n0c_Will apt-get update/upgrade keep me on the lastest developer version?00:07
=== michi__ is now known as michi
=== michi is now known as Guest35414
_lunarcoldanyone mind going over build failure logs with me?01:04
=== Fyodorovna is now known as wilee-nilee
=== salem_ is now known as _salem
=== kriskropd is now known as kriskropd_is_not
=== kriskropd_is_not is now known as kriskropd_isnt_h
=== kriskropd_isnt_h is now known as kriskropd_afk
=== kriskropd_afk is now known as kriskropd
savedjulihows going02:49
=== chihchun_afk is now known as chihchun
Traffickinghi everyone04:53
TraffickingI need some help please04:53
TraffickingHow do i get Ubunto on a Galaxy s2 ? Any guides ?04:53
_lunarcoldTrafficking: which model04:56
=== Namidairo`bnc is now known as Namidairo
Traffickingi9000 Sorry for delay05:13
_lunarcoldTrafficking: https://wiki.ubuntu.com/Touch/Devices/galaxysmtd05:13
_lunarcoldso, a lot of stuff is broken05:13
Traffickinglol yup05:15
Traffickingwish I could get an edge :(05:15
TraffickingDeveloper form the Caribbean but didn't get the change to donate since most of the chepear ones were claimed05:16
_lunarcoldTrafficking: you can work on porting05:16
_lunarcoldmy particular device isn't supported and I am trying to port :>05:16
Traffickinglol good idea05:17
Traffickingwhat device do you have ?05:17
_lunarcoldHTC One (m7spr)05:18
_lunarcoldstill blocking on compiling, need to parse through the errors some more05:18
_lunarcoldthe build on the wiki isn't actually for m7spr I think05:18
=== jono is now known as Guest5052
=== tvoss_ is now known as tvoss|flaky_wire
=== tvoss|flaky_wire is now known as tvoss|test
MirvSaviq: do you know who could look at the http://pad.lv/1207269 ? I think Kaleo is looking at the webbrowser font bug, so that's the other one.07:09
ubot5Launchpad bug 1207269 in Unity 8 "Indicator animation slope wrong with Qt 5.1" [Undecided,New]07:09
SaviqMirv, that'd be dednick07:10
SaviqMirv, let me see if I can reproduce with the ppa07:11
MirvSaviq: ok, thanks07:12
dpmgood morning all07:38
SaviqMirv, confirmed07:49
MirvSaviq: thanks07:53
AskUbuntuUbuntu or linux does not reconize sim slot card (hp 24un..) | http://askubuntu.com/q/32962707:53
Mirvalso just updated qtwebkit there, to contain the same private headers that were added back in saucy as well as webbrowser-app depends on them07:53
dholbachgood morning07:57
=== tvoss|test is now known as tvoss|flaky
=== seb128_ is now known as seb128
diwicogra_, hi, so crosscompiling for android using gcc-arm-linux-androideabi worked fine for a simple example, but what if I want to link to libhardware?08:28
ogra_i guess then you need the full tree ...08:29
xnoxdiwic: if you drop android target libraries into: /usr/arm-linux-androideabi/lib/ you can link against them, and headers into /usr/arm-linux-androideabi/include/. possibly libhardware should be packaged as arch:all and ship those files there under name "libhardware-dev-android" to ease cross-compilation.08:34
diwicxnox, so for the lib part it is as easy as copying the relevant .so files from /system/bin/ on the Nexus 4 to /usr/arm-linux-androideabi/lib/ on the x86 desktop?08:36
xnoxdiwic: correct.08:38
diwicxnox, all right, thank you08:38
ogra_xnox, hmm, does that deserve a wikipage perhaps ?08:44
xnoxogra_: no, idea =) to me it's trivial, since that's how all non-multiarched cross-compilation happens on debian/ubuntu. Plus that's where the libc.so is for android, thus easily discoverable.08:46
=== schwuk_away is now known as schwuk
xnoxogra_: i guess we can document this, but i'd rather hear success from diwic first =) if above actually works. ;-)08:47
diwicxnox, compiling works, but when trying to run I get this error:08:51
diwicxnox,  soinfo_relocate(linker.cpp:989): cannot locate symbol "__exidx_end" referenced by "./a.out"...CANNOT LINK EXECUTABLE08:52
xnoxdiwic: running on nexus4 that is?08:52
diwicxnox, yes08:52
xnoxdiwic: ... and in android lxc container08:52
diwicxnox, I have not set one up08:53
xnoxdiwic: hm? android cross toolchain is to compile for android lxc container / unflipped image system portion.08:54
xnoxdiwic: to run against ubuntu, one should crosscompile for normal ubuntu / armhf.08:54
diwicxnox, I just run "arm-linux-androideabi-gcc test.c -lhardware"08:54
xnoxdiwic: that symbol in in libc.08:56
xnoxdiwic: $ arm-linux-androideabi-gcc test.c -lhardware -lc ?08:57
xnox(well the bionic libc)08:57
diwicxnox, \o/ now it runs fine on the Nexus 408:57
xnoxdiwic: =)))))08:57
diwicI have never had to add -lc before08:57
diwicthat's a new one :-)08:58
* xnox says excellent in Mr. Burns voice from Simpsons08:58
SuperMattNow that I've finally decided to jump in to touch dev, I'm buying a nexus 10 :D08:59
xnoxSuperMatt: hm... would be more intersting to get the new nexus 7 and start a port to that =)09:01
SuperMattnaw, I wanted something bigger09:02
JamesTaitGood morning all, happy Lighthouse Day! :-D09:14
mptrachelliu, seb128: https://wiki.ubuntu.com/Appearance?action=diff&rev2=7&rev1=609:16
seb128mpt, thanks09:16
SaviqMirv, btw, do we have dbg symbols for qt5-beta-proper?09:20
SaviqMirv, like ddebs or something?09:20
asacogra_: moin09:22
cjwatsonplars: use "pkcon -p install-local foo.click" (or --force-missing-framework if you must use the low-level tools, but you might be doing something else wrong if you are)09:23
asacogra_: where are our images at?09:25
seb128mpt, page 19 in the content picker spec has the background capplet case with a popover list rather than a sub-screen, just pointing it out in case that's an issue09:26
ogra_asac, building i'd guess09:34
ogra_ah, no, still 30min to go09:35
=== chriadam is now known as chriadam|away
MirvSaviq: Qt5 packages build debug pkgs always so the same as usual, qtdeclarative5-dbg etc09:46
SaviqMirv, right, thanks09:53
xnoxogra_: diwic: now i have a question for you  =)))10:16
xnox"$ echo "CANNOT LINK EXECUTABLE: could not load library "libubuntu_application_api.so" needed by "./ubuntuappmanager"; caused by cannot locate symbol "_ZN7android14SurfaceControl11setPositionEii" referenced by "libubuntu_application_api.so"..." | c++filt10:16
xnoxCANNOT LINK EXECUTABLE: could not load library libubuntu_application_api.so needed by ./ubuntuappmanager; caused by cannot locate symbol android::SurfaceControl::setPosition(int, int) referenced by libubuntu_application_api.so...10:16
xnoxi guess i need to recompile libubuntu_application_api against goldfish build.....10:16
davmor2Morning all10:17
ogra_libubuntu_application_api.so shouldnt have HW specific bits i think ... thats weird10:18
xnoxogra_: do you happen to know which library usually provides that symbol? i can check the signatures.... unless it's at all not available in that case ubuntu_application_api might need porting to goldfish.10:18
diwicxnox, no idea10:18
ogra_xnox, not really, but surface control might be surface flinger10:19
xnoxogra_: http://paste.ubuntu.com/5958390/ that's all the libs that platform-api is linked against.10:20
ogra_i'd take a look at libhardware then10:21
xnoxogra_: looks liek libgui.so10:26
xnoxhmm... ubuntuappmanager/(hybris, application api?!) doesn't compile against AOSP 4.3 http://paste.ubuntu.com/5958448/10:51
asactvoss_: ^^10:52
tvoss_asac, yup, known issue. We have a patch for that for aosp 4.210:53
asactvoss_: maybe xnox can try to forward port?10:53
tvoss_xnox, there is a patch somewhere in our patchset that makes the copy c'tor available10:54
asacso yeah. give him pointers i guess10:54
tvoss_asac, yup, should be straightfoward to forward port10:54
xnoxtvoss_: hm. in the phablet-ubuntu repo forest or somewhere else?10:54
=== MacSlow is now known as MacSlow|lunch
xnoxtvoss_: or in platform-api project?10:54
tvoss_xnox, in the phablet-ubuntu forest10:55
tvoss_xnox, it is against the android source tree10:55
xnoxtvoss_: ack. i'll look for it.10:55
Chocantofginther: Hey ! :) How can I ask for a rebuild of the poppler-qml-plugin package ? (
tvoss_xnox, remind where we have those patches?10:55
asacfginther: do you have a mako?10:56
asaccould you see if you can reproduce the issues from yesterday? we saw a bunch of failures there, but not on maguro10:56
asacfginther: but maybe wait till todays image is tested10:56
xnoxtvoss_: well, I have a local checkout =) but it should be http://phablet.ubuntu.com/gitweb10:56
tvoss_xnox, thanks10:58
asacxnox: http://phablet.ubuntu.com/gitweb?p=CyanogenMod/android_frameworks_base.git;a=commit;h=00b43634dcbc476bdcd1fffdca7ed74ba0a7b9c3 ?10:58
asactvoss_: ? that one?10:58
tvoss_asac, +110:58
tvoss_asac, just found it, too10:58
xnoxasac: looks good. let me try that =)10:59
* asac likes diving sometimes :)10:59
asacsee where clicks can lead me10:59
ogra_to click packages in the end :P10:59
asacmight also be related10:59
asacbut not sure... sounds important at least :)10:59
asacxnox: ^10:59
dholbachdo we have anyone in here who can easily test one of the images on https://wiki.ubuntu.com/Touch/Devices?10:59
asacguess all in base.git is important10:59
asacogra_: lol... commits and patches in click packages11:00
asacnice picture11:00
asacno more git11:00
asacmaybe that will solve our archive problems as well :)11:01
ogra_asac, new image is up btw11:04
asacok cool11:05
asaclets cross our fingers :)11:05
* xnox launches a recompile & off for lunch11:08
=== marcoceppi_ is now known as marcoceppi
ogra_gema, asac ... maguro doesnt look happy (2 of the default test failed) ... someone should re-run11:37
asacdoanac: gema: plars: fginther: ^^11:38
ogra_looks like a netwrokign issue11:38
diwicgaah, where are those properties again11:39
diwicro.hardware and those11:39
diwicogra_, if I just call /system/bin/getprop the result is nothing11:40
ogra_no, use the ubuntu getprop :)11:40
ogra_its in path11:40
kalikianarsalveti: ping you're involved with lp:qtubuntu-sensors presumably11:41
diwicogra_, why does the android getprop not work? I'm suspecting that's why my android program doesn't work.11:41
ogra_the android getprop works on the androiid side11:42
ogra_but not on the ubuntu side11:42
diwicogra_, you mean it only works inside the android chroot?11:42
ogra_(it needs to talk through hybris to the container, the android one wants to talk to androids /init directly which you cant from outside the container)11:42
ogra_no, it works only inside the container ... not via chroot11:43
ogra_you would actually need to enter the namespace of androids /init11:43
diwicogra_, so if I make an android binary, and that program calls hw_get_module_by_class, and the implementation of hw_get_module_by_class needs properties to function correctly...11:44
ogra_then you should be able to use getprop as long as your app is executed inside the container11:45
ogra_(which you cant reallly do from the ourside, copy it to /system and have an init.rc entry, then restart the container)11:45
SaviqMirv, any update on bug #1181176 ?11:45
ubot5bug 1181176 in Ubuntu UI Toolkit "should depend on the Ubuntu font" [Medium,Confirmed] https://launchpad.net/bugs/118117611:45
diwicogra_, argh11:46
ogra_diwic, there is a way to start adb inside the container instead of in ubuntu ... that way you can exec stuff inside... you have to disable the ubuntu adbd and enable the adbd in the container for this11:47
diwicogra_, so if I use hybris I run into hybris related bugs and if I try to compile android binaries I can't call them when I want to11:47
MirvSaviq: good point, it has been buried lower and lower on my list, but is a small thing. I have no idea though which would be better, the meta package or ui-toolkit? if it's assumed that the UI toolkit is used together with the Ubuntu font, maybe that?11:48
