| === chriadam|away is now known as chriadam | ||
| === LarrySteeze is now known as LarrySteeze|Away | ||
| === chihchun is now known as chihchun_afk | ||
| === LarrySteeze|Away is now known as LarrySteeze | ||
| === vying is now known as Guest97265 | ||
| === phunyguy_ is now known as phunyguy | ||
| === cul_ is now known as cul | ||
| === chriadam is now known as chriadam|aay | ||
| === chriadam|aay is now known as chriadam|away | ||
| === gnu is now known as Guest7562 | ||
| JamesTait | Good morning all; happy Ferris Bueller Day! :-D | 08:34 |
|---|---|---|
| === vying is now known as Guest65626 | ||
| ogra_ | cking, any reason to not seed cpustat, eventstat, powerstat and health-check too ? | 10:20 |
| cking | no reason I can think of | 10:21 |
| ogra_ | ok, adding them ... | 10:21 |
| cking | and forkstat perhaps? | 10:21 |
| ogra_ | lets get suspend-blocker into the archive asap too | 10:21 |
| ogra_ | yeah, forkstat was in the discussion yesterday | 10:22 |
| ogra_ | already got it | 10:22 |
| cking | ack | 10:22 |
| cking | ogra_, my only concern with suspend-blocker is that it is tied to the specific kernels we use on the phone devices and it's not a generic tool | 10:23 |
| ogra_ | hmm | 10:23 |
| ogra_ | i would have said just add a dependency ... but we dont install the kernel packages at all | 10:23 |
| cking | yep, that's part of the issue really | 10:24 |
| ogra_ | you could depend on the specific initramfs-tools-ubuntu-touch package ... but that will likely break on amd64 and i386 ... so not a good option either | 10:25 |
| cking | mmm, i've not got the packaging know-how to figure this out | 10:26 |
| ogra_ | i dont know a solution of the top of my head but will think about it | 10:26 |
| cking | ok, thanks :-) | 10:26 |
| ogra_ | hmm, why did slangasek's change not make it out of proposed last night | 10:28 |
| ogra_ | oh | 10:28 |
| ogra_ | unity-scope-click failed | 10:28 |
| ogra_ | hmm | 10:28 |
| ogra_ | cjwatson, did the click API change recently ? http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#ubuntu-touch-meta seems to be stuck due to unity-scope-click being ftbfs | 10:35 |
| === Ursinha is now known as Ursinha-afk | ||
| === Ursinha-afk is now known as Ursinha | ||
| === alan_g is now known as alan_g|afk | ||
| === alan_g|afk is now known as alan_g | ||
| === MacSlow is now known as MacSlow|lunch | ||
| north | abeato: Apart from rilmodem, does ofono work like normal RIL - RIL daemon + RIL driver ? If so, how do the concept of atoms and plugins chip in ? | 11:42 |
| === alan_g is now known as alan_g|lunch | ||
| oSoMoN | Mirv, hey, I’m looking at bug #1328841, what version of oxide did you rebuild against qt 5.3 ? | 12:00 |
| ubot5 | bug 1328841 in webbrowser-app "webbrowser slowed down on Qt 5.3" [Undecided,New] https://launchpad.net/bugs/1328841 | 12:00 |
| oSoMoN | is it the version in trusty? | 12:00 |
| oSoMoN | well, actually the version in utopic is the same as in trusty, until we do a new release | 12:01 |
| oSoMoN | i.e., did you just do a simple rebuild of the version in the archive against qt 5.3? | 12:02 |
| oSoMoN | if so, then it’s likely that compositing is not enabled, hence the general slowdown and clipping errors | 12:02 |
| Mirv | oSoMoN: check the PPA https://launchpad.net/~ci-train-ppa-service/+archive/landing-005/+packages - a rebuild of the lataest one in archive | 12:13 |
| Mirv | a simple rebuild indeed | 12:13 |
| oSoMoN | Mirv, ok, so that explains it, when a new version of oxide hits the archive (hopefully later this week), another rebuild should make the issue go away | 12:14 |
| oSoMoN | I’ll comment on the bug | 12:14 |
| Mirv | oSoMoN: ah, that'd be great then. thanks for commenting! | 12:14 |
| Mirv | I'll then rebuild it again, too, since it's one of the packages that depends on qtbase-abi-5-2-1 -> required rebuild | 12:15 |
| rsalveti | ogra_: ChickenCutlass: one problem I found yesterday, is that we now have 2 pulseaudio instances running at the same time | 12:39 |
| rsalveti | because of the split-greeter | 12:39 |
| ChickenCutlass | that’s not good | 12:40 |
| rsalveti | it seems we now have 2 maliit-servers running | 12:40 |
| ChickenCutlass | rsalveti: I noticed we have 2 of everything | 12:40 |
| rsalveti | and that's the main process that triggers pulse | 12:40 |
| ChickenCutlass | lots of extra memory | 12:40 |
| rsalveti | yeah =\ | 12:40 |
| === alan_g|lunch is now known as alan_g | ||
| rsalveti | where is mterry? | 12:40 |
| ogra_ | rsalveti, yes, i told you so yesterday | 12:42 |
| ogra_ | we run two sessions | 12:42 |
| ogra_ | one for the lightdm user and one ofr phablet | 12:42 |
| rsalveti | right, but didn't know we had 2 pulse running | 12:43 |
| rsalveti | as that might cause additional issues for me | 12:43 |
| rsalveti | not sure if we really need 2 though | 12:43 |
| seb128 | the lightdm session shouldn't be running when the phone is unlocked though? | 12:43 |
| rsalveti | it's always running | 12:43 |
| seb128 | why? | 12:43 |
| rsalveti | not sure, need to ask mterry | 12:43 |
| seb128 | on desktop we don't run the greeter all the time | 12:43 |
| ogra_ | ChickenCutlass, rsalveti, additionally to running two sessions there is also bug 1325580 | 12:43 |
| ubot5 | bug 1325580 in unity8 (Ubuntu) "after split greeter landing unity8 and the greeter consume a lot more memory" [Undecided,New] https://launchpad.net/bugs/1325580 | 12:43 |
| seb128 | rsalveti, well, split greeter or lock-in-session both have tradeoffs | 12:44 |
| ogra_ | seb128, on desktop we dont run the greeter at all once autologin is enabled | 12:44 |
| seb128 | the main motivation for the split greeter was to have only 1 implementation of the lock screen | 12:44 |
