=== chriadam|away is now known as chriadam [00:53] Quick question: I've built the current (raring) image, and now I'm trying to move over to the flipped image. How to I get phablet-dev-bootstrap pointed at the saucy code? Doesn't seem to be an argument for which branch to use... [00:54] in other words, what command do I run to init a saucy phablet repo? [01:09] hi [01:14] Hello. === roasted_ is now known as roasted [01:24] RAOF: Ubuntu Touch is simklar to Android since based on that mod ? [01:24] yes ? [01:25] Not really. [01:25] Ubuntu Touch is about as similar to Android as Android is to Ubuntu (ie: they use the same kernel) [01:26] RAOF: ok fine [01:26] RAOF: so what about that Ubuntu Edge then? the Ubuntu on their is usign Ubuntu Touch [01:26] well if those phoens actasuly get made [01:26] Correct [01:27] Oh. Ubuntu Touch is like Android in that it's usable on phones and tablets, if that's what you mean. [01:27] yeah I know a bit about it already, but how differnet is to Android? [01:28] Everything above the kernel is different [01:28] As for user-experience, it's... different? [01:29] It's quite a lot like desktop Unity, if desktop Unity were to be made as a phone/tablet UI? [01:30] yeah seen some screnshots and videos of it [01:30] ,but that's it really [01:30] seems it's like you say Unity for phones [01:36] themikem: repo init -b phablet-saucy -u git://phablet.ubuntu.com/CyanogenMod/android.git [01:36] if you already have a repo sync for phablet-10.1 [01:36] just do repo init -b phablet-saucy and repo sync === _xtr3m3 is now known as xtr3m3 === _salem is now known as salem_ [03:20] hi === Aww_ is now known as Aww === chihchun_afk is now known as chihchun === salem_ is now known as _salem === chihchun is now known as chihchun_afk === Namidairo`bnc is now known as Namidairo === chihchun_afk is now known as chihchun === JBassett_ is now known as JBassett [04:17] So they said the edge will work on sprint and Verizon, is that assuming volte or is it just data only? === JBassett_ is now known as JBassett [04:32] xboxfanj: its 100% assuming volte, it wont work for cdma for phone. [04:33] OK that's what i figured [04:33] Thanks [04:33] Not sure i can gamble on Sprint to do that [04:34] I guess i could buy a used one if they do though === Andro|AFK is now known as andromorphone === andromorphone is now known as Andro|AFK === Andro|AFK is now known as andromorphone === andromorphone is now known as Andro|AFK === Andro|AFK is now known as andromorphone === andromorphone is now known as Andro|AFK === Andro|AFK is now known as andromorphone === andromorphone is now known as Andro|AFK === Andro|AFK is now known as andromorphone === andromorphone is now known as Andro|AFK === Andro|AFK is now known as andromorphone === andromorphone is now known as Andro|AFK [07:13] gusch: hi! [07:14] sil2100: ho [07:14] gusch: we encountered some problems with gallery-app AP tests - those look like something changed in the ui-toolkit [07:15] gusch: on the other hand, it seems not have been introduced just now, since it fails with todays version of ui-toolkit and also with the one from 31th [07:15] gusch: let me show you a link to those, I'll create a bug report [07:16] sil2100: ok [07:18] gusch: here are the test results - this test uses the already released ui-toolkit, which is strange [07:18] gusch: we also got the same failure tooday on SDK tests using the latest daily ui-toolkit, as we run a few Apps tests on SDK stack releases [07:18] (to make sure it doesn't break anything) [07:18] Filling out bug [07:22] sil2100: where are the results? [07:23] gusch: https://bugs.launchpad.net/gallery-app/+bug/1207647 [07:23] Ubuntu bug 1207647 in Ubuntu UI Toolkit "gallery-app autopilot test failures - 'NoneType' object has no attribute 'width'" [High,New] [07:23] gusch: ah! Right, heh ;) [07:23] gusch: I forgot to paste the link, but the bug has those [07:23] * sil2100 needs coffee [07:26] sil2100: omg - SDK changed the tabbar implementation - I'll check/update the gallery autopilot tests [07:28] * sil2100 sighs [07:28] didrocks: morning! [07:28] hey sil2100! [07:29] shouldn't we have some