SaviqMirv, or maybe the theme, even?11:48
SaviqMirv, but I believe there's enough font: "Ubuntu" buried in the toolkit itself11:49
SaviqMirv, that it makes sense there, too11:49
=== melmoth__ is now known as melmoth
MirvSaviq: I'll check with the team and make a merge request11:57
SaviqMirv, thanks11:57
=== alan_g is now known as alan_g|lunch
=== Namidairo is now known as Namidairo`bnc
KaleoMirv: talking about hte webbrowser font bug, I think I'm hitting a bug in Qt while fixing it12:04
=== MacSlow|lunch is now known as MacSlow
KaleoMirv: I'll let you know if we need a patch12:04
MirvKaleo: ok12:12
plarsogra_: minor fix was needed, restarted all the tests, until they all run, you'll still see the red ones but they'll be replaced as the new jobs run12:15
ogra_plars, great12:20
dokoMirv, Saviq: hi12:22
Saviqdoko, hey12:22
dokoapt-get install qtbase5-dev:armhf doesn't work yet12:22
otaonhas a version of Ubuntu Touch for Galaxy Note N7000 ?12:22
davmor2ogra_: any idea what is going on with the images today?  All the tests are red and there appears to be a new run starting?12:23
Saviqotaon, doko, why's that?12:23
Saviqotaon, https://wiki.ubuntu.com/Touch/Devices12:23
Saviqdoko, why can't it be installed?12:23
Saviqdoko, it's there in the archive12:23
dokodepends on qt5-qmake (= 5.0.2+dfsg1-7ubuntu2), qtbase5-dev-tools (= 5.0.2+dfsg1-7ubuntu2), qtchooser12:23
ogra_davmor2, infrastructural issues12:23
dokoSaviq, sorry, can't be installed for a secondary architecture12:23
ogra_davmor2, the failures were network related12:23
otaonThanks Saviq12:24
davmor2ogra_: so is that a hold off on phablet-flash --pending then?12:24
Saviqdoko, meaning that it doesn't do multiarch properly?12:24
dokoso the question is can this be qt5-qmake:any qtbase5-dev-tools:any qtchooser:any dependencies?12:24
dokoSaviq, yes12:24
diwicogra_, sorry for bombarding you with questions, but if I'm supposed to run this inside the container...where on the file system do I put it? Everything inside the container seems to be read-only...12:25
ogra_davmor2, no, why ?12:25
Saviqdoko, I think it should be fine, yeah, they're just generating text files12:25
dokoand a qtbase5-dev-tools for the host won't help you when running on the build12:25
davmor2ogra_: ah sorry you mean on the test array rather than the image sorry12:25
ogra_diwic, remount /system rw and just copy it where you like12:25
Saviqdoko, I think we should just fix multiarch for them12:25
dokoSaviq, the question is, if moc knows about to generate files for the host, not the build12:26
ogra_davmor2, right12:26
dokoand I have no clue about qtchooser12:26
ogra_i meant http://reports.qa.ubuntu.com/smokeng/saucy/12:26
fgintherasac, sorry for not responding earlier, I wasn't in yet. I don't have a mako12:27
Saviqdoko, qtchooser is just a wrapper around /usr/lib/*/qt5/bin12:27
Saviqdoko, that "chooses" between qt4 and qt512:27
Saviqdoko, but I'd say it needs to be per-arch to actually choose the right arch for them12:28
dokois this a script? then maybe it should be extended12:28
Saviqdoko, how should we deal with binaries/executables for multiarch?12:28
Saviqdoko, no, it's a binary I'm afraid12:28
rsalvetidiwic: android getprop doesn't work because it tries to get the property via shared memory12:29
rsalvetidiwic: were you linking with the android libhardware or the one provided by hybris?12:29
diwicrsalveti, hi! As you advised, to rule out libhybris problems I'm trying running it as android entirely12:30
dokoSaviq, the idea is to make qtchooser M-A foreign, and then install it for the build machine. However don't know if it does do the right thing then ...12:31
rsalvetidiwic: right, cool12:31
rsalvetidiwic: but you're trying something I never used yet, which is building with our android toolchain :-)12:31
Mirvhello doko.12:31
Saviqdoko, no idea myself :/12:31
Mirvdoko: I was just about to say maybe qtchooser should M-A foreign...12:31
rsalvetiI usually just build it as part of the android build system, looking forward to see if that works as expected12:31
dokoMirv, sure, if it does do the right thing?12:32
rsalvetiif you get it to link with the android library, and pointing to the right headers, it should just work12:32
rsalvetidiwic: how I usually get inside the android container to test stuff there: http://paste.ubuntu.com/5958699/12:33
Mirvdoko: not sure either. the Qt packages now do install arch specific configs for qtchooser, so the answer is strong maybe.12:33
dokoMirv, same for qt5-make and qtbase5-dev-tools12:34
Mirvdoko: they are already12:34
cjwatsonSaviq: If you haven't read https://wiki.ubuntu.com/MultiarchSpec, you should12:34
dokoalready what?12:34
cjwatsonSaviq: Also https://wiki.ubuntu.com/MultiarchCross12:34
Saviqcjwatson, thanks, will do12:34
rsalvetijdstrand: we're moving the audio call setup logic as part of telepathy-ofono, which is part of the user session12:35
rsalvetijdstrand: as a consequence, we need to access binder directly, as the audio plugin will later use binder to call audioflinger12:35
dokoMirv, ahh, ok, but then the deps need to have an :any as suffix12:35
rsalvetiwhen testing yesterday, we got the following:12:35
rsalveti<5>[  411.065246] type=1400 audit(1375836837.082:67): apparmor="DENIED" operation="open" parent=1 profile="/usr/lib/telepathy/telepathy-*" name="/dev/binder" pid=1405 comm="telepathy-ofono" requested_mask="rw" denied_mask="rw" fsuid=32011 ouid=012:35
dokoMirv, also is qtchooser really needed as a dependency?12:36
diwicrsalveti, okay, so now I managed to open the audio HAL from inside android, but it fails because it can't access the sound card anymore...12:37
jdstrandrsalveti: yeah, telepathy-mission-control-5 needs to have its profile updated12:37
jdstrandrsalveti: do you have packages I can test?12:37
Saviqdoko, without qtchooser you'd be required to always use /usr/lib/*/qt5/bin/$command12:38
rsalvetijdstrand: should have in a few minutes, jenkins is still building it12:38
rsalvetidiwic: hm, after a clean boot?12:38
jdstrandrsalveti: k-- once you do, I'll update telepathy-mission-control-512:38
rsalvetijdstrand: awesome, thanks12:38
dokoMirv, Saviq: what about qt5-base? that's not m-a foreign or same12:38
Mirvdoko: ok, building now a test package at https://launchpad.net/~canonical-qt5-edgers/+archive/qt5-staging/+packages12:38
Mirvof qtchooser12:38
Saviqdoko, I'll leave you in Mirv's capable hands :)12:39
Mirvdoko: qtchooser is not strictly needed, but in the end it was chosen to be dependency instead of recommendation in Debian12:39
dokoMirv, should qtbase5-dev-tools-dbg be M-A: foreign too?12:39
diwicrsalveti, hmm, maybe the "shell" user is not "system" or "audio"12:39
rsalvetidiwic: try running as root12:40
rsalvetifrom your host, do 'adb root'12:40
rsalvetithen adb shell again, and you should be root12:40
saiarcot895In some references to Ubuntu for Android, there are mentions of the phone docking to the computer; what if the phone can only be connected through USB?12:40
Mirvdoko: qtbase5-dev itself contains arch specific files in addition to eg /usr/include files, so it can't be. I don't know about -dbg.12:41
diwicrsalveti, right, I have to do it your way then perhaps, I'm currently using tcp:5555 bridge to get inside the container and when I restarted that as root, the adbd process ceased to exist I think12:42
rsalvetidiwic: right, you'd need to restart adb over tcp for that to work, trying directly via usb is indeed easier12:43
diwicrsalveti, /var/lib/lxc/android/pre-start.d/10-no-adbd  <- pre-start.d does not exist12:46
rsalvetidiwic: hm, which image are you using?12:46
rsalvetidiwic: probably from last week12:46
=== alan_g|lunch is now known as alan_g
rsalvetiif so, change /var/lib/lxc/android/pre-start.sh12:47
ogra_huh ?12:47
ogra_thast in since a while12:47
rsalvetidiwic: http://paste.ubuntu.com/5958749/ this is what you need to comment out12:47
ogra_and use a fresher image :P12:48
diwicrsalveti, ogra_ I think it is from this or possibly last week12:48
rsalvetiwe move fast12:48
ogra_more likely last12:48
ogra_i think i uploaded that end of last week12:48
ogra_or mid ...12:48
* ogra_ is to lazy to check12:48
rsalvetidiwic: can you paste the code you were using that does the init over hal?12:50
rsalvetidiwic: interested to check the hybris crash in there12:50
sergiusensogra_: it was last week12:51
ogra_heh, you know my changelogs better than me :)12:51
=== _salem is now known as salem_
diwicokay, so now my program - running as root inside the container - seems to setup the audio HAL for a voice call, of course the GUI in the phone app won't work12:55
diwicah stupid me, forgot to enter sim pin code12:58
diwicall right13:01
=== greyback is now known as greyback|lunch
diwicrsalveti, I'll paste some information for you in a while13:02
rsalvetidiwic: thanks13:03
seb128do we have a way to tell from qml if we are running on a touch image or not?13:10
ogra_there is a bunch of env var stuff in /etc/environemnt13:11
ogra_(Qt related)13:11
seb128ogra_, I was rather expecting something in the sdk telling at least the form factor ;-)13:12
ogra_not sure if there is something13:12
seb128ogra_, I want to make system settings use gnome-settings-daemon rather than powerd on desktop, to make testing easier13:12
seb128I'm going to hack with a if screen.width > 1024 I guess13:13
seb128indeed :p13:13
ogra_in landscape my SGS4 uses 192013:13
ogra_and the n4 128013:14
seb128give me a better way then ;-)13:14
diwicseb128, can't you check if powerd is running, and if not, talk to g-s-d instead?13:15
ogra_yeah, something like that13:15
seb128diwic, qml is mostly UI/js, I'm not sure you can do that sort of things easily13:16
rsalvetisalem_: jdstrand: http://people.canonical.com/~rsalveti/jdstrand/13:16
rsalvetiyou need both ofono and telepathy-ofono to be able to test it13:16
diwicseb128, well, you're supposed to talk to powerd, if it doesn't reply then...13:16
rsalvetithen just try doing a call13:16
seb128diwic, rsalveti: I'm not, I'm reading a gsettings key13:17
seb128and to add to the fun, gsettings abort() if you try to use a missing schemas13:17
seb128well, larsu is going to wrap the gsettings api to check if schemas are available13:17
seb128then I can do that13:17
diwicseb128, side question, is system settings able to switch sms notification sound? I still get a small "bing" even though I tried to change it13:18
diwicseb128, but maybe I didn't understand how to "commit" my notification sound change13:18
seb128diwic, no, nobody on the phone app team is reviewing my merge requests13:18
seb128I had changes pending review for 2 weeks13:18
seb128commented on launchpad bug and mp13:18
seb128but I just get silence in return so far13:18
diwicseb128, oh :-/13:19
seb128diwic, https://code.launchpad.net/~seb128/telephony-service/sounds-events-from-gsettings/+merge/17834413:19
rsalvetiogra_: seb128: we indeed need a clear way to identify if we should behave as touch, and which form factor during runtime13:19
AskUbuntuubuntu phone daily updates | http://askubuntu.com/q/32972313:19
ogra_rsalveti, right, but definitely not based on resolution :)13:19
diwica resolution without resolution13:20
rsalvetisettings somewhere? env variable? some utility to call?13:21
diwicrsalveti, so http://paste.ubuntu.com/5958853 <- that's how the code for talking to the HAL currently looks like, but it was smaller when I tried the hybris variant.13:21
dholbachsergiusens, can we do something about https://lists.launchpad.net/ubuntu-phone/msg03645.html?13:23
dholbachsergiusens, shall he file a bug about it?13:23