| ogra_ | but here the greeter is the lockscreen | 12:44 |
| seb128 | rather than one in lightdm and one in session | 12:44 |
| ogra_ | right | 12:44 |
| seb128 | ogra_, no, the lock screen is unity on desktop | 12:44 |
| ogra_ | so you need to run it all the time in case the user locks the screen | 12:44 |
| ogra_ | ah | 12:45 |
| rsalveti | I know the reasoning, just don't like much the side effects | 12:45 |
| seb128 | right, it's all tradeoffs | 12:45 |
| seb128 | we could decide to change back the approach | 12:45 |
| ogra_ | i think we should at least be able to sigstop the stuff | 12:45 |
| ogra_ | that will still eat your RAM though | 12:45 |
| seb128 | well, I don't see why it needs to run when the screen is not locked | 12:45 |
| seb128 | on desktop we close it on login | 12:45 |
| seb128 | and open it back on switch user | 12:45 |
| ogra_ | probably because spawining all the stuff takes to long | 12:45 |
| seb128 | which is sort of a lock case | 12:45 |
| rsalveti | right, spawning everything every time the user locks the screen might not be ideal | 12:46 |
| ogra_ | i guess only mterry can asnwer this properly though | 12:46 |
| rsalveti | might consume too much cpu then | 12:46 |
| rsalveti | and that will eat your battery | 12:46 |
| ogra_ | right, sigstop should work though | 12:46 |
| ogra_ | but still ... eats your RAM ... | 12:46 |
| ChickenCutlass | ogra_: I am worried about the extra memeory overhead | 12:46 |
| rsalveti | not sure we can have a workaround for the amount of ram consumed though | 12:46 |
| rsalveti | specially on bq, that we have ~900mb of ram | 12:47 |
| ChickenCutlass | rsalveti: already, do a ps and its like way too many processes | 12:47 |
| ogra_ | ChickenCutlass, yes, it got pretty noticeable n dogfooding ... i can definitely only have half of the webapps open i usually use here | 12:47 |
| ChickenCutlass | its bad | 12:47 |
| ogra_ | before they start respawning | 12:47 |
| pitti | hello | 12:47 |
| pitti | ah, freenode is back :) | 12:47 |
| rsalveti | right | 12:47 |
| ogra_ | (and losing their history ... i wouldnt mind the respawning if i wouldnt lose that all the time) | 12:47 |
| pitti | does anyone know how I can run a click app on the phone from teh command line? | 12:47 |
| ogra_ | ubuntu-app-launch | 12:48 |
| pitti | I ssh'ed in, then did cd `click pkgdir com.ubuntu.calculator` | 12:48 |
| pitti | and then what's in the .desktop file: | 12:48 |
| ogra_ | with the full app name (including the version) as the key | 12:48 |
| pitti | $ qmlscene -qt5 ubuntu-calculator-app.qml | 12:48 |
| pitti | QUbuntu: Could not create application instance | 12:48 |
| pitti | Abgebrochen (Speicherabzug geschrieben) | 12:48 |
| ogra_ | ah | 12:48 |
| ogra_ | you need --desktop_file_hint= | 12:48 |
| ogra_ | point it to any .desktop file | 12:48 |
| pitti | well, if that's in the .desktop file, how does Unity start that? | 12:49 |
| ogra_ | it doesnt use the .desktop file | 12:49 |
| seb128 | pitti, using ubuntu-app-launch | 12:49 |
| ogra_ | thats a silly bug in qmlscene ... you need to hand over the option but it wont actually be used for anything | 12:49 |
| pitti | seb128: I did play with that, but I coudln't get it to do anything useful (at least not in a chroot) | 12:49 |
| seb128 | pitti, tedg might be able to help you there | 12:50 |
| pitti | $ ubuntu-app-launch com.ubuntu.calculator | 12:50 |
| pitti | ** (process:4489): WARNING **: Unable to find keyfile for application 'com.ubuntu.calculator' | 12:50 |
| pitti | com.ubuntu.calculator1.3.279 | 12:50 |
| pitti | but click list has that ^ | 12:50 |
| rsalveti | ogra_: wonder if we can at least not start 2 maliit-servers | 12:51 |
| pitti | err, with a tab between name and version | 12:51 |
| ogra_ | pitti, ubuntu-app-launch com.ubuntu.gallery_gallery_2.9.1.989 | 12:51 |
| rsalveti | ogra_: how would you use the keyboard on the lock screen? | 12:51 |
| seb128 | pitti, https://wiki.ubuntu.com/Process/Merges/TestPlan/ubuntu-app-launch might have useful info? | 12:51 |
| seb128 | pitti, " upstart-app-launch com.ubuntu.clock_clock_$(version from click list)" | 12:51 |
| ogra_ | pitti, you need the version (you can assemble the actual name from "click list" output) | 12:51 |
| pitti | ah, that works: | 12:51 |
| pitti | ubuntu-app-launch com.ubuntu.calculator_calculator_1.3.279 | 12:51 |
| ogra_ | yeah | 12:51 |
| pitti | ok, so I have two questions: | 12:52 |
| ogra_ | rsalveti, i have no idea | 12:52 |
| pitti | - what's the _calculator_ and how do I find out what to put there? | 12:52 |
| pitti | - why does the .desktop file have a non-working Exec= ? | 12:52 |
| ogra_ | rsalveti, i know that password input, number input and pattern are on the RTM list though | 12:52 |
| ogra_ | iirc | 12:52 |
| ogra_ | pitti, Exec= thats the app confinement stuff mangling it when you "click install" the package | 12:53 |
| rsalveti | ogra_: oh, right =\ | 12:53 |
| rsalveti | that's annoying | 12:53 |
| ogra_ | right | 12:54 |
| ogra_ | on BQ specifically i guess | 12:54 |
| davmor2 | rsalveti: telepathically you just think your password and the chip in your head transmits it to the device......Next you'll be saying that you don't have a chip in your head won't you :D | 12:54 |
| ogra_ | with the lower ram | 12:54 |
| rsalveti | ogra_: my problem is now that I need to coordinate 2 pulses | 12:54 |
| rsalveti | not sure if the current droid code supports that nicely | 12:54 |
| ogra_ | davmor2, he told us it was "eye lasering" back then when they programmmed it ;) | 12:54 |
| pitti | ogra_: thanks; any idea what the _calculator_ bit is and where I can get it from? | 12:55 |
| ogra_ | rsalveti, well, for single user systems we *could* run a system pulse ... couldnt we ? | 12:55 |
