[01:15] sooo - does anyone know when is the next ubunntu phone gonna come out? [01:16] I just broke my android - but I want something more serious than bq aquaris. [01:17] did not hear anything more than "soon" [01:17] k1l, ok - then how about best supported phone so far? nexus 5? [01:18] erm. both nexus 4 and 5 seem to be ok? [01:22] nexus5 has no bluetooth and won't play videos [01:23] oh. [01:23] nexus4 was the original dev device. [01:24] k1l, so that leaves only nexus 4? [01:24] the meizu mx4 is the next phone [01:29] damn. mx4 is gsm, im on verizon :(. === chihchun_afk is now known as chihchun [02:52] Hello People :D Hay alguien que hable español? [07:35] guys I have a serious issue with my BQ, this morning it froze and I powered it off then on and now I am stuck at boot loop at the BQ logo and powered by Ubuntu screen. [07:43] now attempting to enter recovery, still waiting at the ubuntu logo [07:45] a few bq logo flashes and now it boots "normally"?? [07:46] EdwardMorbius, https://lists.launchpad.net/ubuntu-phone/msg11819.html [07:47] ogra I think this has something to do with changed system language [07:47] from what to what did you change it ? [07:47] I switched from english us to english uk and I read on the Internet about someone else getting bootloop after they changed language [07:48] could you comment on the bu that is mentioned in the above mail ? [07:48] *bu [07:48] bah [07:48] *bug [07:48] will do so now, because I had one phone freeze before but I powered it off and on normally [07:48] but now I got bootloop [07:49] oh, you sounded like it booted again above [07:49] the option in the gui erase&reset everything will make a full factory reset right? [07:49] It did boot but looks shaky to me [07:49] yes :/ [07:49] define "looks shaky" [07:49] indicator date time is bugged and I got to gui some time ago too, attempted a new reboot from gui and again got the bootloop [07:50] so I believe factory reset is the way to get everything in the clear again [07:51] yeah, but at lest you have the ui up to be able to pull down photos etc if needed [07:51] the factory reset will sort it [07:53] didnt have any photos worth saving anyway, restoring all my settings will be a little chore though [07:53] yeah, we really need a backup tool for this ... one that stores and restores all data and settings [07:53] phone resetted and I am commenting on the bug now [07:53] yes that would be very helpful [07:53] thanks ! [07:55] just posted, too bad I had to reset the phone, I am guessing all the logs are gone now too. [07:55] yes, but you are the second person that claims it happened after language switch ... that at least gives a pointer [07:56] (though trying to reproduce it on a development device here doesnt get me the issue) [07:56] development device as in nexus 4? [07:56] no, a pre-production bq [07:57] same device .... but with android buttons on the glass :) [07:57] maybe someone with the production device should attempt it, gremlins inside may be different than in the pre production one. [07:57] yeah [07:58] when device froze I was like "oh I will post some logs as bug report later on" but then :O :D [07:58] yeah, if you cant boot that wont help :) [08:00] I am not touching the system language again :D [08:00] heh [08:00] there might be a differnce in how the welcome wizard sets it and how the system settings do [08:01] true, someone should check that, it is possibly an oversight [08:01] riht, i will ... after easter vacation though ... i'm not officially here :) [08:02] lets hope no more people get ideas about changing the system language :D [08:03] yep :/ [08:04] and that backup tool would be veeery nice if someone can make it [08:09] updating apps... [08:10] Is it just me or is the Ubuntu Store sometimes a bit slow? === thomas_ is now known as Guest55350 [10:22] i rootstrock-ng only useable for the supported devices? what would i do if i had a freshly build system.img for my galaxy s4 [10:22] ? [10:54] good morning all [10:54] shall I assume that everyone is on holiday today? [10:55] except us poor saps from the USA, and then I guess I assume I am the only one crazy enough to be up already? [10:56] rickspencer3: real part of France is working as well :) [10:56] hey! [10:56] hey ;) [10:56] bon courage! [10:57] merci, à toi aussi ! [10:57] I was not expecting a response :) [11:01] the