| === kickinz1|afk is now known as kickinz1 | ||
| Saviq | ogra_, it was actually meant to be much better (when scrolling within a scope) | 06:57 |
|---|---|---|
| Saviq | ogra_, can you tell what scopes you're swiping between? or make a video even? | 06:57 |
| mardy | marcustomlinson: hi! I need some help to understand how scopes talk to OA, do you have a few minutes? | 07:27 |
| marcustomlinson | mardy: I got a meeting in 2min, after? | 07:27 |
| mardy | marcustomlinson: sure, please ping me back when you are done | 07:28 |
| marcustomlinson | mardy: ping | 08:39 |
| marcustomlinson | mardy: hey, sorry long meeting. | 08:39 |
| mardy | marcustomlinson: np | 08:41 |
| marcustomlinson | mardy: what would you like to know? | 08:42 |
| mardy | marcustomlinson: it's about the email from dobey, I don't understand where in the code the request to add an account is made | 08:42 |
| mardy | marcustomlinson: looks like his code is calling unity-scopes-api, but even there I can't find what I'm looking for | 08:42 |
| marcustomlinson | mardy: so the scope puts its service name, type and provider into a result item thats returned to the shell | 08:43 |
| marcustomlinson | mardy: when the user taps that item, the shell checks to see if that result has OA details in it | 08:43 |
| marcustomlinson | mardy: if it does, the shell then instantiates a OnlineAccountsClient::Setup object | 08:44 |
| marcustomlinson | mardy: sets the service name, type and provider and then calls exec() | 08:44 |
| marcustomlinson | mardy: going back a bit. the scope checks the status of its services (using SIGNON_POLICY_NO_USER_INTERACTION), if there are no services available, or there is no enabled service, the scope returns that result. | 08:47 |
| JamesTait | Good morning all; happy Fire Prevention Day! :-D | 08:48 |
| marcustomlinson | mardy: the shell then checks the status of its services (using SIGNON_POLICY_DEFAULT), if still no service is enabled, OnlineAccountsClient::Setup is used to either bring up the "Allow" / "Don't allow", or the log-in page | 08:48 |
| marcustomlinson | mardy: sorry, substitute "no enabled service" with "no authorised service" in the last 2 messages. As the service can be enabled but no access token granted | 08:50 |
| mardy | marcustomlinson: OK, looking at the code it seems that unity-scopes-shell is setting the service-name as applicationId | 08:52 |
| mardy | marcustomlinson: and dobey's code is using "ubuntuone" as service-name... now I need to find their .application file... | 08:53 |
| mardy | marcustomlinson: I cannot find it, so I guess that's the problem | 08:55 |
| marcustomlinson | mardy: He did create one yesterday, perhaps not committed. | 08:56 |
| marcustomlinson | mardy: The scope showed up under the Ubuntu One account apparently | 08:56 |
| marcustomlinson | mardy: (when he added the .application file) | 08:57 |
| marcustomlinson | mardy: but same issue | 08:57 |
| mardy | marcustomlinson: under which apparmor label is unity-scopes-shell running? | 09:00 |
| tsdgeos | pitti: dpm: do you know why in unity-scope-click there is a "Confirm" in the .pot that doesn't show up for translation in laucnhapad? | 09:01 |
| marcustomlinson | mardy: it runs unconfined. Its only the click scope that doesn't work by the way | 09:02 |
| marcustomlinson | mardy: if you click package your scope and add the relevant hooks etc it all works | 09:02 |
| marcustomlinson | mardy: the click scope is not click packaged (installed via deb) | 09:03 |
| mardy | marcustomlinson: btw, did you also reproduce this problem? | 09:03 |
| marcustomlinson | mardy: nope | 09:04 |
| marcustomlinson | mardy: I could try, but I'm battling with something else atm | 09:04 |
| marcustomlinson | mardy: here's a question | 09:04 |
| marcustomlinson | mardy: So I have the shell using SIGNON_POLICY_DEFAULT when calling signon_auth_session_process_async(). | 09:05 |
| marcustomlinson | mardy: OA then does the whole login background stuff, then I get an access token | 09:06 |
| marcustomlinson | mardy: the scope in the meantime gets no callback because the scope enabled state didn't change | 09:06 |
| marcustomlinson | mardy: is there by any chance a "authorized" or similar callback I can listen on for on an account service. I assume not | 09:07 |
| marcustomlinson | mardy: I need the scope to react when the shell has successfully logged in | 09:08 |
| marcustomlinson | mardy: so that it can then re-query with SIGNON_POLICY_NO_USER_INTERACTION again, and get the access token | 09:08 |
| mardy | marcustomlinson: are you still talking about the u1 scope, or others? | 09:09 |
| marcustomlinson | mardy: others | 09:09 |
| mardy | marcustomlinson: weird... as soon as the scope is authorized, the AgAccountService should turn enabled | 09:10 |
| marcustomlinson | mardy: ok, so the service is enabled, but not authorized from the scopes POV | 09:10 |
| marcustomlinson | mardy: its enabled, but was not given an access token | 09:11 |
| marcustomlinson | mardy: thats when using SIGNON_POLICY_NO_USER_INTERACTION | 09:11 |
| mardy | marcustomlinson: ok; can you reproduce this issue easily? If so, can you first enable logging in /etc/signond.conf? | 09:11 |
| marcustomlinson | mardy: ok, where does the log go then? | 09:13 |
| mardy | marcustomlinson: syslog | 09:13 |
| mardy | marcustomlinson: if you could paste it after reproducing the issue, that would help -- just make sure that no password is shown | 09:14 |
| marcustomlinson | mardy: don't know if this is good enough: http://pastebin.ubuntu.com/8525879/ | 09:17 |
| marcustomlinson | mardy: so I created a google account. enabled youtube | 09:17 |
| marcustomlinson | mardy: then loaded the youtube scope, the scope sees the service as enabled but no access token returned | 09:18 |
| marcustomlinson | mardy: that is the log of starting the youtube scope | 09:18 |
| mardy | marcustomlinson: here I see only the part where you use SIGNON_POLICY_NO_USER_INTERACTION | 09:21 |
| marcustomlinson | mardy: yes | 09:21 |
| marcustomlinson | mardy: then a separate process that is the shell uses SIGNON_POLICY_DEFAULT | 09:21 |
| marcustomlinson | mardy: that process gets an access token | 09:22 |
| marcustomlinson | mardy: how does the scope process now know to try again? | 09:22 |
| mardy | marcustomlinson: I'm slowly getting it | 09:23 |
| mardy | marcustomlinson: so, first the shell uses the Setup class to get access to an account (with data provided by the scope) | 09:23 |
| mardy | marcustomlinson: then it authenticates, using SIGNON_POLICY_DEFAULT | 09:24 |
| marcustomlinson | mardy: yes | 09:24 |
| mardy | marcustomlinson: and that succeeds | 09:24 |
| marcustomlinson | mardy: yes | 09:24 |
| mardy | marcustomlinson: but that's not in the logs you pasted | 09:24 |
| marcustomlinson | mardy: true | 09:24 |
| marcustomlinson | mardy: so once that completes, if the scope tries again with SIGNON_POLICY_NO_USER_INTERACTION, then it gets the access token | 09:25 |
| marcustomlinson | mardy: problem is, how does the scope know to try again | 09:25 |
| mardy | marcustomlinson: there's no way it can know it by signond, you (the shell) must tell it | 09:26 |
| marcustomlinson | mardy: I was worried you'd say that | 09:26 |
| marcustomlinson | mardy: synchronisation headache | 09:26 |
| mardy | marcustomlinson: if the shell is doing the authentication on behalf of the scope, it would be better to return the authentication data straight away, instead of having the scope do the same | 09:28 |
| marcustomlinson | mardy: yeah, then confinement is a danger | 09:29 |
| mardy | marcustomlinson: or actually: why can't the scope use SIGNON_POLICY_DEFAULT itself? | 09:29 |
| marcustomlinson | mardy: because it checks the states on start up | 09:30 |
| marcustomlinson | mardy: so if the scope is starting up in the background, you get the Online Accounts loading screen pop up in weird places | 09:30 |
| mardy | marcustomlinson: but you know, the token it gets could expire during the scope's lifetime, so I think we must be ready for this case to happen at unexpected times | 09:31 |