| ogra_ | pitti, nops | 12:55 |
| rsalveti | that's a workaround, yes, but trying to avoid that initially | 12:55 |
| ogra_ | seems to always just be a duplication of the former word | 12:56 |
| rsalveti | running pulse as root is not cool :-) | 12:56 |
| ogra_ | nope | 12:56 |
| ogra_ | well, we would need a pulse user | 12:56 |
| ogra_ | or some such | 12:56 |
| rsalveti | could just run one as lightdm though, and make phablet user to use that | 12:56 |
| rsalveti | or something along that line | 12:56 |
| pitti | ogra_: and finally (sorry, --help doesn't work nor is there a manpage), any idea how to run it in the foreground? | 12:56 |
| ogra_ | then we wouldnt be able to stop the lightdm session though | 12:56 |
| pitti | it just seems to spawn off into the background, which is rather inconvenient for testing and debugging | 12:57 |
| pitti | i. e. like with qmlscene | 12:57 |
| ogra_ | pitti, sorry, no, wait for tedg | 12:57 |
| pitti | ogra_: ok; thanks for your help! | 12:57 |
| ogra_ | there is surely a debug option or some such | 12:57 |
| pitti | seb128: thanks to you, too | 12:58 |
| seb128 | pitti, yw! | 12:59 |
| tedg | pitti, Good morning | 12:59 |
| pitti | seb128: with --desktop_file_hint it indeed at least starts, just whines bitterly about "WARNING **: Unable to register app: GDBus.Error:org.freedesktop.DBus.Error.InvalidArgs: Invalid application ID" and other things | 12:59 |
| pitti | hey tedg, how are you? | 12:59 |
| jibel | pitti, I think it's the key of the hook in click info | 12:59 |
| tedg | Yeah, don't get used to desktop_file_hint, it should die soon :-) | 12:59 |
| ogra_ | pitti, you are the phablet user, right ? | 12:59 |
| ogra_ | (usng phablet-shell instead of adb) | 13:00 |
| tedg | pitti, Good, what are you trying to do? Start calculator? | 13:00 |
| pitti | ogra_: "ssh phablet@nexus" over forwarded ssh port, yes | 13:00 |
| pitti | tedg: right; I'm trying to figure out how to start a generic .click app for testing | 13:00 |
| ogra_ | k | 13:00 |
| pitti | tedg: I first tried with the Exec= line in the .desktop file, but that doesn't work (that needs this --desktop_file_hint= thing) | 13:01 |
| tedg | pitti, Specifically for testing (i.e. testing env vars) or just to test it. | 13:01 |
| davmor2 | ogra_, rsalveti: Thinking outside the box when you select the login option can that phablet account not then spawn the keyboard? That would then help with multiple accounts too right? | 13:01 |
| pitti | tedg: then I got told about upstart-app-launch | 13:01 |
| tedg | pitti, Yes, the desktop files are only for legacy desktops, they're a misnomer. | 13:01 |
| pitti | tedg: well, what would e. g. autopilot or an autopkgtest call? | 13:01 |
| rsalveti | davmor2: problem is that we might have user specific configs for the keyboard as well | 13:01 |
| ogra_ | davmor2, yeah, prob is that it is not that easy ... | 13:01 |
| ogra_ | right | 13:02 |
| rsalveti | unless the keyboard can react to that dynamically | 13:02 |
| tedg | pitti, Autopilot is using the GI bindings to call the start function in libual. | 13:02 |
| ogra_ | and the indicators are started by default for all sessions ... that would need some config hackery | 13:02 |
| pitti | tedg: u-a-l doesn't have --help or manpages; is there an option to run the app in the foreground? | 13:02 |
| rsalveti | but duplicating everything doesn't seems right though | 13:02 |
| tedg | pitti, We have two start functions, one for testing flags and one with out. | 13:02 |
| tedg | pitti, No, it always executes under Upstart | 13:02 |
| tedg | pitti, Signals about what it's doing and the log file path, etc all are in the lib. | 13:03 |
| davmor2 | ogra_, rsalveti: how does lightdm do it now if you have the a11y on screen keyboard up? | 13:03 |
| ogra_ | davmor2, thats not using a server in the backend | 13:04 |
| ogra_ | the desktop OSK is a standalone app ... malitt is a server/client thing | 13:04 |
| ogra_ | *maliit | 13:04 |
| rsalveti | wonder what would happen if during a call, you call greeter again and decide to start another user session | 13:04 |
| rsalveti | or just calling greeter | 13:04 |
| rsalveti | the mic/speaker would come from the phablet session | 13:04 |
| pitti | tedg: so what does u-a-l do with that app ID? | 13:04 |
| ogra_ | rsalveti, well, it should stick to the currently open connection | 13:05 |
| rsalveti | right, and would that be right? | 13:05 |
| rsalveti | in the desktop use case the audio stream is corked | 13:05 |
| ogra_ | if there is a call running ? sure | 13:05 |
| pitti | tedg: oh, and wrt. to its argument: that's com.ubuntu.calculator_calculator_1.3.279 ; the first part is from "click list", the third the corresponding version from "click list", but what's the second (_calculator_)? | 13:05 |
| ogra_ | yeah, thats logind | 13:05 |
| rsalveti | right, because otherwise the audio will go over the other session | 13:06 |
| ogra_ | if you lose your active seat it supends pulse | 13:06 |
| davmor2 | ogra_: :( | 13:06 |
| rsalveti | yeah, there are quite a few specifics that we need to handle here | 13:07 |
| tedg | pitti, We take the app id and create an upstart job instance with it. | 13:07 |
| tedg | pitti, The second is the application, there can be multiple apps in a click package. | 13:08 |
| tedg | pitti, There's upstart-app-triplet that'll help build them. | 13:08 |
| pitti | tedg: ah, does com.ubuntu.calculator_calculator.desktop map directly to these first two parts? | 13:08 |
| pitti | $ ubuntu-app-triplet com.ubuntu.calculator | 13:09 |
| pitti | com.ubuntu.calculator_calculator_1.3.279 | 13:09 |
| pitti | tedg: ah, I see | 13:09 |
| tedg | pitti, For most of our packages there's only one app though. | 13:10 |
