=== infinity_ is now known as infinity [01:10] get the error assert failed: getprop("ro.product.device") == "manta" || getprop("ro.build.product") == "manta" when installing trusty-preinstalled-touch-armel+manta.zip in recovery on nexus 10 === jhodapp|afk is now known as jhodapp === jhodapp is now known as jhodapp|afk [05:22] hello [05:22] anyone here? === Mike is now known as Guest12462 === chihchun_afk is now known as chihchun [08:01] 2014/03/28 07:55:56 Rebooting into recovery to flash [08:01] 2014/03/28 07:57:07 Failed to enter Recovery [08:03] \o/ It worked after all === GURU is now known as Guest33188 [10:00] Good morning all! Happy Friday, and happy Something On A Stick Day! :-D === jaywink_ is now known as jaywink [10:27] nik90: in addition to the tabs failures that zsombi mentioned in the AP tests, I got another one: http://paste.ubuntu.com/7167683/ [10:28] nik90: any ideas? We are testing changes in UITK including https://code.launchpad.net/~elopio/ubuntu-ui-toolkit/tab_selection_timing/+merge/210219 which should actually fix it [11:02] t1mp: looking at it now [11:04] nik90: do you have a nexus4? [11:04] t1mp: yes [11:04] do you happen to have image 264 installed on it? [11:04] I like to figure out if any of these errors can be reproduced with a stock 264 that doesn't have our changes for ubuntu-ui-toolkit [11:04] hmm [11:05] I can re-flash it myself after I finish these tests of course [11:06] t1mp: I am waiting for my phone to start up..but I am definitely on the proposed images for sure [11:06] t1mp: what exactly do you want to test on the phone? [11:06] t1mp: the AP failure above? [11:08] nik90: phablet-test-run ubuntu_clock_app [11:08] that's all :) [11:09] uh I got lucky Ran 20 tests in 516.560s [11:09] OK [11:09] some times the tests pass [11:11] t1mp: could be a timing issue that caused it to fail before [11:11] t1mp: that's usually the suspect when you get StateNotFoundError: Object not found with name [11:11] yes. Then https://code.launchpad.net/~elopio/ubuntu-ui-toolkit/tab_selection_timing/+merge/210219 was not enough to fix it [11:12] no no [11:12] the above MP is for the tab selectio [11:12] selection* [11:12] the failure is due to the listview not appearing on time [11:13] t1mp: the failures that zsombi pointed out at http://pastebin.ubuntu.com/7166706/, those are the true tab timing failures which the above MP is supposed to fix [11:13] the results seem a bit random [11:13] nik90: yes, I got those before also, with the same UITK version installed that should include the fix [11:14] I'm running the tests 10 times now, to see how many pass with the changes [11:14] t1mp: if you got zsombi's failures, then the MP doesn't fix the tab timing issues reliably. [11:14] t1mp: as for the other failure you linked above, that is localised in the clock app and something I need to take a look at my end :) [11:16] t1mp: I just update image 265 on my phone. Should I install it and then test as well? [11:16] nik90: clock app tests look very reliable in the result history here http://ci.ubuntu.com/smokeng/trusty/touch/mako/264:20140327.1:20140304/7399/ubuntu_clock_app/ [11:16] nik90: so I wonder why I am getting errors, and the smoke tests all pass.. am I just unlucky? [11:17] nik90: we're trying to figure out if we introduced a regression, so I think it is better to first see if you can reproduce the failures without the updated UITK, and then add the PPA and check with the new version [11:18] t1mp: ok...I had just a more deeper look at the failure you linked. It seems that the timing issue is related to the toolbar. The AP is trying to click on a toolbar button that is not visible yet. [11:18] results for image 265 look weird http://ci.ubuntu.com/smokeng/trusty/touch/mako/265:20140328:20140304/7406/ so I don't know what's up with that image [11:18] t1mp: and I used the UITK toolbar emulator function here [11:19] yes the toolbar emulator should take care of that [11:20] there is a wait in there to wait for the toolbar to be opened [11:20] t1mp: yes. I have *never* had timing issues with the toolbar before [11:26] t1mp: to flash a specific image do I do "system-image-cli -c trusty -v -b 