=== chriadam|away is now known as chriadam [01:14] robert_ancell, I have another patch I'd like to squeeze into lightdm. This time having it search XDG_DATA_DIRS for config info (to let OEM team customize image) [01:14] whoops, not here === bfiller is now known as bfiller_afk [01:25] is it possible to put ubuntu touch on my samsung galaxy temporarily? [01:25] like a live cd sort of way? === salem_ is now known as _salem === Ursinha-afk is now known as Ursinha === vying is now known as Guest3169 === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [07:43] Hii .. I have compiled phablet kernel with Framebuffer console .. [07:43] the nexus 10 is not coming up with new kernel [07:43] to enable framebufffer console what it require apart from enabling framebuffer console in .config === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [08:04] stgraber: there? [08:04] stgraber: seems our iamges dont work in infra anyumore (e.g. dont install() [08:04] stgraber: have you pipecleaned the system-image update as i suggested yestetrday? [08:06] ogra_: ^^ [08:06] help :) === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chriadam is now known as chriadam|away [09:16] so is phablet-flash finally unsupported? [09:17] tsdgeos: no updates are distributed for a while already [09:17] tsdgeos: dont ask me why we didnt provide a compatibilty link :) [09:17] asac: ?¿ i used it yesterday and got me two different images at least [09:17] tsdgeos: yeah, the new one is ubuntu-device-flash :) [09:18] hwoveer, i would have preferred to just keep shipping the phablet-flash command [09:18] as well [09:18] the command is here [09:18] it just doesn't work === alan_g is now known as alan_g|reboot [09:19] asac, stgraber discussed it with plars and doanac` last night, they even shared code for a fix [09:19] not sure why that didnt lnad [09:19] *land [09:22] ogra_: i think something landed [09:22] ogra_: but noone reran the jobs? === vying is now known as Guest42661 [09:22] e.g. we offlined the device [09:22] and now its working [09:22] ah, k [09:22] anyway, discussed was that if there is an issue we revert [09:22] because CI team is all into something customer [09:22] we had three builds this night, intresting that nobody noticed it over three images :/ [09:23] ogra_: noone is there [09:23] hence i wanted it to be reverted if something doesnt go well [09:23] they wre ehere to discuss about it [09:23] they are all on-site somewhere [09:23] ogra_: yes, but they were in constant firedrill mode for something else [09:23] reverting isnt done easily [09:23] well, i asked if we could revert before doing it [09:23] i didnt hear a NO [09:24] right, but you need stgraber to do it [09:24] right, hence i asked stgraber and you to pipeclean and if anything goes bad revert right away [09:24] i am not asking these things for fun :) [09:24] anyway [09:24] > [09:25] also not to you reall [09:38] Good morning all; happy Biodiesel Day! :-D [09:52] alecu: ping [09:59] hey everyone, i'm trying to get multitouch working on desktop via a magic touchpad. unity seems to get the 3- and 4-finger tap gestures to show dash and window switcher, but i can't do multifinger swipes or pinches. any idea where i should start looking ? === dandrader is now known as dandrader|afk [10:13] nerochiaro, i think #ubuntu-x is the better place for this [10:17] ogra_: ok, thanks [10:17] popey, am i blind or is our topic gone ? [10:17] i see it [10:17] (is is my client messed up [10:17] ) [10:18] aha, k [10:18] silly focus issues on 12.04 === netcurli_ is now known as netcurli [11:03] Morning all === dandrader|afk is now known as dandrader [11:40] cjwatson: Hellor, do you know when the 0.4.19 click will land? [11:40] bzoltan1: hopefully today - I want to sort out https://code.launchpad.net/~mardy/click/lp1245826/+merge/204674 first [11:41] if that looks like it's taking too long I'll try to land click first [11:42] cjwatson: I wold be happy to see it as soon as possible... we have a fix for the click chroot cmake problem and that needs a healthy chroot. [11:42] bzoltan1: your comment on that MP implies that you are creating ubuntu-sdk-13.10 chroots using trusty, which is extremely worrying. please stop [11:42] if you create them using saucy you shouldn't have this problem [11:43] (https://code.launchpad.net/~xnox/click/v8-is-gone/+merge/211485) [11:43] cjwatson: one can not create saucy