| mardy | marcustomlinson: OA currently shows a notification when there's need to authenticate, it doesn't immediately pop up the webview | 09:31 |
| marcustomlinson | mardy: I wonder if you've even seen this | 09:32 |
| marcustomlinson | mardy: if you use SIGNON_POLICY_DEFAULT, you get a black screen displaying "Online Accounts" and a spinner | 09:32 |
| marcustomlinson | mardy: like it does when you load an app | 09:33 |
| marcustomlinson | mardy: It seems that you probably shouldn't be seeing that | 09:34 |
| mardy | marcustomlinson: oh, it may be that things broke when we started using prompt sessions... and then, what happens? | 09:34 |
| mardy | marcustomlinson: the window closes itself and you see a notification? | 09:34 |
| marcustomlinson | mardy: yes, or if you're signing in with SIGNON_POLICY_DEFAULT, that page closes and you then get the callback to finish signon_auth_session_process_async | 09:35 |
| marcustomlinson | mardy: dude, if that loading screen doesn't show up and it just happens in the background. You'll solve my issue! | 09:35 |
| mardy | marcustomlinson: that's what should happen, can you please file a bug about this screen appearing out of the blue? | 09:37 |
| mardy | marcustomlinson: ubuntu-system-settings-online-accounts | 09:37 |
| marcustomlinson | mardy: so you call signon_auth_session_process_async with the SIGNON_POLICY_DEFAULT flag. The black Online Accounts loading screen appears with a spinner. Then after about 4s it disappears again, and you get the callback from signon_auth_session_process_async | 09:37 |
| marcustomlinson | mardy: sorry, I typed the whole thing, I might as well press enter ;) | 09:37 |
| marcustomlinson | mardy: ok will do | 09:38 |
| mardy | marcustomlinson: so, in the best case, you should see absolutely nothing | 09:38 |
| dpm | tsdgeos, it is in the template now: https://translations.launchpad.net/ubuntu/utopic/+sources/unity-scope-click/+translations | 09:38 |
| marcustomlinson | mardy: yes. cause this is my bug: https://bugs.launchpad.net/ubuntu/+source/unity-scopes-shell/+bug/1374394 | 09:38 |
| ubot5 | Ubuntu bug 1374394 in unity-scopes-shell (Ubuntu) "When logged into YouTube scope the video aggregator some sometimes pops up the OA loading screen, then the results go blank" [Critical,In progress] | 09:38 |
| tsdgeos | dpm: ok, what was the problem? | 09:38 |
| dpm | tsdgeos, this comes back to our longstanding issue that templates in the upstream projects don't get updated | 09:38 |
| mardy | marcustomlinson: if the OAuth requires us to show a webview (for instance, if the scopes uses some application key which needs to be authorized), then you should see a snap decision which requests the user to go through the OAuth thingie | 09:39 |
| dpm | tsdgeos, we've got translations enabled both upstream and on the source package | 09:39 |
| tsdgeos | dpm: ah | 09:39 |
| dpm | on the source package, we're good, as the template gets updated | 09:39 |
| mardy | marcustomlinson: well, then just reassign it, no need to file a new bug | 09:39 |
| dpm | tsdgeos, but on upstream, it needs to be updated manually and committed | 09:39 |
| marcustomlinson | mardy: so yeah, that bug is due to the fact that I am using SIGNON_POLICY_DEFAULT from the scope | 09:39 |
| dpm | tsdgeos, for this particular project, it's an annoyance, but not a problem in terms of shipping translations | 09:40 |
| dpm | tsdgeos, as we're shipping the translations from the source package as language packs | 09:40 |
| tsdgeos | dpm: well still a problem of "translation isn't there" :D | 09:40 |
| dpm | tsdgeos, I know, I know, this has been an annoyance for me for years | 09:41 |
| marcustomlinson | mardy: if you open the youtube scope (this is landed), scroll to the bottom of the surfaced results and you'll see a log-in button. Tap that button and you'll be taken through the log-in process. At the end of the process you'll see that loading page I'm talking about | 09:42 |
| mardy | marcustomlinson: actually, I'm not sure how easy it is to fix that, it involves Mir | 09:43 |
| mardy | marcustomlinson: the problem is that Mir pops up a window as soon as you instantiate a QGuiApplication, and not as soon as you call show() on a window | 09:44 |
| mardy | marcustomlinson: and I was told it's not an easy change... | 09:44 |
| vitimiti | Hi | 09:44 |
| marcustomlinson | mardy: can you delay the instantiation of the QGuiApplication object? | 09:45 |
| ev | rsalveti: would https://bugs.launchpad.net/ubuntu/+source/android/+bug/1266747 be your team? Any chance I can get some eyes on that? I'm seeing it with both nested-kvm and on bare metal (well, inside lxc) | 10:03 |
| ubot5 | Ubuntu bug 1266747 in android (Ubuntu) "emulator GL translator errors during autopilot tests" [Medium,Confirmed] | 10:04 |
| ogra_ | ev, glxinfo isnt so helpful for gles :) | 10:06 |
| ev | ogra_: you've highlighted the shallow depth of my understanding here :) | 10:06 |
| ev | I'm afraid I'm rather clueless when it comes to GL | 10:06 |
| ogra_ | we have some egl tool for that ... | 10:06 |
| * ogra_ forgot the name | 10:06 | |
| ogra_ | ev, you want the output of es2_info from the mesa-utils-extra package | 10:08 |
| ev | cool, I'll arrange to get that after my next call | 10:08 |
| ev | cheers | 10:08 |
| marcustomlinson | mardy: so how is it looking? | 10:13 |
| === kickinz1 is now known as kickinz1|afk | ||
| anpok_ | hm the glReadPixel call is part of the test verification? | 10:16 |
| nik90 | ogra_: you need this -> https://plus.google.com/+DaveBesbris/posts/7LP8YNGGo32?pid=6063052381222537778&oid=100817211548713875249 | 10:28 |
| ogra_ | nik90, lol | 10:29 |
| E524 | hey guy's, the meizu phone is comming soon. but what's about the bq aquaris? i can't even find rumors. does anyone know more? my problem is that i can't afford a meizu but my current phone is slowly diying :( | 10:35 |
| ogra_ | who said a Meizu is coming soon ? | 10:36 |
| ogra_ | did Meizu do an announcement ? | 10:36 |
| E524 | no not officially but there was a blog post of the italian meizu office that it would come in december | 11:02 |
| E524 | orga: http://www.androidheadlines.com/2014/10/meizu-ubuntu-touch-landing-meizu-mx4-december.html | 11:03 |
| E524 | orga: this one is better: http://news.softpedia.com/news/Meizu-Confirms-MX4-Ubuntu-Touch-Release-for-December-2014-458782.shtmlv | 11:06 |
| popey | hehe | 11:06 |
| popey | i love how so many news sites have picked that up | 11:06 |
| popey | ogra_: get cracking! | 11:06 |
| mardy | marcustomlinson: I'll comment on the bug | 11:07 |
| ogra_ | popey, lol | 11:13 |
| ogra_ | E524, i think thats rather a fanboy blog filled with a bit of wishful thinking :) (but who knows) | 11:14 |
| popey | Not us, clearly! | 11:14 |
| E524 | popey: who is us if i may ask? | 11:15 |
| popey | we work for canonical on ubuntu | 11:15 |
| E524 | ah ok. well i think a lot of ppl have wishful thinking and can't wait. but good things shall take it's time... | 11:16 |
| ogra_ | well, see the last word in the channel topic ... :) | 11:16 |
| popey | haha, forgot about that | 11:17 |
| E524 | ^^ | 11:17 |
| popey | wow, I set that 2 months ago! | 11:17 |
| simosx | Hi All! I am trying to build Ubuntu Touch on a fresh installation of 14.04.1. During linking, I get a "/usr/bin/ld: cannot find -lGL". I have libgl1-mesa-dev installed. | 11:18 |
| simosx | I used 'lunch aosp_x86-userdebug'. | 11:18 |
| mardy | greyback, Saviq: can you please have a look at bug 1374394 and see if my suggestion makes sense (and is easily implementable)? | 11:19 |
| ubot5 | bug 1374394 in unity-scopes-shell (Ubuntu) "When logged into YouTube scope the video aggregator some sometimes pops up the OA loading screen, then the results go blank" [Critical,In progress] https://launchpad.net/bugs/1374394 | 11:19 |
| mardy | marcustomlinson: I commented ^ | 11:20 |
| marcustomlinson | mardy: thanks man | 11:21 |
| Saviq | mardy, the problem is the splashscreen *is* the app, in terms of the window stack and focus | 11:21 |