| pitti | tedg: what would u-app-triplet do if there's more than one? print a list of all triplets? | 13:11 |
| tedg | pitti, It chooses the first one. You can adjust it to by passing "first-listed-app" or "last-listed-app" for the app name. | 13:15 |
| === rickspencer3_ is now known as rickspencer3 | ||
| dobey | whom can we lobby to, in order to get proper official support for N5 (since its hardware more closely matches what the Meizu phone will ship with)? | 13:15 |
| pitti | tedg: ah, splendid; many thanks for your help! | 13:16 |
| davmor2 | popey, pmcgowan: https://bugs.launchpad.net/music-app/+bug/1324105 I knew there was a bug :) | 13:19 |
| ubot5 | Ubuntu bug 1324105 in Ubuntu Music App "music-app crashed with SIGSEGV in GriloDataSource::removeModel() when using with Qt 5.3" [Undecided,Confirmed] | 13:20 |
| davmor2 | popey: so it's grilo causing the crash so I guess once they move to the new version they lose grilo maybe? | 13:21 |
| === pete-woods is now known as pete-woods-lunch | ||
| pmcgowan | davmor2, ah so indeed looks like scanning | 13:22 |
| popey | k | 13:22 |
| davmor2 | I remember timo telling me that musicplayer was the only hold back on multimedia which certainly seems to be the case although I haven't try camera so I'll do that today | 13:24 |
| pitti | rsalveti: is it possible to pass qemu options to ubuntu-emulator? I'm particularly interested in -snapshot, i. e. I want a transient overlay | 13:25 |
| rsalveti | pitti: yes, just not with ubuntu-emulator run, but there's a way to give qemu arguments | 13:26 |
| davmor2 | popey, pmcgowan: so the camera app doesn't seem to fill the screen on flo in qt5.3 there is a black boarder on both sides | 13:26 |
| rsalveti | we also use snapshots internally | 13:26 |
| pmcgowan | davmor2, how about media player | 13:26 |
| ogra_ | davmor2, like on mako today ? | 13:26 |
| pmcgowan | or gallery slideshow | 13:26 |
| pitti | rsalveti: oh, I suppose "ubuntu-emulator snapshot --revert-pristine trusty | 13:27 |
| pitti | rsalveti: pretty much does what I need | 13:27 |
| rsalveti | pitti: see goge-ubuntu-touch/ubuntu-emulator/run.go | 13:27 |
| rsalveti | oh, great then | 13:27 |
| pmcgowan | ogra_, oh? | 13:27 |
| rsalveti | :-) | 13:27 |
| ogra_ | trusty ... heh | 13:27 |
| pitti | well, I copied that from https://wiki.ubuntu.com/Touch/Emulator :) | 13:27 |
| ogra_ | pmcgowan, i got dark grey stripes in the camera app on the sides ... like ... since forever | 13:28 |
| ogra_ | davmor2, ^^ | 13:28 |
| pmcgowan | ogra_, yeah, I never noticed that | 13:29 |
| davmor2 | http://davmor2.co.uk/~davmor2/screenshots-phone/device-2014-06-11-142821.png | 13:29 |
| davmor2 | ogra_: that wasn't always the case I'm sure | 13:29 |
| ogra_ | davmor2, it is like that since an eternity though | 13:29 |
| ogra_ | but yeah, there were versions where it wasnt like that | 13:30 |
| davmor2 | maybe I wasn't looking that closely and was concentrating more on the transfer to gallery :) | 13:30 |
| pmcgowan | davmor2, bug it for ugo then if its not logged | 13:30 |
| ogra_ | pmcgowan, camera-app bugs currently get the response "it is being re-written anyway" ... | 13:31 |
| davmor2 | pmcgowan: it's on mako but it is far more noticeable on flo | 13:31 |
| davmor2 | indeed | 13:31 |
| ogra_ | we should just make sure the new app doesnt have it also | 13:31 |
| pmcgowan | ogra_, could be appropriate for this bug but god to log | 13:31 |
| pmcgowan | good | 13:32 |
| davmor2 | omg it's really bad on manta the thinkness scales with the screen increase | 13:32 |
| ogra_ | thinkness ? | 13:33 |
| ogra_ | funny word :) | 13:33 |
| davmor2 | thickness even | 13:33 |
| cjwatson | ogra_: unity-scope-click> No, that's a new warning emitted by gcc-4.9; pitti has already forwarded it to dobey | 13:33 |
| ogra_ | ah, cool | 13:33 |
| pitti | yay (not!) -Werror in production builds.. | 13:35 |
| ogra_ | yeah | 13:36 |
| sil2100 | dholbach: hey! | 13:39 |
| pitti | hm, seems the emulator still doesn't like me; 10 minutes, and it's still black, and no progress any more on the terminal | 13:39 |
| dholbach | hey sil2100 | 13:40 |
| sil2100 | dholbach: so, our session is in an hour, right? | 13:41 |
| ogra_ | pitti, did you build an i386 one ? | 13:41 |
| ogra_ | the armhf one is no fun ... | 13:41 |
| pitti | ogra_: no, armhf; I'll try with x86 now | 13:41 |
| dholbach | sil2100, yes, 1h20m | 13:41 |
| dholbach | sil2100, with a majority of the session covered in xnox session, I think I could just quickly summarise the workflow for getting something in the distro the usual way, and then resort to asking you questions about the citrain process and everything and relay questions - what do you think? | 13:42 |
| sil2100 | dholbach: since xnox has his in 20 mintues - should we also join his session? | 13:42 |
| sil2100 | dholbach: ok ;) | 13:43 |
| sil2100 | dholbach: anyway, since 'our' session is on the client track, I guess we'll simply cover things there | 13:43 |
| sil2100 | I have some slides so I can quickly overview what CI Train is about and how to release through it | 13:44 |
| dholbach | nice! | 13:44 |
| dholbach | yeah, I had to put it into the client track because everything else was full yesterday when I looked | 13:44 |
| === olli_ is now known as olli | ||
| davmor2 | rickspencer3: you are live I know it's the wrong channel I don't care though | 14:00 |
| cwayne | davmor2: hahahah | 14:03 |
| === chihchun_afk is now known as chihchun | ||
| === pete-woods-lunch is now known as pete-woods | ||
| === alan_g is now known as alan_g|tea | ||
| === alan_g|tea is now known as alan_g | ||
| === dandrader is now known as dandrader|afk | ||
| === gnu is now known as Guest23946 | ||
