16:23] <jhodapp> pmcgowan: actually no, I won't do that since it's part of ubuntu-system-settings
[16:23] <jdstrand> jhodapp: 'also affects?' I think it is called
[16:23] <jhodapp> jdstrand: actually we're all good since the bug was filed there and not media-hub
[16:23] <pmcgowan> jhodapp, great that was an easy one
[16:23] <jhodapp> pmcgowan: yep, I like those :)
[16:24] <pmcgowan> jhodapp, I checked the error signal but its not triggered
[16:24] <jhodapp> pmcgowan: ok...that's most likely qtubuntu-media not signaling it then
[16:24] <jhodapp> pmcgowan: would you mind filing a bug for me to signal an error when playback doesn't work for whatever reason?
[16:25] <jhodapp> pmcgowan: for qtubuntu-media
[16:26] <pmcgowan> jhodapp, ok
[16:26] <jhodapp> thanks
[16:26] <Laney> you think reverting this apparmor is going to fix sound?
[16:26] <jdstrand> yes
[16:26] <Laney> 1 line explanation?
[16:26] <jdstrand> for the changelog?
[16:27] <Laney> mainly for my brain
[16:27] <jdstrand> jhodapp: what is the bug number?
[16:27] <Laney> https://bugs.launchpad.net/bugs/1342321
[16:27] <jhodapp> jdstrand: bug 1342321
[16:27] <Laney> building a test pkg to check this
[16:30] <jdstrand> * revert lenient AppArmor profile since we won't be limiting access to ofono in this manner. Furthermore, due to a naive check in media-hub-server for the connecting process' profile name, this allows system-settings to playback files via media-hub-server again (LP: #1342321)
[16:30] <jdstrand> Laney: ^
[16:31] <Laney> thanks jdstrand
[16:31] <jdstrand> Laney: basically, media-hub-server does a very simple check for the connecting process. the label used for the lenient profile failed this check. we don't need the lenient profile anymore, so just get rid of it
[16:31] <Laney> no profile still passes this check then?
[16:32] <jdstrand> yes
[16:32] <jdstrand> that is "unconfined"
[16:32] <jdstrand> jhodapp, Laney: you can easily verify this by doing "sudo apparmor_parser -R /etc/apparmor.d/usr.bin.system-settings"
[16:33] <jdstrand> that removes the profile from the kernel, setting it back to "unconfined"
[16:33] <jhodapp> jdstrand: sure
[16:34] <jhodapp> jdstrand: verified
[16:34] <pmcgowan> jdstrand, yep that fixes it
[16:34] <jdstrand> cool
[16:34] <Laney> How come there's no message in dmesg in this case?
[16:34] <jdstrand> Laney: it isn't an apparmor denial
[16:35] <jdstrand> Laney: it is media-hub-server making a call via libapparmor to see the connecting process' profile name, and making a decision based on the name
[16:35] <jdstrand> Laney: the decision in this case is to not play the file
[16:35] <jdstrand> s/is/was/
[16:36] <jdstrand> ('was', assuming the change is reverted)
[16:36] <jdstrand> :)
[16:36] <Laney> oh right, and this call doesn't create a log entry either
[16:39] <Laney> https://code.launchpad.net/~laney/ubuntu-system-settings/revert-apparmor/+merge/227225
[16:39] <Laney> you might recognise the changelog msg
[16:40] <jdstrand> hehe
[16:40] <jdstrand> Laney: thanks!
[16:41] <jdstrand> popey: does './bin/click-check-desktop <path to click>' work?
[16:43] <popey> jdstrand: yes
[16:43] <popey> http://paste.ubuntu.com/7809863/
[16:43] <jdstrand> popey: ok good, can you put the click somewhere?
[16:43] <popey> ya
[16:43] <jdstrand> this means it is just click-show-files
[16:44] <popey> jdstrand: http://people.canonical.com/~alan/clicks/com.ubuntu.calculator_1.3.291_all.click
[16:45] <jdstrand> oh, interesting
[16:45] <jdstrand> popey: are you running that on utopic?
[16:45] <popey> jdstrand: yes
[16:45] <popey> in a utopic chroot
[16:46] <jdstrand> it works here on trusty
[16:46] <popey> hmm
[16:46] <jdstrand> popey: I wonder if it is your locale settings in the chroot?