rsalvetidiwic: cool, let me give it a try13:24
kenvandinemardy, i pushed more fixes to my branch13:26
kenvandinemardy, but the tests fail in pbuilder because it doesn't write the parameters.txt file13:27
mardykenvandine: let me check13:28
mardykenvandine: jenkins approved it :-)13:30
mardykenvandine: but I missed your change to the .pc file, I'll do that13:30
=== greyback|lunch is now known as greyback
mardykenvandine: BTW, why did you rename the library? Is there something wrong with libCamelCase?13:31
kenvandineto match the package13:31
kenvandinelintian complains13:31
ogra_stgraber, so i have some bad news ... i'm about to move livecd-rootfs over to use the android package ... that measn for changed in the initrd scripts you now also need to upload android and wait for it to have picked up the new ubuntu-touch-generic-initrd13:31
ogra_*that means for changes13:32
kenvandinemardy, we could have also renamed the package13:32
mardykenvandine: weird... Qt libs use camelcase as well13:32
kenvandinebut i don't like caps in package names :)13:32
mardykenvandine: ah, OK :-)13:32
jdstrandrsalveti: ack13:32
sergiusensdholbach: if TWRP doesn't support extendedcommands, there is nothing I can do13:32
kenvandinemardy, true.. there are lots of them, i wonder if there is something different about them or if people just ignore lintian warnings :)13:33
sergiusensdholbach: I could look into openrecoveryscript... but that's a new thing13:33
ogra_sergiusens, instead we should make ports use our recovery13:34
ogra_after all there should be an img that came out of the build13:34
kenvandinemardy, odd... even out of tree builds work fine and successfully create parameters.txt in the right place13:34
sergiusensogra_: well, they'll need to manually flash it13:34
kenvandinebut it fails to get created in pbuilder13:34
ogra_sergiusens, aww, no fastboot ... well, still, i would make it a requirement13:35
sergiusensogra_: but sure... plain clockworkmod recovery works fine for now... but when the time for image based upgrades come it will be a hard req13:35
sergiusensdholbach: ^^13:35
dholbachI can relay the message, but I don't understand anything13:37
ogra_dholbach, we need a minimal set of commands in the recovery ... only CWm and our own recovery offer that ... others may or may not ... if users insist on using their own recovery they wont be able to use phablet-flash and are on their own13:39
sergiusensdholbach: already replied13:39
=== Kyle_ is now known as Kyle
sergiusensit seems TWRP supports openrecovery script which we can look into13:39
dholbachogra_, I don't think we effectively advertise what kind of recovery we want used, do we?13:39
dholbachsergiusens, thanks for the reply13:40
stgraberogra_: hopefully I won't have to change it too often then13:40
ogra_dholbach, no, we should define that somewhere13:40
sergiusensdholbach: it doesn't matter when doing it manually13:40
ogra_stgraber, yeah ... its a drawback that only just struck me now when doing a testbuild here13:40
ogra_it will definitely extend the turnaround time for an initrd change massively :(13:41
stgraberogra_: how long is android taking to build?13:42
Chocantofginther: Hey ! :) Do you know how I can get a rebuild for the poppler-qml-plugin job in jenkin ?13:42
ogra_dunno, i never checked, shouldnt be long (1h perhaps or less)13:42
ogra_stgraber, the issue is more that you need to shovel 500MB source around just for a no change rebuild13:43
* ogra_ is tempted to leave the boot.img generation in livecd-rootfs 13:43
stgraberogra_: thankfully I've got gigabit on the machine I usually upload from13:43
ogra_yeah, still. feels unreasonably hard13:44
fgintherChocanto, I'm doing a test build now13:44
Chocantofginther: Ok, thank you :13:44
Chocanto* :)13:44
ogra_for me thats 30min download time and as much for the upload again13:44
kenvandinemardy, ah, i figured it out... give me a minute13:44
ogra_ah, wait, the android package isnt native13:45
ogra_so its a non issue13:45
ogra_so its all fine13:45
Chocantofginther: Yes, I'm reading the log13:49
fgintherChocanto, is that a missed dep?13:50
fgintherChocanto, ack13:50
mardykenvandine: but the tests are passing in Jenkins; doesn't it use pbuilder?13:50
ogra_Chocanto, note that touch images have recommends disabled, dont rely on them13:50
kenvandinemardy, i just pushed a fix... not sure why it's working in jenkins13:51
ogra_if there is something you need thats coming in through a recommends, either make the dep hard or seed it13:51
kenvandinethe problem is client/tests/system-settings isn't set to executable in the build root13:51
Chocantofginther: "/usr/lib/i386-linux-gnu/qt5/bin/qmake: Command not found" so I think yes13:51
mardykenvandine: looks like a bug in the pbuilder configuration, TBH13:52
Chocantoogra_: The poppler-qt5 package is in release since 2 or 3 days I think13:52
mardykenvandine: the file is executable in bzr13:52
mardykenvandine: OTOH, since the workaround is easy, I'll just merge it in13:53
kenvandineit is puzzling why it's working in jenkins13:53
ogra_Chocanto, well, doesnt matter, if there is a recommends and something relies on whatever that recommends ships you need to seed it in touch ... nothing would pull it in otherwise13:53
ogra_oh, sorry13:55
mardykenvandine: maybe pbuilder is passing some flag to "bzr branch" which disables applying the permissions?13:55
ogra_i thought that was an image test13:55
Chocantoogra_: Oh ok yes I understand what you are talking about ^^'13:55
ogra_just noting now thats a package build test13:55
Chocantoogra_: No, but it will be important to consider in the future :)13:56
ogra_but feel free to ignore me until then :)13:56
fgintherChocanto, the ci job is enabled, so if you create a merge proposal, it will be automatically tested13:57
Chocantofginther: Ok :) I had to ask you for this build because nothing was added13:57
awafaamhall119: ping14:06
mhall119awafaa: pong14:10
awafaamhall119: did you send anything for the santa clara thing i emailed you about?14:10
awafaamhall119: if so did you hear back?14:10
mhall119awafaa: you'll have to refresh my memory, which santa clara thing was that?14:13
awafaamhall119: techcon14:13
mhall119awafaa: ah, yes, jono was going to go to it, since he lives in the SF-bay area14:14
mhall119awafaa: did he not email you?14:14
awafaamhall119: ah, no he didnt14:15
mhall119oh, he was waiting on someone else to get back to him about it, I'll check with him today14:15
awafaamhall119: ok, excellent14:16
Chocantofginther: pbuilder that Jenkins use can access to /usr/lib/i386-linux-gnu right ,14:22
fgintherChocanto, why would it be i386? It's an amd64 build14:22
fgintherChocanto, is that path hard coded?14:23
Chocantofginther: Oh... that's why it don't work14:23
Chocantothe package qt5-qmake should install qmake in the amd64 dir but... it don't do it. Weird14:23
balloonsm-b-o, whenever your about, ping :-)14:32
m-b-oballoons: around in 2 hours would be fine! :)14:33
balloonsuntil then :-)14:33
smartboyhwHey balloons14:33
=== mpt_ is now known as mpt
=== alan_g is now known as alan_g|tea
=== dednick is now known as dednick|afk
=== alan_g|tea is now known as alan_g
Chocantofginther: Ok, it's working. But the plugin only work in saucy, can you remove the precise, quantal and raring build please ?15:01
fgintherChocanto, yes, can do15:02
Chocantofginther: It would be really great, thank you :)15:02
=== dednick|afk is now known as dednick
mhall119a new contest for Ubuntu Touch app developers!15:34
=== jhodapp is now known as jhodapp|lunch
nik90Lottery: go ahead and ask your question.16:01
nik90If anyone knows it then they will answer your question16:01
LotteryOk :) Well, I just installed Ubuntu Touch on my Google Nexus 416:01
LotteryBut I can't mount it on ubuntu16:02
Lotterylsusb detect the device as Google Inc device16:02
LotteryI tried "sudo mtp-detect" but still nothhing16:02
LotteryI got Unable to open ~/.mtpz-data for reading16:02
LotteryAnd then No raw devices found.16:03
ogra_there is no mtp support yet16:03
ogra_use adb push and pull to move files abck and forth16:03
=== jasoncwarner__ is now known as jasoncwarner
LotteryI'll try that, thanks :)16:03
ogra_or insrall openssh-server and use scp over the network16:03
LotteryThats an option! Thanks a lot16:04
dholbachcan anyone help review https://code.launchpad.net/~sergiusens/phablet-tools/flash_change/+merge/177927? (om26er, popey, plars, salveti, doanac are CCed on the MP)16:08
doanacdholbach: i'll look and test.16:09
LotteryI have another question, I have nos service using ubuntu touch. Is it a known bug for some carrier/device?16:15
=== greyback__ is now known as greyback|food
pmcgowanLottery, gsm networks are supported16:17
LotteryOk, so my antenna is probably broken16:17
ogra_pmcgowan, i think the /current image still have a few extra bugs wrt indicator16:18
ogra_(with GSM specifically)16:18
ogra_(its nearly a week old )16:19
mardykenvandine: this is a new project which needs your review :-) https://code.launchpad.net/~mardy/signon-apparmor-extension/implementation/+merge/17890116:23
alecucjwatson: ping. How can I make the click scope get at the primary .desktop for an app that was just installed?16:23
ogra_davmor2, there is a 07.1 image in production ... a test once it comes out would be appreciated (if it doesnt fail)16:27
davmor2ogra_: just give me a ping as long as it is in the next hour off to the lug meeting then :)16:27
balloonsm-b-o, time?16:28
ogra_yeah, should be coming out within the hours16:28
m-b-oballoons: yes! :)16:28
LotteryCan I perform update using "apt-get upgrade" using the shell?16:28
dholbachthanks doanac16:29
balloonsm-b-o, ok great. So time to look at weather again. out of the box, it didn't work for me.16:29
m-b-oballoons, why? what was the error?16:30
pmcgowanLottery, yes but if there is an incompatible change on the android side it may break16:31
LotteryOk, do you know why the update manager loops on update search?16:32
ogra_Lottery, if you spot something like libhybris in the update you rather want to flash a newer image ... else its safe16:32
gemacjwatson: hangout?16:32
ogra_Lottery, note that flashing preserves whats in /home/phablet/ so if you want to keep stuff around across re-flashes, store it there16:33
LotteryOk :)16:33
LotteryBut it is weird cause I flashed the latest rom16:34
balloonsm-b-o, I get 12 failures16:34
balloonsm-b-o, err sorry.. 3 failures :-)16:34
balloonsm-b-o, let me paste them for you I guess16:35
ogra_Lottery, you flashed whatever is the recent blessed one16:35
m-b-oballoons, exactly! :)16:35
m-b-oballoons, 've no errors at all, local and on the device16:36
balloonsm-b-o, http://paste.ubuntu.com/5959480/. I'm just now looking into what could be causing it16:36
balloonsI'd like to try on my device now as well16:36
LotteryIs there a way to check if my antenna is still working?16:36
m-b-oballoons: bzr merge16:36
m-b-oballoons, I guess you're missing the last commits16:37
balloonsm-b-o, let me try.. I'll feel silly. I'd guess your right16:37
cjwatsongema: sorry, totally distracted into writing talk16:38
* balloons runs again16:39
=== greyback|food is now known as greyback
balloonsm-b-o, bingo.. all good :-)16:41
m-b-oballoons: Ha! ;)16:41
m-b-oballoons, thanks for your patience...16:42
=== jhodapp|lunch is now known as jhodapp
balloonsok, so let me review the code and then we'll merge this :-)16:43
m-b-odid you run it on the device too?16:44
balloonsI have a manta I'm running it on yes16:44
dholbachsergiusens, so LP doesn't let folks create teams involving an underscore - could we do some tricks in phablet-flash to test for "-" and "_"?16:45