| pitti | rsalveti: hm, I tried "ubuntu-emulator snapshot --revert-pristine adt", but I still have all of my files in /home/phablet/; is that meant/known to only affect some images/partitions? | 14:52 |
| rsalveti | pitti: hm, no, should revert everything in theory | 14:53 |
| * pitti destroys and re-creates | 14:53 | |
| rsalveti | might be a bug | 14:53 |
| pitti | rsalveti: I'll reproduce and file it | 14:53 |
| rsalveti | thanks | 14:53 |
| pitti | I was actually about to file the crasher | 14:54 |
| pitti | rsalveti: oh, sounds pretty much like bug 1320307 | 14:55 |
| ubot5 | bug 1320307 in goget-ubuntu-touch (Ubuntu) "snapshot command doesn't work with x86" [Undecided,New] https://launchpad.net/bugs/1320307 | 14:55 |
| rsalveti | hm | 14:55 |
| pitti | qemu-img snapshot -l actually does show "pristine" | 14:58 |
| seb128 | mterry, hey | 14:58 |
| pitti | rsalveti: /home/phablet is on sdcard.img? | 14:58 |
| mterry | seb128, heyo | 14:58 |
| rsalveti | pitti: yeah, everything is part of sdcard.img | 14:59 |
| seb128 | mterry, the "wizard isn't in production" was a "landing team is not going to fall on us if there is a bug" ... doesn't mean we should land buggy code, but there is less pressure to be careful | 14:59 |
| mterry | seb128, not sure I follow -- I was just saying that I'd like to consider turning wizard on in production images in the relatively short term | 15:00 |
| seb128 | mterry, yeah, I know ... sorry, I was just clarifying my comment | 15:02 |
| seb128 | or trying to | 15:02 |
| seb128 | +1 to turn it soon | 15:02 |
| mterry | :) | 15:02 |
| seb128 | but I would do "flush the queue" | 15:02 |
| seb128 | test the wizard then | 15:02 |
| mterry | seb128, the queue of wizard branches? Sure | 15:02 |
| seb128 | and fix/iterate if it's not good enough | 15:02 |
| seb128 | then turn it on | 15:02 |
| ogra_ | turn it on !!! | 15:02 |
| mterry | Yar | 15:02 |
| ogra_ | and the boot animation too! | 15:02 |
| mterry | ogra_, boot animation isn't on my RTM list, meh | 15:03 |
| mterry | :) | 15:03 |
| ogra_ | pfft | 15:03 |
| mterry | Plus, it probably needs a tweak to be killed after sessions starts | 15:03 |
| oSoMoN | rsalveti, hey, I’m trying to use the X86 emulator on my laptop, but it hangs at startup (I just created an instance an hour ago or so), here’s the output I’m getting: http://pastebin.ubuntu.com/7628986/ | 15:04 |
| ogra_ | mterry, btw, you missed a discussion this moring about why the greeter session runs all the time | 15:04 |
| mterry | ogra_, oh yeah. OK. Were there people that were able to answer that? Cause I can if it matters | 15:04 |
| ogra_ | it eats quite some amount of RAM ... which we will be very low on on the prod. phones | 15:04 |
| ogra_ | mterry, right, rsalveti, seb128 and i were wondering what we can do about it | 15:05 |
| mterry | ogra_, right. So (A) I imagine 95% of the lifetime of a phone is spent in locked mode, so the 5% case is less interesting in my mind. And (B) it was done that way for instantaneous locking -- otherwise we have to bring up a greeter session. We tried that first, but it was a pain point for Design | 15:05 |
| ogra_ | mterry, well, we will only have 1G for everything on BQ | 15:06 |
| ogra_ | and the N4 already is very low on ram with the recent images | 15:06 |
| ogra_ | (and N4 has 2G ... i cant even run two apps without the bg one being killed all the time) | 15:06 |
| mterry | ogra_, understood. The memory loss is due to some crazy memory usage by maliit-server that needs to be investigated and also unity8's core is very beefy, which needs investigation. But we are probably going to revert split greeter for phone, but keep it for desktop images | 15:07 |
| === dandrader|afk is now known as dandrader | ||
| ogra_ | mterry, we should also have a way to SIGSTOP the whole session imho ... | 15:08 |
| mterry | ogra_, does that free the memory? | 15:08 |
| ogra_ | so it at least wont eat your CPU | 15:08 |
| mterry | ogra_, Ah | 15:08 |
| ogra_ | no | 15:08 |
| ogra_ | but the CPU cycles | 15:08 |
| mterry | ogra_, it would be nice if we could swap the whole thing to disk | 15:08 |
| ogra_ | battery life became horrid with split greeter | 15:08 |
| mterry | ogra_, I didn't think the CPU cycles were much | 15:09 |
| ogra_ | my phone doesnt survive half a day anymore when i use it as i did before | 15:09 |
| mterry | never seemed like much in top anyway | 15:09 |
| ogra_ | (it used to last a day) | 15:09 |
| mterry | but maybe it adds up. Wonder what it is doing all that time. | 15:09 |
| rsalveti | oSoMoN: the output is fine, see if you can access with adb shell | 15:09 |
| rsalveti | one thing I noticed: | 15:09 |
| mterry | But anything the greeter is doing, the shell is too. Bugs to be fixed | 15:09 |
| rsalveti | emulator: No kvm device file detected | 15:09 |
| rsalveti | without kvm it will be a bit slow | 15:09 |
| oSoMoN | rsalveti, how do I enable kvm? | 15:10 |
| rsalveti | oSoMoN: it tries to find /dev/kvm, maybe that is not enabled in your bios | 15:10 |
| ogra_ | mterry, well, does the greeter stop everything if the phone suspends ? | 15:11 |
| mterry | ogra_, as much as the shell does? | 15:12 |
| oSoMoN | rsalveti, after running "ubuntu-emulator run", am I not supposed to be dropped in an SSH shell prompt to the device? I’m not getting anything, and the emulator screen is black | 15:12 |
| ogra_ | oSoMoN, did you build an i386 emulator ? | 15:13 |
| ogra_ | or an armhf one | 15:13 |
| ogra_ | armhf is painfully slow | 15:13 |
| rsalveti | oSoMoN: no, you need to use adb shell or phablet-shell to have access to it | 15:14 |
| oSoMoN | ogra_, I tried armhf first, then fell back to i386, but no luck still | 15:14 |
| rsalveti | that window you started it will remain open with the output from qemu | 15:14 |
