[01:28] Hi, I am stuck at installing ubuntu phone on my galaxy nexus [01:28] I am running $ ubuntu-device-flash --channel=trusty --device=maguro --device-tarball=Downloads/takju-imm76i-factory-e8c33767.tgz [01:28] dies with 2014/08/25 03:27:51 Cannot cleanup /cache/recovery/ to ensure clean deploymentexit status 255 [01:29] i cleaned up the /cache directory and that didn't work [02:11] So, I got Touch installed ok...booted up and put in my 4 digit pass...it goes gray like it is logging me in...sits there about 10 minutes so far. [02:27] 20 minutes now :) This is on a Nexus 10(manta) [02:34] nondem, I would try a reinstall things happen. [02:38] when running ubuntu-device-flash, what does it mean "Recovery image not found, cannot continue with bootstrap"? [03:23] i have succeeded installing ubuntu [03:24] all it needed is to reset android and rerun ubuntu-device-flash --channel=trusty --device=maguro --bootstrap [03:25] however ubuntu didn't start. i can open its bootloader, but no idea where to go from there. [03:26] the options are "reboot system now", "install zip from sdcard/sideload", wipe data... [03:26] how can i proceed from this state? [03:29] a few reboots actually made it work! :) superb! [05:08] Wellark: ack silo12> err, where do I need to do that? === chihchun_afk is now known as chihchun [05:31] pitti: thanks, that's enough :) [05:32] Mirv: what is enough, asking the question where to ack it? :-) [05:32] pitti: like, on IRC [05:32] Mirv: I don't know what to ack yet [05:32] pitti: oh, sorry, I thought it was "ack 12". the debian/rules part of https://ci-train.ubuntu.com/job/ubuntu-landing-012-2-publish/lastSuccessfulBuild/artifact/packaging_changes_indicator-network_0.5.1+14.10.20140824-0ubuntu1.diff [05:33] Mirv: ah, thanks; ack! [05:33] thanks! :) [05:33] Mirv: aah, now I know where to see the changes [05:33] I missed the artifact on https://ci-train.ubuntu.com/job/ubuntu-landing-012-2-publish/8/ [05:33] Mirv: btw, llvm got reverted, so my silo 7 built now; I'll test it now [05:34] Mirv: I'm not sure why it still says "ready to build packages" while everything got built, but *shrug* [05:35] pitti: not sure, possibly it got reconfigured after the build or something like that. if you want to keep the current packages, it's useful to now run build with 'watch only' checked. [05:36] oooh! thanks to whoever finally added asking for the SIM pin right at startup [05:36] this was a major pain [05:36] Mirv: ah yes, that was the case (one MP was moved to another silo) [05:36] Mirv: ack, will do that [05:41] urgh, ./phablet-config writable-image reboots [05:41] * pitti adds TODO item to fix that [05:42] ah well, can't call mount -o remount once we switch to user adb [05:47] Mirv: while I'm at it, core-dev ack for silo 5 [05:48] pitti: thanks for that, too, veebers will be happy. [05:49] Mirv: where do you set that status? in the spreadsheet? [05:51] pitti: core-dev ack? sadly the only papertrail on that is IRC logs :( to be fixed by CI Airlines... [05:51] Mirv: ah, so you just click "publish" [05:51] pitti: I only click "publish" with "ack_packaging" [06:28] Mirv: ppa 7 is no good any more, telephony-service is broken (another big change landed in trunk which isn't yet in my branch/PPA) [06:28] Mirv: I locally merged trunk now; can I push that to the same MP, and request a rebuild? [06:37] * pitti does [06:37] pitti: yes [06:38] hm, I figure I need to again do the same dance for rtm now [07:03] good morning === thomasm is now known as Guest23501 [07:22] Mirv: PPA 7 is good to go now; want me to publish, or shouldn't I (and only sil2100/robru/you do that)? [07:22] Hi all! I have a quick question. How hard to porting Ubuntu Touch to an other device, like Sony Xperia Z, L or SP? [07:27] pitti: in normal landings mode, there's no reason why a core-dev wouldn't publish himself :) so, feel free. [07:29] ack [07:35] Mirv: hm, the publish job succeeded, but nothing on https://launchpad.net/ubuntu/+source/messaging-app (and the dashboard rightfully complains about it); is that normal? [07:38] pitti: launchpad just didn't catch the package in -proposed yet (most probably) [07:40] sil2100: hmm, copies betwenn archives are pretty much instantaneous [07:40] ah, it's there now [07:42] pitti: CI Train takes the info about where the package is from LP itself, so if something would be broken it would mention that the package is 'in no known space' for a longer time :) [07:43] If it said it's in -proposed, then this was the info it got from LP API calls [07:44] It's usually accurate, although sometimes hasty [07:50] pitti: "pretty much" instantaneous :) [07:50] Mirv: with sync-package it's literally < 5 seconds (i. e. smaller than the time to load the +source web page) [07:51] oh, ok, that's fast then [07:51] I had thought it was the same for train copies [07:51] but anyway, nevermind === benunter_ is now known as benunter [08:29] can someone top approve https://code.launchpad.net/~aacid/messaging-app/update_pot/+merge/231534 ? [08:29] looking [08:29] +msgid " " [08:30] 19+msgstr "" [08:30] err? [08:31] tsdgeos: reviewed [08:31] tsdgeos: would be nice if you could fix that space along with the new string, while fixing i18n issues [08:32] pitti: yeah, not going to do that [08:32] tsdgeos: oh, why not? it doesn't make sense to have a single space marked as translatable, it's only going to cause trouble [08:32] pitti: because people complain you change their code [08:33] * pitti looks at the code [08:33] title: selectionMode ? i18n.tr(" ") : i18n.tr("Chats") [08:33] title: selectionMode ? i18n.tr(" ") : i18n.tr("Chats") [08:33] yes it's wrong [08:33] i know [08:34] it's not like i care if that's wrong [08:34] similar in Messages.qml -- the i18n.tr() should just be dropped [08:34] what i care is that we have an untranslatble phone [08:35] which is not even my job so i shouldn't be caring for it [08:35] I know this wasn't your fault, but slipping this into this MP will take an extra 10 s [08:36] making an entirely new MP and landing that will take three magnitudes more than that, with our current process (which absolutely doesn't work for small fixes..) [08:36] fix the process then :) [08:36] three magnitudes isn't even enough [08:36] i mean you may as well complain there's a missing ; in some line and that should also be fixed [08:37] well, it's that MP which introduces that space into the .pot [08:37] which is going to cause headaches to translators [08:58] ogra_: do you happen to have an i386 deb for the non-root adbd (for the emulator)? [08:59] pitti, hmm, no ... i'll try to build one during the day (i dont even have the source on an x86 machine ... dead architecture ... :) ) [08:59] ogra_: I'm happy to build it myself, if you have a pointer to the source/MP/branch [09:00] i can push the source package somewhere ... [09:00] * pitti pats his utopic-i386 schroot [09:02] ogra_: so with my phablet-network branch things mostly work, except for having to enter the PIN manually (same what plars was pointing out) [09:03] that'd be an unity8 bug, I think -- we need a D-BUS CLI to unlock the screen [09:03] pitti, right, the --password option to ubuntu-device-flash will handle that [09:03] we didn't yet need it without a pin, starting an app when the screen was locked unlocked it [09:03] oh, the autopilot issue with unlocking you mean [09:03] ogra_: right [09:04] ogra_: everything else works fine now (I made enough adjustments to the adb setup script) \o/ [09:04] yay [09:05] ogra_: it now tries "phablet" and "0000" as sudo passwords, and also has a --password option; and most tests shouldn't need root (e. g. calculator tests run fine without sudo, too) [09:05] yeah [09:05] but all tests fail while the lock screen is still running (i. e. waiting for pin) [09:05] the plan was that in future releases the image doesnt need to be writable and the pahblet user can do all tests [09:05] it seems with a pin a starting app doesn't just get rid of the lock screen [09:06] ogra_: yes, that has worked for some time already (with autopkgtest, not with phablet-test-run) === tvoss is now known as tvoss|test [09:06] right mterry should be able to help you === tvoss|test is now known as tvoss [09:08] ogra_: yeah, I'll file a bug against unity8; I already asked for this a month ago, but it wasn't necessary yet then [09:08] janimo`, hmm you didnt add udev rules for the new arch when you added the device id to adb, did you ? [09:09] * ogra_ upgraded his desktop to trusty on the weekend and cant connect anymore with "no permissions" [09:09] pitti: i've removed those i18n.tr(" ") [09:09] tsdgeos: cheers [09:10] should we not use that in general ? [09:10] (i know the app-dev documentation talks about it everywhere) [09:10] tsdgeos: top-approved now [09:10] now someone just needs to land it [09:11] who said the process was not agile? :D [09:11] *smirk* [09:11] that's the kind of change one would usually directly commit to trunk and just let it flow in with the next update [09:12] same like simple mass-packaging changes or fixing typos, etc. [09:12] tsdgeos, it will get better... [09:12] i. e. ci-quick-walk-to-the-bakery, not everything needs the overhead of an airline :) [09:12] pitti, thats the prob, with the current model (which i heard asac asekd for now) you have to have two branches for *every* project [09:12] but that had been discussed in the airline designs, I think [09:13] which kind of contradicts the initial idea of "handle it like SRUs" we had [09:13] ogra_: yeah, with rtm it's harder (I also need to land 7 more MPs now for an utterly trivial change) [09:13] pitti, we are trying to get a model in place where you can automatically PPA copy the source package into an rtm silo when you land in utopic [09:14] hmm, it seems with every other install the nice numeric keypad "enter your pin" gets replaced with a "Hello" input line and I have to use a letter keyboard for PIN entry [09:14] so that the distro handling happens on a package level and doesnt force you to try to keep two constantly diverged branches in sync [09:14] pitti, i think that happens onyl if your pw is non numeric and 4digits long [09:14] ogra_: that would only work as long as rtm still ships the trunk branch, though? [09:15] pitti, right, with is 80% of our landings [09:15] ogra_: no, I always use "0000" [09:15] there are a few projects that will target rtm directly [09:15] pitti, weird ... there is another dbus settings for the account stuff [09:16] swipe/passcode/passphrase ... i have to look up how exactly it is called [09:16] ah indeed, system settings says password, not pin [09:16] right you can set that via dbus [09:17] so it seems sometimes it mis-sets the pin as password; I'll try to reproduce that [09:18] ogra_: what I was actually doing was a fresh reinstall to confirm that we now have an r/w image by default; that's not intended, is it? [09:18] /dev/loop0 on / type ext2 (rw,relatime,errors=continue) [09:18] which component would that be for filing a bug? [09:18] oh no, definitely not !!! [09:18] pitti, initramfs-tools-ubuntu-touch [09:18] yeah, I was wondering why my test run instaslled all the autopilot bits with dpkg [09:18] ogra_: cheers, filing bug [09:19] that seems to be the partitioning changes causing it [09:21] bug 1361124 [09:21] bug 1361124 in initramfs-tools-ubuntu-touch (Ubuntu) "Image is now writable by default" [High,New] https://launchpad.net/bugs/1361124 [09:22] janimo`, ^^^^ [09:25] janimo`: ah, would you be the appropriate assignee for that? [09:27] ogra_: sorry for keeping bothering you -- who should I ask for reviewing the phablet-tools MP? [09:28] pitti, i can approve it ... just busy with getting the packaging support into phablet-config too ... i'll make sure your change goes in with the same landing [09:28] ogra_: ah, ok; danke! [09:35] jibel, trying to reproduce the lock-up you're getting here, obviously can't... but anyway, does it play with both speakers for you? at least the case suggests there's two speakers at the bottom, but I'm only getting sound through the right one (can cover it and it goes almost completely quiet) [09:37] Saviq, it plays only from the right one. [09:37] jibel, right, wonder if that's expected... [09:39] jibel, but yeah, no lockup here ;| [09:40] Saviq, yeah, it's really difficult to reproduce. I had it twice yesterday but nothing today. asac and victorp also had it several times a day last week. [09:40] jibel, and it's not bug #1295623 for sure? [09:40] bug 1295623 in Unity 8 "Sometimes input breaks and only edges are responsive" [Critical,In progress] https://launchpad.net/bugs/1295623 [09:41] Saviq, yes I'm sure. it's really frozen. For example, in