m-b-oballoons: so what do you think about https://bugs.launchpad.net/ubuntu-weather-app/+bug/1207315 ?16:46
ubot5Launchpad bug 1207315 in Ubuntu Weather App "Autopilot test failures when run on device" [Undecided,New]16:46
balloonsm-b-o, let's land this and close it :-)16:46
balloonspopey is running a nexus 4 btw16:47
sergiusensdholbach: sure16:48
balloonsso m-b-o a couple minor things I just did for the clock app that we might want to do in here to keep things working well.16:49
balloonsm-b-o, check out this line:         self.pointing_device.click_object(searchField)16:49
balloons        self.keyboard.type("London")16:49
balloonsanytime we enter text we could have a timing issue, so let's add an assert there16:49
balloonsself.assertThat(searchField.focus, Eventually(Equals(True)))16:49
balloonsthen start typing.. make sense m-b-o ?16:50
m-b-oballoons, what else?16:51
dholbachsergiusens, do you think something like http://paste.ubuntu.com/5959532/ would work?16:51
janimodiwic, does Ubuntu Touch rely on Android blobs or daemons for audio as well?16:51
dholbachsergiusens, it'd just be interesting for the "checkout" case and there just change the branch name but leave everything else in place - I could ask András if he wanted to test it16:51
balloonsm-b-o, in addition, add asserts before using any object. So just about you have         self.pointing_device.click_object(searchField). Before clicking it add an assert to verify you got it. self.assertThat(lambda: searchField, Eventually(NotEquals(None)))16:52
janimodiwic, I am trying to see why an OMAP4 device has no sound (only lots of error messages in syslog) and first I'd like to know whether the android ROM device/ dir needs to have all the ALSA and sound related bits that a running Android ROM would need on the same hw16:52
balloonsm-b-o, I know it seems a bit tedious, but for the most part your tests are doing that. but adding those asserts will ensure no weird timing bugs occur on the devices while running.. well, in theory anyway ;-)16:53
m-b-oballoons: no problem, let me add this asserts to the MP16:53
m-b-oballoons: anything else? Now is the time! :)16:54
balloonsm-b-o, perfect.. I think I need to write a post up on these.. I'll try and do that today. There's lots of little tips like that16:54
w-flodholbach, I've noticed 2 small issues with "phablet-flash community": 1) the instructions in your mail are missing "./" for the phablet-flash call. 2) unpacking the ubuntu rootfs completely fails on the vision because the rootfs is about 700mb (extracted) while /data has only 1gb of storage, so anything on /data must be erased before flashing the zip. Not sure if I should modify the zip to do rm -rf /data before unpacking the tar.gz or16:54
w-flo if phablet-flash could take care of that if there's not enough free space?16:54
m-b-oballoons: okay16:54
balloonsm-b-o, off the top of my head, no, nothing else stands out. Add those, and let's merge16:54
dholbachw-flo, I'll redirect 2) to sergiusens16:54
sergiusensdholbach: check call raises an exception16:54
w-flodholbach, ah, thanks :)16:55
dholbachthanks w-flo :)16:55
dholbachsergiusens, all rightie - looking into it16:55
m-b-oballoons: okay, give me some minutes. I'll ping you!16:55
balloonsm-b-o, I didn't see you using any strings, but that's another thing to make sure of. Use objectName, not english strings as that will cause it to fail when run in german. But you already know that :-)16:55
sergiusensdholbach: if launchpad doesn't support undersocres I say we replace it regardless of a first check16:55
dholbachsergiusens, haha, yes - you're right16:56
m-b-oballoons, yes ran already into this...16:56
sergiusensdholbach: just do branch = branch_template.format(device.replace("_", "-")) on function enter16:56
Hashcodedholbach: I saw the updates, I'll try and build some tests here in a bit.16:58
dholbachsergiusens, although that'll replace the team name and the branch name, hmmmm16:58
dholbachHashcode, awesome - if you could give some feedback on how phablet-flash worked, that'd be great16:58
dholbachimage updates are of course great too :)16:58
HashcodeWill do :)16:58
sergiusensdholbach: so what is the restriction, team or project?16:58
sergiusensw-flo: add a --wipe to the command16:59
dholbachthe new code in phablet-flash would assume that lp:~pollux-windy-image-dev/phablet-image-info/pollux-windy is correct16:59
dholbachwhile users would probably pass "pollux_windy" as --device17:00
dholbachnot sure - I mean we can just replace everything and document that we want "-" no matter what17:00
sergiusensdholbach: let me think/look a bit,17:01
w-flosergiusens, testing with --wipe now. I was surprised to see the flash process work with the custom 4ext recovery I'm using, great work17:01
dholbachsure sure17:01
=== alan_g is now known as alan_g|EOD
sergiusensdholbach: can you create the pollux_windy entry in phablet-image-info ?17:03
sergiusensdholbach: create it with the format allowed and I'll adapt17:03
sergiusensw-flo: thanks... but it's mostly extendedcommands :-)17:04
dholbachsergiusens, so shall we just ask everyone to replace "_" with "-"?17:04
dholbachthat might be the easiest17:04
sergiusensdholbach: sounds good to me .... are there device codenames with a '-' in them? I'm almost sure they won't overlap if there are17:05
dholbachI haven't seen any17:05
cjwatsonalecu: ~/.local/share/applications/$APPID.desktop, where $APPID is (currently) pkgname_appname_version17:05
cjwatson(I might drop version from that)17:05
sergiusensdholbach: let's make the _ -> - switch for those devices17:05
=== alesage is now known as alesage|lunch
cjwatsonalecu: lool is right that we should be using upstart to launch the app as soon as that's supported17:06
dholbachsergiusens, András is just asking on the list, so I'll let him know and let you know once the branch is there17:06
sergiusensdholbach: ah, I don't use email as realtime communication :-)17:06
dholbachsergiusens, it just happened at the same time ;-)17:06
dholbachor well17:07
alecucjwatson: the thing is that after installing, I only have the package name (like com.ubuntu.ubuntu-weather)17:08
alecucjwatson: where should I take the default appname and the installed version from ?17:08
cjwatsonthe appname is in the manifest, first level below "hooks"17:09
cjwatson(package may have >1 app)17:10
=== alesage|lunch is now known as alesage
cjwatsonin fact you can get everything from "click list --manifest"17:10
cjwatsonbit cumbersome with lots of apps, I should probably improve that, but workable for now I think17:11
alecucjwatson: so, should I be spawning "click" to get at this? I've still got a bug to replace the calls to pkcon with dbus17:11
alecu*the spawning of pkcon, that is17:11
cjwatsonRight now yes - the plan is absolutely to expose this over the dbus packagekit api17:12
cjwatsonWait, maybe I already did17:12
* cjwatson 's past self is smarter than present self17:12
cjwatsonI hooked up GetPackages, but it doesn't expose the app name17:12
cjwatsonSo you're probably still best off calling click for now until I have that hooked up somewhere17:13
* cjwatson files a bug for himself17:13
cjwatsonbug 120932917:14
ubot5bug 1209329 in click (Ubuntu) "expose more information from "click list --manifest" in PK API" [High,Triaged] https://launchpad.net/bugs/120932917:14
slangasekhmm, what's this ubuntu-touch image build failure today? panda weirdness again?17:14
dholbachsergiusens, want me to file a bug for the 'additional_zip' thing?17:15
sergiusensdholbach: yes please17:16
sergiusensslangasek: ogra_ was experimenting/enabling the deb package of the android build, might be that17:16
slangasekah, could be17:17
cjwatsonslangasek: Yep, looks like another untested livecd-rootfs change17:18
cjwatson+       echo "deb ${MIRROR} saucy main universe multiverse restricted" >$sourceslist17:18
dholbachsergiusens, bug 120933017:18
ubot5bug 1209330 in Phablet Tools "Add additional_zip entry for community ports manifest definition" [Undecided,New] https://launchpad.net/bugs/120933017:18
slangasekogra_: ^^ are you working on this already?17:18
ogra_slangasek, yeah, its my fault17:19
ogra_i was trying to get $MIRROR handed through but at that point its not in the environment anymore17:20
cjwatsonogra_: I think that should be $LB_PARENT_MIRROR_CHROOT17:20
ogra_so the sources.list doesnt have a server17:20
sergiusensdholbach: should I do that in the mega branch or a new one?17:20
cjwatsonogra_: But could you please test it before the next upload?17:20
ogra_cjwatson, oh thanks ! i just rilled back to hardcoded ... let me  fix :)17:20
dholbachsergiusens, as you like it17:20
ogra_cjwatson, i did like 50 builds locally today17:20
ogra_cjwatson, but i have to export MIRROR to have it use my internal mirror ... so indeed it worked here17:21
cjwatsonAh, I see17:21
diwicjanimo, at least today, we rely on the audioflinger and the audio HAL for audio on the Touch images.17:23
diwicjanimo, we're trying to get rid of at least the audioflinger daemon, but that change has not landed yet.17:24
janimodiwic, do current images use Pulse only to talk to audioflinger or is ALSA involved?17:24
diwicjanimo, ALSA is not on the kernel side of things.17:25
dholbachsergiusens, you should be able to try --device pollux_windy (and pollux-windy) now17:25
janimodiwic, not on the kernel side? Is ALSA not used by Android so in the device kernels?17:26
diwicjanimo, sorry, you're right. It's up to the audio HAL implementation how they want to speak to the kernel, but ALSA is a common method.17:26
janimodiwic, I get SDP4430 related errors in dmesg, similar to pandaboard issues two years ago. I am not sure the problems are the same or not, but those issues seem to have been at least partly fixed by UCM profiles17:26
diwicjanimo, what errors btw?17:27
janimodiwic, [15160, pulseaudio] [10541.492706]  SDP4430 Media: dsp: no backend DAIs enabled for SDP4430 Media17:27
janimoand [15160, pulseaudio] [10541.485290]  SDP4430 Media: asoc: SDP4430 Media no valid capture route from source to sink17:27
janimodiwic, and in syslog  pulseaudio[3815]: [pulseaudio] module-alsa-card.c: Failed to find a working profile.17:28
davmor2ogra_: how we looking?17:28
janimodiwic, this is a device using a SoC similar to maguro (GNex) but using 4.0 Android blobs17:28
ogra_davmor2, build failed17:28
ogra_davmor2, so dont worry, i'll test myself17:29
ogra_once i can trigger the next17:29
diwicjanimo, out of curiousity, what device is it?17:29
janimoI don't know whether our audioflinger compat layer cares or knows about ICS vs JB Android , I am told Cyanogenmod abstracts those away but I don't know any detals17:29
janimodiwic, a Huawei Ascend P117:29
ogra_diwic, FYI i see the same on my galaxy S217:29
ogra_(the same janimo sees)17:29
dholbachw-flo, does --wipe make anything better?17:29
dholbachsergiusens, will the jenkins bot automerge https://code.launchpad.net/~sergiusens/phablet-tools/flash_change/+merge/177927 now?17:30
janimodiwic, these errors generate tens of MiB of logs per hour so they fill up the data partition in a couple of days :)17:30
ogra_diwic, all apps that use sound (camera, media player) dont start at all as long as pulse is running ... if i make pulse unexecutable the device works fine and these apps start17:30
diwicjanimo, ogra_ ok, let me see if I can find those messages in the kernel code to see what causes them to appear17:30
ogra_diwic, and my logs show the same messages janimo reports17:30
diwicogra_, oh, that's an interesting finding. Maybe they're both trying to use the device?17:30