| rsalveti | https://wiki.ubuntu.com/Touch/Emulator#Using_the_emulator | 15:14 |
| oSoMoN | ok | 15:15 |
| * oSoMoN RTFM | 15:15 | |
| === vying is now known as Guest55492 | ||
| pmcgowan | oSoMoN, ironically there is a session right now on using the emulator ;) | 15:17 |
| oSoMoN_ | rsalveti, ok, I can adb shell into the device, but still nothing on the emulator screen | 15:18 |
| rsalveti | oSoMoN_: run ps and top, and see what is happening in there | 15:19 |
| brendand | Mirv, do you know anything about QtContacts module? | 15:20 |
| oSoMoN_ | rsalveti, init consumes 10% CPU constantly, unity8 is running | 15:22 |
| rsalveti | might just be too slow because kvm is disabled | 15:23 |
| rsalveti | see if you can enable that first | 15:23 |
| Mirv | brendand: no, but renato does (he contributes to upstream development as well) | 15:27 |
| seb128 | mterry, is the wizard supposed to run on unity7 session? | 16:01 |
| mterry | seb128, no | 16:01 |
| seb128 | mterry, it aborts on libmirplatformgraphics.so code | 16:01 |
| seb128 | :-( | 16:01 |
| mterry | seb128, it listens for unity8-greeter upstart signals | 16:01 |
| seb128 | mterry, well I tried to run the command manually | 16:02 |
| mterry | seb128, though I suppose that will need editing now that we're going to make the greeter optional on phones | 16:02 |
| mterry | seb128, aborted... with any particular message? Are you running under mir? | 16:03 |
| seb128 | mterry, no, I'm running unity7 :p | 16:04 |
| seb128 | I was trying to test it on my desktop | 16:05 |
| seb128 | that used to work | 16:05 |
| mterry | ah yeah | 16:05 |
| mterry | seb128, you people and your "Desktops" | 16:05 |
| mterry | Well I suppose I should say you people and your X desktops | 16:06 |
| mterry | Mir desktop would in theory work | 16:06 |
| seb128 | mterry, I'm trying to run it on unity8-desktop-mir, but I guess I need to hack the upstart job there | 16:07 |
| seb128 | no luck starting it from the dash either :/ | 16:08 |
| seb128 | hitting an exception in libmirserver.so | 16:09 |
| * seb128 needs to unbrick phone to test | 16:09 | |
| seb128 | mterry, I'm going to trust your testing and put that in a silo | 16:11 |
| seb128 | one day those things are going to run on unity8-mir-desktop! | 16:11 |
| mterry | seb128, hrm | 16:11 |
| mterry | seb128, I only tested on phone | 16:11 |
| seb128 | mterry, yeah, which is the right place to test it | 16:12 |
| seb128 | it's only that my phone is bricked atm | 16:12 |
| seb128 | so I was trying to test somewhere else | 16:12 |
| === gatox is now known as gatox_lunch | ||
| === Tassadar_ is now known as Tassadar | ||
| === mhr3_ is now known as mhr3 | ||
| === alan_g is now known as alan_g|EOD | ||
| === chihchun is now known as chihchun_afk | ||
| === gatox_lunch is now known as gatox | ||
| sarnold | I don't know how far along our mobile pdf viewer is, but this might be interesting to consider if we're not very far along: http://www.foxitsoftware.com/blog/?p=641 | 17:26 |
| ogra_ | sarnold, it is pretty much stalled afaik | 17:29 |
| ogra_ | since it need deep integration with the underlying architecture that nobody did yet | 17:29 |
| ogra_ | (content hub etc ... you cant just open pdfs ... the app needs to be written for the architectural design) | 17:29 |
| sarnold | ogra_: I figured it'd be a fair amount of work.. | 17:30 |
| ogra_ | yep | 17:30 |
| === dandrader is now known as dandrader|lunch | ||
| ogra_ | we already ship all bits and pieces (poppler mainly) | 17:30 |
| ogra_ | just missing all the integration | 17:30 |
| dobey | i wonder if reverting the split greeter will fix the screen not waking up issue on n5 | 17:47 |
| * ogra_ doubts that | 17:47 | |
| === Ursinha is now known as Ursinha-afk | ||
| === dandrader|lunch is now known as dandrader | ||
| === Ursinha-afk is now known as Ursinha | ||
| mhall119 | ogra_: theibaud is giving you a run for most apps published | 19:02 |
| ogra_ | haha, yeah, i noticed | 19:02 |
| ogra_ | after i landed ten games he landed 10 french apps | 19:03 |
| mhall119 | on a related note, beuno we sure could use a way to filter apps by language | 19:03 |
| ogra_ | mhall119, you can already ... | 19:03 |
| ogra_ | up to the developer | 19:04 |
| mhall119 | where? | 19:05 |
| ogra_ | in the submission form | 19:05 |
| ogra_ | final page before you provide the click app foor review | 19:05 |
| ogra_ | *for | 19:06 |
| beuno | mhall119, filte where? | 19:06 |
| mhall119 | beuno: in the results from the store | 19:06 |
| ogra_ | beuno, by target audience/language | 19:06 |
| mhall119 | I have a lot of apps available to install in a language I don't speak | 19:07 |
| ogra_ | because the dev didnt restrict them | 19:07 |
| beuno | right | 19:07 |
| beuno | correct | 19:07 |
| mhall119 | ogra_: I see how to limit my app by country, but not by language | 19:07 |
| beuno | and they also may be punching in a foreign language in the english description field | 19:07 |
| ogra_ | right ... i thought they were the same ... | 19:07 |
| beuno | which may be a UI issue | 19:07 |
| ogra_ | but it is indeed by country | 19:08 |
| mhall119 | beuno: quite often, yes | 19:08 |
| beuno | so language should be automatic | 19:08 |
| mhall119 | so I think this wouldn't so much be "Don't let non-French speakers download my app", but rather "My apps is available in French and German" and then if the user says "I only speak English" it won't show that one | 19:09 |
| beuno | if your device is in english, you should only see apps in english | 19:09 |
| ogra_ | i dont really want that ... | 19:09 |
| beuno | the apis are certainly there for the scope to allow users to select languages | 19:09 |
| ogra_ | i might want to keep my phone in english but still use german apps | 19:09 |