| Saviq | mardy, so I'm not sure it's safe to just hide it, we'd need to defer adding it to the window stack | 11:22 |
| Saviq | greyback, mzanetti, can you please comment on mardy's request ↑ | 11:22 |
| greyback | yep am thinking | 11:23 |
| mardy | Saviq: eh, I was afraid it wouldn't be so trivial :-/ | 11:23 |
| === MacSlow is now known as MacSlow|lunch | ||
| mzanetti | oh dear... | 11:25 |
| mzanetti | yeah... greyback I guess a X-Ubuntu-Splash=False would make sense in any case | 11:25 |
| mzanetti | which i guess would do for the <hack></hack> proposed by mardy | 11:26 |
| mzanetti | how to properly solve this... not sure yet | 11:27 |
| greyback | mzanetti: I'm not fond of that, as there's little to stop any app settings that, but it would do the job I guess | 11:27 |
| mzanetti | greyback: I think this is a legit feature for an app | 11:27 |
| mzanetti | greyback: if an app wants a visual user experience where neither of our predefined splashes fit I think it's legit to allow disable the splash screen for an app | 11:28 |
| greyback | mzanetti: that flexibility is at a cost that the app appears to launch very slowly | 11:29 |
| greyback | which is the whole reason we're using splash screens | 11:29 |
| mzanetti | greyback: sure, usually you do want the splash | 11:29 |
| mardy | greyback: I guess that only trusted helpers would set X-Ubuntu-Splash=False | 11:30 |
| greyback | I'm not a fan of allowing apps flexibility on this matter. We want our platform to feel speedy | 11:30 |
| mzanetti | greyback: but for instance... launch cylinder-3d from the store. it launches so fast, you can't really see the splash anyways | 11:30 |
| greyback | mardy: that'll be work to police | 11:30 |
| mzanetti | greyback: noone will disable it unless there is a reason... after all app developers want their apps to feel speedy too | 11:31 |
| mzanetti | and if there's a reason, like for example this one. | 11:31 |
| mzanetti | why disallowing it? | 11:31 |
| greyback | mzanetti: it makes our app story more confusing. When you launch an app, it immediately slides in and appears in the spread as a splash screen, that then is replaced with actual UI. User focus is on that new app. | 11:34 |
| greyback | If no splash screen, do we wait until app has drawn frame to slide it in & add it spread? And what is focused in the mean-time? | 11:34 |
| anpok_ | user needs feedback | 11:34 |
| greyback | I agree. This is one of those times I don't want to trust app developers - they might think their app launches super-quick, when they test on their super-fast phone - but on a slow phone | 11:35 |
| mzanetti | I'm not saying we should remove the splash for good... | 11:35 |
| anpok_ | on actions that are expected to change large portions of the screen it should be within 300ms .. otherwise below 100ms .. otherwise people might start hammering the phone | 11:35 |
| mzanetti | why do you guys think app developers try to make their apps the worst possible? | 11:36 |
| greyback | because I see android apps | 11:36 |
| mzanetti | I'm not saying we should make it any harder to have the splash screen, its still there, it is default and everything | 11:36 |
| greyback | anyway, discussion went a bit off-topic | 11:36 |
| simosx | so, these commands were needed: | 11:36 |
| simosx | 1. sudo ln -s /usr/lib/i386-linux-gnu/libX11.so.6 /usr/lib/i386-linux-gnu/libX11.so | 11:36 |
| simosx | 2. sudo ln -s /usr/lib/i386-linux-gnu/mesa/libGL.so.1 /usr/lib/i386-linux-gnu/libGL.so | 11:37 |
| ogra_ | simosx, what are you building for ? | 11:37 |
| ogra_ | usually there is no GL on phones | 11:37 |
| simosx | I used 'lunch aosp_x86-userdebug'. | 11:37 |
| ogra_ | ah | 11:37 |
| ogra_ | yeah, the x86 emulator is a bit special | 11:38 |
| ogra_ | any reason to not just use the binaries ? | 11:38 |
| simosx | i want to test out a new keyboard layout. | 11:38 |
| ogra_ | err | 11:39 |
| ogra_ | and why do you rebuild the android container for this ? | 11:39 |
| ogra_ | you just need to re-package the ubuntu--keyboard package and install it | 11:39 |
| ogra_ | *ubuntu-keyboard (one dash) | 11:39 |
| ogra_ | the android container only priovides some access to binary kernel drivers ... there is no keyboard stuff in it | 11:40 |
| simosx | you mean with 'debuild', etc? | 11:40 |
| ogra_ | sudo apt-get build-dep ubuntu-keyboard; apt-get source ubuntu-keyboard ... cd into the package dir, add your changes and build the deb (yes, debuild works) | 11:41 |
| ogra_ | then use the emulator and install your package in there to test it | 11:42 |
| simosx | ogra_, nice. Thanks, I'll try that. | 11:42 |
| ogra_ | (or use a device if you have one) | 11:42 |
| === _salem is now known as salem_ | ||
| === dandrader is now known as dandrader|afk | ||
| gatox | seb128, hi... do you know if there is any problem landing my branches? i was told the silo with my changes was not building | 12:24 |
| === salem_ is now known as _salem | ||
| seb128 | gatox, there was an issue with gdb, doko fixed it earlier | 12:28 |
| seb128 | it should build fine now | 12:28 |
| seb128 | that's from what I see | 12:28 |
| seb128 | cjwatson, can you retry https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-020/+build/6446982 ? | 12:28 |
| seb128 | it hit the gdb multiarch issue | 12:28 |
| seb128 | gatox, oh, that's different from your landing | 12:28 |
| seb128 | not sure then | 12:29 |
| seb128 | kenvandine should know when he gets online | 12:29 |
| cjwatson | seb128: no, https://launchpad.net/ubuntu/+source/gdb/7.8-1ubuntu4/+build/6446385 hasn't finished yet | 12:32 |
| seb128 | cjwatson, oh, good point, thanks for catching that | 12:33 |
| cjwatson | looks very close though | 12:33 |
| seb128 | I saw the upload accepted this morning, I didn't think much about the build time | 12:33 |
| cjwatson | I think this might be a case where processes are left over from the test suite and stall the build ... | 12:34 |
| === _salem is now known as salem_ | ||
| === salem_ is now known as _salem | ||
| === dandrader|afk is now known as dandrader | ||
| matv1 | so i updated to rtm image 4 the other day. When I did there were also a bunch of app updates listed. I updated the ubuntu image and didnt bother about the apps. But to my surprise they are still listed after the update | 12:59 |
| matv1 | all core apps btw | 12:59 |
| matv1 | so it seems there is something out of sync in the app versioning | 12:59 |
| matv1 | probably a bit of cache somewhere? | 13:00 |
| matv1 | but it doesnt look pretty | 13:00 |
| matv1 | anyone ideas how to get rid of that? | 13:00 |
| matv1 | or if this is a known bug | 13:01 |
| dobey | mardy, marcustomlinson: the scope (well i presume it's the scope, as all i see is "com.canonical.s" in the list), is added to the access list regardless of whether there is an .application file or not. | 13:17 |
| mardy | dobey: what list? in the OA UI, when you click on the account? | 13:18 |
| dobey | mardy, marcustomlinson: and the scope itself is unconfined, so even with the .application file in the system, the pop-up is still asking to allow access, presumably for an "unconfined" app, and thus it doesn't have the appid or app's name in the dialog | 13:18 |
| marcustomlinson | dobey: can you try make your service name and application id the same | 13:18 |
| dobey | mardy: yes | 13:18 |
| ted | mpt, Are you aware of the sound UX bugs that got filled yesterday? (I can't assign them) bug 1378961 and bug 1378618 | 13:19 |
| ubot5 | bug 1378961 in Ubuntu UX "Volume notifications over multimedia content are annoying" [Undecided,Confirmed] https://launchpad.net/bugs/1378961 | 13:19 |
| ubot5 | bug 1378618 in unity8 (Ubuntu) "volume notification looks bad when an indicator is open" [Undecided,New] https://launchpad.net/bugs/1378618 | 13:19 |
| mardy | dobey: that cannot be, it's either some other app, or you have a stale .application file somewhere | 13:19 |
| gatox | seb128, i see the logs for the silo for my branches says: dpkg-checkbuilddeps: Unmet build dependencies: gdb:any | 13:20 |
| gatox | seb128, does the build need to be retrigger or something? | 13:20 |