janimodiwic, these are the two bugs that come up on google most frequently related to these messages https://bugs.launchpad.net/linaro-ubuntu/+bug/816638 and https://bugs.launchpad.net/ubuntu/+source/alsa-lib/+bug/74602317:31
ubot5Launchpad bug 816638 in Linaro Ubuntu "Pulseaudio consumes 100% of the cpu when trying to play a sound with natty's linaro LEB and 3.0.0-1402-linaro-lt-omap " [High,Fix released]17:31
ogra_janimo, did you try a chmod -x /usr/bin/pulseaudio and reboot yet ?17:31
ubot5Launchpad bug 746023 in alsa-utils (Ubuntu Natty) "No sound on omap4" [High,Fix released]17:31
dholbachsergiusens, for me the fix seems to work for the pollux-windy thing17:31
ogra_janimo, that makes everything work for me17:31
w-flodholbach, it's definitely better, after flashing /data is now 600mb instead of 300mb, so that's good. It still doesn't work though, still trying to figure out why (manually wiping data, then flashing device.zip and ubuntu.zip manually  made it work.. x.x)17:31
ogra_i can even play the movies17:31
diwicogra_, in /etc/pulse/default.pa, do you have the line "load-module module-udev-detect" (possibly followed by parameters)? If so, try commenting that line out and reboot17:31
dholbachsergiusens, are you going to push it or shall I create a MP to be merged into your branch?17:31
sergiusensdholbach: great, I'll apply17:31
dholbachthanks sergiusens17:31
dholbachsergiusens, I'll inform András17:31
w-flodholbach, "still doesn't work" => device won't boot (not even adb comes up)17:32
diwicand if that works for ogra_ it might work for janimo too :-)17:32
dholbachcan somebody help w-flo? he's working on the HTC Desire Z port?17:32
ogra_rebooting, lets see17:33
sergiusensw-flo: can you get the /cache/recovery/log and /cache/recovery/last_log files after deploy?17:33
w-flosergiusens, sure!17:34
janimodiwic, btw I have no modules loaded at all. I wonder if my kernel is configured correctly (same as Android's plus our UTouch options only)17:34
diwicjanimo, kernel modules or pulseaudio modules?17:35
diwicjanimo, if it is kernel modules, I'm not the right person to answer17:35
janimodiwic, kerne modules. Ah I see, the file you pointed to deals with PA modules17:35
ogra_diwic, well, my phone comes up with a load of 917:36
ogra_pulse seems to die and start over and over17:36
seb128is qtsystems5-examples's /usr/lib/i386-linux-gnu/qt5/examples/sysinfo/qml-battery/qml-battery.qml giving working battery infos for anyone?17:36
janimoogra_, that is what I see too even before this change17:36
seb128on my n7 it gives -1 for all the battery settings17:36
janimoogra_, I have 3 pulse processes with close PIDs17:36
ogra_janimo, well, making pulse -x helps here17:36
janimothat keep getting renewed17:37
ogra_same here17:37
janimoogra_, well that is not a proper fix though :)17:37
ogra_with the change17:37
janimoalthough 100% compatible with the way of fixing sound on desktop a lot of time :/17:37
ogra_janimo, no, but it shows that sound generally works17:37
ogra_and you can test camera and mediaplayer :)17:37
diwicogra_, janimo, okay, https://wiki.ubuntu.com/PulseAudio/Log if you want me to investigate further17:37
janimoI think camera will not really work here, I remember when porting UTouch preview, I had troubles with the camera module17:38
sergiusensdholbach: revno 178 allows ./phablet-flash community --device pollux_windy17:38
* dholbach hugs sergiusens17:38
m-b-oballoons: there you go!17:38
ogra_janimo, ah, sad, works fine for me once sound started working17:38
balloonsm-b-o, yay!17:39
janimodiwic, some progress after commenting out that udev-detect line17:40
janimodiwic, dmesg is now quiet17:40
janimoaplay still crashes17:40
dholbachsergiusens, documenting on the wiki17:40
janimosyslog is much quieter17:40
ogra_for me it got worse after commenting it17:40
ogra_before pulse just hung and didnt madly restart in a loop17:40
m-b-oballoons: http://paste.ubuntu.com/5959705/ this occassionally pops up17:41
nik90balloons: I am unable to confirm https://bugs.launchpad.net/ubuntu-clock-app/+bug/120902417:42
ubot5Launchpad bug 1209024 in Ubuntu Clock App "Toolbar prevents text preset from accepting input" [Undecided,New]17:42
janimodiwic, here's the new syslog, now I'll do what the Log wikipage says http://paste.ubuntu.com/5959707/17:42
balloonsnik90, hmm17:42
balloonsnik90, m-b-o since your both here; http://www.theorangenotebook.com/2013/08/autopilot-best-practices.html17:42
nik90balloons: nvr mind, I can confirm the bug. But I swear this did not happen before updating to new sdk version17:44
diwicjanimo, "socket-server.c: bind(): Address already in use" that sounds bad, have you installed the pulseaudio-esd-compat package?17:44
janimodiwic, no I have not installed any sound package17:45
janimodiwic, I have 3 pulse processes that keep respawning17:45
balloonsnik90, it is interesting.. we can start a conversation about if it should be that way or not17:45
janimoeven with autospawn=no17:45
balloonsthe autopilot tests work around it.. and perhaps the toolbar shouldn't autoclose17:45
diwicjanimo, ah, it's the native protocol. Can you check who is using the files in /run/user/<nr>/pulse17:46
nik90balloons: I think by design when you press add preset, it should show the new page and hide the toolbar automatically. So a user can click the textfield once and start typing stuff.17:46
diwicjanimo, are all pulse processes the same user?17:47
balloonsnik90, yes in that case.. but should the toolbar autoclose for you. or should you close it/17:47
nik90balloons: however the toolbar behavior has been changed literally 3-4 times, hence the bug. I will fix it17:47
balloonsmake sense? I'm not sure, but it seemed the sdk used to autoclose it17:47
nik90balloons: for me as of now, I have to manually close the toolbar17:47
balloonsnik90, right me too.. but I believe in the past it wasn't the case.. hence the bug popping up. but the autopilot test found it, so :-)17:48
nik90balloons: true. Fixing it should be simple17:48
w-flosergiusens, http://pastebin.ubuntu.com/5959716/  -- I've put a "NOTE" in there. It seems like wiping /data didn't work17:48
janimodiwic, all 3 are of the phablet user17:50
w-flosergiusens, maybe /data is not mounted at the time it's supposed to be wiped?17:50
ogra_bah, i typoed when commenting default.pa17:51
* ogra_ starts over again17:51
ogra_diwic, ok, the default.pa change works for me, pulse doesnt freak out, medaiplayer and camer still work, but i dont seem to have volume control and the indicator is empty as well17:53
balloonsm-b-o, good stuff.. I hope this runs like a champ in the lab now!17:53
diwicogra_, is the sound indicator working on the non-pulseaudio images? If so I should see if I get it working on the pulseaudio image too17:54
ogra_diwic, it works on maguro for me17:54
sergiusensw-flo: maybe, but I assume /data is always mounted unless your specific recovery doesn't do that17:54
diwicogra_, all right17:54
w-flosergiusens, http://pastebin.ubuntu.com/5959731/ <-- it's not mounted in my recovery :(17:54
ogra_diwic, it doesnt on grouper iirc, let me check17:54
diwicogra_, I should then test that too17:54
ogra_yeah, grouper has no controls17:55
sergiusensw-flo: I see you have a custom edify, can you add the logic there?17:55
w-flosergiusens, yes17:55
sergiusensw-flo: I'll add a check mount before wiping which should help too17:56
nik90balloons: you said that there were 8 abstractbutton (tabs)?17:57
w-flosergiusens, thanks. :-)17:57
balloonsnik90, yea, we're not finished with the clock app ;) I was speaking with elopio and the sdk folks.. elopio made further changes to tabs to fix things up. We need to fix some more of the bad practicies in clock tests and move to the emulator17:58
balloonsnik90, interally the tabs are duplicated, you can see it in vis..17:58
nik90balloons: yeah I am facing a wierd issue in the clock app. I want to know if this is a upstream fault?17:59
nik90balloons: In the clock tab, if you try adding a new world city you will notice that the tabs are visible despite being inside a pagestack17:59
balloonsnik90, the duplication is intended, but I'm not sure why. and yes we spoke about visible.. visible lies to you17:59
balloonsor so it feels17:59
nik90balloons: let me grab a screenshot to show what i mean18:00
janimodiwic, hmm pulse went away eventually. I had to log out and in from the shell session. I don't think I did anything else to trigger it18:01
diwicjanimo, it's some interesting race condition for sure, they seem to be in different phases of startup.18:02
nik90balloons: http://imgur.com/0ZaVKw718:02
nik90balloons: do you see the tab header colliding with the page header (add city and clock)18:03
sergiusensbzoltan: jppiiroi1en where was the Qt branch that used phablet-flash?18:03
nik90balloons: This is also experienced by the rss reader folks. I talked to timp about this. I have to reproduce this using the examples given in the api docs. But this is only reproducable here.18:04
janimodiwic, these are the pulse logs, it segfaults very quickly though18:04
janimodiwic, http://paste.ubuntu.com/5959765/18:04
balloonsnik90, reading and looking18:05
diwicjanimo, so it looks like it has problems talking to audioflinger or sth18:05
diwicjanimo, because that's where it fails18:05
sergiusensw-flo: can you try revno 179 ?18:06
janimodiwic, so it is likely I need to have all the device/<vendor>/<device>/audio things in place in the ROM build18:06
janimodiwic, my build is based on phablet ROM and blobs from the vendor but the device specifics are mostly from scratch so it is likely I have no audio specific code there. I'll try copying what maguro does as it is the same SoC then18:07
janimodiwic, thanks18:07
w-flosergiusens, looks good, now waiting to see if it reboots into ubuntu (will take 5-10 mins)18:08
w-flodata is now mounted and almost empty :)18:08
balloonsnik90, ok, yes I saw that happen in rss reader too, your right. So timp wants an example using the api? it is perhaps something misconfigured in the apps themselves?18:08
balloonsnik90, the sdk has certainly changed, perhaps this is a bug because of how your tabs are laid out vs other apps18:09
nik90balloons: timp wanted me to validate this using the api examples, but I wasn't able to.18:09
nik90balloons: that said, the structure of the example and the clock app is very similar18:09
diwicjanimo, /system/lib/hw/audio*.so, and possible dependencies of that18:09
nik90balloons: which puts me in a pickle since I cannot figure out why this is then happening in the clock app.18:10
balloonsnik90, I'm sure it is :-) so it's a bit of a mystery then why it occurs18:10
diwicjanimo, e g on the Nexus 4 the audio lib loads completely different so:s too, e g libqmi_cci.so and other fun stuff18:10
nik90balloons: the annoying part is that I am in work in the morning, so very hard to coordinate with sdk devs18:11
janimodiwic, indeed, devices have their own alsa wrappers and helpers from what I superficially noticed18:11
nik90balloons: the best thing would be to ask timp to look at the ubuntu-clock-app.qml file. It is the main file and is not much. He should be able to tell why this is happening18:11
nik90balloons: since he is the one behind all the tabs implementation18:11
balloonsnik90, do you have a bug in clock app for this? let's coordinate via it and ask him to look at it18:13
balloonsis it holding you up?18:13
nik90balloons: I will create one with all the info required18:13
nik90balloons: it isn't holding me, but it is worrying :(18:14
balloonsok.. ;-)18:14
w-flosergiusens, r179 works for me :)18:16
