=== chriadam|away is now known as chriadam === nhaines_ is now known as nhaines [02:05] could someone tell me why in the first step of the porting process several necessary packages are removed from 12.04LTS x64? [02:19] hiho [02:20] question [02:21] what's your question? [02:24] which smartphone should i buy to play with .......to get unbuntu on it ..... cause i want one ..... and i ruled me not to buy one until a independent os run on one [02:25] I don't know. I'm working on a port for the SGSIII (pain in my ass) [02:27] Neverevermind: the best thing to do (unless you're a developer) is to order one from ubuntu.com later this year. [02:28] it only has to be able to make and receive calls and wlan [02:28] yeah i read sthn about that [02:28] nhaines... could I get your help with this bro? [02:29] anonynimity: I do a tiny bit of SDK advocacy but I don't know anything about the system level stuff, unfortunately. :( [02:29] I know the ubuntu-phone mailing list has some updated porting instructions. [02:29] during compilation it gave me some errors most of which i was able to figure out except for com.android.nfc_extras.xml is required by ..... [02:30] and I found the .xml file, but the jar file was missing in the "firmware" folder... which I had downloaded, and set in the firmware folder... but it wouldn't take it... [02:30] :/ [02:31] right now I'm following this guide: https://wiki.ubuntu.com/Touch/AndroidDevel then will go back to the porting guide tomorrow. === chihchun is now known as chihchun_afk [02:34] using the aosp and phablet mirrors. I'm determined to get the networking working, as well as the gsm, audio, camera, etc... [02:42] @nhaines,anony ;-) no im not a developer but i like it to play with this things/progs/osses since i got my first own pc when i was a kid .... so my question is (exept buying an unbuntu - phone which i would and will do if it will/would come out;-) : ..... which smartphone with ubuntu f. ph. can make and receive calls/sms (hdy-speaker) and manage wlan with the rest i can play ;-) [02:50] check ubuntu.com/phone? [02:50] or ubuntu.com/touch === chihchun_afk is now known as chihchun [03:23] Hi. [03:23] Cyanogenmod just committed a "quickboot" feature, which essentially hibernates the phone when shutting down instead of doing a complete shutdown. [03:23] Would it be possible that Ubuntu Touch would include some similar functionality? [04:08] can someone confirm this tutorial is good to go, to install ubuntu touch on nexus7: http://itsfoss.com/install-ubuntu-touch-nexus-7-2013/ === fginther` is now known as fginther [07:20] good morning [08:11] make bootable usb stick for ubuntu touch | http://askubuntu.com/q/444376 === chriadam is now known as chriadam|away [08:47] Good morning all; happy Monday, and happy No Housework Day! :-D [09:23] Saviq, so i was doing bootcharts over the weekend, and was wondering if it would be possible that unity8 re-orders its own start phase a bit ... if you look at http://people.canonical.com/~ogra/touch-bootcharts/ubuntu-phablet-trusty-279.png and scroll down to where the indicators start, [09:24] you see that it takes 8 seconds of the boot until the panel service gives them the upstart event to start ... that really holds up the boot process, would it be possible to start it earlier ? [09:25] (if i put the event into a post-start script in unity8's upstart job they start about 3 seconds after unity and i dont see any ill effects, so i assume it could be emitted earlier) [09:28] lool, ping re: ubuntu-touch-customization-hooks [09:32] ogra_, yeah, it could be emitted earlier indeed, right now it's the indicators QML plugin that kicks those, but that won't happen until late in unity8 startup [09:32] it would be really helpful in speeding up the boot if we could start it earlier ... [09:33] (and if indicator-messages could finally be ported to upstart startup ... ) [09:36] ogra_, I'll have a chat with dednick on what'd be required there (i.e. would we need to parse things or something), but ultimately maybe we just need to