| kenvandine | gatox, there was a gdb upload | 13:20 |
| seb128 | gatox, k, same gdb issue I mentioned then | 13:21 |
| mpt | ted, I had not | 13:21 |
| kenvandine | without multiarch | 13:21 |
| seb128 | kenvandine, gatox: that gdb fix is still building on arm64 though | 13:21 |
| dobey | mardy: how so? all i see in the logs is "unconfined" everywhere, and nothing even shows up in syslog until after i tap on allow/don't allow in the pop-up; and nothing in the dbus.log is telling me what it's actually looking for. just that the ui was activated and the qml for that is logging that access was granted; no appid is logged anywhere afaict | 13:21 |
| gatox | seb128, kenvandine ah ok, thx | 13:21 |
| kenvandine | gatox, i was trying to land your branches yesterday, but hit that | 13:21 |
| ted | mpt, Ah, glad I pinged you then :-) | 13:21 |
| ted | mpt, If you could instantly come up with genius ideas, that'd be great. KTHXBYE ;-) | 13:21 |
| mpt | k | 13:21 |
| ted | mpt, They're now the list of bugs blocking audio notifications, got failed by QA when trying to land. | 13:22 |
| dobey | mardy: and why am i getting the dialog at all anyway, for an unconfined app? | 13:22 |
| dobey | marcustomlinson: changing it to "ubuntuone" in the application file didn't help | 13:22 |
| mardy | dobey: please uncomment the LoggingLevel line i /etc/signond.conf, and try again | 13:24 |
| marcustomlinson | dobey: try copy the name of your ini file (without the ".ini") as the service name | 13:24 |
| dobey | marcustomlinson: in the code or in the .application? | 13:24 |
| dobey | mardy: it is already uncommented | 13:24 |
| marcustomlinson | dobey: actually just give mardy your log | 13:24 |
| mardy | dobey: make sure that no passwords are visible | 13:27 |
| robotfuel | cyphermox: ping /usr/share/urfkill/scripts/flight-mode 0 doesn't toggle flight mode from off to on anymore, do you know anything about a change to urfkill? | 13:28 |
| === MacSlow|lunch is now known as MacSlow | ||
| dobey | mardy: http://pastebin.ubuntu.com/8526788/ | 13:30 |
| mardy | dobey: thanks, indeed that's not too informative -- but at least I don't see anything wrong | 13:38 |
| mardy | dobey: next step: | 13:38 |
| mardy | export OAU_LOGGING_LEVEL=2 | 13:38 |
| mardy | export OAU_DAEMON_TIMEOUT=9999 | 13:38 |
| mardy | online-accounts-service | 13:38 |
| mardy | dobey: type this from a shell running in the phone ^ | 13:39 |
| jgdx | Wellark, what r in rtm proposed has the urfkill fix? | 13:41 |
| dobey | mardy: http://pastebin.ubuntu.com/8526859/ | 13:41 |
| mardy | dobey: you clicked the "denied" button, right? | 13:42 |
| mardy | dobey: but in any case, it looks like the .application file is missing or not found | 13:43 |
| dobey | mardy: i did. | 13:43 |
| dobey | mardy: but in any case, why am i seeing the dialog at all, for an unconfined app? | 13:43 |
| mpt | ted, done | 13:46 |
| mardy | dobey: because unity-scopes-shell is probably using OnlineAccountsClient to request access <- right, marcustomlinson? | 13:48 |
| dobey | no | 13:48 |
| Wellark | jgdx: don't know | 13:48 |
| dobey | it's using libaccounts/signon-glib | 13:48 |
| Wellark | jgdx: please ask ci-eng | 13:48 |
| marcustomlinson | mardy, dobey: the shell does use OnlineAccountsClient to request access yes | 13:49 |
| ted | mpt, So for no alert sounds, is that only for the multimedia role or for instance the phone role as well? | 13:49 |
| ted | mpt, No visual either, or just no sound? | 13:49 |
| mardy | dobey: actually the PID of the requestor is logged, check who owns PID 2460 | 13:49 |
| mpt | ted, still visual, just no sound | 13:50 |
| dobey | oh the dash | 13:50 |
| mardy | dobey: now, what should happen is that once the account is created, your app/scope is added to the ACL, and next time the dash won't request the access anymore | 13:50 |
| mardy | dobey: but in order for this to work, you need to have proper .application and .service files | 13:51 |
| mpt | ted, separately, there should be an API for an app to embed a system volume slider into its own UI. The app developer wouldn’t need to do anything special to alter the volume, the toolkit would handle it all. And whenever the slider was visible, volume notification bubbles wouldn’t appear, because you could see the slider moving instead. Video player apps, System Settings, and indicator-sound would all use this special slider. | 13:51 |
| dobey | mardy: so what the heck am i supposed to be put in the .application file? because afaict it's asking for access for "unconfined" which is already in the ACL, so it shouldn't be asking | 13:51 |
| mardy | dobey: is your scope running under the "unconfined" apparmor label? | 13:52 |
| dobey | mardy: yes | 13:52 |
| mpt | ted, I wish there was a version of <https://launchpadlibrarian.net/186796512/volume2.png> that included a fixed volume notification (one that doesn’t just say “Volume”) | 13:52 |
| mpt | Then I could show it to visual designers and they would be confused only by the thing I wanted them to be confused by | 13:53 |
| mardy | dobey: ship an .application and a .service file with the same name | 13:53 |
| mardy | dobey: write the id of the .service file inside the .application file | 13:53 |
| ted | mpt, You can ask MacSlow, but I'm not sure if he has one, but he might be able to generate one. | 13:54 |
| mardy | dobey: make the .application file refer to a valid .desktop file (if you want to see a proper icon and name in OA) | 13:54 |
| dobey | mardy: i still don't understand what exaclty i'm supposed to put in the .application file for the appid | 13:54 |
| mardy | dobey: I think that that should make it work (bugs allowing :-) ) | 13:54 |
| ted | mpt, I don't think we give apps today the permission to adjust volume. | 13:55 |
| ted | mpt, I'm not sure that they should have that. Why would an app want to adjust the volume? | 13:55 |
| MacSlow | mpt, ted: one sec... | 13:55 |
| marcustomlinson | dobey: the desktop file is generated by the scopes runtime, so just make sure the desktop file name is the same as your scope id (ini file name) | 13:55 |
| mpt | ted, just about every media player I’ve ever seen lets you do that | 13:55 |
| marcustomlinson | dobey: with the .ini replaced with .desktop | 13:55 |
| ted | mpt, To be fair, everything gets routed through indicator-sound today. It's an action. System settings, the indicator, everything goes through that action. | 13:56 |
| dobey | bah stupid pastebin | 13:56 |
| MacSlow | mpt, ted: btw... if that indicator (sound) not sync. (volume) notification should be triggered in the first place... seems very redundant to me. | 13:56 |
| dobey | it won't let me paste xml | 13:56 |
| mardy | dobey: check /usr/share/accounts/applications/account-polld.application, for example (just do s/service-type/service/g) | 13:56 |
| ted | mpt, Is that just because they expect to be full screen? | 13:57 |
| MacSlow | mpt, ted: doh.... "is shown" is missing there somewhere in my last sentence | 13:57 |
| dobey | well i don't have that file here | 13:57 |
| mpt | MacSlow, right, if indicator-sound used the magical volume slider I talked about, volume notification bubbles would be suppressed | 13:57 |
| dobey | but anywya, i have an application file and it isn't working | 13:57 |
| marcustomlinson | dobey: examples are in the scopes api doc | 13:57 |
| dobey | nothing changed when i added it | 13:57 |
| dobey | marcustomlinson: yes i copied the youtube scope .application and adjusted it for clickstore, but it doesn't work | 13:58 |
| mpt | MacSlow, but as I said in the bug report, this visual design problem would still exist if, for example, you used the volume buttons while the Time & Date indicator was open, or the Bluetooth indicator, or any of the others | 13:58 |
| marcustomlinson | dobey: the application id and service name need to be the same in scopes case | 13:58 |
| marcustomlinson | dobey: are they? | 13:58 |
| dobey | no | 13:59 |
| dobey | why do they? | 13:59 |