264" ? [11:28] nik90: huh? I never used system-image-cli [11:28] nik90: I use ubuntu-device-flash [11:29] t1mp: ah the above command is to switch from trusty-proposed to trusty as told by ogra_ a long time ago [11:29] nik90: that should work..but you need trusty-proposed for the channel [11:29] popey: ah ok [11:29] as 264 isnt in trusty [11:30] why do we have multiple commands to flash the device? [11:30] and where does system-image-cli come from? the command doesn't exist on my laptop [11:32] nik90: I had a few runs of the AP tests for clock now without any failures [11:32] zsombi: ^ [11:32] t1mp, thats on the device [11:32] we dont have multiple commands [11:32] one is on device, one is from pc [11:33] ahh [11:33] I never did that from the device [11:33] thanks, it makes sense now :) [11:33] ☻ [11:37] * zsombi back [11:38] t1mp: so, everything is cool now then? [11:38] zsombi: no, it is flaky [11:38] zsombi: I had two full test runs without problems, and now another one that gave an error [11:39] t1mp: excellent [11:39] zsombi: if the tests show the same flakiness in image 264 without the changes that we try to land, then we can land [11:39] zsombi: but we must make sure that we did not introduce them [11:39] zsombi: I am still testing with the landing ppa enabled [11:39] t1mp: ok [11:42] t1mp, 264 had no errors http://ci.ubuntu.com/smokeng/trusty/touch/mako/264:20140327.1:20140304/7399/ [11:45] ogra_: I know. I get errors with UITK changes *some times* [11:45] ogra_: so I cannot reproduce them consistently. I wonder if that is the case with 264, and we were just lucky [11:46] might be, didrocks usually sends a summary in the landing mails about known flaky tests [11:47] t1mp: no flaky tests are known as of now, but there are some unity8 crashers, maybe look if they are linked [11:54] t1mp: ok... I'm continuously getting unlock_screen() failures on AP tests on any MR I create [11:55] didrocks: is the unity8 crash related to the unlock_screen()? [11:55] zsombi: no, this is a bustage in powerd [11:55] zsombi: I revert it in the distro and an image is going to be kicked [11:55] didrocks: okay, thx [11:55] same than dashboard results [11:55] zsombi: if you want to speed it up, install latest powerd [11:55] zsombi, install the powerd from the archive [11:56] orga_ you mean the powerd to get the ulock_screen() issue workaround? [11:57] the fix ... [11:57] :) [11:57] zsombi: yeah, just apt-get install it [11:57] reboot [11:57] and you should be fine [11:57] didrocks: okay :) [11:57] zsombi: see, everyday is fun here :) [11:58] didrocks: I have ~2 MPs failing on Jenkins because of this unlock thingie [11:58] didrocks: I have no doubth about that :D [11:58] bugs happen [11:58] zsombi: yeah, next image is going to be published in ~50 minutes [11:58] orga_ no, that's sh*t, that happens :D [11:58] so then, you should be fine on jenkins [11:59] :) [11:59] didrocks: awesome! I anyway have loads of other stuff to do, so I can wait till then ;) [11:59] heh :) [11:59] didrocks: I know I'm a lazy bastard :D [12:00] zsombi: well, if you can avoid loosing this time and focus on something else, that's fine :) [12:00] didrocks: the bug backlog is never empty, so... ;) [12:01] zsombi: better hurry up and clear the bug backlog then :) [12:01] zsombi: I'm not surprised :p [12:01] t1mp: look who's talking :D [12:01] you mean there are bugs in our UITK ?!? [12:01] orga_: uhm... no, there're loads of features :D [12:01] ogra_: no, only people who use it wrong ;) [12:01] :) [12:02] we need some toxins for the bugs :D [12:02] ok, /me goes back to dig into Layouts [12:03] zsombi: speaking of bugs, do you know what is expected from this? https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1187335 can I close it? [12:03] Ubuntu bug 1187335 in Ubuntu UI Toolkit "Tests are missing a short description from the beginning of the file" [Undecided,Incomplete] [12:06] t1mp: well, there was a sheet (yet another one) back then to fill in about which test has what. I guess it can be closed, the sheet is not in use, at least I haven't seen it afterwards... [12:06] ok I closed it, finally :) [12:07] :) [12:07] Ran 20 tests in 518.940s [12:07] OK [12:08] zsombi: ^ after purging the landing ppa... [12:08] I'll repeat it a few times [12:10] ogra_, lool, rsalveti, I identified two files that need cleanup due to dropping sflinger support: [12:10] lxc-android-config: /etc/init/boot-hooks/set-display-mir.conf [12:10] ubuntu-touch-session: /etc/profile.d/qpa_plugin.sh [12:10] /etc/environment could lose the QML2_IMPORT_PATH, not sure where that comes from, though [12:10] Saviq, from live-build ... [12:11] Saviq, there are more files [12:11] (which is why it scares me to remove the .display-mir file ... easy to miss something) [12:11] ogra_, well, we need to drop it somewhen anyway... [12:12] indeed [12:12] you want to drop it from usc-wrapper too [12:12] and most likely we need to rework the lightdm session [12:13] ogra_, so, are you saying you'd rather do that separately? [12:13] no, would be good to do it all in one silo [12:14] but when we do it we should drop the special casing that enables the back/forward switching for lightdm in ubuntu-touch-session as well [12:15] mzanetti, please coordinate with ogra_ ↑ [12:15] sure [12:15] hi Oli :) [12:16] hey [12:16] mzanetti, there are a few bits in ubuntu-touch-session that mterry put in place to enable the switching in lightdm ... [12:16] we should probably just ask him to clean it up since he knows best what can go [12:17] ok. will take care of that when he shows up [12:17] he's away today :| [12:17] bah [12:17] yay timing [12:18] ok. at least we can prepare everyhting else today I'd say [12:19] yep [12:19] so those are to be cleaned up for sure: https://code.launchpad.net/~unity-team/unity8/right-edge-2/+merge/213172/comments/504199 [12:20] "cleanly shot down" [12:20] lol [12:21] Saviq: ? [12:22] looks like http://bazaar.launchpad.net/~phablet-team/ubuntu-touch-session/trunk/revision/163 and http://bazaar.launchpad.net/~phablet-team/ubuntu-touch-session/trunk/revision/133 are the related revs [12:22] it doesn't look like there'd really be a lot to change [12:23] type=unity;surfaceflinger [12:24] hmm, i thought there was more [12:24] * Saviq greps for MIR_SOCKET, too [12:25] ubuntu-touch-session: /usr/bin/ubuntu-touch-session [12:25] https://code.launchpad.net/~mzanetti/ubuntu-touch-session/drop-sf-support/+merge/213243 [12:26] mzanetti, usc-wrapper, too [12:26] Mirv, popey any thoughts on that saucybacon issue with qmlscene? [12:26] should we put the QT_QPA_PLATFORM somewhere else (/etc/environment), now that it's not dynamic? [12:27] pmcgowan: fixed, new version in store already [12:27] ogra_, ↑? do we need it in both profile.d and separately in u-t-session? [12:27] popey, did you just change the app? [12:27] Saviq: not exactly sure what to do with usc-wrapper [12:27] mzanetti, just drop the "return 1" [12:27] popey, is it not a real issue with qmlscene? i.e. will other apps hit it [12:28] Saviq, well, profile.d is used by adb/sudo [12:28] mzanetti, the whole if [] actually [12:28] yep [12:28] pmcgowan: i notified the developer, gcollura, who updated the app [12:28] Saviq, we need to make sure that we have the same variable available in both, session and adb sudoed session [12:28] pushed [12:28] popey, I am still concerned that qmlscene is complaining [12:29] do we understand why? [12:29] mzanetti, also, 52-ubuntu-touch.conf === jhodapp|afk is now known as jhodapp [12:29] mzanetti, drop ;surfaceflinger [12:29] ack [12:29] ogra_, yeah, doesn't /etc/environment give us that? or should we move stuff out from /etc/environment? [12:29] i think we should move it out there [12:30] as part of this change? [12:30] if you feel like :) [12:30] its in livecd-rootfs [12:31] pmcgowan, what are you referring to if I may know? :) [12:31] gcollura: bug 1298659 [12:31] ogra_, I'm just not sure what's The Right Thing™ to do there, I only know a few of the vars there, and not sure we only want them in the touch session [12:31] bug 1298659 in qtdeclarative-opensource-src (Ubuntu) "qmlscene reports invalid option -I" [Undecided,New] https://launchpad.net/bugs/1298659 [12:31] gcollura, the issue with the -I option [12:31] ogra_, I know we won't need QML2_IMPORT_PATH at all, we're cleaning that up [12:32] so