emit a custom signal (indicators-ui-starting) in unity8 post-start that the indicator jobs would start on? [09:38] Saviq, right, thats what i'm doing here for the above test ... i just added an initctl emit for the event the indicators listen for to post-start script ... but i was execting that unity8 actually needs to have the service ready so that seemed suboptimal [09:38] *expecting [09:39] ogra_, nope, they can be started early [09:39] cool [09:40] ogra_, so just MP a new event from unity8 pre-start, even [09:40] that and fixing the messaging indicator might bring us towards a 25sec boot :) [09:40] will do [09:40] dbarth: hello! [09:40] dbarth: are you around? [09:40] ogra_, huh, apparently there's actually a custom event already [09:41] Saviq, right [09:41] and unity seems to emit it from its code too ... just a little late [09:41] sil2100: yes [09:42] whatś up? [09:42] dbarth: ah, just PMed you ;) [09:43] ogra_, indicator-services-start and indicator-services-end [09:43] Saviq, right [09:44] ogra_, we should probably yank most of that code out of unity8, too http://bazaar.launchpad.net/~unity-team/unity8/trunk/view/head:/plugins/Unity/Indicators/indicatorsmanager.cpp#L220 [09:46] Saviq, yeah [09:47] ogra_, think you could do that as part of the same MP? [09:48] the whole "IndicatorsManager::setLoaded" function ? [09:49] or the whole file ? [09:49] * ogra_ has not much clue about that code === zequence_ is now known as zequence === cjwatson_ is now known as cjwatson [10:05] oSoMoN: hi! [10:05] sil2100, hey [10:05] oSoMoN: silo for your webbrowser landing assigned o/ [10:05] sil2100, awesome, thanks! [10:15] Morning all [10:41] regarding the converged desktop experience: is there any way to install plain ubuntu and then add ubuntu-touch related packages afterwards? Any other hints to bring ubuntu-touch to x86 (http://askubuntu.com/questions/444376/make-bootable-usb-stick-for-ubuntu-touch ? ) Thanks! === MacSlow is now known as MacSlow|lunch [11:27] sergiusens: hello [11:27] sergiusens: do you know the status of python3 in AP? [11:27] sergiusens: didrocks believed it landed but was disabled [11:27] but i had to fix a python3 issue ... so seems it got enabled? [11:27] is that through phablet-test-run? [11:28] asac: yes; it's through phablet-test-run; it pics it depending on the packaging [11:28] the clicks need to change for py3 [11:28] * sergiusens will brb [11:29] sergiusens: so i see it tries to be smart, falling back to py2? [11:29] if imports blow up [11:29] i think we should somehow figure how to keep those highlighted that are still py2 [11:29] hello [11:30] sergiusens: can you please look at https://code.launchpad.net/~dpm/ubuntu-filemanager-app/include-plugin/+merge/213368 - dpm has a q for you.. [11:31] ogra_, is lxc-console or similar available to run inside the android container? [11:32] I have a nexus 4 and when I try to flash it with ubuntu-device-flash --channel=stable --bootstrap I get error autodeploy.zip not found error in cwm recovery [11:32] janimo, sur [11:32] e [11:32] lxc-console -nandroid -t0 [11:33] needs an explicit "enter" once you are in the console [11:33] ogra_, ah, the explicit enter tripped me up, thanks :) [11:33] yeah, i think lxc-console should just send a newline here [11:34] but didnt strike me as important enough to file a bug for it yet :) [11:34] Well I have been trying this feature for a while, and typing ctrl-a and other combinations it prints as a greeting [11:34] yeah [11:34] so maybe just saying 'press return' is good enough without actually sending it [11:34] ogra_, actually no, the -t0 matters too [11:35] by default tt1 does not work [11:35] yup [11:35] so the default setting for lxc=cosnole is not optimal I'd say [11:35] well, for our container [11:35] yes [11:35] open a bug and discuss it with stgraber :) [11:36] ogra_, against the lxc package [11:36] ? [11:36] dunno, i think so [11:36] dpkg -S $(which lxc-console) [11:36] :) [11:37] ogra_, sure I just wsnt' sure whether there's some lxc-touch-android package ship[ping config