| mpt | ted … Have you ever used iTunes? Or Rhythmbox? Or Banshee? Or Windows Media Player? Or QuickTime Player? Or even RealPlayer? They all have a volume slider | 14:00 |
| ted | mpt, Yes, and I've seen that. But I've never used it. Usually I start the media player and minimize it. | 14:00 |
| dobey | marcustomlinson: and what name should they be exactly? because the logs aren't telling me what they should be (ie, i don't see any "No service found for 'blah.blah.blah'" or such in the logs) | 14:00 |
| ted | mpt, I know it's there, just not sure why. | 14:00 |
| marcustomlinson | dobey: the .ini file without the .ini | 14:01 |
| mpt_ | ted: …regardless of whether they’re full screen | 14:01 |
| ted | mpt_, Yes, and I've seen that. But I've never used it. Usually I start the media player and minimize it. | 14:02 |
| ted | mpt_, I know it's there, just not sure why. | 14:02 |
| dobey | and why can't i tell it to just use an existing service? | 14:02 |
| dobey | well that didn't work | 14:04 |
| === simosx_ is now known as simosx | ||
| dobey | marcustomlinson, mardy: i finally got a different behavior, but it's a bit worse. installed an ubuntuone.application using "ubuntuone" for app/service ids, and now i just get an empty blank window popped up. on top of the dash. in a trusted session. | 14:07 |
| === mpt_ is now known as mpt | ||
| mpt | MacSlow, by “one sec…” did you mean that you were making another screenshot? If so, could you do it using an indicator that isn’t indicator-sound? :-) | 14:07 |
| marcustomlinson | dobey: with "Online Accounts" and a spinner? | 14:07 |
| dobey | marcustomlinson: no | 14:08 |
| cyphermox | robotfuel: it works here | 14:08 |
| dobey | marcustomlinson: entirely blank. just the standard grey-white background for a qml window | 14:08 |
| robotfuel | cyphermox: what image are you using? | 14:08 |
| MacSlow | mpt, is the first one not ok? | 14:09 |
| cyphermox | robotfuel: 91 | 14:09 |
| mpt | MacSlow, sorry, if you already posted one I missed it when I disconnected | 14:09 |
| MacSlow | mpt, I can do another one with a different indicator... | 14:09 |
| robotfuel | cyphermox: it worked for me in 91, 93+ is where the issue is | 14:09 |
| mardy | dobey: the online-accounts-service logs will probably help | 14:13 |
| mardy | dobey: can you also paste your .application and .service files? | 14:14 |
| dobey | mardy: no (pastebin doesn't like xml) | 14:14 |
| dobey | and don't know where else to paste them | 14:15 |
| mardy | dobey: just private query | 14:16 |
| dobey | oh adding non-xml helps | 14:17 |
| dobey | mardy: http://pastebin.ubuntu.com/8527054/ | 14:17 |
| dobey | mardy: that's the application file, and the online-accounts-service log | 14:17 |
| mardy | dobey: what's the name of the application file? | 14:18 |
| dobey | mardy: ubuntuone.application | 14:18 |
| dobey | err | 14:18 |
| dobey | oops, i pasted the one that didn't work | 14:18 |
| dobey | mardy: http://pastebin.ubuntu.com/8527063/ | 14:19 |
| mardy | dobey: that looks right; unfortunately I don't see any output from the U1 plugin, do you know how to make it more verbose? | 14:22 |
| dobey | mardy: U1_DEBUG=1 | 14:22 |
| cyphermox | robotfuel: I'm not aware what changes might have happened in 93 and later.. | 14:22 |
| cyphermox | robotfuel: maybe awe_ knows? | 14:23 |
| mardy | dobey: try exporting that as well, and actually, I think that a "strace -f" of online-accounts-service will also be interesting, so we see what files it's opening | 14:23 |
| robotfuel | cyphermox: did you upgrade and have it fail? | 14:23 |
| cyphermox | not yet, in a meeting | 14:23 |
| dobey | mardy: why would there be any extra output from the u1 plug-in? | 14:23 |
| dobey | mardy: is online-accounts-ui displaying the blank "existing account" page for some reason? | 14:24 |
| === ubuntu is now known as Guest73210 | ||
| mardy | dobey: you don't have any other U1 account configured, right? | 14:29 |
| dobey | mardy: just the one and we only allow having one account | 14:29 |
| mardy | dobey: I mean, you started the test when you didn't have any U1 accounts at all | 14:29 |
| dobey | mardy: i've tested both with and without an account configured already, so i can test both flows | 14:30 |
| dobey | mardy: i'm also trying to test the case where an account exists on the phone, but the token has been revoked on the server | 14:30 |
| mardy | dobey: and the blank screen case happens when you don't have an account configured, or have one? | 14:31 |
| dobey | mardy: when i have one | 14:31 |
| mardy | dobey: OK, I need to investigate this, it happens because we are hitting an unusual code path | 14:36 |
| mardy | dobey: normally the access shouldn't get requested in this case | 14:37 |
| mardy | dobey: can you please delete your U1 account, and try again? | 14:37 |
| mardy | dobey: what version of ubuntu-system-settings-online-accounts do you have, BTW? | 14:38 |
| dobey | 0.4+14.10.20141006-0ubuntu1 | 14:40 |
| dobey | mardy: ^^ | 14:40 |
| mardy | dobey: does this blank window disappear almost immediately, or does it stay? | 14:41 |
| dobey | mardy: it stays | 14:42 |
| Z3 | Hi, do you know what's the release date for ubuntu phones in Europe? | 14:43 |
| dobey | mardy: if i delete my account in system-settings, then try to install from the scope, i get the login/register page as expected, but after logging in, nothing happens. the window just goes away and it's sitting at the dash again | 14:43 |
| dobey | mardy: and then tapping install brings up the blank window again | 14:44 |
| ogra_ | Z3, except that we know they will release in europe first, there is no date or anything yet | 14:44 |
| dobey | Z3: when the manufacturer is ready to release them :) | 14:44 |
| mardy | dobey: OK, now if you go to the system settings, is your scope listed? and enabled? | 14:44 |
| dobey | mardy: there is now "com.canonical.scopes.clickstore" and "ubuntuone" both in the list, and both are enabled, yes | 14:45 |
| mardy | dobey: OK, I guess that the first one is from a stale file, but otherwise it looks fine | 14:45 |
| Z3 | ogra_ dobey ok, thank you :) I'm impatient they are released | 14:46 |
| mardy | marcustomlinson: so, if the ubuntuone scope is successfully enabled, access should not be requested again; any idea why it is? ^ | 14:46 |
| dobey | mardy: well, except for the blank window and nothing working :-/ | 14:46 |
| mardy | dobey: I mean, the blank window should not appear, because access has already been granted and shouldn't be requested again | 14:47 |
| Z3 | When will be release the official stable version for Nexus 4? (No developer versions, the "user" version) | 14:47 |
| ogra_ | once the RTM release goes out | 14:47 |
| ogra_ | (same time when the manufacturer getss the images) | 14:48 |
| Z3 | ogra_ is the same as the current RTM version ? | 14:48 |
| dobey | ogra_: for nexus4? will we actually call it stable for end users on n4 at that point, or will n4 always be treated as a development device? | 14:48 |
| ogra_ | (read: there is no definitve date for that) | 14:48 |
| ogra_ | dobey, does that matter ? it will be the same RTM release :) | 14:49 |
| dobey | ogra_: well, it matters for the purpose of managing expectations | 14:49 |
| ogra_ | rtm is our stable thing ... it will be released for all phones (likely at the same time for all of them) | 14:49 |
| Z3 | before the end of the year? | 14:50 |
| ogra_ | for sure | 14:50 |
| Z3 | great ! :) | 14:50 |
| dobey | sure but i wouldn't necessarily call something that is "you must jailbreak your phone to install this" a "stable for end user" thing | 14:50 |
| dobey | anyway | 14:51 |
| ogra_ | dobey, so you call preinstalled phones that you need to "fastboot oem unlock" any different ? | 14:51 |
| marcustomlinson | dobey, mardy: we only request access if we don't get an access token back from signon | 14:51 |
| dobey | ogra_: why would you need to unlock them to get ubuntu, if ubuntu is already on them? | 14:51 |
| marcustomlinson | dobey: do you get an access token from within the scope? | 14:51 |
| === VargaD_ is now known as VargaD | ||