deprecation methods? [07:29] especially on the sdk [07:29] didrocks: we have #1207647 blocking the apps and sdk stacks [07:29] LP: #1207647 [07:29] Launchpad bug 1207647 in Ubuntu UI Toolkit "gallery-app autopilot test failures - 'NoneType' object has no attribute 'width'" [High,New] https://launchpad.net/bugs/1207647 [07:29] sil2100: this did block sdk as well, right, as we run their tests? [07:29] thanks for opening the bug :) [07:29] asac: not sure if you want to track that one ^ [07:29] didrocks: strange thing, since it didn't block SDK last time! [07:30] sil2100: was it failing? [07:30] didrocks: although now SDK clearly fails on gallery-app tess [07:30] *tests [07:30] I think we added those in between, right? [07:30] (like since last time to detect that sdk breaks apps) [07:30] didrocks: last time no, now it does, but it already released the changed one it seems - since as I wrote in the bug, Apps stack uses the released version of ui-toolkit and fails there ;/ [07:31] didrocks: which is strange as it should fail [07:31] didrocks: since now SDK fails properly on the gallery-app tests [07:31] But why didn't it during last release? [07:32] lool: where are you? [07:33] sil2100: btw, you can look at unity8 now as well :) [07:33] didrocks: i cant track anything :) [07:33] sil2100: let me check if we didn't had to set some relax on the treshold in past [07:33] i will ask every day i guess [07:33] asac: come on, you like doing this ;) [07:33] asac: anyway, we'll push upstream [07:34] and now there is a bug opened [07:34] didrocks: already poked Saviq ;) [07:34] sil2100: Saviq is travelling [07:34] sil2100: please try bzoltan [07:34] sil2100: yes sir [07:35] sil2100: no relax, the treshold is 0 [07:35] sil2100: didrocks: Before anybody shouting "SDK fails" it would be fun to see the logs... [07:36] Kind of tired of watching failures what has nothing to do with the sdk :) [07:37] bzoltan: there is a bug I pasted to gusch about that [07:37] sil2100: details please [07:37] One moment [07:37] bzoltan: https://bugs.launchpad.net/gallery-app/+bug/1207647 [07:37] Ubuntu bug 1207647 in Ubuntu UI Toolkit "gallery-app autopilot test failures - 'NoneType' object has no attribute 'width'" [High,New] [07:37] bzoltan: I don't assume anything! I just fill in bugs ;) [07:38] sil2100: I am watching ot [07:38] it [07:38] bzoltan: seems like SDK changed the nav-bar implementation, so I'll have to update the autopilot tests [07:39] gusch: seems or you know? [07:39] bzoltan: I looked about 1 minute on it - so for now I'm only assuming ... [07:41] gusch: cool... just for the sake of good spirit :) would you be so kind to make sure that you point to the right place before tossing the SDK under the bus :) Recently it become very fashionable to blame the SDK without being sure that the problem is not caused by something else... [07:41] sil2100: ^ [07:42] bzoltan: agreed, but as I said, I don't blame anyone! I just observe and fill in bugs so that everything can be recognized, so I said 2 bug targets ;p [07:44] sil2100: No worries :) Noone got hurt :) I just would like to be very sure that it is really and SDK issue... before we jump on it and spend time to fix some other project's bug [07:48] why is the compiled kernel suddenly too big for the boot partition on galaxysmtd flipped? [07:51] gusch: so, who is looking at it a little bit closer? [07:51] gusch: as this is blocking all stacks as SDK won't publish (as this test failed) [07:51] didrocks: give me 5 minutes, then I'll investiage [07:51] gusch: ok, thanks, please keep us posted [07:52] rather than finding who is to blame exactly, let's try to first get the things fixed as top priority [07:52] unblocking everything [07:52] then, we can see what we can do :) [07:57] is a boot.img always a kernel and ramdisk? [07:57] yes [07:58] gusch, bzoltan I changed the TabBar implementation [07:58] I updated the emulators.py in UITK autopilot tests, I was assuming that apps use those, but I'm not really up-to-date on the tests [07:58] ok, well the last working boot.img i have is 5.8MB... the new one is 10MB and the kernel alone is 8MB [07:58] gusch, sil2100: Let's arrange a public humilation session for timp :D [08:00] timp: ok - I'll have a look at your emulator.py and update the gallery tests [08:01] gusch: you can see the changes here https://code.launchpad.net/~tpeeters/ubuntu-ui-toolkit/tabs-restructure/+merge/177131 (search for emulators.py). they are not very big [08:01] gusch: I think we should provide something like this as a library that the apps can call [08:02] elopio: hi. We moved the discussion about the emulators here [08:02] hey. [08:03] elopio: are the emulators from UITK ready to be used by the app autopilot tests? [08:03] elopio: my MR that change the tab bar broke the tests in apps [08:03] timp: yes, the filemanager and dropping letters are samples of what needs to be done to use the emulators. [08:05] gusch: do you need to do a quick fix now? the better solution seems to use the UITK emulators [08:06] timp: quick fix now (some people seem to be nervous here ;), and then I'll try to use your emulators [08:06] ogra_: do you have a maguro? [08:07] gusch: while using the emulators, you can be sure that all the methods you call are going to work always, because they have selftests on the ui-toolkit branch. If a change on the ui toolkit breaks the emulator, it will not land. [08:07] gusch: ok, sounds good. get_tabs_bar() seems straightforward to update (remove the New from "NewTabBar"), but I'm figuring out now how you use the result [08:07] timp, gusch: I'm sleeping now :) But I'll be up in ~6 hours in case you need a hand. [08:08] elopio: sleep well :) [08:08] gusch: it's not a question of being nervous or not, it's just how reactive can we be to get issues fixed and having the flow opened again [08:08] so let's open the gate again! :) [08:09] didrocks: did more apps tests break besides gallery? [08:11] didrocks: the problem is that autopilot tests of apps make use of internal implemetation details of UITK, which do not need to be stable, but luckily elopio did work to solve that by starting an autopilot library in UITK that other apps can use [08:11] before I go, I think that a good fix for the tests is better in the long run, even if we need to close the daily releases while we get it implemented. I would use this failures as a sign that we have a high priority to use the emulators on the toolkit on all apps. [08:11] ...just sayin' :) [08:12] timp: not more tests, but maybe more apps :) [08:12] ogra_: should we be testing on --ubuntu-bootstrap now? [08:12] timp: it's something that in any way apps can get access to? [08:13] if so, sil2100 can unblock the stack and still publish [08:13] then, next run will ensure that the tests are now running again [08:13] but that would make the dashboard red [08:13] so I can see an angry asac next to me :p [08:15] didrocks: 10:03:55 < elopio> timp: yes, the filemanager and dropping letters are samples of what needs to be done to use the emulators. [08:15] asac, indeed i do, what do you need ? [08:15] timp: thanks, I was disconnected while you wrote that [08:15] didrocks: ^ so I guess apps can get access to it, but how to do it I don't know, let's have a look at filemanager and dropping letters [08:15] yeah [08:15] so let's not release [08:15] didrocks: I don't know all the details. elopio wrote that [08:15] in case an app do that [08:18] popey, we should test what utah tests to enhance the automation imho ... === alan_g|EOD is now known as alan_g [08:18] it is a larger MR than I was hoping for, but for filemanager, it is here https://code.launchpad.net/~elopio/ubuntu-filemanager-app/emulators/+merge/177070 [08:19] popey, (which means just flipped atm i guess) [08:20] ogra_: folks said maguro is boot looping? [08:20] hmm... just installed aug 1 [08:20] that seems to boot fine [08:20] asac, rsalveti and i tested that last night and couldnt reproduce [08:20] right [08:20] gema: cant confirm its an issue [08:20] gema: installed 0801 and it just worked [08:20] gema: guess we need the exact scripts [08:25] ogra_: rsalveti: so yeah, last night's android upload has sqlite back in - both binary and .so [08:25] xnox, great, thanks [08:27] ogra_: ok [08:27] asac: did you use phablet-flash? [08:27] gema: sure [08:28] phablet-flash --pending [08:28] asac: ack, will ask rfowler and plars to work together in triaging this, then [08:28] sil2100 bzoltan didrocks timp here the quick fix: https://code.launchpad.net/~schwann/gallery-app/gallery-autopilot-navbar/+merge/178221 (timp can you maybe review it?) [08:28] gusch: looks good, approving [08:29] great! please ping us when the 2 are merged [08:29] didrocks: 2? what's the second [08:29] ? [08:29] the filemanager one? [08:30] didrocks: that was already merged, elopio referred to that MR so we can see how to use the uitk emulators in apps [08:30] ah, great :) [08:30] sil2100: so once the gallery-app is merged, can you please rekick the stack? [08:30] if it publishes, then we can publish sdk + gallery [09:01] ricmm: you there? [09:06] didrocks sil2100 gallery is merged [09:08] Good morning all, happy Friday and happy Ice Cream Sandwich Day! :-D [09:08] yay [09:08] right on the spot for this weather :) [09:08] gusch: thanks! sil2100 still around? [09:08] ok, seems not, let me rerun it [09:09] (done) === hikiko is now known as hikiko|afk|brb [09:48] rsalveti: so I'm trying to figure out whether the NM delta can't really be limited to the policykit policy after all; but with my dpkg-diverted && hacked-up policy, chewie isn't working, and I'm also not seeing any traffic with dbus-monitor --system. Do you have any hints how to debug this - at least to the point that I can actually see the dbus traffic? === hikiko|afk|brb is now known as hikiko [09:56] elopio: I'm trying to use the SDK autopilot emulators in gallery - but need some help, can you? [10:01] gusch: 10:07:36 < elopio> timp, gusch: I'm sleeping now :) But I'll be up in ~6 hours in case you need a hand. [10:01] timp: ah - ok - thx [10:03] Ubuntu SDK: qmlscene shows a blank image - Ubuntu 12.04 | http://askubuntu.com/q/327778 [10:21] timp: do you know how the MainView of the toolkit emulators is created? I only get a NoneType ... :( [10:23] timp: or do you know anyone who knows about how to use it (filemanager does not hep me :( ) === chihchun is now known as chihchun_afk [10:25] YAY ! [10:25] my first kernel patch since 2.0 times and it works :) [10:33] asac, todays image build failed with a weird python-minimal errors ... i suspect the buildd is borked === hikiko is now known as hikiko|lunch [10:35] yes, it's the usual panda damage (#ubuntu-release) [10:35] yeah [10:35] re-running, lets see [10:35] if it fails again i'll ask for a reboot [10:55] did any of the port maintainers (or other people who use a non-official) succeed flashing their devices with lp:~sergiusens/phablet-tools/flash_change ? [10:58] sforshee, mfisch, hi, so I think PulseAudio needs to talk to powerd to make sure the system does not go into S3 while playing back audio. Do you have pointers to how to accomplish this? [11:01] mardy: please merge lp:~laney/ubuntu-system-settings/new-icon-paths [11:02] that or make those files not hardcode the directory [11:03] diwic: I can help you out, but I'm not sure I agree that any audio playback should block suspend [11:04] sforshee, well, if you play back audio, you need to wake the CPU up every now and then to read more data and feed it into the DMA buffer? [11:04] sforshee, if the system suspends, audio playback will stop [11:05] diwic: I would expect this to be a higher-level policy decision that we're playing some class of audio that needs to continue, even if it's to the detriment of battery life [11:07] sforshee, PulseAudio could check what type/class of audio is being played back and only request a wake lock on certain types of audio [11:07] diwic: so long as PA has that sort of information I guess that's fine [11:08] sforshee, it seems like PA is the one that has the most information about what it is playing back at least [11:08] sforshee, since streams can carry classifications [11:08] Laney: oops, done [11:08] diwic: so it can tell the difference between e.g. background mp3 playback and sound effects from some app? [11:09] sforshee, yes, at least if the app is well-behaving [11:09] Laney: did