chroot either due to an other problem [11:43] what problem is that? [11:43] cjwatson: it fails ... i do not have the logs in hand [11:43] we should fix that separately [11:43] cjwatson: just try click chroot -a armhf -s saucy create [11:43] my internet is very slow [11:44] cjwatson: let me produce the logs then [11:44] it would be most helpful if somebody with more than 2Mb/s could try it [11:45] [also, the correct invocation is click chroot -a armhf -f ubuntu-sdk-13.10. the -s option is really only for debugging] [11:46] in particular if you leave out -f then you'll get different results once click chroot's default framework changes [11:47] which will have to happen at some point soonish [11:47] cjwatson: OK, I will use only the arch and the fw [11:49] cjwatson: http://pastebin.ubuntu.com/7113628/ [11:50] hm, interesting [11:50] thanks, I'll see if I can disentangle that somehow [11:52] cjwatson: that might ease some pain [11:53] the cross toolchains are a bit delicate [11:55] :) that is a rather mild statement [11:56] I have been infected by the British habit of understatement [11:57] ogra_, any idea why between the last two promoted images (Feb 20 and Mar 14) adbd would no longer start and the device seems offline? I see the mtp server config file had been added, you did some adbd/ssh work, but any other clues are helpful [11:57] ogra_, this is not a nexus device [11:57] janimo`, adb was moved to start after the container [11:58] to prevent the numerous disconnect issues we had when it gets dynamically reconfigured all the time [11:58] ogra_, but once unity8 is up and working, adbs should also be ready right? I have a GUI just device offline from adb [11:59] (we leave the whole configuration to android now and only run the adbd on the ubuntu side, that saves from getting like 1000 popups on your desktop and about 4 reconnects, which breaks automation in various places) [11:59] cjwatson: well, it is better than being born with the continuously exaggerative Hungarian habit [11:59] janimo`, yeah, it should just come up [12:00] mtp should not have any influence [12:01] janimo`, could it be that your android container misses the forcing of the persistent property for adb/mtp in init.rc ? [12:02] # adb force hack [12:02] setprop persist.sys.usb.config mtp,adb [12:02] your init.rc should have this one [12:04] Regarding https://bugs.launchpad.net/touch-preview-images/+bug/1293797 can someone tell me what service the device contacts to check for updates? [12:04] Ubuntu bug 1293797 in touch-preview-images "nexus 4 "COMMAND ERROR" when updating apps" [Undecided,Confirmed] [12:05] Click package updates, that is. [12:05] * ogra_ sighs, why do poeople not file bugs against the packages [12:06] JamesTait, i think beuno might know [12:06] (they should query the store server) [12:07] JamesTait, click index, yes [12:07] JamesTait, it sends a list of packages it has, and the server replies with what's on the server, and the client compares [12:07] IIRC [12:07] alecu knows the details [12:08] Ah, it's querying click index? Interesting. [12:09] Hm, I don't recall writing any code that accepts a list of packages. :-/ [12:15] JamesTait, I forget who did, maybe it was pindonga? [12:15] JamesTait: I can answer that bug, they are on the 14.04-dev version to make them work for the QT5.2.1 version so they shouldn't install on a system that isn't on QT5.2.1 [12:16] beuno, I thought so, but he wasn't that sure - I think we've figured out what the problem is now though. [12:16] cool cool [12:16] davmor2, at a high level, yes. :) [12:28] bzoltan1: OK, I think I've fixed saucy chroot creation - testing that (sloooowly) and then I'll see about a landing [12:29] cjwatson: is there anything I can test? [12:29] (it needed a dist-upgrade of the chroot before trying to install the cross-compiler etc.) [12:29] bzoltan1: you can grab lp:click, do debian/rules build, then sudo ./run-click chroot -a armhf -f ubuntu-sdk-13.10 create [12:29] cjwatson: Ohh... that one. Super [12:30] also fixed an embarrassing bug where we weren't actually suppressing daemon startup in the chroot as was always intended [12:31] which I confidently asserted in some UDS session worked fine :) [12:32] bzoltan1: going out to do some admin, hopefully my test will have finished by the time I get back [12:32] cjwatson, hi, is the click frontend going to be rewritten in Vala at one