[16:46] <popey> could well be, yes
[16:46] <popey> balls, sorry
[16:46] <popey> yes, locale, posix
[16:46] <popey> well, thats good news ☻
[16:47] <jdstrand> :)
[17:08] <seb128> jdstrand, ok, I'm off for tonight/tomorrow but can have a look on monday if nobody beats me to it
[17:10] <mterry> slangasek, so... one solution would be to add support to passwd for falling back to the extrausers locations.  I took a look at how hard it would be -- looks not super difficult, something like http://paste.ubuntu.com/7809993/  (untested, checking now how it works in practice).  On a scale of 1-10, how acceptable would such a thing be?  :)
[17:21] <PreSSion> excuse me, somebody know if the screen of the ubuntu tablet will have got digitalize/capacitive screen or something like that to write in the tablet like in the samsung galaxy note 10.1 (sorry for my "engrish")
[17:23] <popey> jdstrand: getting errors with url-dispatcher now for a number of apps - have you tightened those tests?
[17:23] <popey> ERROR: url-dispatcher json unparseable: com.ubuntu.shorts_shorts.url-dispatcher (Expecting value: line 1 column 1 (char 0)):
[17:23] <popey> for example
[17:24] <popey> -rw-rw-r-- 1 alan alan 0 Jun 16 21:53 com.ubuntu.shorts_shorts.url-dispatcher
[17:25] <sarnold> PreSSion: so far as I know there are no planned tablets from any manufacturers; bq and meizu are both building ubuntu phones ATM...
[17:25] <jdstrand> popey: I added a bunch of tests, yes
[17:25] <popey> bugger ☻
[17:25] <jdstrand> can you give me the click?
[17:26] <jdstrand> no, that is a good thing :)
[17:26] <popey> not when I'm trying to get apps which previously passed into the store ☹
[17:26] <jdstrand> this sounds like a legitimate failure, but let me see the click
[17:27] <jdstrand> popey: right, but if the json fails to parse, ir probably wasn't working right
[17:27] <popey> its a zero byte file
[17:27] <jdstrand> even though we let it pass
[17:27] <popey> so yeah, it will fail to parse as json
[17:27] <jdstrand> ah, well, they should remove it
[17:27] <jdstrand> and remove "urls" from the click manifest
[17:27] <popey> I'll file a bug for it and get it fixed
[17:27] <PreSSion> sarnold: i am sorry, but i am sure to understand you, are you saying bq and meizu only are building the phones and not the tablets?
[17:28] <jdstrand> they are declaring a url-dispatcher but not providing the necessary configuration
[17:28] <popey> jdstrand: indeed, I'll fix it, thanks
[17:28] <popey> PreSSion: bq and meizu are making phones yes, thats the focus for the moment
[17:28] <popey> tablets will come later
[17:28] <jdstrand> popey: speaking of shorts-- do you use it regularly? it stopped updating feeds for me some time ago
[17:28] <popey> oh, not seen that happen
[17:28] <PreSSion> ok! thanks
[17:28] <popey> will check
[17:29] <jdstrand> I haven't been able to figure out why yet
[17:29]  * popey goes to make burgers, will fix these later
[17:29] <jdstrand> eg, I click 'refresh', but I have only ancient stuff and the ancient stuff doesn't update anymore
[17:29] <jdstrand> I haven't filed a bug yet
[17:29] <jdstrand> pfft
[17:30] <jdstrand> popey: of course, I tried it just now and it worked
[17:30] <jdstrand> but it had been weeks
[17:30] <jdstrand> weird
[17:30] <sarnold> ah, the good old "show someone else" debugging trick :)
[17:31] <jdstrand> yes
[17:31] <jdstrand> :)
[17:46] <mterry> slangasek, OK, that patch seems to function.  I'm curios for your opinion. That plus the pam_extrausers patch seem to be all we need for RTM
[17:49] <ogra_> mterry, and the adduser fix
[17:51] <mterry> ogra_, again, I think that while we need that down the road when we want to allow creating users, for now we can just move the phablet entries into their own files manually right after creating the user
[17:52] <ogra_> mterry, that will be a weeks work to change the image build scripts then :P
[17:52] <slangasek> mterry: hey, so for this piece I'd punt back to mdeslaur for review :)
[17:52] <mterry> ogra_, ?  in livecd-rootfs we can just do some sed work
[17:52] <ogra_> since they create the user and add it to a gazillion groups ... using adduser
[17:53] <ogra_> mterry, well, i'd prefer that we use the safe and proven tools here instead of adding such awful hackery
[17:53] <mterry> ogra_, adding an extrausers user to a group in /etc/group works fine with existing adduser
[17:54] <mterry> ogra_, sure, I'd prefer it to be already done too.  I'm just saying, it doesn't seem that hard to move one line from passwd and shadow to a different file
[17:55] <mterry> slangasek, understood.  :)  mdeslaur, can you take a look at http://paste.ubuntu.com/7809993/ when you get a chance?  It's a patch to shadow's passwd to support falling back to extrausers locations for shadow/passwd file modifications (like passwd -d USER)
[17:55]  * mdeslaur looks
[17:56] <slangasek> mterry: "adding an extrausers user to a group in /etc/group" - you mean setting the user's primary group to an existing group?  or something else?