one of the case the greeter stopped in the middle of the screen. In the other case the dash was displayed and I couldn't reveal the launcher or indicators [09:42] jibel, right [09:42] Saviq, it's really like when unity8 is crashing, excepted that it is not crashing [09:43] jibel, yeah, sounds like a deadlock somewhere [09:43] jibel, the gdb trace will help a lot next time you see it [09:46] Saviq, sure, I'll add this information next time it happens. Thanks for looking. [09:49] ogra_: I did an initial review of your phablet-config branch; I debug this further why it doesn't work yet [10:06] pitti, hmm, your bug above ... (writable root) is that on a device or emulator ? [10:06] could be that the emulator still mounts it rw [10:06] ogra_: on mako; it has always been writable on the emulator (which is another bug) [10:07] ogra_: bug description updated to clarify [10:07] thanks [10:18] ogra_: do you have a branch to make system-image-cli -i work as user? [10:19] (PermissionError: [Errno 13] Permission denied: '/root/.cache/system-image') [10:19] or does that get wrapped by phablet-config somehow? [10:19] ah, it apparently tries /var/log/system-image/client.log first, then falls back to the above, but both are root-only [10:24] pitti, http://people.canonical.com/~ogra/ubuntu-touch/adbd-no-root.patch for your x86 build ... [10:24] ogra_: cheers! [10:24] pitti, system-image-cli can not work as user [10:24] barry would kill me if i changed that [10:24] ogra_: ok, I'll adjust the test for "is that an ubuntu phone"; thanks [10:25] there is some way around the log issue that the autopilot test for system-settings uses i think [10:25] ogra_: don't worry, I replaced it with "type unity8" [10:26] adt-run can't determine/log the image version then, but *shrug* [10:27] you could just grep it from /etc/system-image/channel.ini [10:42] nik90, could you find anything about my problem? I couldn't fix it yet [10:43] vitimiti: no I couldn't. The XmlListModel seems to be defined correctly [10:43] so I cant think why it doesnt unless I dive deeper [10:43] I am a bit busy atm to do that [10:43] I see, thanks for trying, anyway [10:44] I split the pages to make it more readable, though [10:45] are there known issues in image #207 (on mako) that the screen doesn't turn on? [10:45] the device appears to boot (I can connect to it with adb), but the screen stays black [10:46] the tests in the lab ran fine ... so the devices there must work [10:47] http://ci.ubuntu.com/smokeng/utopic/touch/mako/207:20140825:20140811.1/9930/ [10:48] t1mp, do you see the google logo or actually a black screen [10:49] ogra_: hold on, I am re-flashing the device. I'll tell you in a few minutes [10:49] vitimiti: yup I noticed that [10:50] vitimiti: so is the xml listmodel data displayed correctly in HospitalPage0.qml but not the rest? === dandrader is now known as dandrader|afk [10:51] nik90, exactly, the name appears no matter how much I change it, but the news, staff and so only appear on the console.log test I put there [10:52] vitimiti: may be then the XML ListModel is faulty [10:52] pitti, answered your MP commants for phablet-config ... [10:53] vitimiti: how about you create a single page when you show the hospitalName, hospitalNews and see if they are displayed correctly? [10:53] s/when/where [10:53] ogra_: ah, I'm just testing a fix for the sync issue, I'll post it there once confirmed to work === tvoss is now known as tvoss|lunch [10:53] nik90, I will try to put all of them where the name is and not display the other pages [10:53] pitti, heh, i meant your comments on my MP :) [10:53] ogra_: yes, so do I :) [10:54] ah [10:54] vitimiti: yeah just see if the xml model even returns the hospitalNews variable [10:55] nik90, yeah, still a blank label [10:56] vitimiti: why do you have @ infront of the queries? [10:57] Because it's instead of stuff [10:57] Just like in the developer's page tutorial [10:58] The console.log() does I funny thing I just saw, it adds a lot of empty lines that shouldn't be there [10:58] ah yes [10:58] why do you define the xml file like it is? [10:58] cant you change to stuff ? [10:58] I found it easier to understand for me, but it does the same with stuff, I tried it