heathen land of Ireland also defiant [11:02] greyback: hey! you keep wanting to be different than the UK, I see ;) [11:03] didrocks: you're lucky there's an ocean between us! [11:03] greyback: j/k ;) [11:03] :D [11:05] greyback! [11:05] rickspencer3: at your service *hat tip* :) [11:05] o/ [11:05] just saying "hi" [11:06] I got up early because I was so worried about the bootloop issue I couldn't sleep [11:06] I imagined hundreds of people factory resetting their phones this morning :/ === cwayne-afk is now known as cwayne === dkessel_ is now known as dkessel [11:34] how do i get my compiled images and the rootfs to my s4 in order to get it up running? [11:40] Isotop7, hey, I feart that the channel may be slightly laggy for you today, I know I lot of people are afk b/c of Easter :/ [11:40] dudes, I'm getting "invalid cross-device link" when using the citrain thingy to install a silo [11:40] how go problem fix I? [11:56] Laney, you might want #ubuntu-ci-eng rather [11:57] Laney, i think robru fiddled with the code very recently ... (last week) [11:58] probably check if there is a newer version [11:58] where? [11:58] phablet-team PPA [11:58] I used that Ubuntu archive thing [12:17] ogra_: how did you test that lxc-android-config fix? [12:18] i made the file writable through writable-paths and rebooted [12:18] installing lxc-android-config is rather painful so i only tested the chnage [12:18] so touch /userdata/.wriatble_paths; reboot; edit; rm /.../.; reboot; test ? [12:19] i never touch writable_paths... [12:19] you can just edit it? [12:19] wget the deb ... reboo to recovery ... [12:19] woah [12:19] what is this? [12:19] install lxc-android-config chrooted into the rootfs [12:20] then reboot ... [12:20] O_O [12:20] it's only a file change, right? [12:20] thats how to install this package [12:20] as i said ... painfull [12:20] thanks to our bindmount farm spanning across partitions [12:20] you can mount -o remount,rw …, do the change, reboot [12:20] not to test the package [12:21] but yeah, to test the change [12:21] yeah, not the package, but just the 8 char changes in a string file [12:21] right, that wont tell you if the package did the right thing though ... but is enough o check the change works [12:22] given that you think it does and jibel says it does not ... [12:22] yeah, i told him i'll research it after the holidays [12:23] is there a reason why we dont use the same .override mechanism we use in vivid ? [12:23] seems that all frenchies are trying to make Laney's life hard :) [12:23] conspiracy! [12:23] ogra_: well, it needs some changes to multiple components to work [12:23] whoopsie, whoopsie-preferences… [12:23] not sure you want those changes for rtm [12:23] (handling transitions…) [12:24] didrocks, ah, no [12:24] ogra_: look at the postinsts, you will have fun :p [12:24] but I think those kind of changes are good, no more "ENABLE" in conffiles [12:25] i dont really care about conffiles on the phone :) [12:25] ogra_: well, you have to find a way to transition though? [12:25] like, how would you transition to the new system with the .override [12:25] why ? we just replace [12:25] once moved to vivid? [12:25] how do you preserve the user setting? [12:25] ok, so the user disabled whoopsie [12:25] /etc/default/whoopsie have ENABLED=false [12:25] in vivid-based image, this reflect in the /etc/init/whoopsie.override with "manual" [12:26] how do you handle that in the touch world? [12:26] ah, that ... yeah, a simple upstart job does that ... in our upgrade management [12:26] ok, this happens in recovery, after the unpack? [12:26] we have that upstart subdir wheer you can drop transition jobs [12:26] (and before removing files that are not anymore in the new image?) [12:26] on first boot after upgrade [12:26] hum [12:26] but /etc/default/whoopsie would disappear, wouldn't it? [12:26] or by developer chice on every boot) [12:27] or we keep it and let the transition job deleting it? [12:27] yep [12:27] ogra_: that's the thing based on session-migration or something else? [12:27] something else [12:27] (because my session-migration is only executed by session) [12:27] right [12:27] ok, was afraid it was going to start too late [12:28] search for boot in /etc/init on a phone [12:28] sounds good then, I'm just afraid we forget about it though [12:28] (and that's just one