[17:57] <mterry> slangasek, I didn't test with primary group, just a random group
[17:57] <mterry> specifically, nopasswdlogin as a test
[17:57] <mterry> Not something that will work on a RO image of course
[17:57] <slangasek> right, that's what I was wondering
[17:57] <slangasek> :)
[17:57] <mterry> But should be sufficient for our build scripts that add / remove groups
[17:57] <slangasek> which brings you back to the question of whether /etc/nsswitch.conf will merge group information from extrausers+files
[17:58] <mterry> slangasek, for production, we don't *need* to be able to add / remove groups to users on the fly
[17:58] <slangasek> mterry: surely we do for nopasswdlogin
[17:58] <ogra_> long term we do ... and i would really really appreciate if we could do this right from the beginning
[17:58] <mterry> *for RTM rather
[17:58] <slangasek> oh
[17:58] <slangasek> so that's phase 2?
[17:58] <ogra_> instead of rush it in this week
[17:59] <slangasek> when does the requirement for the nopasswdlogin group come into play?
[17:59] <ogra_> (with hacks and sed'ing of files at build time and whatnot)
[17:59] <ogra_> i thought nopasswdlogin was only a split greeter thing
[17:59] <mterry> slangasek, actually we don't *need* nopasswdlogin since we are deleting the user's password
[18:00] <mterry> slangasek, so they get to login for free anyway
[18:00] <mterry> ogra_, it's an 'anything-that-logs-in-via-PAM' thing
[18:00] <slangasek> not really
[18:00] <slangasek> nopasswdlogin is only used by lightdm
[18:00] <slangasek> $ grep nopass /etc/pam.d/lightdm
[18:00] <slangasek> auth    sufficient      pam_succeed_if.so user ingroup nopasswdlogin
[18:01] <mterry> slangasek, fair.  An "anything-that-logs-in-via-lightdm's-PAM-stack" which is unity8, unity7, and any actual lightdm greeter
[18:02] <VivisClone> IS there any intention of making this compatible with SGS5?
[18:02] <mterry> slangasek, point being that since the phone's 'no-password' mode is a REAL no password mode, users log in fine without being in that group
[18:02] <ogra_> VivisClone, if someone wants to ... its all open srouce
[18:02] <ogra_> *source
[18:02] <renat__> popey, I have add comments on the MR: https://code.launchpad.net/~pkunal-parmar/ubuntu-calendar-app/CalManagement/+merge/213355
[18:03] <renat__> popey, looks like the problem is in the way that Kunal Parmar is using the filter
[18:03] <mterry> slangasek, long term, I'm not sure what we want to do for dynamically adding/removing from a system group.  We could list extrausers first in nsswitch.conf and then have their entry live there... but that has its own issues
[18:04] <VivisClone> How would I go about compiling this for the samsung galaxy s5?
[18:04] <mdeslaur> mterry: hrm, you're going to face difficulties with locking there...I think it's using glibc to create the lock file and that's hardcoded to /etc
[18:05] <ogra_> VivisClone, see the channel topic ... there is a (very outdated and not really complete) porting guide
[18:06] <mterry> mdeslaur, hmm, darn hardcoded everything
[18:06]  * mterry makes his phone RO to check
[18:06] <ogra_> yeah, thats not really FHS compliant :P
[18:06] <mdeslaur> mterry: yeah, this is turning into a can of worms isn't it :)
[18:08] <ogra_> we should just set up a local ldap server :P
[18:08] <PreSSion> somebody here will buy the ubuntu phone in autoumm?