| mhall119 | it's decent enough for the first release | 19:10 |
| beuno | but maybe we need to think about this a bit | 19:10 |
| ogra_ | and because i'm weird i would want to use a french keyboard in that setup | 19:10 |
| popey | +1 | 19:10 |
| popey | (to ogra_ being weird) | 19:10 |
| rickspencer3 | uh, even I would want to see French apps, but would certainly keep my phone in English | 19:10 |
| ogra_ | heh | 19:10 |
| brendand | plars, hey | 19:11 |
| rickspencer3 | what's wrong with just showing all the apps? | 19:11 |
| popey | we dont show all apps | 19:11 |
| popey | only a subset | 19:11 |
| beuno | well, the idea is that people would translate french apps, at least the description to english | 19:11 |
| popey | only 100 are returned, which is IMO a problem | 19:11 |
| ogra_ | rickspencer3, that mhall119 needs to learn french ;) | 19:11 |
| popey | I can't even see my own apps listed on my phone | 19:11 |
| rickspencer3 | ogra_, or he could just not install apps that have no English translation? | 19:11 |
| ogra_ | or that | 19:11 |
| mhall119 | rickspencer3: I do need to learn french, yes, but in the mean time the first ~20 apps being suggested to me in the Dash I can't understand | 19:12 |
| ogra_ | well, specifically the webapps (which are essentially just better bookmarks atm) are tricky | 19:12 |
| rickspencer3 | mhall119, well, wait until the apps page has the new design before we decide we need to make changes | 19:12 |
| mhall119 | I don't recall Android ever presenting me with French-only apps to install | 19:12 |
| mhall119 | ok, if there's *some* solution coming, I'm happy to wait | 19:12 |
| rickspencer3 | Android has 100s of Ks of apps | 19:12 |
| rickspencer3 | there are many problems to solve when we hit those numbers :) | 19:13 |
| mhall119 | rickspencer3: a large number of which I assume are not in English, but I still don't see those | 19:13 |
| rickspencer3 | mhall119, right, that's point | 19:13 |
| ogra_ | rickspencer3, how to spend our money and the like :P | 19:13 |
| rickspencer3 | we'll see | 19:13 |
| rickspencer3 | :) | 19:13 |
| ogra_ | if we have 100s of Ks of apps ... canonical will have 5000 employees too | 19:13 |
| rickspencer3 | I think when the apps are displayed in a sensible order (rather than in the reverse order they were uploaded) and are sensibly categorized ... a solution may present itself | 19:14 |
| beuno | agreed | 19:14 |
| rickspencer3 | or the problem may not seem so bad | 19:14 |
| beuno | but we'll have to think about a good balance here | 19:14 |
| ogra_ | yeah | 19:14 |
| rickspencer3 | meantime, my apps should be on top .. always | 19:14 |
| rickspencer3 | that's the only firm rule I insist on | 19:14 |
| beuno | otherwise maybe the phone will be popular in china and none of us will be able to install an app | 19:14 |
| beuno | rickspencer3, that's why we have "if openid == 'rickspencer3'" everywhere | 19:15 |
| brendand | "order by price" | 19:15 |
| brendand | "order by name" | 19:15 |
| brendand | "order by rickspencerishnish" | 19:15 |
| === dandrader is now known as dandrader|afk | ||
| rsalveti | ogra_: shouldn't we fix the crash files to use a timestamp in the name as well? | 19:26 |
| rsalveti | said that in #ubuntu-devel yesterday but got no reply | 19:26 |
| ogra_ | rsalveti, they have a creation time in ls | 19:27 |
| ogra_ | isnt that enough ? | 19:27 |
| rsalveti | ogra_: nops | 19:27 |
| rsalveti | ogra_: once you get a crash file, you don't get any other from the same process | 19:27 |
| rsalveti | until the crash is removed | 19:28 |
| rsalveti | see on your desktop | 19:28 |
| rsalveti | susres.2014-06-06_19:45:19.608452.crash | 19:28 |
| rsalveti | there's a timestamp as well | 19:28 |
| ogra_ | ogra@styx:~$ ls /var/crash/ | 19:28 |
| ogra_ | ogra@styx:~$ | 19:28 |
| ogra_ | :P | 19:28 |
| rsalveti | so you can get multiply crashes | 19:28 |
| ogra_ | my system is stable | 19:28 |
| ogra_ | :) | 19:28 |
| rsalveti | *multiple | 19:28 |
| rsalveti | haha, right | 19:28 |
| ogra_ | well, we could indeed add a timestamp | 19:29 |
| ogra_ | i thought they append an integer for foollowup craches | 19:29 |
| ogra_ | *crashes | 19:29 |
| rsalveti | I raised this to bdmurray during the sprint, and he thought we were already using timestamps | 19:29 |
| ogra_ | rsalveti, we should drag bdmurry into a hangout next week ;) | 19:29 |
| rsalveti | yup | 19:29 |
| ogra_ | since it is his baby now | 19:29 |
| === artmello is now known as artmello|afk | ||
| tedg | rsalveti, I brought this up before and pitti brought up the concern that we'd over report crashes. | 19:54 |
| tedg | i.e. running the same app and crashing the same way 20 times isn't useful data. | 19:55 |
| rsalveti | right, that's true, but we're also losing useful crashes | 19:55 |
| tedg | Personally, I think we should let the stack trace dedup handle that. | 19:55 |
| rsalveti | as once you got a single crash, and keep using the phone, you'll never get any other crash for the same process | 19:56 |
| tedg | We have a bunch that are things like qmlscene that *aren't* the same app, even if the same path. | 19:56 |
| tedg | rsalveti, I think they clear at some point, but I'm not sure the mechanism there. | 19:56 |
| rsalveti | I think we just enabled cron | 19:56 |
| rsalveti | not sure yet if enabled on latest image though | 19:57 |
| rsalveti | ogra_ was working on that | 19:57 |
| rsalveti | but without cron we don't have anyone cleaning that up automatically | 19:57 |
| ogra_ | rsalveti, thats still in discussion | 19:57 |
| ogra_ | if we ship a cron daemon it will be anacron, not cron itself | 19:57 |
| ogra_ | and it will need s good bunch of diverts so the std. desktop cron jobs dont kick off | 19:58 |