point too? [12:33] ugh [12:33] janimo`, i thought C all the way :) [12:33] cjwatson: Thanks a lot. I am testing... [12:33] ogra_, well seeing the libclick is in Vala... [12:33] ogra_, which is C after all === MacSlow is now known as MacSlow|lunch [12:46] cjwatson: I could create an armhf chroot with ubuntu-sdk-13.10 fw without any trouble. Thank you. === _salem is now known as salem_ [12:53] cjwatson: how will the new splitted frameworks actually work. Will you need a own chroot for -html, -qml, -papi? [12:53] cjwatson: or will a 14.04 framework just support all of them [12:53] erm chroot i mean === alan_g is now known as alan_g|lunch === MacSlow is now known as MacSlow|lunch === dandrader is now known as dandrader|afk === MacSlow|lunch is now known as MacSlow === phantom is now known as Guest37914 [13:25] is anyone else's flo turning off after some idle time? [13:25] it's the umpteenth time I thought it discharged, but it's fully charged... just turned off [13:26] Saviq: hm, a kernel crash would probably reboot it [13:26] rsalveti, not reboot [13:27] rsalveti, it powers off [13:27] Saviq: just noticed mine did the same [13:27] was completely off here [13:27] unless my cats are having fun with me, this is crazy [13:27] Saviq, i noticed it too, but it didnt happen for a while now [13:27] Saviq, was your device wired up ? [13:27] rsalveti, ogra_, sounds like a broken deep sleep mode or something [13:27] ogra_, no [13:27] nothing changed in the kernel side [13:28] rsalveti, i noticed it randomly since i run flo [13:28] Saviq, yes; rsalveti I thought I told you about that one [13:28] but it only seemed to happen to me when the device was kept wired over night [13:28] havent seen it in over a week now [13:29] (and i didnt have to charge in over a week either !!!) [13:29] ogra_, maybe it did power off, then ;) [13:29] Saviq, because it was finished charging ? [13:29] :P [13:29] ogra_, no idea, something like that [13:29] * Saviq files a bug [13:30] janimo`: I'm thinking about it at some point, but (a) I want "click build" to remain pure-Python for portability, (b) the performance-sensitive paths should all be rewritten to use libclick rather than fork/exec anyway [13:30] bzoltan1: great, thanks [13:30] bq27541_get_psp temperature was quite high in my syslog [13:30] it definitely drains significantly less power than all other devices i have [13:30] before it decided to shutdown [13:30] sergiusens: ping, ubuntu-device-flash reboots my maguro in to recovery mode, it doesn't flash my phone like phablet-flash did. is there a flag I am missing? [13:30] rsalveti, add a fan then :P [13:30] ogra_: and it also finished charning [13:30] *charging [13:30] cjwatson: is there a simple way to ask from a click chroot what frameworks are available [13:30] ? [13:31] zbenjamin: I'm open to feedback on that; I suspect it will generally be more convenient to have a single chroot, and have the ABIs enforced by other things like the review tools. What do you think? [13:31] robotfuel, what are you trying to do? [13:31] bq27541_get_capacity = 0% ret= 0 [13:31] bq27541_get_psp temperature= 349 [13:31] wow [13:31] bzoltan1: No, it's a bit light on query interfaces right now. Suggest an interface, maybe? [13:31] sergiusens: flash the phone automatically like phablet-flash did. all of the phones in ci are broken, because they are still trying to use the phablet flash command. [13:31] is that in farenheit ? [13:32] Saviq: check yours as well [13:32] ogra_: might not be in any scale we know [13:32] robotfuel: maguro isn't supported passed the 4.4.2 android update [13:32] yeah [13:32] tedg, what's the bug # for that FFe for indicator-sound? [13:33] robotfuel, ubuntu-device-flash --wipe channel trusty-proposed ... like that? [13:33] ogra_: actually, bq27541_get_capacity = 0 means it was completely out of battery === dandrader|afk is now known as dandrader [13:33] sergiusens: yes, I have a maguro which is probably the issue [13:33] mterry, bug 1293677 [13:33] bug 1293677 in indicator-sound (Ubuntu) "FFE: Export data to accounts service" [Undecided,New] https://launchpad.net/bugs/1293677 [13:33] rsalveti, yeah, but the 349 $whatever for the temp are weird [13:34] rsalveti, 250...260, doesn't get much higher [13:34] robotfuel, I don't think the phablet-flash being broken is related to ubuntu-device-flash [13:34] robotfuel, I don't think there's devel-proposed