[18:08]  * ogra_ will buy 5 
[18:08] <ogra_> (at least)
[18:09] <mdeslaur> or perhaps pam_userdb
[18:10] <VivisClone> (Boost mobile)
[18:10] <VivisClone> I'd like one but my sevice provider probably won't work on it
[18:25] <mdeslaur> mterry: your also won't support when the user runs passwd on the command line, as when that happens, it uses pam instead of setting it directly
[18:25] <mdeslaur> mterry: or are you _also_ modifying pam_unix?
[18:25] <mterry> mdeslaur, I have a fork of pam_unix called pam_extrausers
[18:25] <mterry> :-/
[18:25] <mdeslaur> ah, ok, cool
[18:26] <mterry> mdeslaur, I originally thought that was all I'd need, but was surprised to learn about passwd -d being shadow-specific
[18:27] <mdeslaur> huh, that's weird
[18:32] <mdeslaur> mterry: what happens if you just change to a blank password instead of using -d?
[18:32]  * mdeslaur tries
[18:32] <sarnold> good luck there
[18:32] <sarnold> passwd throws a royal fit
[18:34] <mdeslaur> passwd doesn't care, pam_unix is what's rejecting the blank password
[18:34] <mterry> ogra_, you mentioned hacks.  I know you don't like the sed'ing, do you have other concerns?
[18:35] <ogra_> nope, as long as the tools we use work (so that we get the defaultgroups as in normal ubuntu etc)
[18:36] <mterry> ogra_, we should have them, yeah.  Adding/removing from groups is a bit of an open question right now but should be solvable
[18:36] <mterry> (I mean post install)
[18:37] <ogra_> the only thing that scares me is the possibility that we have to migrate something on the fly for existing users out there ... thats why i want that planned and thought through properly before RTM
[18:37] <ogra_> even if it takes a week more to do that ...
[18:38] <mterry> ogra_, you mean migrating passwords for users testing with development images and such?
[18:39] <ogra_> mterry, no, i mean releasing RTM with a certain setup and a month later recovering that we cant migrate the RTM users to the proper solution we will implement post RTM
[18:39] <ogra_> what we do here needs to be upgrade safe
[18:43] <mterry> ogra_, I get ya.  Well I feel like our options are relatively constrained.  Security want PAM.  The RO image requires that we not use /etc.  So extrausers it is.  And the last time we had a think about it (the sprint), we figured on extrausers too.  We just realized how poorly our tooling supports that
[18:43] <mterry> ogra_, although maybe there are other exotic PAM modules that we could use?
[18:43] <mdeslaur> pam_userdb
[18:43] <mdeslaur> but you'll have to re-tool anyway
[18:44] <mterry> mdeslaur, do you happen to know if that has nsswitch support too?
[18:44] <mdeslaur> sorry, don't know
[18:44] <mdeslaur> right, that needs to be supported too
[18:45] <mterry> mdeslaur, there doesn't seem to be anything even close to a turnkey solution for our needs here
[18:45] <mterry> And forking pam_unix into pam_extrausers is probably easier to maintain than a new nsswitch, maybe.  Well, especially since I already did it
[18:46] <mterry> And tooling would be harder to adjust for a database file than just pointing at a different shadow path
[18:46] <mdeslaur> mterry: if you kill the null check in pam_extrausers, you can simply change to an empty password instead of requiring -d and then handling it all manually
[18:47] <mterry> mdeslaur, interesting...  as a matter of policy, passwd -d requires root, but that method wouldn't
[18:47] <mterry> mdeslaur, not sure if you have feelings on that matter
[18:47] <mdeslaur> well, we're certainly allowing our users on the phone to set a blank password if they want
[18:48] <mdeslaur> and yeah, we need to discuss all of this after rtm
[18:49] <mterry> mdeslaur, do you feel squicky about going forward with this now for rtm?
[18:50] <mdeslaur> I'm not entirely comfortable with the idea of users having accounts with no passwords, but I don't have a better suggestion, or a better idea
[18:50] <mdeslaur> requiring that users have _2_ passwords, one on boot, and one to unlock isn't a good solution either
[18:51] <mdeslaur> while the phone has a limited attack surface, a converged device in the future will approach the desktop
[18:51] <mdeslaur> at which point having users set blank passwords is risk
[18:51] <mdeslaur> risky
[18:52] <ogra_> we could just set it to "password" ... thats at least well known to everyone :P
[18:52] <mdeslaur> ogra_: right...having it blank is no worse than shipping a device with "phablet".