sergiusensw-flo: great!18:17
sergiusensogra_: I'm most likely going to look into supporting openrecoveryscript as well, for TWRP18:21
ogra_as long as it doesnt interfere with system images or android rollback :)18:21
sergiusensogra_: it shouldn't18:22
ogra_yeah, i would think so too ....18:22
sergiusensogra_: experience tells you different :-P18:22
ogra_we'll see18:23
ogra_i was 100% sure the livecd-rootfs change wouldnt break as well :)18:23
ogra_did 100 tests ... :)18:23
ogra_broke anyway18:23
* ogra_ shakes his fist towards murphy 18:24
ogra_you and your law ... go away !18:24
ogra_<infinity> Alright, so we have 23 highbank nodes building armhf now.18:24
ogra_in case you didnt see it in #ubuntu-devel18:24
* ogra_ looks forward to 15min buildtime for libreoffice :)18:25
sergiusenscalxeda's finally made it to prod infra18:25
rickspencer3ogra_, first things first, I am using those builders to build photobomb for 11.04 ARM Desktop18:25
rickspencer3I'll need every node for at least 3 days18:25
ogra_stop the line !18:25
ogra_rick is building a package18:25
rickspencer3not *a* package18:26
rickspencer3*the package18:26
ogra_ported to touch already ?18:28
ogra_else it doesnt count18:28
rickspencer3fair enough18:28
rickspencer3I withdraw my demands :)18:28
rickspencer3ogra_, in all seriousness, this could be a sweet thing for Touch (and other ARM) development18:28
ogra_it will !!18:29
rickspencer3thanks to infinity and everyone for running this down18:29
ogra_we have been waiting for this  for like 2 years now18:29
sarnoldkgunn: I'm curious about vnc support in Mir; one of the features of Mir we're really looking forward to having is preventing windows from snooping on mouse, keyboard, and video for other windows, but that's more or less the whole point of vnc and screen recorders and some automatic qa systems (for poking buttons, reading feedback..)18:30
sarnoldkgunn: is there some documentation I could read about this?18:30
Chocantomhall119: Hey :) The poppler-qml-plugin package is ready, do I have to ask to make available in the core apps PPA ?18:30
kgunnsarnold: do you mean vm or vnc ?18:31
=== schwuk is now known as schwuk_away
sarnoldkgunn: well.. I did mean vnc, but now I'm curious, what you mean about 'vm' :)18:32
mhall119Chocanto: did you make the package or is it from the upstream release?18:33
Chocantomhall119: I made it18:33
Chocantomhall119: And it's working with jenkins18:34
mhall119Chocanto: I can put it in the Core Apps PPA for now, but eventually we'll want it in the archives18:34
cjwatsonrickspencer3: When we ran the numbers at the release sprint, this was by far the biggest single potential improvement to pipeline speed18:34
mhall119Chocanto: where can I get the package source?18:34
cjwatson(source upload to image build)18:34
rickspencer3cjwatson, \o/18:35
cjwatsonSo I have high hopes.  We'll see18:35
=== chihchun is now known as chihchun_afk
Chocantomhall119: https://code.launchpad.net/~ubuntu-docviewer-dev/ubuntu-docviewer-app/poppler-qml-plugin if that's what you are asking for*18:35
mhall119Chocanto: it is :)18:35
mhall119Kaleo: bzoltan: Mirv: ^^ can you guys look at including that in the SDK soonish?18:36
kgunnsarnold: sorry...distractions...yeah, after a re-read i see what you mean18:37
sarnoldkgunn: hehe, yeah, I'm familiar with distractions :)18:38
kgunntvoss_: thots on vnc applications with mir ?...^ ..... i suppose at least with Xmir that'll still be possible, but for unity8...you'd almost have to have a "special" app18:38
kgunnwith server side accees18:39
tvoss_kgunn, right, it will have to integrate with the compositor, with the benefit that both screenrecording and vnc-like apps can leverage the same api/hook inside the server18:40
sarnoldtvoss_: will someone be able to run a vnc server and a screen recorder simultaneously?18:40
tvoss_sarnold, well, nothing should prevent that given the architecture I have in mind. But it needs to be detailed18:41
tvoss_sarnold, any specific use-case you have in mind?18:41
sarnoldtvoss_: no, I'm trying to better understand how we'll still provide features users love while also closing off the worst of the X security problems :)18:42
mhall119Chocanto: ok, pushed the source package to the coreapps PPA, give it a bit to show up18:42
=== chihchun_afk is now known as chihchun
Chocantomhall119: Ok, thank you :)18:42
Chocantomhall119: So I'll wait before making this package a dep for the ubuntu-docviewer-app :)18:43
tvoss_sarnold, fair point :) so from my pov, vnc and screen recording apps need privileged access to mir. Essentially, they need to be able to see the final composited frame before it is sent off.18:44
sarnoldtvoss_: any chance you've got documentation / code that you can point me towards?18:46
tvoss_sarnold, at this point? no, sorry18:51
tvoss_ogra_, tried it the other day18:52
ogra_did it work ?18:53
tvoss_ogra_, at least what I was looking at: yes18:59
tvoss_ogra_, ah sorry, wrong abbreviation :)18:59
ogra_i was already wondering :)18:59
* ogra_ sighs ... so why does live-build not do on the buildd what it does here 19:05
rsalvetijdstrand: hey, were you able to test the ofono/telepathy-ofono changes?19:06
sergiusensmamenyaka: hello19:07
=== b0bben_ is now known as b0bben
mamenyakaboth pollux_windy and pollux-windy working. great job!19:08
mamenyakasergiusens, can you help me with phablet saucy?19:09
jdstrandrsalveti: hi! yes, I did. I wasn't able to reproduce. I installed your packages, rebooted and could make and receive calls19:10
jdstrandrsalveti: what are the steps to reproduce?19:10
ogra_cjwatson, http://paste.ubuntu.com/5959959/ thats what i get for a local live build with the same code change in lvecd-rootfs, do you have an idea why the links wouldnt be created on cadejo ?19:12
rsalvetijdstrand: hm, just install latest image with phablet-flash --pending, then both ofono and telepathy-ofono from http://people.canonical.com/~rsalveti/jdstrand/19:12
rsalvetiand try to make a call19:12
rsalvetijdstrand: make sure you're also using the ofono package as provided by this link19:12
jdstrandrsalveti: yeah-- so, I didn't phablet-flash-- I dist-upgraded, then installed your packages, then rebooted, then made the call19:12
jdstrandrsalveti: so I may have an old kernel. I can try phablet flash, but it will take a while19:13
mamenyakaogra_, hi, can you help me?19:13
rsalvetijdstrand: but should still work, in theory...19:13
jdstrandrsalveti: but you see this problem right now? I'm on mako19:14
rsalvetijdstrand: yeah, also tested with mako19:14
rsalvetiI just reproduced before giving you the packages with a clean image19:14
jdstrandrsalveti: not that it would make a difference, but I also installed ofono-scripts from your url19:14
jdstrandbecause it was also installed19:14
rsalvetiyeah, that's fine19:14
* jdstrand reflashes19:14
jdstrandrsalveti: oh, what does --pending do?19:15
mamenyakajdstrand, grabs latest pending ubuntu image, I guess19:15
sergiusensjdstrand: pending grabs the non qa'd build19:15
sergiusensjdstrand: so the not 'current' link on cdimage19:16
rsalvetiyeah, latest of latest19:16
jdstrandI see19:16
rsalveticurrent is more than one week old I guess already19:16
* jdstrand flashes19:16
sergiusensmamenyaka: missed your ping, depending on the question I can help, but have not been really involved in flipping19:16
jdstrandrsalveti: I'll get back to you after testing19:16
mamenyakasergiusens,  well, than I guess I just ask my other question: is it normal that on tablet, apps that do not display on fullscreen (calculator) are black, but with fullscreen app opened (browser) they render fine?19:18
rsalvetijdstrand: cool, thanks19:18
sergiusensmamenyaka: that would be a bug19:21
mamenyakasergiusens, in the aug. 1 irc chat I provided some screenshots here, can you somehow retrieve them?19:21
mamenyakaor is the bug known?19:22
sergiusensmamenyaka: moer of a Saviq question if it's known or not19:24
mamenyakasergiusens, who can help me with flipping? samsung note 10.119:27
sergiusensmamenyaka: where are you stuck?19:27
mamenyakasergiusens, I posted on the mailing list too, device boots, black screen19:28
mamenyakaadb and ubuntu shell working19:28
sergiusensmamenyaka: did you add the required udev rules19:28
mamenyakalong time ago19:28
sergiusensmamenyaka: any logcat errors?19:29
mamenyaka[1] http://paste.ubuntu.com/5941329 [2] http://paste.ubuntu.com/5941366 [3] http://paste.ubuntu.com/594168719:29
mamenyakasergiusens, dmesg, logcat, some weird SN19:30
=== chihchun is now known as chihchun_afk
mamenyakasergiusens, I think they are older ones, sorry19:34
sergiusensmamenyaka: well do the hybris tests work?19:34
mamenyakasergiusens, I remembered completely wrong, device state is: stuck at bootscreen, no bootloop; [1] http://paste.ubuntu.com/5948192/ [2] http://paste.ubuntu.com/5948193/19:37
mamenyakasergiusens, what are the hybris tests?19:38
sergiusensmamenyaka: type test- and TAB in adb shell :-)19:39
sergiusensmamenyaka: sorry -> test_19:40
mamenyakasergiusens, okay, flashing in a moment to try out19:40
mamenyakasergiusens, can it be that the tests fail on a working flipped device? (should I test on every device?)19:40
mamenyakasergiusens, okay, again wrong, device boots, black screen, SN error no more19:45
mamenyakasergiusens, http://paste.ubuntu.com/5960094/  http://paste.ubuntu.com/5960096/19:46
sergiusensmamenyaka: I'll need to look into this later tonight, I'm on the hook for getting something done right now19:46
mamenyakawhich test should I run?19:47
=== chihchun_afk is now known as chihchun
sergiusensmamenyaka: this seems to be your issue W/libEGL  ( 2272): eglInitialize(0x1) failed (EGL_BAD_ALLOC)19:47
sergiusensmamenyaka: the gl ones19:47
mamenyakasergiusens, yep, I have seen that19:47
mamenyakasergiusens, can you tell me about that error?19:47
=== chihchun is now known as chihchun_afk
mamenyakasergiusens, the test give this output: __pthread_gettid -219:48
mamenyakasergiusens, test_glesv2 displays some animation19:49
sergiusensmamenyaka: then gles works19:50
mamenyakasergiusens, great :)19:50
* sergiusens goes away to concentrate19:50
mamenyakasergiusens, thank you for your help!19:50
dmj_novamfisch: nice job on the AMA20:06
nixkidhey guys, can I install UT on acer-e350?20:11
thomisergiusens: can you join us in a hangout call please?20:13
sergiusensthomi: when?20:17
thomisergiusens: 10-15 minutes time?20:19
sergiusensthomi: ok20:19
thomisergiusens: just waiting for plars to  have a spare moment20:19
thomisergiusens: thanks20:19
rsalveti_jdstrand: able to try the packages with latest image?20:41
jdstrandrsalveti_: still downloading the image (sorry, slow link)20:41
rsalveti_jdstrand: np20:42
thomisergiusens: it looks like plars is held up.20:46
thomisergiusens: so I may as well type to you now20:46
thomisergiusens: the plan with click packages & autopilot tests (encode bzr revno in click packages and grab tests from bzr branches) simply will not work.20:46
plarsthomi, sergiusens: sorry, don't wait for me20:47
thomisergiusens: I keep saying this to people, and it seems no one cares, but I haven't had a chance to talk to you yet20:47
thomisergiusens: the problem is that test suites have dependencies that won't be in the image20:47
thomisergiusens: for example, test suites will depend on each other, to get the emulators packaged there20:47
sergiusensthomi: so what is your solution?20:48
thomisergiusens: some concrete examples: anyone who wants to use the OSK as an autopilot Keyboard backend will need to depend on the maliit-autopilot package.20:48
thomisergiusens: we talked about this in the IOM sprint.20:48