you debug that QNetworkAccessManager issue? [11:09] sforshee, and tags its streams accordingly [11:09] mardy: no, but I do have that plugin [11:09] diwic: I don't want our power management predicated on apps behaving appropriately [11:09] mardy: looks like I might not need it if the design gets simplified [11:12] sforshee, do you have a better suggestion? [11:13] this guide https://wiki.ubuntu.com/Touch/Porting says that the ubuntu filesystem is kept within a container [11:13] is that still the case? [11:15] diwic: I would expect whatever implements our media API for apps to make these decisions. Will apps actually talk to PA directly? [11:16] I'm trying to reply to http://forum.xda-developers.com/showpost.php?p=44197859&postcount=5 - is there any reason calls, sms and 3g should not work on a galaxy nexus (using verizon)? [11:16] diwic: My expectation would be that apps would have some kind of play_background_audio() API, and part of the implementation of that API would be to request the active state [11:17] dholbach, verizon uses cdma which we do not yet support, if this is in reference to touch images [11:17] pmcgowan, gotcha [11:18] sforshee, hmm. Do all apps use the SDK? What about "regular" ubuntu software, will they not run on the phone? [11:19] sforshee, I mean, if all apps use the SDK we could pull in some policy layer between the app and PA, but that just makes life harder for applications not using the SDK. [11:21] I'd imagine games and apps that do their own UI components from scratch wouldn't necessarily use the SDK [11:22] diwic: I really don't know. That's probably a question for tvoss_ [11:23] sforshee, well, from the discussions with tvoss_ he said it was okay to do it in PulseAudio === MacSlow is now known as MacSlow|lunch [11:23] sforshee, but if you like we could set up some meeting about it if you have concerns about that [11:23] diwic: I've got to run, but the documentation for powerd's dbus api is at https://wiki.ubuntu.com/powerd [11:23] diwic: you want requestSysState and clearSysState [11:23] sforshee, thanks [11:24] ogra_, I have a mako stuck at the boot screen [11:24] ogra_, adb sees the device but I cannot shell in, anything to check? [11:25] ogra_, nm it just booted, hmmm === hikiko|lunch is now known as hikiko [11:31] what is the "container" that ubuntu touch runs in? [11:31] lxc [11:32] rah, android components are int he container [11:32] in the latest images [11:32] pmcgowan: so the arrangement has changed recently? [11:32] rah, around 1 month ago, the so called flip [11:33] I see [11:33] it looks like just control groups and whatnot [11:33] the previous images had ubuntu in a chroot [11:33] I see [11:33] so nothing virtual [11:33] no [11:34] good :-) [11:34] sent another call for help with updating ports and testing the new phablet-flash to http://forum.xda-developers.com/showthread.php?p=44206124#post44206124 (just in case anyone here is interested in helping out as well) [11:39] rah, we boot natively into the ubuntu touch rootfs and fire up androids hardware layer in an lxc container to initialize the hardware ... [11:39] ogra_: I see [11:39] ogra_: why? :-) [11:39] I mean, why use a container? [11:39] would you prefer the old model of using a chroot for ubuntu ? [11:40] why is the choice container or chroot? [11:40] (running on top of a minimal android) [11:40] why not just: execute? [11:40] because the two systems function totally different [11:40] for example? [11:40] gusch: sorry, I don't know. I know that elopio created them [11:41] ubuntu uses udev for managing device nodes (and has a certain standardized linux naming scheme apps rely on) ... android uses ueventd [11:41] I see === chriadam is now known as chriadam|away [11:41] dholbach: why are you still posting the in progress guide? it's not materially different to the old guide in what it tells you to do, and does not work! [11:41] you cant really use the same /dev for both at the same time without introducing massive issues you would have to hack around [11:41] lxc can change mount points? [11:42] lxc operates like a vm ... even though it isnt one [11:42] interesting [11:42] do the android components need modification to work with this setup? [11:43] well, depends what you define as modified [11:43] we rip out 90% of the android tree anyway since we only use the hardware abstraction layer [11:44] but yeah, there are some minor additional adjustments [11:44] I see [11:44] i.e. we keep just enough android to make the binary drivers usable :) [11:47] ogra_: whats with the touch images? have they just not run enough all the 2.1 images are red on install and boot? [11:48] davmor2, patience :) [11:48] seems they just havent started yet [11:49] ogra_: so just not run long enough then? [11:49] * popey has just flashed 20130802.1 here [11:49] davmor2, the dashboard works asynchronous ... it only updates every n munites [11:50] popey: I was about to do the same so checked the smoke test first and the see of red didn't fill me full of hope :) [11:50] ogra_: ah fair enough [11:53] davmor2, hmm, though looking at the logs it seems like it actually failed to install, seems the tests changed [11:53] gema, ^^^ [11:53] gallery app looks broken if launched from the camera app [11:53] but works if launched on its own [11:53] gema, apparently the new "install-and-boot" test failed for all 0802.1 images [11:53] popey: works for me [11:54] gusch: on 20130802.1? [11:54] popey: 20130801 ... [11:54] 02.1 is only 20min old :) [11:54] I'm fast ☻ [11:55] yeah [11:55] it works after gallery has been launched at least once [11:55] but if you flash, open camera, press button in bottom right to launch gallery, it fails [11:56] now, is that a camera-app bug or a gallery-app bug? ☻ === om26er is now known as om26er|internet_ === om26er|internet_ is now known as om26er [11:58] popey: write it as a camera app bug and add gallery to the bug [11:58] yeah [12:00] https://bugs.launchpad.net/camera-app/+bug/1207720 [12:00] Ubuntu bug 1207720 in gallery-app "gallery wont start after clean flash, from camera app" [Undecided,New] [12:00] popey: open clock, set city to wolverhapmton, what time does it display? [12:00] 1830 [12:00] (the year) [12:00] (kidding) [12:01] popey: for me utc is saying 12:01 wolves is saying 04:01 [12:01] davmor2: 13:01 [12:01] http://popey.com/~alan/device-2013-08-02-130142.png [12:01] not quite sure what happened there :) [12:02] popey: ta I'll add it again and see what happens [12:03] i added then tapped on it [12:03] and now it's right :) [12:03] magic [12:03] popey: must of just been a minor glitch [12:03] fat finger syndrome [12:04] popey: possibly except that under wolverhampton there are 5 listings all of which are wolverhapton in the uk [12:05] popey: I might of clicked before it had refreshed fully and got some random timezone [12:06] popey: if I open camera and click on gallery I get the gallery but it takes a couple of attempts at opening [12:07] popey: flashes grey, closes back to camera, then opens properly [12:09] linker.c:1095| ERROR: Library 'libmaliinstr.so' not found [12:09] GRMPF ... [12:14] HM?! [12:14] why would we need maliit on android side? or is this inside ubuntu? === alan_g is now known as alan_g|lunch [12:16] davmor2: i tried 3 times before it would launch [12:17] popey: is that from a --pending or a --pending --wipe? [12:17] its --pending, but it's a clean device [12:18] popey: okay once we find out what is happening with the m-i-l I'll try a --wipe and see if that replicates it. [12:22] xnox, android side ... on a galaxy S2 [12:23] cjwatson, heh, i liked "mantra" :) ... thanks for reminding me of the WI ... (/me closes, long done) === _salem is now known as salem_ [12:25] xnox, seems the camera-app triggers something that wants to use this lib (and indeed falls over with white screen) [12:33] ogra_: that means the image doesn't install [12:33] gema, well, it surely does for popey and me [12:33] ogra_: install-and-boot is our test for flashing and doing something small [12:34] ogra_: looking [12:34] i just did "phablet-flash --pending" [12:34] alan@deep-thought:~$ adb shell cat /var/log/installer/media-info [12:34] Ubuntu Saucy Salamander (development branch) - armhf (20130802.1) [12:34] ogra_: maguro is