[18:53] <mterry> ogra_, what if a user wants to use the password "password"?  :)
[18:53] <mdeslaur> at least the user gets to decide
[18:53] <ogra_> mterry, then he needs to use windows phone :P
[18:55]  * mterry is currently tracking down where something is breaking with a RO image
[18:56] <mdeslaur> mterry: so, if you do disable the null password check in pam_extrausers, you probably want to tie it into "nullok"
[18:56] <mdeslaur> I'm not sure why pam_unix has 'nullok', but it doesn't work for password
[18:57] <mdeslaur> oh, nullok is for validating the old password
[18:57] <mdeslaur> I guess
[18:57] <mdeslaur> anyway, off
[18:57] <mdeslaur> s/off/odd/
[19:01] <mterry> I'm getting a LOT of apparmor failures in syslog where mediascanner is trying to read /var/lib/extrausers/passwd
[19:01] <mterry> Why would it read that?
[19:02] <jdstrand> probably getent() type stuff
[19:02] <jdstrand> I imagine apparmor will need to have an update to /etc/apparmor.d/abstractions/nameservice
[19:03] <mterry> jdstrand, would it be acceptable to add extrausers locations to that?
[19:03] <jdstrand> I think so
[19:03] <jdstrand> tyhicks: can you add that to your list of things to add to the apparmor upload? ^
[19:04] <mterry> tyhicks, /var/lib/extrausers/{shadow,passwd,group}
[19:04] <jdstrand> note, I meant "*I* think so", not "I *think* so" ;)
[19:04] <mterry> :)
[19:04] <jdstrand> well, not shadow
[19:05] <jdstrand> passwd and group
[19:05] <jdstrand> mterry: is mediascanner2 trying to access /var/lib/extrausers/shadow too?
[19:05] <mterry> jdstrand, right
[19:05] <mterry> jdstrand, no
[19:05] <jdstrand> ok, good
[19:05] <mterry> jdstrand, i.e. is mediascanner2 suddenly malicious?  :)
[19:07] <jdstrand> mterry: no, but it processes untrusted input
[19:09] <jdstrand> granted, DAC should prevent the shadow access, but still, the nameservice abstraction only allows /etc/passwd and /etc/group now, and that is all any application should need unless it is performing authentication or updating it
[19:09] <mterry> right
[19:14] <tyhicks> jdstrand, mterry: ack - I've added a todo to allow access to extrausers passwd and group when I do the upcoming apparmor upload
[19:15] <mterry> tyhicks, thanks!
[19:15] <tyhicks> np
[19:27] <jdstrand> tyhicks: thanks!
[19:57] <pmcgowan> Laney, any idea on https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1343553
[19:57] <pmcgowan> are those icons needed or out of date?
[21:05] <awe_> pmcgowan, do you know why the cellular data switch went from a toggle to a checkmark???
[21:05] <awe_> -1
[21:19] <awe_> sergiusens, hey went to test the mms silo, but am confused what needs to be installed?
[21:19] <awe_> is golan-udm-dev required?
[21:19] <awe_> sorry, goland-udm-dev?
[21:22] <sergiusens> awe_: nope
[21:22] <awe_> ok, so just nuntium + udm
[21:22] <awe_> thanks
[21:22] <sergiusens> awe_: what's easiest is to install the citrain package on desktop and do citrain device-upgrade 13
[21:22] <mterry> mdeslaur, you mentioned removing the blank password check in pam_extrausers rather than patching shadow's passwd.  I think I'd prefer to patch passwd anyway because that way we are slightly more converged (don't rely on pam_extrausers behavior, can use ubuntu-system-settings to set your desktop password state)
[21:23] <sergiusens> awe_: can you please run sudo tcpdump -w mms.pcap -i rmnet_usbX where X corresponds to the one associated to the MMS context
[21:23]  * sergiusens adds that to the test plan
[21:24] <awe_> sure
[21:24]  * awe_ goes to checkout the test plan
[21:37] <sergiusens> awe_: https://wiki.ubuntu.com/Process/Merges/TestPlans/nuntium sort of updated
[21:39] <awe_> sergiusens, ack
[21:59] <awe_> sergiusens, gonna let the ofono silo testing continue into tomorrow... and we can land tomorrow
[22:00] <sergiusens> ok
[22:16] <brunogirin> popey: are you trying the redirect from #ubuntu-phone?