options for lxc :) [11:37] bfiller: hi! [11:37] bfiller: are you already around? [11:37] janimo, nope, there isnt [11:37] I think changing the default is not a good idea if it affects all other lxc uses [11:37] Publish a QML/C++ application as a click package | http://askubuntu.com/q/444446 [11:39] ogra_, filed https://bugs.launchpad.net/ubuntu/+source/lxc/+bug/1303756 [11:39] Ubuntu bug 1303756 in lxc (Ubuntu) "make lxc-console work non-confusingly by default on touch images" [Undecided,New] [11:40] confirmed [11:40] ogra_, thanks === FuLgOrE is now known as Georg [11:46] mardy, https://code.launchpad.net/~mardy/ubuntu-system-settings-online-accounts/app-access2/+merge/211711 doesn't have the account-plugin click hooks -- is that expected? [11:53] cwayne_: uh... let me double check, looks like a directory is missing from that branch... [11:54] cwayne_: I can blame the CI train, anyway :-) [11:55] mardy, :) i blame the CI train for everything, it's awesome [12:02] cwayne_: ah, lol, the click-hooks dir is already in trunk [12:03] yeah, the dir is, but the hooks themselves are missing right? [12:03] cwayne_: the files are there, but the directory is not listed in the qmake file, so they won't be installed [12:04] ah [12:04] cwayne_: the story was that CI merged this branch, then reverted it (but apparently not completely) [12:04] cwayne_: and this app-access2 branch is the second attempt [12:05] ah, right [12:12] ogra_. popey: about https://bugs.launchpad.net/ubuntu/+source/webbrowser-app/+bug/1303676 [12:12] Ubuntu bug 1303676 in webbrowser-app (Ubuntu) "[webapp-container] when using multiple webapps they crash randomly, if there is only one app remaining, this one starts being replaced when new ones start" [Critical,Confirmed] [12:13] i can reproduce, and suspect this could be the app lifecycle stopping instances because of memory pressure [12:13] thre is an open bug to handle memory pressure callbacks, which the container does not handle yet [12:13] dbarth, well, it should only SIGSTOP them, not kill [12:13] asac: what's the problem, that change landed 2~ weeks ago? [12:14] ogra_: unless the signal is not handled, and so the app stops, right? [12:14] dbarth, to me it looks more like the apps all end up under the same appid or some such [12:14] oh, indeed the app needs to know to stop on SIGSTOP and to start again on SIGCONT [12:14] sergiusens: the problem was that everyone believed that autopilot didnt swithc to python3 yet [12:15] they thought it was landed disabled [12:15] ogra_: so that could be it [12:15] so the issue that was not caught by the import fallback mechanism [12:15] great :) [12:15] came unexpected and also folks didnt know they should start using autopilot3 now [12:15] for pre-landing testing [12:15] right, apps are killed onc i have ~4 of them, whichever i start first [12:15] (which i dont agree with to be clear - they should use phablet-test-run) [12:16] sergiusens: so for me the biggest thing left is that its not making clear where we still run python2 [12:16] e.g. too much magic [12:16] without warning/error/feedback [12:16] maybe adding a big WARNING that counts as an error unless you pass --allow-py2 [12:17] would be the solution [12:17] dbarth: I don't think that's it. I still see lots of webapp-container processes, and oxider-renderers but the apps all disappeared (bar one) from unity [12:17] dbarth, the thing that made me think about the app id was that the last remaining app gets constantly replaced once you start new ones [12:17] asac: oh, warnings could be fine [12:17] and easy to add [12:17] but that might be just a coincidence [12:18] sergiusens: did you see my comment earlier about https://code.launchpad.net/~dpm/ubuntu-filemanager-app/include-plugin/+merge/213368 ? [12:18] ogra_: btw, the ofono/rild job seems racy, it wasn't that start stanza; it sometimes comes up, sometimes doesn't [12:18] sergiusens, btw, mind to review my phablet-bootchart MP (before asac pokes me about bootcharts again :P ) [12:18] popey: no, not really, sorry [12:18] sergiusens: really think we should be failing unless you epxlicitely set it to ignore, so that folks using phablet-test-run get reminded that they are not done yet. [12:18] np [12:18] and in infra continue running with ignore flag with big warnings [12:18] ogra_: saw it yesterday, looked fine. But didn't test it out yet [12:18] or something [12:18] will do [12:18] ok have to cook lunch [12:18] bbiab [12:18] sergiusens, thanks ... i'll take a look at the ofono thing ... probably we need to start later [12:20] whee ! [12:20] http://people.canonical.com/~ogra/touch-bootcharts/ubuntu-phablet-trusty-280.png [12:20] 27seconds :) [12:30] ogra_, nice! [12:30] cwayne_, i want to reach 20 :) (and the screen is up around 18 already in that bootchart btw) [12:31] ogra_, that's awesome [12:31] yup [12:32] poor mterry and MacSlow did all that work for the boot animation, and you'll never even be able to see it if we boot too fast :) [12:33] well, the last time i tried the animation silo it added another 5 sec to the boot [12:33] :P [12:33] that should be enough to see some animation :) [12:33] lol === mandel is now known as mandel|lunch [12:41] cwayne_, :) [12:41] cwayne_, well see it this way... if anything messes up our boot-time... we're at least ready with nice bling ;) === jhodapp is now known as jhodapp|brb [12:42] haha yep! im excited to see it land in the image MacSlow [12:42] cwayne_, going to push a few more tweaks later today. [12:45] MacSlow, the problem is that the boot *before* u-s-c starts is way longer than the few seconds it takes from u-s-c to the shell [12:45] * ogra_ would love if we could start u-s-c earlier [12:46] ogra_, I see... are the finer details for the boot-process documented somewhere? [12:46] not finer than in the bootchart above i fear ... point is that we need the android container up for the graphics driver [12:47] which takes about 10-12 seconds alone [12:47] (if you look at the chart you see a second init being started by lxc-start, thats the container startup) [12:50] ogra_, there's nothing coming that'll give us full native (non-android) startup, right? We need that because of all the hardware-drivers?! [12:51] yeah [12:51] specifically for the graphics driver === _salem is now known as salem_ [12:52] ogra_, it's not just the graphics-driver I assume... but also other SoC-parts too... DSPs, WiFi-chips etc [12:53] wifi is handled on the ubuntu side, but yeah. the rest [12:53] what i'm currently working on is to get the container start earlier and faster ... so we can cut down that time [12:53] but there is only so much i can do :) [12:54] with luck i can get it to 8 seconds or so [12:57] ogra_, I understand... good luck hunting for those fractions of seconds! :) [12:58] :) [13:00] popey: what was your comment? Don't see it in the MR [13:02] sergiusens: not mine, dpm's [13:03] sergiusens, I'm here now if you prefer to chat directly instead of via comments (re: the filemanager MP) [13:03] popey: yeah, he answered my question, so I think it's fine; not entirely fine, but fine; my original question implied that if they knew what they were doing and wasn't an accident it was ok === mhr3__ is now known as mhr3 [13:04] dpm: want me to do a full review? [13:04] dpm: is https://code.launchpad.net/~dpm/ubuntu-terminal-app/enable-translations/+merge/187986 still needed? [13:05] sergiusens, that'd be great [13:06] ok, will add to my TODO [13:06] mhall119: is https://code.launchpad.net/~vthompson/ubuntu-terminal-app/fix-1284602/+merge/208949 good to go? [13:07] popey, yes, it just never got reviewed or merged. I guess it'd need to be slightly changed now, as it's an old branch [13:07] right [13:07] awesome, thanks sergiusens, that'll allow us to land file manager to the store, as the version there now is a few months old [13:08] dpm: if you get a chance can you confirm bug 1303763 [13:08] bug 1303763 in Ubuntu Reminders app "File: qml/reminders.qml does not exist at any of the standard paths!" [Undecided,New] https://launchpad.net/bugs/1303763 [13:08] popey, looks