thomisergiusens: we keep producing .deb packages for test suites20:48
sergiusensthomi: how do you install those?20:48
thomisergiusens: and we install them as usual (after remounting the FS rw20:48
thomiso.. remount,rw, install packages, remount,ro... and go!20:49
sergiusensthomi: that's a no go20:49
thomisergiusens: why?20:49
=== interloper is now known as drsuess
sergiusensthomi: we don't want to alter the system and we also want to provide a way for people to test without entering developer mode which would break image based upgrades20:49
thomisergiusens: according to slangasek, installing those test packages shouldn't break image based updates20:50
thomisergiusens: it boils down to the fact that you need a mechanism for dependency resolution20:50
thomisergiusens: as it happens, apt-get works quite nicely for this20:50
thomisergiusens: I don't understand why we're so keen to throw that away20:50
slangasekwe don't *want* to throw it away, but there are limits to what we can do with apt-get+system-based updates20:51
slangasekthanks to The Overlayfs Problem20:51
sergiusensthomi: I'm imagining a more android like approach where I either bundle this in one place or slap it in the image20:51
thomisergiusens: another thing you could do is use an overlayfs, and install the debs on that20:51
slangasekhowever, for the case of needing to install packages for testing purposes, I think "mount rw, apt-get install, remount ro" is sensible and low risk wrt breaking further system image upgrades20:52
slangasekthomi: there's no overlayfs that's supported across our BSP kernels, AFAIK20:52
thomislangasek: ahh, ok20:52
thomiit's not like we even care that much about breaking image based updates on the test phones anyway20:52
sergiusensthomi: I'm not thinking of only the test phones20:53
sergiusensI'm thinking about every other app only developer20:53
thomisergiusens: anyway, apt-get may not be the perfect solution, but I can tell you right now that the current proposal won't work at all20:53
sergiusensthomi: it will if I do it right20:53
sergiusensthomi: I'll tell you this, I'll own click packages in the image and you can own the testing strategy20:54
sergiusensI'm only offering a helping hand after all20:54
thomisergiusens: Wwe're happy for the help! It was a bit of a surprise to me to hear that you're implementing something contra to the plans we made in the IoM, that's all.20:55
thomisergiusens: if we're going to do this, we should make sure that everyone involved at least hears about it before the code is written20:56
sergiusensthomi: where was that plan communicated? I talked with lool and slangasek yesterday and this wasn't mentioned at all20:56
thomisergiusens: I figured you'd rather hear about these problems *before* you wrote the code, rather than after ;)20:56
sergiusensthomi: I know about the dep problem20:56
thomisergiusens: so how will you adress that?20:56
sergiusensthomi: the solution today needs to be migrated away from anyways20:56
thomisergiusens: I agree that we cannot continue doing things exactly as they are today. It's what we're moving *to* that concerns me20:57
sergiusensthomi: I plan to solve it like android, base testing requirements are part of the image, so any aplication developer can benefit from it20:57
sergiusenseither that or...20:58
thomisergiusens: so that means you'll need many of the autopilot test suite packages as part of the base image. Is that really what you want to do?20:58
bzoltan mhall119:  with pleasure.... please ask the dev to cover it with API specs, docs, examples and tests.20:58
sergiusensthomi: no, just the emulators20:58
thomisergiusens: OK, so you'd package all emulators in the image, but not the test code?20:58
sergiusensthomi: that's my first thought, yes20:59
thomisergiusens: OK, I think you'll need to find a clever way to make sure that you can split the python package like that21:00
thomisergiusens: I mean, what's to stop some code in 'unity8.emulators' from depending on code in 'unity8.tests'?21:00
sergiusensthomi: whatever choice we make will find problems, I just want to make it once.21:00
sergiusensthomi: I can also not put them on the image, grab the bzr branch and pull in everything we'd call base emulator into a something we'd push to pythonpath21:01
thomisergiusens: that might work as well, so you'd need to keep a list of 'base' emulators, and maintain that list somewhere21:02
sergiusensthomi: that's the easiest part ;-)21:03
thomisergiusens: right, this is now starting to sound like it might work21:04
thomisergiusens: one more thing I can see might be an issue: are you assuming that *all* apps will be hosted on launchpad?21:04
thomiand is that a valid assumption?21:04
* ogra_ fires off a new build and crosses fingers ...21:05
ogra_(this is starting to get annoying)21:05
sergiusensthomi: not, not all apps will be hosted on launchpad... for those that are not, we'd need hooks into qtc21:06
sergiusensthomi: well those two options were heavily discussed, and I was well aware of the emulator dep before hand21:06
thomisergiusens: OK, when I spoke to plars, I got the sense that this hadn't been discussed21:07
thomisergiusens: it'd be great if you could keep me in the loop on this please!21:07
sergiusensthomi: are you subscribed to the blueprint?21:07
thomisergiusens: which one?21:07
sergiusensthomi: https://blueprints.launchpad.net/ubuntu/+spec/qa-s-image-based-updates21:08
sergiusensthomi: that's where I'll add everything... I'm focusing on upstream merger first though21:08
=== drsuess is now known as interloper
thomisergiusens: OK, I'm subscribed now21:12
heynoqDoing a port right now and am currently building, though I skipped some steps/modified and was wondering if anyone was there that could answer some (hopfully easy) questions :)21:13
=== claydoh_ is now known as claydoh
w-floheynoq, maybe I can.21:22
w-floheynoq, *maybe*. don't thank me yet, I'm having some difficulties with porting as well :D21:23
heynoqNo problem21:23
heynoqSo i am following the basic guide...21:24
w-flobetter use this: https://wiki.ubuntu.com/Touch/PortingFlippedInProgress21:24
heynoqMy phone is officially supported by cyanogen, so that is a plus21:24
heynoqOk well...21:25
w-floheynoq, maybe the old guide is still okay, but if you want a "flipped" port (which is the future), you should use the new guide. but someone really needs to rewrite most parts, so.. yeah21:25
heynoqwell, see how this goes...if it fails, i'll try the other guide21:25
heynoqIn the step Retrieving the proprietary blobs from Android21:25
heynoqAfter running the script, it tells me to make sure to download the qualcom blobs, jf, and the google ones21:26
heynoqit gives me URLs for all but the "JF" stuff21:26
heynoqdon't know what that jf means, but dont care at this point21:26
heynoqso my question21:26
heynoqthey said to put the blobs in a specific folder21:26
w-flohm, never heard of that JF thingie.. luckily my inofficial port has all the blobs in a git repository21:26
heynoqyeah, lucky ;)21:26
heynoqso the google blobs that the url took me to had a zip21:27
heynoqi looked in zip and it was a .sh file21:27
heynoqi put that in the directory they indicated21:27
heynoqthen for the qualcom crap, the zip had a folder and in that folder there were system, app folders21:27
heynoqi put the main folder in the root of where they indicated21:27
heynoqbut i feel that was wrong21:27
heynoqlet me get more details..., one sec21:28
w-flohm, it doesn't feel right21:28
w-floI think usually the blobs are .so files or binaries21:28
w-floor some config file21:28
heynoqAdreno200..etc folder in /phab/device/motorola/msm8960-common  Also google crap in that folder under extract-qcom-mako.sh21:28
w-flothe extract-qcom-mako.sh sounds like a script you are supposed to run in order to extract more files from your phone (like the extract-files.sh ?)21:29
w-flooh, and mako is the nexus 421:29
w-floare you sure this is related to your device?21:29
heynoqin the extract-files.sh, it noted to get the mako, let me provide more details on that google thing, one sec21:30
w-flomaybe you could try building the android parts and see if it works. If it compiles and then boots, hopefully logs will indicate if there are any missing files...21:30
tedgricmm, Uhm, so ubuntu-touch-session.  It installs a bunch of stuff into /etc/init...21:30
heynoqYeah, i suppose...there is a build that already works for my phone, but it is missing the stuff to make the radio work...bs qualcom crap, so I am trying to include all that stuff, which i think is one thing i did wrong21:31
heynoqbut, to your point, that is what i am doing now, if i can get it to build and boot21:31
heynoqthen I am very close to getting the other stuff working...21:31
w-floah, cool.21:31
heynoqSo this "blob" i need to add to that directory21:32
heynoqits a zip containing21:32
w-flowhat's the device code name? I'd like to take a look at the cyanogen extract-files script21:32
heynoqYeah, at the end of the extract-files script, you will see the comments about these other blobs you need...that is basically what I'm not sure about if I put them in the correct folders...21:34
ricmmtedg: and?21:34
tedgricmm, And it changes the XDG_RUNTIME directory... which causes basically everything to be angry.21:35
w-floheynoq, yes.. it's not very clear what to do with those files. :/ you probably put them in vendor/motorola/msm8960-common or something?21:35
tedgricmm, It really should be installing things in /usr/share/upstart/sessions21:35
heynoqw-flo, yes its clear, but I am cloudy :)...   So do I put the ZIP files in that folder, or extract them and put the extracted in that folder?21:36
tedgricmm, As an example my gnome-settings-daemon log file was 250MB of errors.21:36
=== salem_ is now known as _salem
ricmmtedg: https://code.launchpad.net/~ricmm/session-manager-touch/migrate-to-upstart-session/+merge/172936/comments/38797121:38
w-floheynoq, to be honest, I have absolutely no idea :(21:38
ricmmalso, it installs the right stuff in the right places, the things you see in /etc/init/ are for the current non-lightdm way of firing the user session21:39
heynoqlol yeah, it sounds straight forward, but it seems like more clarity is needed...21:39
ricmmwhile the right way lands21:39
w-floricmm, you could check the .mk files for copying the vendor blobs into the target image and see which files they expect21:39
w-flosorry, heynoq ^^21:39
heynoqi ended up extracting them...if things dont work I can try again.21:40
heynoqNext question :)    In the section "Device changes" it talks about this fstab file21:40
heynoqThe default fstab for the new device will have its /data partition mounted with nosuid, which needs to be removed, as certain Ubuntu applications/daemons require the use of setuid. The path to fstab is: • device/[manufacturer]/[codename]/fstab.[codename] If you don’t find any files there, grep for “nosuid”.21:41
ricmmtedg: why are you installing gnome-settings-daemon?21:41
tedgricmm, okay, but it's in archive.  I installed it, others could as well.  It needs to conflict with something then.21:41
tedgricmm, I installed it on my desktop.21:41
heynoqI could not find that file and grep didn't give results21:41
timpyeah first enterprise bundle on igg :)21:41
heynoqI'm ok with skipping that step at this point, so I'll move on to next question that I think is the most important for now21:42
w-floheynoq, https://github.com/CyanogenMod/android_device_motorola_msm8960-common/blob/cm-10.2/setup-makefiles.sh that file probably has some/most/all (??) of the props you need21:42