case) [12:28] there is a subdir and a job executing th bits indside that subdir [12:28] both called boot-$something [12:28] * ogra_ has no phone cnsole around [12:29] ogra_: no worry, but I think as long as we have the pre/post* world and this image-based one, we might miss those transitions on one side or the other [12:29] we surely will [12:29] * ogra_ looksforward to snappy [12:29] ;) [12:30] getting rid of all the deb mess :) [12:30] Suppose I want to replace Ubuntu with Android on the Aquaris e4.5 (I need PGP support for email), can I use fastboot flashall with imgs taken from bq's firmware zips for an hard reset? [12:31] lemmster, i fear thats something you need to ask bq [12:31] not sure you can install android at all, since the partitioning scheme is different [12:32] ogra I believe it can be done because I think I saw someone on Twitter that said he reflashed the phone with Android while he waits for bugs to be fixed in UT [12:32] and ... well ... there are no android buttons on the glass ... [12:32] How can I enable ttyUSB/ttyACM usb serial support for host-mode usb (on a Nexus 5)? [12:32] ogra_: There are manuals on the web to replace android with ubuntu. I would assume the reverse should be possible as well (not that I have tried the former). [12:33] EdwardMorbius: You still happen to have a ref to the tweet? [12:33] ogra that can be solved by using softkeys if firmware allows it and it should I believe [12:33] lemmster, well, i know that there is a different partitioning scheme applied to ubuntu phones ... it might or might not work [12:33] lemmster I will check if I have it somewhere [12:33] this is definitely not something we exercise in development :) === tosky__ is now known as tosk === tosk is now known as tosky [12:35] also keep in mind that the flashing of android might perhaps change the partitioning ... i have no idea how the android zips work in this case ... it might prevent you from re-installing ubuntu then [12:35] lemmster https://twitter.com/sharkdrink/status/581786668480503808 [12:36] EdwardMorbius: Thanks [12:37] lemmster no problem [12:41] so I heard on the last Ubuntu live session that the UT actually has another theme inside, the Ambience but currently there is no way you can switch to it, that correct? [12:45] ogra_: do you know of any documentation for this writable-paths file? [12:45] I just checked and after modifying it (even with 'transition') /etc/default/apport is empty on reboot [12:46] apport [12:46] ? [12:47] whoopsie! [12:47] ah [12:47] :) [12:47] wait WTF now it's not [12:47] now I'm questioning the nature of reality itself [12:48] http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/initramfs-tools-ubuntu-touch/vivid/view/head:/scripts/touch [12:48] line 336ff [12:49] hex line numbers [12:49] heh [12:50] How do I chrt into the lxc container? [12:50] lxc-console -t0 -nandroid [12:50] from the running ubuntu [12:51] Cool [12:51] (you need ctrl-a-q to exit it) [13:06] ogra_: Could not write configuration: Failed to create file '/etc/default/whoopsie.0NIIWX': Read-only file system [13:06] it's trying to do an atomic write [13:06] yeah, that wont wrk [13:06] snazzy [13:08] Laney, that will be a bigger change ... we have a dir for such special cases, but the file needs to be pre-created during build [13:08] what you want is /userdata/system-data/etc/writable/whoopsie and a bind mount like hostname and localtime use [13:09] I remember this hack [13:12] going to go climbing instead :) [13:12] enjoy :) [13:14] rsalveti: ping [13:19] How do I compile a .ko for the hammerhead kernel? [13:20] hey guys, i hope someone can help me. Im working on the port for the one plus one and im struggling to get the /firmware partition to mount correctly. From what i have read there is a script that runs that updates the fstab on boot? I have tried to correct add some code to out the correct mount options for /firmware in the script /usr/lib/lxc-android-config/update-fstab but have had no luck. Any ideas [13:28] spazzymoto, that script just grabs the fstab from android and processes it, if you have a /factory entry in your android fstab it will be automatically created [13:28] err [13:28] /firmware [13:31] hmm [13:31] i do have it in my fstab.bacon [13:31] but if i check /etc/fstab once booted i dont see an entry for /firmware [13:32] and doing an ls /firmware