[22:17] <popey> i was ☻
[22:17] <popeytest> and it worked \o/
[22:17] <brunogirin> \o/
[22:30] <mhall119> awe_: sergiusens: just wanted to give you guys an update, the stop ofono/clear gprs context1/start ofono was enough to allow me to receive MMS
[22:30] <sergiusens> great
[22:30] <awe_> mhall119, cool
[22:43] <thomi_> tedg: you around? Got a second?
[22:44] <thomi_> tedg: does upstart / libUAL install a qt message handler to produce the logs in ~/.cache/upstart/ or does it just read stdout / stderr and redirect?
[22:45] <mhr3> thomi_, redirect
[22:46] <mhr3> we wouldn't want upstart to only work with qt apps :)
[22:46] <thomi_> mhr3: hmmm - so I 'd like to add timestamps to those log messages, which I think means adding a message handler in Qt. I'd like to write that code once and have it work everywhere. Any idea where I can do that?
[22:46] <thomi_> I could distro-patch the qt libs, but I was hoping to avoid that
[22:47] <mhr3> hmm, not sure
[22:48] <awe_> sergiusens, MMS never received by T-Mobile/Android phone...  do you want the nuntium log, or should I look for anything in particular?
[22:48] <mhr3> some apps are already adding a timestamp, you'd be duplicating it
[22:48] <awe_> sergiusens, also noticed that photos weren't appearing in the gallery right away either
[22:48] <thomi_> mhr3: which ones?
[22:49] <brunogirin> popey: it looks like re-flashing to stable and then back to devel fixed my off button problem: interesting...
[22:49] <awe_> ...and that the camera icon for mms is mis-leading, as you can only pick existing photos
[22:49] <popey> shouldn't have needed stable brunogirin, stable is v olde
[22:49] <brunogirin> popey: yes I noticed but by doing that I effectively re-flashed the device so that may have fixed it
[22:50] <popey> just going to devel should have done it
[22:50] <mhr3> thomi_, dont have a phone connected, but i think i saw some... but maybe i'm wrong
[22:51] <thomi_> mhr3: ok, I haven't seen any, and none of the AP clients I have on my HDD install a mesage handler - they might do it some other way, which is why I  was curious :)
[22:51] <thomi_> mhr3: I'll keep asking around, thanks
[22:53] <mhr3> thomi_, i suppose you could install a qt msg handler in the platform plugin, but not sure that's a great idea
[22:53] <thomi_> mhr3: why's that?
[22:53] <thomi_> rm -rf *
[22:53] <thomi_> oops :)
[22:53] <mhr3> thomi_, feels quite invasive to do that
[22:54] <thomi_> I agree
[22:54] <thomi_> ahh well, maybe I'll just patch each app separately
[23:00] <sergiusens> awe_: the pcap file if it was sent from ubuntu touch
[23:01] <sergiusens> awe_: photos don't appear in the gallery; you need to import them through the content hub
[23:02] <sergiusens> awe_: I do agree about the camera icon; but that is a _salem issue :-)
[23:03] <awe_> sergiusens, ?  the photo *did* show up in the gallery, it just too awhile to propagate
[23:03] <awe_> sergiusens, where do I find the pcap file?
[23:03] <sergiusens> awe_: hmm, ui might of changed; we had to import it before
[23:04] <sergiusens> awe_: pcap is from running tcpdump :-)
[23:04] <sergiusens> awe_: nuntium.log is fine if you don't have that
[23:18] <awe_> sergiusens, got it, let me try again and grab the pcap file...hadn't read the test plan thoroughly enough
[23:19] <awe_> sergiusens, ever thought of adding raw io tracing directly to nuntium ( ie. kinda like OFONO_RIL_TRACE or OFONO_RIL_HEX_TRACE )?
[23:20] <sergiusens> awe_: well I don't deal with the upload or download; that's the download manager; it's more of dealing with files
[23:27] <awe_> sergiusens, ack
[23:28] <awe_> sergiusens, where do I find the nuntium log again?
[23:29] <sergiusens> awe_: ~/.cache/upstart/nuntium.log