like a duplicate of elopio's bug last week [13:09] maybe [13:10] dpm: but mine was built on jenkins, not click-buddy - so *should* work [13:10] (lol) [13:10] famous last words? :) [13:10] popey, what does jenkins use to build the click package? [13:11] as far as I know, magic. [13:12] dpm: cmake - http://s-jenkins:8080/job/reminders-app-click/6/console - build looks sane to my eye [13:12] fginther, could you join us today as well at the core apps call? [13:13] popey, I don't seem to be able to see that URL [13:13] are you on the vpn? [13:13] if not, thats why [13:13] ok, yeah, I try to stay away from setting up VPNs :) [13:13] in any case, de-duplicated, it seems to be another bug [13:14] Spironal [13:14] oops [13:14] dpm, I have a meeting conflict, I could 5-10 minutes early === jhodapp|brb is now known as jhodapp [13:15] fginther, np, shall we jump on the hangout then, say 15 minutes before time? [13:16] popey, balloons, would that work for you too? I.e. start the core apps review 15 mins earlier just for today? [13:16] dpm: sure [13:16] balloons is on vacation dpm [13:16] ah, I thought he'd be back today [13:17] done [13:17] oh, maybe, i thought he was out till wednesday at least [13:17] yes, his calendar confirms this [13:20] dpm, yes, works for me [13:20] perfect, thanks [13:21] popey, you're probably right, I've not checked the calendar, I just thought he'd be back Monday [13:32] rsalveti, sergiusens, i just noticed we still ship /var/lib/lxc/android/pre-start.d/15-no-uchroot ... i assume we can drop that nowadays ? [13:37] ogra_: I think so [13:37] great [13:37] the original file is not even there [13:37] that is really old stuff [13:38] right [13:38] and wastes a sed call on startup [13:40] rsalveti, when did you last look at logcat output of one of our images ? [13:40] W/Adreno-ES20( 2074): : GL_INVALID_OPERATION [13:40] W/Adreno-EGLSUB( 2074): : PMEM_INV_CACHES undefined [13:40] i cant remember seeing that last time i looked ... but thats admittedly weeks ago [13:41] now it seems to be a frequent error [13:41] (like every few seconds) [13:41] hm, not sure if that normal [13:41] let me flash latest [13:41] flo? [13:42] mako === eikeon_ is now known as eikeon [13:43] flo has it too though [13:43] just checked [13:45] rsalveti, seems to happen every time i switch apps or slide back to the applications scope [13:45] right [13:45] smells like something caused by the new app switcher [13:46] just flashing an older image to check [13:46] 250 should be fine [13:46] was the last promoted one [13:46] as I remember we had such warnings before, but not sure if that frequently [13:46] without the new switcher iirc [13:47] not sure i have ever seen "INVALID_OPERATION" [13:47] i know we had stuff there before ... [13:47] but iirc only on load of the driver [13:47] janimo: the reason why lxc-console won't attach to tty0 but always to tty1 or higher is that attaching to tty0 only works for backgrounded containers [13:47] janimo: containers which aren't backgrounded have tty0 attached to the original lxc-start process and so can't be opened by lxc-console [13:48] didrocks, hey, I have packages in silo 4 that build-depend on the latest oxide-qt that is currently in trusty-proposed, is there a way to have the PPA depend on proposed to build the packages? [13:48] stgraber, well, probably having a config file for lxc-console would help :) [13:48] ogra_: not sure if you saw in scrollback/e-mails but it looks like I've got cgmanager working reliably on touch now. I'll give you a debdiff for the lxc-android-config change in a bit. [13:49] stgraber, yeah, i saw that ! [13:49] thanks a lot ! [13:49] lxc-android-config is currently blocked in a silo so we cant "just upload" atm [13:51] ogra_: so in theory if lxc-console could query LXC to know whether a given container was started on the background or not, then it could default to tty0 when it's backgrounded. However there's no way to query that at the moment and I believe adding that feature would require an API break of the internal