| ogra_ | s/s/a/ | 19:58 |
| rsalveti | right | 19:58 |
| rsalveti | so yeah, short answer is that we're not erasing any crash file | 19:58 |
| rsalveti | :-) | 19:58 |
| ogra_ | right | 19:59 |
| ogra_ | until we either have anacron seeded or have whoopsie immediately send them | 19:59 |
| asac | i think whoopsie doesnt work either, right? | 19:59 |
| asac | or does it autosubmit now? | 19:59 |
| ogra_ | it could if we had a cron daemon | 19:59 |
| asac | can we please remove cron from the image? | 19:59 |
| ogra_ | all bits and pieces are in place afaik | 19:59 |
| asac | oh its gone? | 20:00 |
| ogra_ | asac, it isnt on there yet .... which causes tons of issues | 20:00 |
| asac | misread what you said: ) | 20:00 |
| asac | its great | 20:00 |
| ogra_ | no its not | 20:00 |
| asac | cron needs to die | 20:00 |
| rsalveti | yeah, we're trying to add it | 20:00 |
| rsalveti | not remove it | 20:00 |
| asac | its the tool of the devil | 20:00 |
| ogra_ | it makes our logs grow eternally | 20:00 |
| rsalveti | haha, right | 20:00 |
| asac | encourages hacky solutions for hacky problems :) | 20:00 |
| ogra_ | including app logs | 20:00 |
| asac | we can make an api that uses alarms type of stuff to do things on a schedule | 20:01 |
| ogra_ | asac, we wouldnt ship cron but anacron ... so you dont wake up the device ... and we would have to remove all existing cron jobs | 20:01 |
| ogra_ | only keeping logrotate and whoopsie active | 20:01 |
| asac | does android ship cron or anacron? | 20:01 |
| ogra_ | no idea | 20:02 |
| asac | check what our competitors are doing | 20:02 |
| ogra_ | probably their own implementation of something similar | 20:02 |
| tedg | systemd will solve all these problems of choice. | 20:02 |
| asac | that might inspire us | 20:02 |
| ogra_ | asac, well, inventing something from scratch right now doesnt sound like a good idea for RTM | 20:02 |
| asac | tedg: does systemd support crontab format? | 20:03 |
| tedg | asac, Haven't looked that deeply, but I would guess not. Not their MO. | 20:03 |
| ogra_ | asac, and we need a solution for logs and whoopsie ... both are hooked in with cron already | 20:04 |
| ogra_ | i would say lets think about a cool solution at the next UOS ;) | 20:04 |
| ogra_ | and go with anacron for RTM ... and some clever logrotate setup | 20:04 |
| asac | the hooks are super cheap; so their existance shouldnt be taken into the equation | 20:04 |
| asac | what is not that cheap is the daemon; but we have many things running that could just call out | 20:04 |
| ogra_ | well, at this point of the release i'd like to rely on proven technology | 20:05 |
| ogra_ | we can still rip it out in an update and ship something shiny | 20:05 |
| ogra_ | but have something that we know works reliable for now | 20:06 |
| asac | in practice, whatever we add will stay in there | 20:06 |
| asac | for a while | 20:06 |
| ogra_ | sure | 20:06 |
| asac | we dont even have tools to track technical debt and remember to go back | 20:07 |
| ogra_ | but there will be updates at some point | 20:07 |
| ogra_ | and i would like the phone to work til that point ;) | 20:07 |
| asac | rsalveti: why are you trying? what is holding you back? | 20:09 |
| asac | oh getting rid of all the cron files :) | 20:09 |
| asac | hehe | 20:09 |
| asac | i think i get it | 20:09 |
| ogra_ | asac, right | 20:10 |
| asac | how about we chat with tvoss tomorrow about this problem | 20:10 |
| ogra_ | thats a pretty ugly bunch of diversions we need to add | 20:10 |
| ogra_ | other packages ship cron files ... | 20:10 |
| ogra_ | like dpkg ... | 20:10 |
| ogra_ | asac, we had a call with lool already | 20:10 |
| asac | what did he say? | 20:10 |
| ogra_ | there are some workitems ... | 20:10 |
| ogra_ | if i could find the google doc now :P | 20:11 |
| asac | who was on that call? | 20:11 |
| ogra_ | https://docs.google.com/a/canonical.com/document/d/1PKqrxzJbHqsHfsSnAqRdV1F42u-fXlh8nvTh8QFXNfY/edit?usp=drive_web | 20:12 |
| ogra_ | geez. stop jumping around with your cursor ! | 20:13 |
| * ogra_ thinks thats asac secretl plan for distracting people | 20:13 | |
| * asac deletes everything :) | 20:15 | |
| asac | j.k. | 20:15 |
| asac | damn there is history feature | 20:15 |
| asac | so that doc starts off very good | 20:15 |
| asac | but then it becomes a mess. i think i can envision how the meeting went :) | 20:15 |
| ogra_ | very coordinated ... the notes arent perhaps | 20:16 |
| asac | anyway, dont feel distracted by me | 20:16 |
| ogra_ | asac, i think it is a good idea to use /dev/alarm as a long term solution and we should look deeper into that | 20:16 |
| asac | the main reason i am still asking/looking is to understand how our grown decision making process works | 20:16 |
| ogra_ | just not for RTM | 20:17 |
| asac | and how to improve the way we take architecture decision | 20:17 |
| asac | s | 20:17 |
| asac | but ignore that :) | 20:17 |
| ogra_ | well, the blocker for "true innovation" is always "but hey, will that work in convergence too" | 20:17 |
| asac | please realize that we define the rules of what convergence mean; there is nothing saying that in convernge everything of the old world will be preserved in the same manner and all of the new world will also be there :) | 20:18 |
| asac | anyway, not on topic here | 20:18 |
| ogra_ | no, but you will want to not reinvent all the wheels | 20:19 |
| === abhijit is now known as Guest66457 | ||
| === dandrader|afk is now known as dandrader | ||
| === fmasi is now known as Guest75017 | ||
| DX3M | Hi everyone. | 23:55 |
| DX3M | Anyone know if the HTC M7ul port is still being worked on? | 23:56 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