avail for maguro anymore [13:34] sergiusens: phablet-flash looks for a json file that does not exist on the server anymore [13:34] cjwatson: I will figure out something. For now I will rely on the name of the chroot [13:34] sergiusens: we should have it still [13:35] robotfuel, for that, do --channel ubuntu-touch/devel-proposed [13:35] rsalveti, it's just that robotfuel is mixing up two problems [13:36] there are still both channels for maguro ... but devel/trusty only has 188 as the last promoted image [13:36] (and we dont promote any for it anymore) [13:36] rsalveti, yeah maguro is there; [13:36] ogra_, rsalveti bug #1294130 [13:36] bug 1294130 in android (Ubuntu) "Nexus 7 (flo) powers down randomly" [Undecided,New] https://launchpad.net/bugs/1294130 [13:37] Saviq, thx [13:37] Saviq: can you add your /var/log/syslog in there? [13:37] rsalveti, sure, but I've rebooted twice since [13:37] rsalveti, dmesg.? [13:37] Saviq: no worries [13:37] Saviq: want the messages before it went off [13:38] Saviq: you should have it unless you flashed your device [13:38] robotfuel, and the full descriptions for ubuntu-device-flash and installation should be in the wiki [13:38] 47MBs syslog - sure, it'll be there... [13:39] ouch, but gzip should help quite a lot [13:39] robotfuel, https://wiki.ubuntu.com/Touch/Install#Step_4_-_Downloading_.26_Deploying_Image_to_Device [13:40] rsalveti, https://bugs.launchpad.net/ubuntu/+source/android/+bug/1294130/comments/1 [13:40] Ubuntu bug 1294130 in android (Ubuntu) "Nexus 7 (flo) powers down randomly" [Undecided,New] [13:40] Saviq: thanks [13:52] asac: hmm, so on my side everything went well including the end to end tests on devices here. As for QA, my understanding was that sergiusens figured out that they were using broken code ahead of time and provided a fix, I wasn't told of any problem... [13:53] * stgraber continues to read backscroll [13:54] ogra_: hey ogra, thanks for the help yesterday. I followed the instructions to prepare the ubuntu touch filesystem and got the following when booting up: 4.258965] init: Failed to bind socket 'property_service': No such file or directory [ 4.268258] init: init startup failure [ 4.272900] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000100 . Does anyone have thoughts on what's going on? [13:54] stgraber, so after a couple I recalled that ci used the device specific channel's json to figure out if there was a new build available; that spanned into a couple of branches that needed updates [13:54] stgraber, all fine [13:55] oh, ok, because they also used the hardcoded URLs instead of parsing channels.json... [13:55] stgraber, I'm going to see if I can provide an alternative for them; exactly === alan_g|lunch is now known as alan_g [13:55] stgraber, seems they were confused as I was by the definition of 'links' where we thought they were at the file layout level :) [13:56] as in ln -s ubuntu-touch/saucy saucy [13:56] well, at least with that change we managed to spot all the tools that aren't spec-compliant in one go ;) [13:57] cjwatson: yes single chroot sounds way better [13:57] zbenjamin: yeah, I think otherwise we end up creating a combinatorial explosion of chroots, it sounds weird [13:58] stgraber, agreed; [13:58] cjwatson: completely agree, thats exactly why i'm asking because each chroot will be a own target in creator, that would be totally confusing :) [13:59] ogra_, found the adbd issue. The change in the init script to not poke sysfs much but use the functionfs mount broke it here [13:59] zbenjamin: feel free to play with lp:click r403, which is what I'm trying to get landed now [13:59] apparently that other branch of code was not preserved at all. This device has functionfs but since I first heard of it today I'll need to see why it does not behave [13:59] cjwatson: ok [14:00] janimo`, oh ? do you have a device with functionfs ? [14:00] ogra_, well, mount -t functionfs did mount something and that f_ffs entry is there in syslog so I guess [14:01] as I said I do not know what fs this is or what exactly that init script change was meant to fix [14:01] janimo`, and your android doesnt configure functionfs for the gadget ?