communication protocol, so probably not something we can do until the next major upstream release (as we try not to break our API or commun [13:52] you cut off after "API or commun" [13:52] ogra_: I don't think lxc-android-config is terribly urgent, but we probably want to make sure this gets done this week, quite likely before final freeze [13:52] yeah [13:52] "or communication protocol within the same SOVER)" [13:52] stgraber, well, i personally know about -t0 and my fingermemory uses it ... i dont think it is urgent :) [13:53] janimo might disagree though [13:53] for me its in the "nice to have" category [14:03] oSoMoN: it does depend on proposed normally [14:03] oSoMoN: let me recheck [14:03] didrocks, ok, so it might be that the oxide build wasn’t completed in proposed when I triggered the build in the PPA [14:03] hello all [14:04] didrocks, so the next question would be: if I hit the build button again on the silo page, will the packages be rebuilt in the PPA? [14:04] oSoMoN: yep, checked again, it does dep on proposed [14:05] rsalveti, do commands like breafkast not work on the new aosp branch?? you gave me to checkout last friday?? [14:05] oSoMoN: better to not rebuild a source package for nothing, if it does say it build-deps, there is a cron every 30 minutes on launchpad PPA retrying a build [14:05] annerajb: you need to use lunch and then make [14:05] but the device i am porting too is not listed there [14:06] rsalveti, I am porting (or trying) the lG G2 which should be d802 but is not listed since it has no notion of it existing. [14:06] didrocks, it’s been more than 30min since oxide-qt hit proposed (lp says 4hours), and no rebuilt has happened [14:06] oSoMoN: more than 30 minutes that it's published in proposed? [14:07] oSoMoN: so built + published? [14:07] didrocks, yes, I think so [14:08] oSoMoN: it's because you not build-dep [14:08] oSoMoN: did you forgot to bump the build-dep on latest oxide-qt? [14:08] you just FTBFS [14:08] annerajb: you need to add the target to lunch [14:09] didrocks, ok, got it, there’s no explicit version number requested in the build dep, there’s only a builddep on liboxideqt-qmlplugin [14:09] oSoMoN: yeah, it should have been with the explicit version number [14:09] didrocks, I’ve hit build again, it should fix the ftbfs [14:09] oSoMoN: can you fix the version number in next build? [14:10] oSoMoN: you surely want to provide "webbrowser-app" in rebuild only package list [14:10] to not rebuild the qml component [14:10] didrocks, it doesn’t make sense for webbrowser-app to depend on an explicit version number, as oxide is a moving target, it would only be a maintenance burden [14:10] didrocks, I did [14:10] rsalveti, any hints on the syntax i searched on google and tried doing which lunch to find the source of lunch but I had no luck [14:10] also tried lunch -? --help -h [14:10] annerajb: check build/envsetup.sh [14:11] annerajb: you have all the used functions in there [14:11] oSoMoN: well, don't expect magic happening then from LP, you need to rebuild everytime :) [14:11] you can also source it with bash -x, and see the output when calling lunch [14:11] didrocks, that’s alright, thanks for explaining :) [14:12] oSoMoN: yw ;) [14:12] didrocks, I like magic though, when it makes my life easier ;) [14:13] oSoMoN: well, there is no magic info though to know you need latest oxide-qt ;) [14:13] yeah, I was just kidding :) [14:16] rsalveti, found out how you have to add it to vendersetup.sh the add_lunch_combo and when you run build/envsetup.sh it updates the menu [14:16] annerajb: right [14:16] rsalveti, should this be documented on the porting guide? [14:16] but I guess you might need a few additional files and defined variables in your device specific repository [14:16] annerajb: yes, if you have the time, please do so [14:19] ogra_: E/MP-Decision( 879): Error 13 setting online status to 0 for cpu2 [14:19] E/MP-Decision( 879): Error(-19) changing core 2 status to offline [14:19] guess this is expected? [14:19] logcat is now basically useless [14:19] the entire ringbuffer is full of those messages [14:19] i was wondering if it makes sense to put mpdecisionn into late start [14:19] only for the first 60 sec [14:20] not sure if that will help, let me check that [14:21] rsalveti, rsalveti echo manual >/etc/init/no-cpu-hotplug.override [14:21] and reboot [14:21] popey, did you file a bug on some webapps not starting properly [14:22] pmcgowan: can you be more specific? [14:22] popey, I run facebook and get a whitescreen [14:22] pmcgowan: wikipedia is the only one (in the store, not ours) that didnt start [14:22] wfm [14:23] facebook is out, use G+ [14:23] http://popey.com/~alan/phablet/device-2014-04-07-152300.png [14:23] hmm [14:23] will reboot and try again [14:23] pmcgowan: facebook works fine for me too [14:23] ogra_: you can only share to facebook from the gallery ;-) [14:23] thanks [14:24] * ogra_ wonders if any of our app developers care about migrating app data etc ... [14:24] so facebook wins now [14:24] pfft [14:24] oSoMoN: ping [14:24] ogra_: as a user I was bothered to need to relog in given that I don't remember all my passwords :-) [14:25] mhall119, pong [14:25] see that as an opportunity to memorize them better ;) [14:25] ogra_: yeah, it's fine after you override that job [14:25] rsalveti, sure [14:25] let me see if late start helps, but I don't think it'll help much [14:25] i just wonder how we could make mpdecision be quiet about it [14:25] oSoMoN: does UbuntuWebView support the UrlSchemeDelegate overrides that WebKit has in experimental? [14:26] yeah, it will cut a few messages [14:26] but surely not all [14:26] as the ro/rw switch for the sysfs node only happens after 60sec [14:28] mhall119, no [14:29] ogra_: why 60sec? [14:29] rsalveti, to have all cores up during the whole boot [14:30] 45 would suffice for us i guess ... but i want it to be fast on slower devices too [14:37] ogra_, hey, seen my last comment on bug #1303676 ? [14:37] bug 1303676 in webbrowser-app (Ubuntu) "[webapp-container] when using multiple webapps they crash randomly, if there is only one app remaining, this one starts being replaced when new ones start" [Critical,Confirmed] https://launchpad.net/bugs/1303676 [14:38] oSoMoN, nope ... will check and comment ... [14:38] ogra_, thanks [14:40] oSoMoN, [14:40] ogra@styx:~/Devel/apps$ adb shell ps ax |grep -c webapp-container [14:40] 9 [14:40] ogra [14:40] thats with only one app visible in the scope [14:40] so yes, i can confirm [14:40] ogra_: changing to late_start doesn't help much [14:41] ogra_, so it seems like a bug in unity8, right? [14:41] oSoMoN, also a good bunch of "[oxide-renderer] " [14:41] and mpdecision is a binary blob afaik [14:41] so not much we can do [14:41] just annoying that while you fixed the kernel to have a useful dmesg, now logcat is useless lol [14:41] rsalveti, hmm, then we have to live with it i fear ... i wont give up the 4seconds that cgains us [14:42] well, you can disable the upstart job ... [14:42] its just useless if you didnt do that [14:42] right, but I mean for debugging [14:42] without making it rw [14:43] could we redirect mpdecision at startup ... to /dev/null ? [14:44] ogra_: nops, because this is part of the androd log system [14:44] not stdout [14:44] (i dont want to add more sed commands to pre-start.d though) [14:44] but guess not much we can do, unfortunately [14:44] if we want to have a super fast boot :-) [14:44] yeah [14:44] this was one of the more noticeable speedups [14:44] ogra_: so where's the latest lxc-android-config? you mentioned it being in a silo somewhere. I'd like to base my debdiff on that so it's easier to apply once the current one has landed. [14:45] i'm curious what the next one will give ... once we rebuild android my new handling of the container unpacking will land [14:45] stgraber, silo 19 [14:45] stgraber, but i want to have at least a few people test it ... i'm a bit scared about breaking the lab for days again [14:46] and testing that change is time consuming ... === brendand_ is now known as brendand [14:53]