returns nothing either [13:32] if i manually mount it with /dev/mmcblk0p1 /firmware vfat ro,shortname=lower,uid=1000,gid=1026,dmask=227,fmask=337,context=u:object_r:firmware_file:s0 wait i can browse the partition [13:33] http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/initramfs-tools-ubuntu-touch/vivid/view/head:/scripts/touch#L408 [13:33] and http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/initramfs-tools-ubuntu-touch/vivid/view/head:/scripts/touch#L82 for the function [13:35] thank you let me have a look at that [13:39] hmm, looking at this code, i wonder what happens if you have multiple fstab files in that location :) [13:39] spazzymoto, do you have more than one file that starts with "fstab" in yoour android rootfs ? [13:41] ogra, only 1 file hey [13:50] i do see the message [ 16.174895] initrd: mounting /dev/disk/by-partlabel/modem as /root/android//firmware [13:50] ogra_: FYI: I successfully flashed Android with this MKT Flash Tool on Windows and the bq firmware reset image. [13:54] ah mount: special device /dev/block/platform/msm_sdcc.1/by-name/modem does not exist :( will try change this to /dev/mmcblk0p1 [13:55] lemmster, right, as i said, you might not be able t go back ... better test that before you start usingthe phone ;) [13:57] spazzymoto, wont that break android then ? [13:57] ogra_: I will try to go back the moment the features I'm missing are there. [13:57] you usually dont have /dev/mmcblkX there [13:57] Is there an Ubuntu Touch that doesn't require Android? [13:57] dbeal, there is the desktop-next image for PCs [13:57] orga :/ ah i didnt know that [13:58] spazzymoto, you rather want to make sure your kernel provides /dev/block/platform/msm_sdcc.1/by-name/modem ... [13:59] ogra, thanks for the help. Ill go research on how to make that happen [13:59] :) [13:59] spazzymoto, are you sureat all that you need /firmware on the ubuntu side ? [14:00] usually the modem is completely handled inside the container ... ubuntu talks to it via rild [14:01] i might be on the wrong tack. Im seeing these messages in my dmesg [14:01] [ 20.829270] msm_ipc_load_default_node: Failed to load modem [ 20.858718] pil-q6v5-mss fc880000.qcom,mss: mba: Failed to locate mba.mdt [14:01] well, that is something your init.rc in the container should handle i guess [14:02] the kernel prints messages for both systems ... it is sometimes hard to allocate them to the right one ... but i guess this one comes from inside the container [14:04] ok cool :) let me poke around a bit more. Thanks for putting me on the right path [14:36] Is there a phone with Ubuntu that doesn't require Android? [14:36] [02:27] dbeal, there is the desktop-next image for PCs [14:36] dbeal, no [14:36] if you want to make any use of the hardware yu wont get around using the binary drivers [14:37] In my mind, a phone is a small computer. [14:38] sure [14:38] So is the problem that, for some specific hardware (e.g., Nexus 5 camera), the only drivers available are within Android? [14:38] well, if it would just be the camera :) [14:39] modem, all sensors, gps, camera, graphics chip, audio and video codecs [14:39] Are those all binary only drivers? [14:39] yes [14:40] Is there one for host-mode usb serial support? [14:40] that really depends on the device [14:40] Say for example I wanted to connect a Bus Pirate or Aardvark? [14:41] Arduino also has a usb serial interface. [14:42] as i said, that competely depends on the device you run on [14:43] What about this device? [14:43] [ 1245.529600] usb 1-1: Product: FT232R USB UART [ 1245.529625] usb 1-1: Manufacturer: FTDI [14:43] no idea [14:43] what about it [14:45] I would expect a /dev/ttyUSB* or /dev/ttyACM* entry to appear when connecting a usb serial device. [14:45] With a module like this. [14:45] if your kernel supports it, sure [14:45] usbserial 45100 1 ftdi_sio [14:45] Would I need to recompile the kernel? [14:45] and if the USB interface isnt blocked by something the kernel ships [14:46] i have no idea, really [14:46] totally depends on your device and how the kernel is set up by default, what it already does with the usb port etc etc [14:47] I'm pretty new to the Ubuntu Touch architecture, so I'm not entirely sure how the system works with the lxc and how kernel module compilation works. [14:47] this is a bit like you coming to a car shop, showing a random