| ogra_ | dobey, the preinstalled phones wont differ much from n4 ... except that the bootloader is locked by default and ubuntu is already installed | 14:51 |
| dobey | marcustomlinson: this is becasue the scopes-api code only works with oauth plug-ins? | 14:51 |
| dobey | marcustomlinson: we aren't using the scopes-api to actually get the token | 14:52 |
| dobey | ogra_: right, but the bootloader doesn't need to be unlocked to get ubuntu | 14:52 |
| ogra_ | dobey, you need to unnlock them for development ... like any other phone ... since you dont want that if an enduser forgets his phone in a bar anyone can just access the data from recovery | 14:52 |
| marcustomlinson | dobey: ah | 14:52 |
| ogra_ | dobey, it does need to be unlocked to develop on it | 14:53 |
| ogra_ | and the unlocking will behave like on AOSP/nexus ... and wipe the phone | 14:53 |
| marcustomlinson | dobey: yes, you are not considered authenticated if you do not have an access token | 14:53 |
| dobey | ogra_: yes. but i thought we were talking about end users, not development :) | 14:53 |
| dobey | marcustomlinson: i'm not sure what that means | 14:53 |
| ogra_ | i would expect 80% of our first endusers to be developers :) | 14:53 |
| Z3 | I am a developer :) | 14:53 |
| ogra_ | and the majority to actually unlock them as a first step | 14:54 |
| ogra_ | to take a deeper look etc | 14:54 |
| dobey | ogra_: there's a difference between being a developer, and developing on the device itself. one can write an app and never need to unlock the device (or even need a device, really) | 14:54 |
| dobey | but anyway | 14:54 |
| ogra_ | sure | 14:55 |
| ogra_ | app devs wont *need* it | 14:55 |
| ogra_ | but i'm sure they will unlock anyway ;) | 14:55 |
| ogra_ | in any case they will be locked down by default | 14:55 |
| marcustomlinson | dobey: if signon does not return an access token, the shell will think its not authenticated, so everytime you tap install it'll request access | 14:55 |
| ogra_ | which was the initial question | 14:55 |
| dobey | marcustomlinson: what do you mean by "access token" exactly? | 14:56 |
| Z3 | thank you, bye ! | 14:56 |
| dobey | marcustomlinson: scopes-api only works with accounts that are using the oauth plug-in in online acocunts? | 14:56 |
| marcustomlinson | dobey: yes | 14:57 |
| dobey | marcustomlinson: well that just won't do | 14:57 |
| marcustomlinson | dobey: otherwise, use OnlineAccountClient to get services statuses | 14:57 |
| dobey | u1 isn't oauth | 14:57 |
| marcustomlinson | dobey: if there is an enabled service, don't call register_account_login_item | 14:58 |
| marcustomlinson | dobey: calling that will tell the shell to check for authorisation when tapped | 14:58 |
| marcustomlinson | dobey: bare in mind that enabled != authenticated | 14:59 |
| marcustomlinson | dobey: so if you check for an enabled account, and you find one, don't call register_account_login_item on your "Install" button | 14:59 |
| dobey | marcustomlinson: as it should; but i think it shouldn't require the secret to be stored in a particular manner in signon (as it apparently does now) | 14:59 |
| === oSoMoN_ is now known as oSoMoN | ||
| nik90 | ted: Hey, I saw UAL landed in the recent images. Will you app for the startup times work correctly now (actually will the server show the correct data) ? | 15:04 |
| marcustomlinson | dobey: ok, thinking about it if the shell thinks its a failure when no oauth token is received, your ContinueAction will not work | 15:05 |
| marcustomlinson | dobey: I need to fix that | 15:05 |
| marcustomlinson | dobey: I'll have the shell consider service enabled as good | 15:05 |
| dobey | marcustomlinson: is that in the shell, or in the scopes api? | 15:05 |
| marcustomlinson | dobey: unity-scopes-shell / unity-plugin-shell | 15:06 |
| marcustomlinson | dobey: unity-scopes-shell / unity-plugin-scopes | 15:06 |
| marcustomlinson | dobey: there is no unity-plugin-shell | 15:06 |
| dobey | marcustomlinson: well, what about just having it check if it can see the secret, rather than trying to parse the secret? | 15:06 |
| marcustomlinson | dobey: which secret? client secret? | 15:07 |
| ted | nik90, just a sec, otp. | 15:07 |
| cyphermox | robotfuel: I just upgraded to 94; and the flight mode script works | 15:08 |
| dobey | marcustomlinson: the signon "secret" | 15:09 |
| dobey | marcustomlinson: in the IdentityInfo | 15:09 |
| robotfuel | cyphermox: ok I'll update to 94 and see if that fixes it for me | 15:11 |
| dobey | anyway, i need to get lunch and i have an appointment to get to. :-/ | 15:11 |
| dobey | bbiaw | 15:12 |
| mhall119 | dbarth: when do we expect Facebook push-notifications to work on the phones? | 15:13 |
| mhall119 | IIRC, there was something blocking it, missing token or somthing, in the past | 15:14 |
| popey | pffft facebook... who uses _that_!? | 15:14 |
| robotfuel | cyphermox: it works for me in 94, so it must have been an issue in just image 93 | 15:16 |
| mhall119 | popey: us luddites who haven't switched to Ello yet | 15:17 |
| popey | Do keep up! Facebook -> Status.net -> Diaspora -> G+ -> Ello | 15:18 |
| robotfuel | cyphermox: it works in 94 for me too. | 15:20 |
| === gatox is now known as gatox_lunch | ||
| ted | nik90, Mostly, the problem now is that the allocation of devices changed and the startup tests haven't been ported over to the new architecture. | 15:32 |
| ted | nik90, So we're not getting startup results right now. | 15:33 |
| ted | nik90, But that's getting fixed… | 15:33 |
| ted | nik90, So close, but yet so far :-) | 15:33 |
| robotfuel | bfiller: can you tag this with a touch-<date> that you estimate this bug will be fixed by https://bugs.launchpad.net/ubuntu/+source/camera-app/+bug/1376495? | 15:37 |
| ubot5 | Ubuntu bug 1376495 in camera-app (Ubuntu) "autopilot3 run camera_app.tests.test_zoom.TestCameraZoom.test_slider_zoom_out crashes camera-app" [High,Confirmed] | 15:37 |
| matv1 | my updateable apps list is filled with apps that are allready updated with that version. I have some wonky cash probably. Anyone know what i should clear out? | 15:39 |
| matv1 | or whatever else might be causing that? | 15:40 |
| === dandrader_ is now known as dandrader|lunch | ||
| bfiller | robotfuel: tagged, is omer26 around? he has a branch to fix it | 15:42 |
| robotfuel | bfiller: does that fix the crash or just the pinch? | 15:42 |
| robotfuel | bfiller: the test will say it passes when there is a crash unfortunately | 15:43 |
| popey | matv1: you're seeing multiple copies of the same app in update list? | 15:43 |
| bfiller | robotfuel: fixes the test | 15:43 |
| matv1 | popey no not that. | 15:43 |
| bfiller | robotfuel: so it actually works | 15:43 |
| popey | matv1: odd | 15:43 |
| matv1 | i am seeing a list of core apps that need updating, but going by the version number they allready are updated | 15:44 |
| matv1 | popey it happened after i updated to rtm #4 | 15:44 |
| robotfuel | bfiller: I haven't seen omer all week, there was a Monday - Wednesday holiday in Pakistan this week. | 15:44 |
| matv1 | popey: at that time there a set of apps needing updating but i didnt bother and just did the image upgrade | 15:45 |
| matv1 | so it seems that the click update manager got out of sync during the image update | 15:46 |
| matv1 | because after i got the new image with latest version of the core apps, they remain in the update list | 15:48 |
| matv1 | of course trying to execute the update fails | 15:48 |
| matv1 | popey: makes any sense to you? | 15:49 |
| === pstolowski is now known as bbl | ||
| === bbl is now known as pstolowski|bbl | ||
| matv1 | popey is there perhaps any kind of local cash related to click update manager that i can flush? | 16:05 |
| === _salem is now known as salem_ | ||
| robotfuel | cyphermox: I assigned a new bug to you https://bugs.launchpad.net/ubuntu/+source/urfkill/+bug/1379406 the number of crashes has skyrocketed it's the 2nd most crasher on krillin today | 16:17 |
| ubot5 | Ubuntu bug 1379406 in urfkill (Ubuntu) "Top Crasher: /usr/lib/arm-linux-gnueabihf/urfkill/urfkilld:6:g_assertion_message:g_assertion_message_expr:urf_device_register_device:urf_device_ofono_new:urf_ofono_manager_add_modem" [Critical,Confirmed] | 16:17 |