= [14:01] ogra_, I need to dig into that. I tested the functionfs mount in recovery mode. In ubuntu boot mode I cannot adb shell in [14:01] janimo`, its like "a unionfs to make the gadget device have more than two functions" [14:02] usually g_android can only handle two functions [14:02] ogra_, I may be missing some kernel config options or it's just something not tested much in this kernel. Does android use this? [14:02] the functionfs extends that [14:02] you should see it being handled in init.$device.usb.rc [14:03] *if* your device supports it [14:03] ogra_, for now I'll override that script with the old version in our rootfs so we can test recent images, but I need to find the cause [14:03] robotfuel, just show me the code/script branch and I'll propose an MR [14:03] worst case just drop it from the kernel [14:03] we dont even support more functions than adb and mtp yet [14:04] ogra_, do I need extra stuff in init scripts? I just diffed the two ubuntu image fss and this is how I found the change. If you have a pointer to a bug or changelog for this change which may touch other files that would be good :) [14:04] no, we dont have a bug for it [14:04] changelog entry should be in android-tools-adbd [14:04] iirc [14:05] janimo`, http://paste.ubuntu.com/7114228/ [14:06] ogra_, thanks. So on devices where there's no fgunctionfs is the old codepath not needed either? [14:06] right [14:06] those commands poking sysfs/andoird [14:06] sergiusens: they are using phablet-flash in a job, pinky is going on site to switch to ubuntu-device-flash, in case something goes wrong. [14:07] janimo`, we wait for the container to have it configured before we start adbd so the sysfs hacks arent needed anymore ... what *is* needed is the line in init.rc that forces the persist property [14:07] sergiusens: ubuntu-device-flash reboots my phone to the recovery screen, but doesn't flash it. I am using a maguro. [14:07] robotfuel, you can still use phablet-flash for the time being [14:07] robotfuel, just change the channel to ubuntu-touch/devel-proposed [14:07] Hello! I am looking for the kernel sources for 3.0.0.3-maguro for building a kernel module. Where are they, and do I have to do something special to my Ubuntu touch phone in order to build and use kernel modules? [14:07] sergiusens: yes, we just tested your changing the channel and it works :) [14:08] robotfuel, afaik, your box is on raring; I would wait for the switch to saucy [14:08] cjohnston: ^^ [14:09] robotfuel, that was what we discussed with plars and doanac` fwiw [14:09] yup [14:09] sergiusens, robotfuel: that's happening this friday [14:14] sergiusens: I opened a ubuntu-device-flash bug https://bugs.launchpad.net/ubuntu/+source/goget-ubuntu-touch/+bug/1293937 who do I ping to triage? [14:14] Ubuntu bug 1293937 in goget-ubuntu-touch (Ubuntu) "ubuntu-device-flash fails to locate image when using relative revision" [Undecided,New] [14:15] robotfuel, that already has an MR which I asked mandel to review [14:16] sergiusens: ok thanks [14:25] ogra_: Do you have any thoughts on the error? I haven't been able to find anything. [14:26] hedz09, not really, thats just the last line ... the error is most likely somewhere before [14:28] ogra_: is `ubuntu-device-flash --revision -1` broken? (channel trusty-proposed) [14:30] ogra_:hmm, ok. Thanks. [14:30] barry, ask sergiusens .... thats his baby :) [14:30] ogra_: ack :) [14:31] sergiusens: so, without --revision -1, image 243 gets installed. but s-i.u.c seems to have image 242 [14:31] % ubuntu-device-flash --channel=trusty-proposed --revision -1 --bootstrap [14:31] 2014/03/18 10:27:14 Expecting the device to be in the bootloader... waiting [14:31] 2014/03/18 10:27:14 Device is |manta| [14:31] 2014/03/18 10:27:16 Failed to locate image -1 [14:31] [14:32] barry, yeah; there's an mr for that [14:32] barry: command line help says "--revision=" not --revision [14:32] ? [14:32] popey: nope. `--revision -1` also fails (but cli should accept both) [14:32] ☹ [14:33] barry, popey yeah, cli works with both; relative versions against latest need a merge [14:33] sergiusens: ack. thanks, i'll use an absolute number for now [14:34] barry, should land today; getting reviewers is what takes time :-) [14:34] popey, care to take a look https://code.launchpad.net/~sergiusens/goget-ubuntu-touch/relative/+merge/211523 [14:34] hah, saw that coming [14:35] i can test at the moment, i dont want to flash my devices [14:36] popey, just see that it start downloading ;-) [14:38] ack === gatox is now known as gatox_lunch [14:59] sergiusens, hey! do you know why I am seeing