bolt from your car to a mechanic and asking why the radion doesnt work :) [14:47] *radio [14:48] tha also depends on the device ... for the nexus devces we ue normal linux kernel packages in the archive ... the binaries get dynamically pulled in during the build of the android container in the android package ... for all othr devices we build the kernel during the container build instead of using a binary deb [14:53] Hello everybody, I'm trying to figure out if I have a problem with BQ's Aquaris E4.5 Ubuntu Edition. Is this correct place to try it? [14:53] you aretrying to figure out ? [14:53] What is the name of the linux kernel and android container package? [14:54] dbeal, the android package is called android ... the kernel packages are linux-image-$your-devices-codename (i.e. "mako" for N4 or "flo" for N7) [14:56] I don't seem to have an "android" package. [14:56] it is in the ubuntu archive [14:57] I don't have a linux-image package installed either. [14:57] installed ? [14:57] no [14:57] these bits live all in the device tarball [14:57] Denco, most likely yes [14:58] Where is the device tarball? [14:58] My BQ's Aquaris lost 13% of battery in two hours after it was fully charged. I made just 2 phone calls in that time, it was just lying on the table the rest of the time. Is this normal? [14:58] depends on your device once again [14:59] I have an N5. === chihchun is now known as chihchun_afk [14:59] dbeal, we don have device taballs for N5 ... [14:59] dbeal, for that i think you need to talk to Tassadar [14:59] So when I ran a system update this morning, it actually rebuilt the kernel? [14:59] no [15:00] it is rebuilt beofore the device tarball hits the server [15:00] It just downloaded a new container binary? [15:00] Oh, ok [15:00] Is there a package I can install to build my own kernel containers? [15:01] not sure how the N5 is maintained [15:02] might be there is a hammerhead linuc-image package somewhere, no idea [15:02] Does the BQ also use Android? [15:02] yep, there is [15:02] " if you want to make any use of the hardware you wont get around using the binary drivers" [15:02] ^^that applies to all phones ... including the bq indeed [15:03] linux-hammerhead - Complete Linux kernel for the Nexus 5 (hammerhead). [15:04] If I have an N5 device, why don't I need to have linux-image-hammerhead installed? [15:04] because you boot from a boot.img file [15:04] https://code-review.phablet.ubuntu.com/gitweb?p=ubuntu/kernel/trusty.git;a=summary this is the kernel source [15:05] installing the pckage would just waste diskspace ... it wouldnt be used [15:05] *package [15:05] Is there any special reason why the N5 is not officially supported? [15:05] What would I do with the hammerhead kernel package? [15:06] no, no special reason except that canonical doesnt have N5s [15:10] I would be fine with the idea of binary drivers, if they were designed as microkernel-style device drivers. [15:11] well, they are designed as the manufacturer decided ... not much anyone but them can do about that [15:11] its like nvidia on PCs [15:12] e.g., each binary driver operates as an independent process that communicates with a parent kernel solely via message passing. Any process that wanted to talk to the device would then talk directly to the device driver process. [15:12] (except that the binary drivers provided for android are not linked against linux libc but android bionic) [15:13] dbeal, what do you do with the binary daemons that are needed to initialize and configure the driver ? [15:13] It almost seems like the android-as-lxc architecture is a step towards the slippery microkernel slope. [15:13] (note, they are also not linked against libc and wont run on linux out of the box) [15:14] it is the most elegant solution we found to the problem ... [15:15] jolla and firefoxOS have the same issue btw ... [15:15] i don t found anymore the fb messanger...was installed...disinstalled...gone from store :/ [15:16] Good question. [15:17] hey [15:17] hello [15:21] I am developing an Ubuntu Phone application "QML app with C++ plugin (cmake)" but I have so many qml files so I want to separate them by functionality [15:21] ApplicationName [15:21] CMakeListes.txt [15:21] .. [15:21] app [15:21] CmakeListes.txt [15:21] ApplicationName.desktop.in [15:21] tests [15:21] mainComponents [15:21] sub directories of