too [10:59] I will try with all in one page, though [11:01] ogra_: why do we need the non-dbus call code paths? [11:02] ogra_: (there's no followup comment from you except the new commit -- was that eaten by LP from some email reply?) [11:02] pitti, yes, until the stable channel is gone/updated [11:02] ogra_: after re-flashing the device works again. I'll try to figure out if I broke it or if there is a bug that I can reproduce [11:02] ogra_: ah, of course; I see [11:03] nik90, now there's something new, apart from being now "undefined" instead of blank, this error appears: Error XPTY0004 in file:///usr/lib/x86_64-linux-gnu/qt5/bin/qmlscene, at line 2, column 54: Required cardinality is zero or one("?"); got cardinality one or more("+"). (three times) [11:03] t1mp, did you have any click packages installed (beyond the preinstalled ones) ? [11:04] t1mp, that can cause a lot of apparmor stuff on boot and might keep you on the google logo for minutes [11:06] nik90, but if I reduce it to just one new, service and staff, it works. Now I have to find a way to make it read several of them [11:08] ogra_: followup sent; you may vomit now, but at least it works :/ [11:09] ogra_: oh, and yay for LP killing initial space in my patch :) [11:11] ogra_: I haven't tested the writable part; does calling the SetProperty() thing actually do the remount? (hopefully without reboot)? or does that need a manual reboot afterwards [11:11] pitti, the dbus service just touches /userdata/.writable_image and triggers a reboot ... [11:12] ogra_: ah; what's the source project for that? I'd like to eliminate the reboot [11:12] pitti, i missed re-adding the adb.wait_for_device() and wait_for_network() bits ... i'll add that now [11:12] "mount -o remount,rw /" is quite enough [11:12] no [11:12] ogra_: ah, thanks [11:12] thats not the allowed/preferred way for devs [11:13] why not? it's the same result without having to wait for a full boot again [11:13] (i prefer it too btw ... but the file is the "right way" ) [11:13] ogra_: yes, sure -- touch the file, remount rw [11:13] ogra_: so that on next boot it'll stay r/w [11:13] no. it will handle the writable bits mounting differently iirrc [11:13] hm, we've use the remount,rw for a fair while (and ignored phablet-config as it's too slow) [11:13] (i would have to check the initrd scripts, but there is a differentce ... also some tools check for that file) [11:14] ogra_: well, of course it should continue to touch that file [11:14] i dont want to change functionality anyway [11:14] phablet-config writable-image has always needed the reboot ... if you want to remount, fell free :) [11:14] ogra_: yes, not in that MP 9just something for later) [11:15] ogra_: well, s/needed/done/ [11:15] yeah, we can discuss that [11:15] I was just asking which project provides PropertyService [11:15] oh, the dbus part is an ugly shellscript in dbus-property-service [11:15] ah, https://code.launchpad.net/~phablet-team/dbus-property-service/trunk [11:16] thanks [11:16] (that initially only *did* set properties ... i need to find a new name at some point) [11:16] device-munging-service :) [11:16] haha [11:16] well, i might also split it after rtm [11:17] yeah, not critical, it just occurred to me [11:17] I lagged too much [11:17] to keep a clean "peroperty-service" backend for android properties [11:17] (which was the initial idea behind that service) === dandrader|afk is now known as dandrader [11:23] Sorry about that [11:37] vitimiti: np. I will try to use your xmllistmodel to create an app to display the data in a page. But no promises :) [11:40] nik90, ok, thanks again. [11:41] zsombi: if you have a pitch-black screen, but you can connect with adb, then you have the same problem that I just had [11:41] pitti, I'll look at that, since indeed it may be due to recent changes I had been involved with [11:41] zsombi: did you see the google logo when booting? [11:42] ogra_: ^ it seems someone reproduced my issue [11:42] t1mp: yes, I did [11:42] t1mp: but now the screen is on... [11:43] ok, I had to reflash for that [11:43] so it just took long ? [11:43] yeah, thats aa-clickhook running when it finds non default click packages installed ... can take up to 5min [11:46] nik90, I found a way around: if I want more data, instead of using stuff, I should use stuffstuff... [11:46] nik90, that using a listmodel and a function y the xmllistmodel to append every new data it finds [11:46] y=and* [11:47] cool [11:47] vitimiti: I would go different naming [11:47] vitimiti: like the container which holds all the news children could be called [11:48] so something similar [11:48] I'll try [11:48] otherwise it is a bit confusing imo [11:48] * ogra_ curses pep8 ... [11:50] nik90, yeah, that works too and looks better, perfect [11:51] nik90, I the listmodel isn't even needed, changing from name="" to was definitely the solution [11:52] vitimiti: atm too many changes that I cannot imagine the code without seeing it :D. But if it works great === mibofra_ is now known as mibofra === mandel is now known as mandel|lunch [11:56] pitti, did you ddo a fresh flash when you saw the image being writable ? [11:56] or was that an OTA upgrade ... [11:56] pitti, I'll try to reproduce this, for now I dont' see why the initrd change itself would make that writable though. hmm [11:56] also did you use --wipe or --bootstrap [11:57] right, i was just talking with janimo` in the other channel [11:57] I just pushed it, if you want to see, nik90 [11:57] i think if you made the image writable only --wipe or --bootstrap will make it non-writable again (i.e. remove the file) [11:58] ogra_, he says there's no Note that there is no /userdata/.writable_image stamp at all [12:00] vitimiti: so much better. nice work [12:00] Yeah, that feels good, after 4 days [12:00] heh [12:01] :D [12:08] GRRR [12:08] why does the jenkins bot not like me today :( === renato is now known as Guest9436 [12:14] sigh, i dont get that, the indendation is correct ... yet the bot moans [12:16] hi all, I would like to file a bug for mtp not quite working. do I log it here? https://bugs.launchpad.net/ubuntu/+source/gvfs/ [12:16] rickspencer3, no, against the mtp package [12:17] ok [12:19] lol [12:20] so adding a blank line makes the bot not moan anymore ... how silly ... (since i didnt change any indendation for the indendation complains) [12:20] * ogra_ wonders why he always runs into the false positives [12:20] sil2100, I was told that you might want to track this bug wrt new (and awesome) stuff in mtp: [12:20] https://bugs.launchpad.net/ubuntu/+source/mtp/+bug/1361181 [12:20] Ubuntu bug 1361181 in mtp (Ubuntu) "dragging files in nautilus onto an SD card does not work" [Undecided,New] [12:20] rickspencer3, assign to cyphermox [12:21] ogra_, cyphermox if you guys would like me to do anything to help get some info off the phone, just let me know [12:21] sil2100, ^^ ... seems realted to the mtp-server crashers we see on tests [12:22] ogra_, oh, I see, I have _usr_bin_mtp-server.32011.crash [12:22] should I have ubuntu-bugged that? [12:23] well, i guess we already have a few of these on errors.u.c [12:24] * rickspencer3 nods [12:25] rickspencer3, it's likely bug 1360569 [12:25] bug 1360569 in mtp (Ubuntu) "TOPCRASHER: /usr/bin/mtp-server:6:__gnu_cxx::__verbose_terminate_handler:__cxxabiv1::__terminate:std::terminate:__cxxabiv1::__cxa_throw:core::dbus::Bus::send_with_reply_and_block_for_at_most" [Critical,Confirmed] https://launchpad.net/bugs/1360569 [12:25] it's one of the winner on errors.u.c [12:25] nice [12:26] :( [12:26] when I rebooted the phone, my sd card didn't show up [12:26] * rickspencer3 tries again [12:34] ogra_: no, fresh install (I think I wrote that in the bug) [12:34] yeah, saw that [12:34] sorry [12:34] ogra_, janimo`: I used the dual-boot installer under android [12:34] oh [12:35] but that was r/o before [12:35] you should have written that you use an unsupported method :P [12:35] pitti, hmm maybe ondra knows something about that [12:36] ogra_: oh, is that unsupported? sorry [12:36] :) [12:36] we never supprted dual booting officially [12:36] but it was r/o before [12:36] right [12:36] but it might be the dual-boot bit thats broken here [12:36] pitti, I asked ondra, let's see if he's around [12:36] so if it works on a full install, then you can at least downgrade the importance [12:36] guess not, Bank Holiday in the UK [12:36] someone with mako should verify [12:37] but it wasn't clear right away that it would only affect dual boot [12:37] as it's now r/w in both dualboot and the emulator [12:37] yes, but emulator is by design [12:37] uh, why's that? [12:37] I thought it was a bug [12:37] iirc its a workaround for other issues ... yes its a bug ... [12:38] but its known [12:56] I loaded the stable verson instead of dev and it works ok(relatively) [13:07] boiko_: could you comment on: https://bugs.launchpad.net/ubuntu/+source/share-app/+bug/1360670 [13:07] Ubuntu bug 1360670 in share-app (Ubuntu) "drop from archive" [Undecided,New] [13:07] so that I can perform [13:07] https://bugs.launchpad.net/ubuntu/+source/libhud-qt/+bug/1360671 [13:07] Ubuntu bug 1360671 in libhud-qt (Ubuntu) "drop from archive" [Undecided,New] [13:09] pitti: I sent mterry an email about the screen unlock needs but haven't heard back. I wonder if someone else from the unity8 team could adjust their unlocker to account for that? [13:09] plars: yeah, I hope so; I also discussed that about a month ago, but there was not much response to that either [13:10] plars: back then I tried rather hard, but didn't find a way to get rid of this lock screen programatically === _salem is now known as salem_ === tvoss|lunch is now known as tvoss [13:13] plars, well, it is tricky withug ripping a giant security hole [13:13] plars, per security design the lock screen is our major secutity check ... once you are in you can do a ton of stuff you wouldnt be able to normally [13:14] ogra_: err, no -- if you are in an adb shell, you are "in" [13:14] you have all the data and access to phablet's processes, etc. [13:14] you can even kill unity8 [13:14] so adding any kind of way to work around the lockscreen should be handed to the security team for review at least [13:14] Wellark: hmm, I don't now anything about share app, [13:14] so I don't see anything wrong with adding a d-bus method for it [13:15] ogra_: I understand, I'm not complaining, just saying that I pinged mterry on it who did the unlock code currently in there. Hopefully someone will have a way to fix it though, because it's also not possible to get automated results without a way to get the screen unlocked [13:15] pitti, well, that shouldnt be possible anymore once the dev mode is fully implemented [13:15] (or ssh even, which is what I'm more interested in) [13:15] you will need a pw set to get in [13:15] right [13:15] and we have that [13:15] and you will need an unlocked screen to get in too [13:15] per design [13:15] so if you know the pw and can ssh phablet@localhost, you are in [13:16] this is the final barrier to make sure you are safe [13:16] we can absolutely be fine with running the unlocker under sudo though [13:16] if we pierce that i think jdstrand will haunt you at night [13:16] ogra_: you mean there won't be ssh until you unlock the screen? [13:16] there wont be adb [13:16] and you will need adb to enable ssh [13:16] then I guess we need some --testing mode again for that [13:16] well [13:17] once sergiusens has added the --password option to u-d-f you know the password [13:17] ogra_: yes, but that doesn't help if there's neither adb nor ssh :) [13:17] you need to find the dbus runes to hand that tp the screen unlocking [13:17] *to [13:17] and even if you have it, there's no way to unlock through them [13:17] erk, no [13:18] if you can do that, then you can as well do a d-bus call to do that properly [13:18] making it hard is not a security measure [13:18] it's just making it a pain to test [13:18] well, talk to the security team [13:18] * sergiusens thinks the --password option should die in favor of pitti's 'testing-mode' option [13:18] i dont think such a swithc is currently planned [13:18] or just create a channel with the device ready for testing [13:18] sergiusens, i surely wont start over with dev-mode now [13:19] well, I don't see anything wrong with having a d-bus call or other programmatic way to unlock the screen *if you are already on the device* [13:19] (ssh or adb) [13:19] the image is already bastardized enough to consider it not in the original state [13:19] this was all agreed on in malta [13:19] ogra_: you should know I never agree to anything :-P [13:19] haha [13:19]