lets drop what we are sure about [12:32] the rest is for a later review [12:33] +1 :) [12:33] mzanetti, ↑ frop QML2_IMPORT_PATH from /etc/environment, shipped in livecd-rootfs [12:33] ack [12:33] pmcgowan, ok that is weird [12:33] drop [12:33] pushed ubuntu-touch-session btw [12:33] mzanetti, but we also need unity-mir to install in the path from unity-api.pc [12:33] is Mirv around [12:34] mzanetti, and then we need to clean unity8's main() to not try and mangle the existing QML2_IMPORT_PATH [12:34] pmcgowan, if I remember correctly I've noticed that kind of error a week or two ago, but I didn't get worried much actually [12:34] main.cpp:22 [12:34] pmcgowan: < Exec=qmlscene -I modules/ app/saucybacon.qml [12:34] erm [12:34] main.cpp:122 [12:34] pmcgowan: > Exec=qmlscene $@ -I backend ./app/saucybacon.qml [12:34] thats the diff between the broken and working versions [12:35] oh really [12:35] i just diffed between saucybacon 0.14 and 0.15 [12:35] well, not entire diff ☻ [12:35] that looks more like a syntax fix between qt versions or something [12:35] but diff in .desktop file [12:36] popey, ok guess I will stop worrying about it [12:36] pmcgowan, no wait I was referring to another thing [12:36] thanks [12:36] gcollura, ? [12:37] ogra_: Saviq: https://code.launchpad.net/~mzanetti/livecd-rootfs/drop-sf-support/+merge/213249 [12:37] pmcgowan, nothing, forgot the last two messages [12:37] looks fine [12:37] gcollura, what messages [12:39] pmcgowan: hi, it doesn't seem to happen on desktop, which is the odd part [12:40] Mirv, right noticed that [12:40] mzanetti, you're onto lxc-android-config, too? [12:41] so far I don't have a clue how it can be different, but I haven't been able to look at that much either. I'll get someone to help me if I don't find anything myself. [12:41] yeah... autopilot bailed out of course... fixing that [12:41] Mirv, thanks [12:42] ogra_, will you do uploads to the silo? (I don't think livecd-rootfs goes through CI train? does -session, lxc-android-config?) [12:43] session should now, i think robru reworked it (never tried to use it since though) [12:43] the others need to be uploaded ... will happily to that [12:46] Saviq: ogra_: that set-display-mir.conf file can go away completely, right? [12:46] yup [12:46] mzanetti, yup [12:48] Saviq: ogra_: https://code.launchpad.net/~mzanetti/ubuntu/saucy/lxc-android-config/drop-set-display-mir/+merge/213251 [12:48] mzanetti, debian/*install? [12:48] do'h [12:48] mzanetti, not sure it's there, just check, please [12:49] Saviq: says "etc" [12:49] mzanetti, good [12:53] nik90: it looks like I cannot reproduce the failures after purging the landing ppa [13:01] ogra_, ok, whenever you're ready, please push lxc-android-config and livecd-rootfs to silo 015 [13:01] ok === MacSlow is now known as MacSlow|lunch [13:06] popey: pmcgowan: ok one more datapoint: the same error about -I is gotten also with image #237, ie Qt 5.0. so it has seemingly never worked. but it also doesn't prevent trying to start up. [13:06] Saviq, pushing livecd-rootfs to the silo doesnt make sense ... you wont be able to test it until there is a new image [13:07] mzanetti, ^^ [13:07] ogra_, yeah, but don't we want to land it together anyway? [13:07] Saviq, that var is only used in SF ? [13:07] ogra_, both, but not when we move things around [13:07] hrm [13:07] ogra_, and make them proper [13:07] Mirv, yes, I am hearing "we always got that warning" but the apps do run [13:07] how long will your testing take ? [13:08] just saw it in qtc with the emulator [13:08] ogra_, which will happen in the same silo, but we can't remove that before we land that silo [13:08] the issue is that the branch that others use for distro build stuff can go out of sync [13:08] ogra_, we're fixing bits'n'pieces in unity8 and unity-mir still [13:08] ogra_, the env can be there [13:08] livecd-rootfs is used by all flavours too [13:08] ogra_, so let's just leave it until after the silo landed? [13:08] yeah [13:09] sounds better [13:09] ogra_, that's fine, ok /me removes from silo [13:09] i have a ready made source package locally, just ping me once the silo landed [13:09] lxc-android-config is uploaded to the silo [13:09] ogra_, thanks [13:14] ogra_, just make sure that the hook runs always (in case you missed it); or if you want to do it properly, remove appmanager from android and to the least sf from init.rc [13:15] sergiusens, ? [13:15] sergiusens, we are dropping all traces of ~/.display-mir [13:16] uh oh ! [13:16] ogra_, exactly; don't forget /var/lib/lxc/android/pre-start.d/30-no-surface-flinger [13:17] Saviq, kill that öxc.