| === gatox_lunch is now known as gatox | ||
| cyphermox | robotfuel: ah, thanks. | 16:22 |
| === robbiew is now known as robbiew-afk | ||
| === salem_ is now known as _salem | ||
| nik90 | ted: :-) so close yet so far | 16:48 |
| === alan_g is now known as alan_g|EOD | ||
| ahayzen | rsalveti, ping | 17:29 |
| === dandrader__ is now known as dandrader | ||
| slvn | hi ! | 17:49 |
| slvn | After a few month of inactivity, I try to run my apps on ubuntu touch + nexus10. Many thing seems to have changed. | 17:50 |
| slvn | Unfortunatly my apps refused to start | 17:51 |
| slvn | dont know if related, I see this message in /var/log/apport.log | 17:51 |
| slvn | gdbus call error: Error: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.gnome.SessionManager was not provided by any .service files | 17:51 |
| slvn | This is a SDL native application. Three/For months ago, that was working fine. | 17:53 |
| slvn | Any help is welcome ... | 17:53 |
| slvn | I have just update the devices with the latest version (ubuntu 14.10 (r243)) | 17:54 |
| slvn | I have updated my SDL trunk and recompiled every think ... | 17:55 |
| slvn | the only think I have had to change was to add : "--allow-untrusted" to the installation command. (sudo -H -u phablet pkcon --allow-untrusted install-local /home/phablet/Downloads/....click) | 17:56 |
| dobey | slvn: what is trying to connect to the gnome session manager for your app? | 18:09 |
| slvn | dobey. No idea ... The app is game. It requires drawing stuff. not sure if this is related to gnome session manager. (I would say no..) | 18:16 |
| slvn | I just found a log message interesting : | 18:16 |
| slvn | it says : Can't init SDL VIDEO: No available video device | 18:16 |
| slvn | seems like SDL is not talking to mir anymore ... | 18:17 |
| dobey | slvn: is it using gtk+ or any other gnome libs? | 18:17 |
| slvn | Only SDL libs. no gtks, no gnomes libs. | 18:18 |
| slvn | this is a native application that talks to MIR, and do EGL calls | 18:18 |
| slvn | dobey, the message "Can't init SDL VIDEO: No available video device" is for sure the issue. | 18:19 |
| dobey | yes | 18:19 |
| slvn | the application is launch but fails to connect to MIR | 18:19 |
| slvn | I remember there was some MIR_SOCKET stuff ... | 18:20 |
| dobey | is your app statically linked? or does it include the SDL libraries you're linking to? | 18:21 |
| popey | slvn: is it SDL2? | 18:21 |
| slvn | it is dynamically linked. I provide the SDL libs | 18:22 |
| slvn | yes SDL2 | 18:22 |
| popey | i had trouble with this recently | 18:22 |
| popey | it couldn't find a library, I _think_ because SDL had been built against an old libmirclient... | 18:22 |
| popey | bschaefer knows more ㋛ | 18:22 |
| dobey | was the version of SDL you're including not built with mir support? | 18:22 |
| slvn | I have rebuild and update everything | 18:23 |
| bschaefer | which library? | 18:23 |
| bschaefer | for umm SDL2 | 18:23 |
| === VargaD_ is now known as VargaD | ||
| bschaefer | theres an issue with libmirclient changing the return of a function from bool to MirBool | 18:23 |
| slvn | updated my SDL2 trunk, updated ubuntu touch OS, updated my chroot for cross compilation | 18:24 |
| bschaefer | which causes the dynamic loader to fail sooo you'll either need to fix that in sdl/src/video/mir/SDL_mirsym.h | 18:24 |
| bschaefer | or compile SDL2 with | 18:24 |
| bschaefer | --disable-mir-shared | 18:24 |
| bschaefer | (configure) | 18:24 |
| bschaefer | slvn, ^ | 18:24 |
| slvn | bshaefer :) | 18:24 |
| slvn | how can I patch mirsym.h ? | 18:25 |
| bschaefer | i forgot what function needed to be changed | 18:25 |
| bschaefer | umm | 18:25 |
| bschaefer | SDL_MIR_SYM(MirBool,mir_connection_is_valid,(MirConnection *connection)) | 18:25 |
| bschaefer | SDL_MIR_SYM(MirBool,mir_surface_is_valid,(MirSurface *surface)) | 18:26 |
| bschaefer | slvn, IIRC, they are normal bool | 18:26 |
| bschaefer | vs MirBool | 18:26 |
| bschaefer | make sure they are MirBool and all should work | 18:26 |
| dobey | i would disable as many shared libraries and possible and go with static, for click packaged things. | 18:26 |
| bschaefer | dobey, well SDL has it set up to dynamically load the libraries so they are not dependent on the library it self | 18:27 |
| dobey | bschaefer: yeah i know. pluggable modules are all the rage. but static might make the app start faster too, since it won't have to go resolving all the dlopen()s | 18:28 |
| bschaefer | this is true | 18:28 |
| slvn | I have tried the Symbols but does not seem to have changed anything. (I may be wrong as I did that quickly). | 18:29 |
| slvn | I have to go. I will be back | 18:29 |
| slvn | I will try also the build stuff static! | 18:29 |
| slvn | thanks | 18:29 |
| bschaefer | there could be something else | 18:29 |
| bschaefer | yeah try the static i need to get it fixed in main | 18:29 |
| dobey | and it's a click package, so it's not like other things on the platform are going to use the libs you ship inside the click; knowing that you're the only app using the lib, and knowing what platform it is going to run on, one can optimize the build as well, so you don't have all the modules you don't need, along with the one you do | 18:30 |
| bschaefer | dobey, its not click yet IIRC, i still need to talk to Saviq about getting SDL into the sdk framework | 18:30 |
| bschaefer | once that happens, yes i think moving to a static library would be fine | 18:30 |
| bschaefer | dobey, its still a legacy app atm | 18:30 |
| dobey | bschaefer: well right now anyone using SDL has to ship it in their click | 18:30 |
| * bschaefer doesn't know a lot about click packages :) | 18:31 | |
| dobey | there is no sdl lib on the phone image | 18:31 |
| bschaefer | yeah, hopefully at some point! | 18:31 |
| dobey | well, sdl will need a bit of work to have it be on the image i think; so people can ship additional modules they might need SDL to load, inside the click as they may not be part of the sdk/image | 18:32 |
| bschaefer | yeah | 18:32 |
| bschaefer | sdl2 is working pretty well (if things stoped changing! haha) | 18:33 |
| bschaefer | dobey, another issue, is theres nothing that really uses SDL2 atm | 18:33 |
| dobey | on the phone, or in general? | 18:34 |
| bschaefer | that works on the phone, besides small apps and tests | 18:34 |
| bschaefer | phone | 18:34 |
| bschaefer | dobey, i've quite a few steam games working on the desktop through SDL2/mir | 18:34 |
| dobey | yeah, i was just going to say, steam uses it pretty heavily :) | 18:34 |
| bschaefer | yup :) | 18:34 |
| dobey | hmm | 18:36 |
| dobey | didn't someone port vba to the phone? | 18:37 |
| dobey | although it might be broken too now | 18:37 |
| bschaefer | dobey, i think it depends on SDL1.2 | 18:38 |
| bschaefer | but i can test it out if i can get it to work | 18:38 |
| dobey | yeah. qvba-m | 18:40 |
| dobey | it's in the store | 18:40 |
| bschaefer | nice! | 18:41 |
| dobey | if you want to try porting something though, it looks like neverball is using sdl2 in utopic. porting that and adding support to use the accelerometer/gyroscope in the phone, should make it an interesting thing on the phone/tablet | 18:43 |
| bschaefer | dobey, o nice, yeah i can take a look at that | 18:44 |
| bschaefer | i've not checked if anything new got in for sdl2 | 18:44 |
| dobey | hmm, why is my cpu suddenly so hot :-/ | 18:46 |
| dobey | yay c++ :-/ | 18:47 |
| Saviq | bschaefer, me? I've nothing to do with frameworks :) | 18:56 |
| Saviq | TBH I'm not sure who to talk to about that | 18:57 |
| bschaefer | Saviq, yeah i was going to poke you to figure out who to poke haha :) | 18:57 |
| bschaefer | but ill poke around during the sprint | 18:58 |
| Saviq | bschaefer, we'll find out soon enough | 18:58 |
| Saviq | bschaefer, it's ubuntu-sdk-libs, comes from https://code.launchpad.net/~ubuntu-branches/ubuntu/utopic/ubuntu-touch-meta/utopic | 18:58 |