broken [12:34] ogra_: the phone itself [12:34] ogra_: int he lab [12:34] gema, yeah [12:34] ogra_: I need paul and rfowler to look into that later today [12:35] ogra_: and I don't know about grouper [12:35] ogra_: but we enabled this test yesterday, so give us a couple of days to go through the hooks to make it work [12:35] gema, install-and-boot seems to have failed for grouper and manta too though [12:35] ogra_: yep, I see that [12:35] k [12:35] ogra_: plars will fix that [12:36] yeah, i didnt want to be pushy, just report it :) === MacSlow|lunch is now known as MacSlow [12:39] ali1234, how is https://wiki.ubuntu.com/Touch/PortingFlippedInProgress?action=diff&rev2=13&rev1=1 "not materially different"? [12:41] dholbach: the instructions for checkout and building a port are identical. those notes only apply *after* you have done that [12:42] and also they are device specific anyway [12:43] ali1234, well, there are always going to be device-specific parts - thanks in any case for bringing it up on the mailing list [12:51] i'm currently trying to build vanilla cyanogenmod to see if it also builds a bloated 8MB kernel [12:51] but it keeps failing on some proprietary app :( [12:56] ali1234, woah, 8M ? [12:56] yeah i know [12:56] this didn't happen with the unflipped image, when it still worked [12:56] boot is only 7MB and the ramdisk is another 2MB [12:56] yeah [12:56] i havent seen a kernel above 5M yet [12:56] so somehow i need to trim 3MB from the kernel [12:56] in android [12:57] you could repack the ramdisk and remove adbd and libcrypto.so [12:57] that won't help given that the kernel alone is bigger than boot [12:58] ah, your boot partition is 8M ? yeah, then it wouldnt [12:58] i wonder why it got bigger for you though ... its not like we touch kernel configs of ports in any way [12:58] well for one thing samsung uses a non standard boot.img format [12:58] (and the configs for the supported devices comes from packaged kernels) [12:59] they are probably using lzma or something, to fit their massive kernel in it [12:59] yeah, i know ... thats what my last mail to the ML was about [12:59] for the galaxy S2 though [12:59] anyway, if i can see how vanilla CM does it, i will know what the problem is [12:59] they do, but that hasnt changed since before the flip [12:59] if i can get it to build that is [12:59] er... but it has changed [13:00] it can only be a change on the CM side [13:00] the unflipped images also no longer work for me [13:00] as i said, we dont touch kernel configs ... the porter has to do that himself [13:00] in the flipped image it uses mkbootimg, that's where it fails [13:00] but since your kernel gets most likely pulled from CM i would blame them [13:01] mkbootimg doeesnt have anything to do with the kernel config though [13:01] the last time i successfully built an ubuntu touch image, it had a boot.img without the "Android!" magic, ie it's a none standard samsung one [13:01] it is called after your kernel was built [13:01] exactly [13:02] the last time i successfully built ubuntu-touch, it didn't use mkbootimg [13:02] it used whatever tool samsung uses === daker_ is now known as daker [13:02] but since i no longer have the source of that version, i can't compare [13:02] hence trying to build CM [13:04] it may be as simple as hacking the script to use a different version of mkbootimg, one that uses different compression [13:04] like maybe they disabled compression on the kernel build and put it into mkbootimg? [13:04] /home/al/Source/ubuntu-touch/cm/out/target/product/galaxysmtd/obj/STATIC_LIBRARIES/libwebcore_intermediates/Source/WebCore/bindings/V8SVGRectElement.cpp:373:1: internal compiler error: Segmentation fault [13:04] fun [13:10] wow [13:10] gallety app looks really bad on http://reports.qa.ubuntu.com/smokeng/saucy/image/3316/gallery-app-autopilot/ (mako on 0802.1) [13:12] is "make out/target/product/galaxysmtd/kernel" a valid thing to do? [13:13] no [13:13] go into the kernel subdir and call make Image [13:13] *make zImage [13:13] (my Z key is wonky) [13:13] that won't tell me what i need to know [13:13] well, i suppose it might [13:13]