heynoqhmm, ok, I'll look at that21:43
w-floand yeah, I can't find the fstab for your device either. Every device is like a totally different setup it seems..21:43
w-flohttps://github.com/CyanogenMod/android_device_motorola_msm8960-common/tree/cm-10.2/rootdir/etc <-- heynoq, your fstab file is probably in that directory21:44
sergiusenstimp: http://insights.ubuntu.com/news/press-releases/bloomberg-bets-big-on-the-ubuntu-edge/21:44
heynoqok,  i do have that setup-makefiles.sh file in my /device/motorola/msm8960-common/ folder21:45
rsalveti_w-flo: from a quick look, you should avoid using the cm-10.2 branches as well21:45
ricmmtedg: sorry, but it escapes me why you would try to install this on a desktop installation21:45
rsalveti_as that's in heavy dev mode21:45
ricmmtedg: if you really need the protection I can make it conflict with ubuntu-desktop :)21:45
w-flooh.. heynoq, see what rsalveti_ said :)21:45
tedgricmm, 1) yes, that would make sense.  2) Saviq told me it was where I could find the unity8 upstart jobs, so I grabbed it to look.21:46
ricmmyou could try to install libhybris' EGL components... doesnt mean you should21:46
heynoqOk yeah, thanks rsalveti,  I'm using only 10.1 right now21:46
rsalveti_w-flo: for new ports, using the final tag/branch for 10.1 is recommended21:46
w-florsalveti_, yes.. I didn't even notice the default branch for that repo was 10.2 already21:46
tedgricmm, Yes, but probably those wouldn't kill my afternoon to un-foo-bar my machine.21:47
ricmmtedg: he meant lp:session-manager-touch, I'm sure he didnt say install the touch session on your desktop21:47
w-floheynoq, if you wonder where a prop should go, try to find it in a "PRODUCT_COPY_FILES" list in that setup-makefiles.sh file21:47
ricmmyou chose to install a package not meant for a desktop installation, thats the only foobar action that happened here21:47
timpsergiusens: ah nice. that link is better to share :)21:48
ricmmhowever if you feel there are others like you that might follow suit, please MR against lp:session-manager-touch21:48
heynoqI also followed this tutorial initally to get my git setup21:48
heynoqWhere a plob should go? would that have to do with those blobs or you just giving advice on a place I can dig into21:48
heynoqMy next problem I beleive has to to with the extra configs that need to be added.     at kernel/[manufacturer]/[codename]/arch/arm/configs/ cyanogenmod_[codename]_defconfig. Please double check that it is indeed the default config file name in device/[manufacturer]/[codename]/*.mk (look for the TARGET_KERNEL_CONFIG variable).21:50
heynoqNot sure what file, using msm9615_defconfig in the common/arch folder.  Added to end of file with Hash about ubuntu touch.   I got errors right away about target stuff when building21:50
w-floheynoq, If you have a file "libc2d2_z180.so" from that "jf" thing (whatever that is?).. and then check the setup-makefiles.sh file, you will find this line: "\$(LOCAL_PATH)/lib/libc2d2_z180.so:system/lib/libc2d2_z180.so \\", so the file should be placed into vendor/motorola/msm8960-common/jf/lib/21:50
heynoqI removed those lines I added and then its currently building now....21:51
w-flothat error message appears to be cut off?21:51
heynoqAh nice!  Good find, I had no idea what that JF reference was for,  I'll have to do some hunting for that file21:52
heynoqNice, my first "most likely failed" build just finished, I'll try that out, but will also work on that jf thing21:52
w-flogood luck! =)21:52
w-floah, now I see that this is actually from the porting guide, not an error message21:53
w-floTARGET_KERNEL_CONFIG := msm8960_mmi_defconfig   in device/motorola/msm8960-common/BoardConfig.mk, so try to use that file. :-)21:55
heynoqah, again with the good finds!  I'll update my notes for when I rebuild to modify boardconfig.mk!  Thanks!21:56
w-flomaybe it's better to not modify boarconfig.mk and instead modify the msm890_mmi_defconfig21:57
w-floso you have the same kernel config used by cyanogen (only with ubuntu additions)21:57
heynoqnext note(possibly last :) )22:01
heynoqThe main build file needs to be checked if updates are required to it to support new drivers or parts of the build not used before, it's path is:build/core/main.mk The part of interest are the subdirs included in the build and if special treatment for devices need to be made, i.e.; make sure the new vendor subdirs are added. This will eventually not be needed.22:01
heynoqI have no idea what i need to add to teh subdirs section22:01
heynoqwhat I added caused build errors22:01
heynoqi removed them and the build continued on...22:01
w-flothat's a strange step, I've just removed some stuff and added the device/htc/ directory for my HTC device.. I'll show you what my commit looks like, but I have no idea if that's the right way to do it22:02
w-flohttps://github.com/w-flo/android_build/commit/809db1b56231644632b8b22605398354b345ae2e works for me22:03
heynoqI initally added this, which caused failrures22:03
heynoqAdded this: vendor/motorola/mb886 vendor/motorola/msm8960-common22:03
w-flomaybe just "vendor/motorola"22:03
heynoqI'll make that change22:03
heynoqThanks!  and you thought maybe you couldn't help!22:04
w-floheynoq, it's all trial and error for me :) I hope you can get your port up and running with radio!22:04
heynoqLol yeah, thats development!  Trial and error, best thing to keep you from banging your head against a wall is someone that can offer similar solutions that my fit :)22:05
heynoqYeah, my phone is currently stuck at the cyanogenmod boot logo animation....I'll try some other flashing options, but I think I need to implement those fixes and rebuild22:06
=== jono is now known as Guest48557
w-floheynoq, maybe it's related to the boot.img changes for Ubuntu Touch, not sure if those are in the old phablet branches as well22:07
tedgricmm, https://code.launchpad.net/~ted/session-manager-touch/conflicts-lightdm/+merge/17905622:08
w-flocheck the "out/target/product/[device]/" directory.. do you see an ubuntu-root subdir?22:08
=== rsalveti_ is now known as rsalveti
heynoqgotta reboot...brb in a bit22:13
ogra_yay, finally !22:19
ogra_plars, so the image passed the install test on mako, no need to handhold it then (beyond the normal stuff indeed(22:20
plarsogra_: awesome, yes seems to be rolling along22:21
plarsogra_: the maguro one seems to have failed, I think there's a race in phablet-flash still22:22
plarsogra_: should work on the restart22:22
ogra_yeah, i'm not worried22:23
ogra_it wasnt clear if the boot images work and have the right initrd22:23
ogra_thats both confirmed22:23
rickspencer3dang it22:25
rickspencer3wrote an app, and the osk doesn't come up on the phone :(22:25
rickspencer3ogra_, how am I supposed to log a bug these days?22:25
rickspencer3$ubuntu-bug something?22:26
ogra_open the terminal and run ubuntu-bug something, right22:26
rickspencer3at least the osk works for the apps already on the phone22:28
ogra_so likely a toolkit thing (or an app bug ?)22:28
plarswow, it even checks that that dir doesn't exist first, it got into a pretty small window for this race I think22:28
rickspencer3ogra_, I assume it's an sdk bug22:30
rickspencer3can't really see what my app could be doing wrong22:30
ogra_well, i havent seen such a bug with any of the shipped apps22:31
ogra_kbd works everywhere here22:31
ChickenCutlassrickspencer3, I think (don't quote me) is you have to request focus.22:32
ChickenCutlassrickspencer3, on the entry you want the keyboard for22:32
rickspencer3ChickenCutlass, I never had to before22:32
rickspencer3and that's not in any of the documentation22:32
ChickenCutlassrickspencer3, then I am talking crap22:32
* rickspencer3 looks back at docs22:32
rickspencer3ChickenCutlass, are you referring to this:22:33
rickspencer3activeFocusOnPress : bool22:33
plarssergiusens: https://bugs.launchpad.net/phablet-tools/+bug/120940822:33
ubot5Launchpad bug 1209408 in Phablet Tools "os.makedirs race" [Undecided,New]22:33
ChickenCutlassrickspencer3, maybe.  lol I just think I remember something to that effect.22:33
rickspencer3it's supposed to be true by default22:34
plarssergiusens: we could move the makedirs() call above the log message if you think that would be enough, otherwise we could just catch and discard the exception if it's OSError22:34
rickspencer3and also, I just tried it and it didn't work22:34
ChickenCutlassrickspencer3, then sorry22:34
ChickenCutlassrickspencer3, I am of no help22:34
rickspencer3ChickenCutlass, no worries22:34
rickspencer3I am sure it's an SDK bug, then22:34
rickspencer3and it helped to have a sanity check22:34
sergiusensplars: my plan is to discard the error22:37
sergiusensplars: but this should be mostly fixed the new branch (I think)22:38
plarssergiusens: I was going to propose a patch, but I figured it would be pointless given your major rework in flight22:38
plarssergiusens: sounds good22:39
* plars has to run away for a bit, back later to check on the jobs but so far they are looking good22:39
cjwatsonogra_: They seem to be on cadejo just fine.  I think perhaps your image build failure is because you haven't adapted cdimage code to download them properly?23:22
cjwatson(assuming you didn't figure it out since you asked ...)23:23
cjwatsonrsalveti: please run the cdimage test suite before you commit; ./run-tests23:25
cjwatsonrsalveti: (could you fix the test failure you introduced on Monday?)23:25
rsalveticjwatson: sorry, I just followed the procedure described by ogra_ in a wikipage, didn't have that23:26
cjwatsonrsalveti: which wiki page?23:26
rsalveticjwatson: and I just changed the jenkins build job name, but let me check the results23:26
cjwatsonrsalveti: perhaps also it would be appropriate to substitute config.series into jenkins_project, so that we have the minimum possible series dependencies23:26
cjwatsonthe jenkins build job name was also in a test :-)23:27
rsalvetiright, just saw that23:29
rsalvetican't even open the wiki here, over a really bad 3g, my isp decided to be off today23:30
rsalveticjwatson: mind pushing the change in the test as well?23:30
rsalvetiotherwise it'll take a few hours here I believe23:30
sergiusenscjwatson: I think it was this one https://wiki.ubuntu.com/ReleaseTeam/CDImageSetup23:30
cjwatsonI've updated that wiki page now to document running the test suite23:31
rsalveticjwatson: ogra said this could would be removed later this week, so not sure if we need to care much about that23:31
rsalveti*this would23:31
rsalvetimoving away from jenkins and just downloading the stuff from the android package23:31
cjwatsonrsalveti: Sure, I just didn't want to clean up after people without mentioning it23:32
cjwatsonrsalveti: the test suite must not fail; I'll fix it even if it will go away23:32
rsalvetidoesn't even open here :-(23:33
cjwatsonrsalveti: test fixed23:33
rsalveticjwatson: sure, we need to fix the test suite, just saying that we don't necessarily need to improve that code much as it'll be reworked (or was today)23:34
rsalveticjwatson: thanks, sorry for the noise23:34
cjwatsonAnd I've done the series change even if it goes away later.  I have ingrained scepticism about the future tense :-)23:34
rsalvetiright :-)23:37
cjwatsonhighbank nodes: made of awesome23:44
cjwatsonlast three firefox/armhf builds were 20, 38, 20 hours23:44
cjwatsonfirst one on highbank was 5h4023:44
AdrianTobonhello guys, I'd like to know if whatsapp is already available for ubuntu touch23:53
AdrianTobonsomebody there?23:57
AdrianTobondo you know if whatsapp is available for ubuntu touch?23:59

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!