| bschaefer | Saviq, sweet! I can dig around that to figure out how the other are working atm. Thanks! | 18:59 |
| Saviq | bschaefer, looks like between rsalveti and ogra_ someone should be able to help with getting stuff into frameworks, now question is who is to make the call to add it :) | 19:00 |
| bschaefer | bregma, ^ some info | 19:00 |
| bschaefer | Saviq, right, and if it should be added or not :) | 19:00 |
| bschaefer | if i can get neverball working in SDL2 i think that would be a good start | 19:01 |
| bschaefer | Saviq, as atm theres nothing really working for SDL2 and the phone besides small apps and tests | 19:01 |
| bschaefer | though for the desktop theres a lot more | 19:01 |
| bschaefer | soo it could still be worth for it the desktop alone | 19:01 |
| Saviq | well, yeah, but adding it to the phone has implications we need to think of, it needs to be worth it for real | 19:02 |
| bschaefer | yeah | 19:02 |
| === dandrader_ is now known as dandrader|afk | ||
| rsalveti | ahayzen_: pong | 19:09 |
| ahayzen_ | rsalveti, I've been told you were working on this component ... When pressing volume up/down there is the new notification and a 'pop' sound. However the pop sound causes music to be paused, I assume it is using the wrong role? Is there a bug for this already or do you want me to add one, if so against which project? | 19:10 |
| rsalveti | ahayzen_: yeah, that is going to be reverted, and the proper implementation is in progress | 19:10 |
| rsalveti | ahayzen_ ted is taking care of that | 19:10 |
| ahayzen_ | rsalveti, cool thanks | 19:10 |
| rsalveti | we have bugs, just don't know the number | 19:11 |
| rsalveti | -etoomany | 19:11 |
| ahayzen_ | hah lol | 19:11 |
| dobey | bschaefer: just convince valve to write port steam mobile client to ubuntu too ;) | 19:13 |
| bschaefer | dobey, well if they make a steam mobile client it should be in SDL2 sooo it'll work on ubuntu :) | 19:14 |
| slvn | back! So this is not exactly the symbols that fails in my case, but the library name. It looks for libmirclient.so.7, but the tablet has libmirclient.so.8 | 19:16 |
| dobey | bschaefer: well, i don't know if it's in SDL2 itself, but it might provide SDL2 for games to use. it's available for android (and i think ios). | 19:17 |
| bschaefer | dobey, interesting, as their steam client is also written in SDL2.... i slightly assume everything from value is in SDL2 haha | 19:17 |
| bschaefer | slvn, well going to a static library should work... | 19:18 |
| dobey | slvn: so you built it on a 14.04 chroot? | 19:18 |
| slvn | dobey: android, ios, and also winRT, winPhone, Nacl ... | 19:18 |
| dobey | bschaefer: well the module in sdl probably will still want to dynamically load the mir lib itself (as it's part of the system), unless it's also explicitly linked statically | 19:19 |
| bschaefer | dobey, if you disable --shared, it'll load it in during compile. (So everything will be compiled with -lmirclient) | 19:20 |
| slvn | yes, that's fore sur the issue. my chroot is 14.04 | 19:20 |
| dobey | bschaefer: building against the right version might be better here anyway; might be other issues if you pull in libmirclient7 | 19:20 |
| bschaefer | slvn, make sure the confiure | 19:20 |
| dobey | slvn: yes, create a new chroot/kit that's 14.10 | 19:20 |
| bschaefer | configure/rules files is looking at the right version but ... i didn't specify a version there | 19:21 |
| slvn | yes, I took my old script/chroot. I did think about changing the version. I just update it | 19:21 |
| dobey | although that also presents an interesting problem | 19:21 |
| kenvandine | mterry, landing you wizard refresh branch is on my short list... sorry for the delay | 19:22 |
| kenvandine | mterry, had quite a few speed bumps today with landings | 19:22 |
| mterry | kenvandine, no worries, I get it | 19:22 |
| kenvandine | mterry, that'll be my next settings landing though | 19:22 |
| mterry | kenvandine, I'm just happy it's approved, when it lands is only bugging other people :) | 19:22 |
| kenvandine | started the day with ftbfs because of gdb... and gone down hill from there | 19:22 |
| mterry | kenvandine, :-/ | 19:22 |
| kenvandine | like unintended silos getting published :/ | 19:23 |
| kenvandine | mterry, it'll probably be tomorrow morning at this point though | 19:23 |
| dobey | bschaefer: yeah, steam mobile is just android/ios: http://store.steampowered.com/mobile | 19:23 |
| mterry | kenvandine, oh no, a silo got through accidentally :-/ | 19:23 |
| bschaefer | dobey, :( | 19:23 |
| dobey | bschaefer: but it's also proprietary, so i have no idea what they're actually doing in terms of displaying things to the screen | 19:24 |
| nhaines | The only thing the Steam Mobile client does is catalog, purchasing, and instant messaging. | 19:24 |
| bschaefer | dobey, yeah, sdl1.2 had ios support IIRC but i dont see it in sdl2 | 19:24 |
| dobey | bschaefer: they could just be embedding SDL2 inside an NSWhatever on ios | 19:25 |
| dobey | oh | 19:25 |
| bschaefer | dobey, yeah, hmm i can attempt to check it out, i've had to deal with proprietary games from value before | 19:25 |
| slvn | I am confused about the chroot. I actually have a "partial" chroot. I use scripts from MIR (i guess), that are called : "setup-partial-armhf-chroot.sh", "cross-compile-chrool.sh" and "debian/control". I see not trace of 14.10 or 14.04. how to change the version ? | 19:25 |
| bschaefer | dobey, o sorry, it does have coca support | 19:26 |
| dobey | slvn: are you not building your app in qtcreator? | 19:26 |
| bschaefer | cocoa* | 19:26 |
| slvn | dobey. no. I build the SDL with the partial chroot. | 19:27 |
| slvn | then I build my app with SDL | 19:27 |
| slvn | when I tried a few months ago, that was not possible to use qtcreator for this | 19:28 |
| slvn | the partial chroot, has indeed the lib : ./usr/lib/arm-linux-gnueabihf/libmirclient.so.7 | 19:29 |
| dobey | slvn: if you integrate the SDL build as part of buidling your app, rather than building separately, and ensure the lib gets installed to the right place to be pulled into the click, it should work fine in qtcreator | 19:30 |
| slvn | dobey: I would rather use no IDE ... The application now works because SDL has few dependencies. My application work with the hack of finding the correct libmirclient | 19:35 |
| slvn | the line : "fakeroot debootstrap --include=${builddeps} --arch=armhf --download-only --variant=buildd trusty ." | 19:37 |
| dobey | slvn: you don't need to use qtcreator to actually work on the code. but qtcreator does make it much easier to build click packages for deployment to the device | 19:37 |
| slvn | should be replaced with the 14.10 version ? | 19:37 |
| dobey | slvn: yes, s/trusty/utopic/ | 19:38 |
| slvn | I like script because of this :) | 19:39 |
| === VargaD_ is now known as VargaD | ||
| charles | AlbertA2, AlbertA: could you see about getting https://code.launchpad.net/~charlesk/unity-system-compositor/lp-1365557-decrement-display-on-requests-correctly/+merge/237498 into the next system compositor silo? | 19:49 |
| AlbertA2 | charles: yes we are gated | 19:50 |
| AlbertA2 | by the mir landing | 19:50 |
| charles | woo | 19:50 |
| dobey | #landinggate | 19:50 |
| charles | I didn't see that. Thanks :) | 19:50 |
| AlbertA2 | charles: np, as soon as that's done I'll land it | 19:51 |
| charles | AlbertA2, #landinggate ? | 19:51 |
| dobey | bad joke | 19:53 |
| === VargaD_ is now known as VargaD | ||
| kenvandine | mterry, it's building in silo 8, did you confirm the test plan is current with your changes? | 20:17 |
| mterry | kenvandine, yeah... The wiki test plan for the wizard is pretty bare -- it's mostly "run through the wizard" | 20:18 |
| kenvandine | haha | 20:18 |
| kenvandine | ok | 20:18 |
| mterry | kenvandine, speaking of which, it really needs some solid AP tests | 20:18 |
| kenvandine | yes... yes it does :) | 20:19 |
| slvn | Thanks for the help ! bye | 20:40 |
| === dandrader|afk is now known as dandrader | ||
| vitimiti | Bye | 23:47 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