mainComponents [15:21] some files [15:21] some other files [15:21] ApplicationName.qml [15:21] po [15:21] CMakeListes.txt [15:21] ApplicationName.apparmor [15:21] manifest.json.in [15:21] but the "mainComponents" folder dose not appear in the project tree and after some searches I found that I must add CMakeListes.txt in every new folder and this: [15:21] Okay, so you have binary drivers, binary daemons, but you can still build the android lxc kernel? [15:21] add_subdirectory(folder_name) [15:21] instruction in the parent folder CMakeListes.txt file but I don't if there is more to add and I don't know how to create the CMakeListes.txt [15:21] LeousSama, please use a pastebin like paste.ubuntu.com [15:21] thanks [15:22] and you might have better luck in #ubuntu-app-devel with that question [15:22] ok sorry [15:22] dbeal, the kernel luckily desnt care about the libc that is used in the rootfs [15:23] we use a patched android source (see the porting guide from the topic, it explans the kernel changes) [15:23] so the kernel serves both OSes [15:24] But are the binary device drivers all compiled into that kernel? [15:24] no [15:24] No, that wouldn't make sense, would it? [15:24] most of them reside as libs in the container [15:25] or as firmware ... [15:25] or even as daemons [15:25] Ok, so the kernel just has to be binary compatible with the binary kernel modules? [15:25] (which simply include the driver bits) [15:25] right [15:25] and the environment has to provide the files in the right places [15:26] since the binaries usually have hardcoded paths etc [15:26] ifthey read configs or lod other binary stuff [15:26] *load [15:26] Hence the chroot/lxc [15:27] right [15:27] https://developer.ubuntu.com/static/devportal_uploaded/136981fa-6287-49d3-9874-06f40b2e4eb7-cms_page_media/380/ubuntu_touch_architecture.png [15:27] that illustrates it pretty well (a bit outdated though) [15:28] Are the modules in /system/lib/modules/ binary-only, or are they the ones you can build from source? [15:28] (and there is a set of arrows missing, the orange par can indeed always talk directly to the kernel too) [15:28] How do I load or inspect the state of one of the binary-only drivers? [15:28] however you would do it on android :) [15:29] differs per device and driver [15:29] you can enter the running contaner with "lxc-console -t0 -nandroid" [15:29] and poke around in there if you like [15:33] It looks like /system/lib/modules looks like it has linux native modules. I might be able to build something like usbserial or cdc_acm support in. [15:33] sure, you should be able to drop them there and update the module deps === ahayzen is now known as ahayzen_ === ahayzen_ is now known as ahayzen__ === ahayzen__ is now known as ahayzen [15:40] awe_: ping [15:43] peat-psuwit, pong [15:44] awe_: Now I have a working fix for ofono now, what should I do to create a new plugin for it? [15:45] so we maintain ofono for ubuntu in a github repo [15:45] and would be glad to review new pull-requests [15:46] one thing that's different than standard ofono plugins [15:46] # mount -oremount,rw /lib/modules mount: cannot remount /dev/loop1 read-write, is write-protected [15:46] is we use our own plugin to dynamically load a rilmodem type device plugin vs. using the std udevng plugin normally used by ofono [15:47] peat-psuwit, here's the git repo: https://github.com/rilmodem/ofono [15:47] Is it possible to load a .ko from somewhere other than the write-protected /lib/modules? [15:47] sure, you can use insmod ... [15:47] peat-psuwit, and here's the plugin we use to dynamically load device-specific ofono plugins: https://github.com/rilmodem/ofono/blob/master/plugins/rildev.c [15:49] peat-psuwit, maybe the first step would be to create grab a copy of our repo, apply your changes [15:49] have you just changed the existing code directly? [15:50] ( ie. modified the ril.c device plugin, and rilmodem code itself ) [15:51] awe_: I actually edit drivers/rilmodem/sim.c in function sim_status_cb [15:51] to call my function that fire another request [15:52] are you familiar with git/github? [15:52] I also edit gril stuff to support new type of request. [15:52] awe_: I'm pretty good about git. [15:53] ok. then probably the best approach would be to clone our git repo, apply your changes to your own branch, then ask abeato or myself to review [15:53]