-android-config package from the silo please [13:17] sergiusens, yeah, thanks a lot !!!!! [13:17] i *knew* we would miss stuff [13:18] ogra_, I can't [13:18] ogra_, but will wait for another upload then [13:18] ogra_: are Ubuntu Touch's specifications revealed? [13:19] I mean phone [13:21] Saviq, ok, upploaded with the same version, lets see if the PPA likes that [13:21] ogra_, if different contents, it won't, afaict :) [13:22] sigh [13:22] ogra_, and deleting the old one wouldn't help [13:22] double sigh [13:22] ogra_, not for a few hours/days at least === _salem is now known as salem_ [13:24] ogra_, you need someone to do a "real" delete [13:24] webops person [13:24] mandel: looking at it I think it will be best if setLocalPath just sets the directory, either that or the download will need to expose the original path for us to retrieve the filename from (we're not sending any of the original URL info to the content-hub-service, so all its aware of is the download id) [13:24] mandel: I'll do something temporary for now for testing purposes though [13:27] t1mp: so it is a regression in the UITK update? [13:29] popey: so was saucy bacon updated recently? I'm just wondering since in the Qt 5.2 manual testing it was tested to be working, so I'm wondering what has changed [13:29] popey: at least it's not probably the -I that prevents it from working [13:29] Mirv: yes, today [13:29] 0.15 works, 0.14 does not [13:30] ok [13:32] popey, last question on that, do we have any idea what changed between manual testing when it worked and now? [13:33] Mirv, popey why are you using -I? [13:33] popey: um, I have it running for me [13:33] pmcgowan: I don't, no. sergiusens I am not. its saucybacon we're talking about, not my app [13:34] popey: 1.0.15, I installed it on my #266 from app store and started from command line with upstart-app-launch com.ubuntu.developer.gcollura.saucybacon_saucybacon_1.0.15 [13:34] Mirv: does it work from icon? [13:34] popey: ye [13:34] s [13:35] I just assumed it wouldn't so I went for command line [13:35] *boggle* [13:35] confused [13:35] popey, Mirv sure, so if it's a compiled app, use the upstart-app-launch locations; I have something sort of summarized, albeit for go, here: http://discourse.ubuntu.com/t/click-packaging-for-golang-based-binaries/1557 [13:36] Elleo, yes, make sense, I'll take care and of that and will rename th method accordingly [13:37] mandel: great, thanks :) [13:54] nik90: yes it seems like a regression [13:54] nik90: a regression in a *proposed* landing, so nothing broke yet :) [13:57] t1mp: technicalities :P [14:01] popey: my app was using ubuntu-sdk-13.10. Do I change this to ubuntu-14.04-qmldev1 when I push an update? [14:02] nik90: you can update but you need to use the right name and version.. from ... https://wiki.ubuntu.com/Click/Frameworks [14:03] * nik90 reads through it === dandrader is now known as dandrader|afk === sil2100_ is now known as sil2100 [14:28] tedg: hi! Sorry I had a shorter day yesterday and couldn't participate in the discussion about the trust sessions [14:28] mardy, You missed out on all the fun! :-) [14:29] tedg: I have a slightly unrelated question: what happens when the URL dispatcher gets a URL to dispatch, and the target app is already running? [14:29] mardy, It sends the URL to the app over DBus. === MacSlow|lunch is now known as MacSlow [14:30] tedg: would it be possible for the target app to specify that it prefers having another instance launched, instead? [14:32] popey: could you approve my app. I fixed the authentication error bug and uploaded a new update to store. [14:33] ack [14:33] done === mandel is now known as mande-lunch === dandrader|afk is now known as dandrader [14:38] ty [14:38] np [14:40] Hi there, I am planning to buy a tablet and I am thinking about the type. I want Ubuntu there and not sure what kind of tablet I should get. Is Nexus 10 still supported? === SonikkuAmerica_ is now known as SonikkuAmerica [15:16] rsalveti: ogra_ any reason we couldn't add screen/tmux/byobu onto the phone ? [15:17] no specific reason besides image size [15:20] popey, well, people wouldnt be able to use it with the terminal-app and i dont know how the adb tty behaves with it (remember nano ?) [15:20] popey, so some testing in advamce would be nice ... and then you need to distract pmcgowan while i seed it [15:21] hmm? [15:21] (to prevent the complaints about image size waste) [15:21] I'm thinking more for me doing "adb shell start ssh" and then sshing into my phone and poking stuff [15:21] popey, pfft, thats so obsolete [15:22] and having screen / tmux would let me leave stuff runing on the phone [15:22] setprop persist.service.ssh true [15:22] popey, ^^^^ [15:22] suh-weet [15:22] thanks [15:24] ogra_, what's the phablet ssh password? [15:25] phablet [15:25] gcollura, phablet indeed ;) [15:26] thanks popey ogra_ [15:39] Ubuntu touch 13.04 stops clicking | http://askubuntu.com/q/440366 [15:46] ogra_: on a scale of 1 to 10 how much will I break my phone if I change /etc/hostname? [15:46] don't like having 3 devices on my network with the same hostname [15:47] popey: ask cwayne; but it should just work [15:47] popey, check if you can use hostnamed [15:47] (or was that called hostnamectl ?) [15:47] the latter [15:48] root@ubuntu-phablet:/# hostnamectl set-hostname foobar [15:48] root@ubuntu-phablet:/# exit [15:48] ogra@anubis:~/touchbootchart$ adb shell [15:48] root@foobar:/# [15:48] seems to work [15:48] * ogra_ reboots ... lets see if it persists [15:48] Saviq, So has your kill-surface-flinger branches landed? [15:48] Saviq, I'd like to remove the SF code in UAL as well. === Ursinha is now known as Ursinha-afk [15:48] i know we have a mechanism for the pretty hostname ... [15:48] (for bluetooth) [15:49] just changed it here, worked fine [15:49] yay [15:49] thanks chaps [15:49] phablet@blue-phablet:~$ [15:49] \o/ [15:49] thank you for validating that making /etc/hostname writable was a good idea :) [15:50] this makes me very very happy indeed [15:50] oh [15:50] root@foobar:/# hostnamectl set-hostname ubuntu-phablet [15:50] root@foobar:/# exit [15:50] ogra@anubis:~/touchbootchart$ adb shell [15:50] root@foobar:/# [15:51] i cant change it back === chihchun is now known as chihchun_afk [15:52] you can echo "stuff" > /etc/hostname [15:53] sure i can [15:53] but there is something wrong that it doesnt set it now [15:54] i wonder if thats the dash in the name ... but then i would expect hostnamectl to properly issue an error [15:54] tedg, not yet, soon [15:55] * tedg starts flinging surfaces to fill the time === Ursinha-afk is now known as Ursinha [15:57] ogra_: works if you put ubuntu-phablet in quotes [15:57] cwayne, well, it should still spill an error or tell me it couldnt set it [15:58] pitti, is hostnamed in your realm ? [15:58] ogra_: well, that's got nothing to do with us [15:58] blame systemd [15:58] cwayne, right, thats why i pinged pitti ... want to know if he considers it a bug too before i file it :) [15:59] :) [15:59] it shouldnt no-op silently if it fails [16:01] i agree === salem_ is now known as _salem [16:18] how is media-hub going :)? [16:18] ok seems it at least built now [16:18] good [16:19] rsalveti: jhodapp: how is testing going? [16:19] asac: test is on going [16:19] I assume wont be before next week that we see this go through? [16:19] creating a custom image and we'll all test it later today [16:19] no, will land next week for sure [16:19] too many moving parts to land this today [16:19] rsalveti: ok. dist-upgrade isn't good enough to test this quickly? [16:20] asac: no, because we also got changes in the android side [16:20] oh right [16:20] are they in the silo too? [16:20] asac, yeah what rsalveti said [16:20] asac: they will be in a few, but can be tested without it already [16:20] as to test the custom tarball, you can also use a custom android image [16:20]