this error when doing click setup ? http://paste.ubuntu.com/7114344/ [15:00] om26er, you are not using unity8 from trunk is my guess [15:00] sergiusens, its the latest unity that comes on the image [15:02] om26er, which version; I'll flash and check [15:03] sergiusens, its the very latest image 243 [15:03] om26er, need to look at the latest unity8, maybe something changed there [15:03] sergiusens, so the problem was I did not had ubuntu-dev-tools installed due to some reason [15:03] which provides pull-lp-source [15:03] I remember installing phablet-tools with --no-install-recommends that's probably the reason [15:04] oh [15:04] wow, why ? [15:05] stgraber: hi, I'd want to add that ubuntu-touch "flavour" (is that what you call it?) to my server too - are all the code changes required on launchpad already? [15:06] Tassadar: yep, the production server is running the branch as it is on Launchpad, I can copy/paste you the si-shell commands I did yesterday to do the change [15:06] yeah, that would be nice, thanks [15:07] ogra_: here's the full log: http://pastebin.com/ass2A2ws [15:09] hedz09, that looks very much like you are booting into android [15:09] instead of into ubuntu [15:09] stgraber: also, I suppose it should be fine to add flavor as option when selecting the image to my app (it currently has "channel" and "version", but the channel names are bit too long right now). That would require for the naming schema to be consistent, like there can only be just one slash in the channel name, dividing the flavor name and channel name. Is that what you plan to do? [15:11] for channel in pub.list_channels().keys(): [15:11] pub.rename_channel(channel, "ubuntu-touch/%s" % channel) [15:11] pub.create_channel_redirect(channel, "ubuntu-touch/%s" % channel) [15:11] if "alias" in pub.list_channels()["ubuntu-touch/%s" % channel]: [15:11] pub.change_channel_alias("ubuntu-touch/%s" % channel, [15:11] "ubuntu-touch/%s" % pub.list_channels()["ubuntu-touch/%s" % channel]['alias']) [15:11] Tassadar: that should do the trick ^ [15:12] Tassadar: in theory we could add more sub-directories in the path but I can't think of a good reason to do that at the moment, so just splitting on / should be fine for now [15:12] ogra_: I put the contents of out/target/product/root on my root partition. Are these the wrong ramdisk files? [15:12] stgraber: okay, thanks. I'll make it work correctly even when there is more than one slash then [15:12] hedz09, you want the contents of the tarball in your root partition [15:13] hedz09, (the ubuntu rootfs) [15:13] well, in fact you want a file called system.img in /userdata/ with the rootfs in it [15:14] like rootstock creates [15:14] (the initrd will find it and mount it as /) [15:17] ogra_: my current setup is: root partition contains root, /userdata/ has the system.img with the ubuntu rootfs and android-ramdisk in it. From what you're saying it seems that my root partition is incorrect. Which initrd do I put on my root partiton (if the userdata partition contains the ubuntu rootfs in system.img) [15:18] how can I look at the last commit in bzr, like git show [15:18] hedz09, you root partition should only contain /userdata [15:19] Tassadar: bzr log -p -rlast: [15:19] Tassadar: I use bzr alias show="log -v -p -n1 --long" [15:19] thanks, I'll add that too [15:19] so that "bzr show -rlast:" works [15:22] ogra_: ok. I'm using a prebuilt kernel so my build doesn't make a boot.img so I have to make it myself. What ramdisk should be packaged with it? [15:25] bzr is a bit...different from git which I'm used to. And now I have a merge conflict, guess I'll have to learn how to merge properly in bzr :x [15:29] hedz09, ah, you should have an ubuntu-ramdisk file in your build dir that was downloaded during build [15:29] oh, when I merge with some other branch in bzr, I don't get history from that branch into my tree, that's unusual === dholbach_ is now known as dholbach [15:32] ogra_: I have android-ramdisk.img and ramdisk.img. The ramdisk.img has init as a shell script -- is this right? [15:33] yes [15:34] ogra_: file adbd [15:34] adbd: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.24, BuildID[sha1]=dd06a054f09b70ae16fc2637a9ca78d22c0e9a20, stripped [15:35] rsalveti, hmm [15:35] then it should start for him as well [15:36] rsalveti, could you give him a binary initrd ? [15:36] ubuntu/ubuntu-touch-generic-initrd/usr/lib/ubuntu-touch-generic-initrd/initrd.img-touch [15:36] just need to grab that one from the package itself [15:36] ubuntu-touch-generic-initrd [15:37] same as I do for the x86 emulator image [15:37] oh, we build it for x86 ? [15:37] ogra_: I tried using this before and the kernel couldn't find /init : [ 2.903520] Failed to execute /init. Attempting defaults... [ 2.913137] Kernel panic - not syncing: No init found. Try passing init= option to kernel. [15:37] i thought we didnt [15:37] ogra_: we build it now ;-) [15:37] had to change to support x86 [15:37] hedz09, then your initrd is broken ... it does all the mounting of the rootfs before swtiching to init [15:37] rsalveti, ah, that was the missing bit :) [15:37] yeah, I'd guess that the data partition is not mounted [15:38] hedz09, you do *not* unpack the ramdisk anywhere [15:39] it needs to come from the boot.img (or your u-boot setup needs to load it from somewhere) === gatox_lunch is now known as gatox [15:41] hedz09, looking at http://pastebin.com/ass2A2ws you never load the initrd [15:42] find out how to configure your u-boot in a way that it loads it [15:42] and put the img file next to your zImage [15:43] i guess the guys in #beagle can help pointing you to docs worst case [15:43] davmor2: I fiddled with the nexus sheet formatting because a user showed me a screenshot showing it to be hard to read on windows ⍨ [15:44] just ask him to install ubuntu in a VM :P [15:44] :D [15:45] or add: "windows is deprecated, please use a modern OS like Ubuntu" at the top of the sheet ;) [15:45] haha [15:45] is the "unlock sim" dialog kind of broken in image 244? [15:45] Hi, quick question: how can i restart unity8 in the emulator manually? [15:45] it reacts very slow.. so slow that I cannot properly enter the pin [15:46] and then it just dies before I click done [15:46] t1mp, 243 had some bad Mir issues, not sure that already got reverted properly in 244 [15:46] tester4711: adb shell; sudo -iu phablet initctl restart unity8 [15:47] ogra_: okay. could be mir also, I haven't tried much else besides the unlock sim yet [15:47] t1mp, start the system-settings ... if it stays white until you touch the screen, thats the issue [15:47] rsalveti: when will the x86 emulator be released? [15:47] ogra_: I've been following TI's Android image for the structure. In hindsight this was a bad route. For their Android image, they make a boot, rootfs and userdata partition. Boot has u-boot.img, MLO, zImage, uEnv.txt unpacked. Rootfs has the ramdisk unpacked as well as the contents of system (uEnv.txt passes an argument to the kernel to indicate this as the root partition). I was able to build CM and construct a similar struct [15:47] Tassadar: yeah you do, at least after you commit [15:48] Tassadar: you might want to use bzr log -n0 to see the merged history [15:48] cwayne: when it's done :P [15:48] @cwayne: thx. [15:48] ogra_: uEnv also passes to the kernel the location of the init binary [15:48] hedz09, yeah, bad idea ... load the initrd from u-boot, have a system.img in /userdata on your root partition and you should boot into ubuntu [15:49] you shouldnt need to pass the init= arg [15:49] rsalveti: :D [15:49] cwayne: working on it, should have more news later this week [15:49] cjwatson: what happens to the revision numbers then? do they change? [15:49] the kernel usually looks in known places anyway [15:49] (/sbin/init and /init at least) [15:49] ogra_: yes, system settings stays white [15:49] Tassadar: revision ids stay the same; revision numbers depend on the branch you're on [15:49] (--show-ids to see the ids) [15:50] t1mp, then blame Mir for the PIN stuff too i guess ... see if it works on 241 or 242 [15:52] ogra_: ok, so on my boot partition I should put initrd (just ensuring I understand this correctly so I can move away from what TI did) [15:53] you should put the initrd and change your u-boot configuration to load it [15:55] ogra_: put initrd where, sorry? [15:55] next to the zImage file ... wherever that lives [15:56] (most likely a vfat boot partition) [15:56] (and no, i dont know what you need to change for u-boot to make it load it ... you will have to find out yourself) [15:57] ogra_: ok, I'll look into it. Thanks for the help. === chihchun is now known as chihchun_afk [16:03] ogra_, what branches did you have on your machine when you made that boot animation video? [16:03]