=== m_conley is now known as m_conley_away [05:19] Good morning [05:20] good morning [05:25] morning jibel [05:25] morning dpm === Tm_T_ is now known as Tm_Tr [06:33] didrocks: morning. PPA arm builders are being cranky, webops are looking into it. meanwhile there'd be package acks http://pastebin.ubuntu.com/6153293/ (indicator-sound, trivial) + http://pastebin.ubuntu.com/6153294/ (libusermetrics, adds QML plugin package) [06:35] Mirv: hey! looking ;) [06:35] Mirv: both are on the landing list? [06:35] didrocks: yes (whole indicators stack), and assigned to me [06:36] + A set of bindings allowing the use of libusermetrics from QML applications. [06:36] this is less than 80 char? [06:36] * didrocks test [06:36] which is good, because usually after the meeting and luch I don't have that many hours left [06:36] oh yeah 77 ;) [06:36] Mirv: yep ;) [06:36] +1 on libusermetrics [06:36] indicator-sound doesn't explain why, I guess Ted has a reason not mentioned [06:37] oh, that discussion on indicator-sound, yeah, +1 [06:37] ok [06:37] Mirv: it was a veryyyyyy long discussion on #ubuntu-devel [06:37] with drama and all that :p [06:37] a lot of drama for one-line change, ok :) good that it has been discussed. [06:37] yeah, threating of revert and so on. You missed a lot of fun :) [06:37] a summary in changelog entry could prevent someone making it the other way around in a couple of months [06:38] but on the plus side we could get more drama to watch [06:39] yeah, I'm balanced because of the drama side :) [06:39] agreed though that no description is bad… [06:43] to be more precise bradm and wgrant are the people looking at the arm builders [06:45] I noticed libusermetrics needed cu2d-skip on powerpc, and the builds were stuck. I skipped it and then after a while noticed these arm problems. [06:45] asac: ^ you're probably interested too in that at the moment all armhf builds are broken. [06:45] last successful build was 5 hours ago [06:57] Morning [06:57] hi sil2100! are you feeling any better already? [06:59] Mirv: hi! A little bit, yes, decided that I can't be laying around forever [07:00] sil2100: ok, good that you have at least some positive progress === tkamppeter__ is now known as tkamppeter [07:04] Damn, I was so 'detached' from life that I didn't even know what day of the week is today ;) [07:06] oh, news alert, arm builds functional again thanks to wgrant [07:09] Mirv: hi [07:09] whats the problen? [07:10] asac: the PPA armhf builders (kishi*) were broken, I contacted webops and they just managed to fix it. something clashing between production <-> staging. [07:11] wow [07:11] Mirv: well done gettingt his fixed :) [07:11] thanks [07:11] absolutely good news [07:11] Mirv: do we have a very bad backlog because of this? [07:12] asac: started this morning only, most of the stuff are cleaned [07:12] asac: not terribly, the last good builds were 6h ago and I used those for click and indicators stack testing. I'm now waiting for the new builds of phone and sdk stack to test. [07:14] dee-qt was also manually uploaded so I synced its changelog as usual for this process-overriding uploads [07:14] right. i dropped the comment about that in the landing plan/sheet [07:14] was needed to get mir promote to saucy [07:15] it's found easiest from the cu2d jenkins, 'yellow' blob tells something is wrong and details include the tidbit that there's a changelog diff [07:15] Mirv: you mean looking at the dashboard? [07:15] hmm. thats not yellow :) [07:15] http://reports.qa.ubuntu.com/daily/ [07:16] thats redish [07:16] no, the jenkins http://10.97.0.1:8080/view/cu2d/view/Saucy/view/SDK/job/cu2d-sdk-saucy-1.1prepare-dee-qt/ [07:16] it was also visible at the main sdk page http://10.97.0.1:8080/view/cu2d/view/Saucy/view/SDK/ before I fixed it [07:16] asac: I told you many times that the dashboard shouldn't be trusted for daily releases [07:16] didrocks: i know [07:16] http://10.97.0.1:8080/view/cu2d/view/Saucy/ btw, almost all yellow/green [07:16] was just confused because mir said that he is loking for yellow [07:16] (yellow most of the time because of manual publication) [07:17] well, jenkins has colors :) [07:17] right. but i am living in the qa dashboard bubble, so if someone mentions yellow without specifying furtger, my brain just picks that [07:19] asac: I live in the cu2d bubble and handle everything via that since it works well for the packages that are handled there :) [07:19] hehe yeah. i should have known that :) [07:19] the more we can get back to handling cu2d handled packages directly in cu2d (sans device testing), the better [07:19] guess it has to do with still waking up mode as well on my side :) ... /me goes and tries to make a caffeinated drink [07:22] Mirv: do we know what happened with the builders? [07:22] asac: not much more than 20130925|10:05 < wgrant> Mirv: Should be fixed now. staging decided to have an argument with production. [07:22] half an hour before that "yeah, its very odd" [08:04] morning all [08:21] desrt: sorry I didn't get back to you yesterday. ping me when you arrive and let's figure this out [08:24] can i get environment variables from qml? [08:25] not that I know of [08:30] good morning desktopers [08:30] hey Laney chrisccoulson, how are you? [08:30] hey seb128 [08:31] tired, kept up a lot of the night by students(!) [08:31] first week of term this week [08:31] hey seb128 [08:31] lut didrocks [08:31] Laney, did they pay you beers, or just made noise? ;-) [08:31] haha [08:31] I bet they will have left some presents if I go and look at the road [08:31] hehe [08:33] * Laney loads up on tea [09:22] didrocks: could you pre-ack ui-toolkit whie I'm still running a couple of tests? http://10.97.0.1:8080/view/cu2d/view/Saucy/view/SDK/job/cu2d-sdk-saucy-3.0publish/lastSuccessfulBuild/artifact/packaging_changes_ubuntu-ui-toolkit_0.1.46+13.10.20130925.1-0ubuntu1.diff [09:23] Mirv: acked ;) [09:23] thanks :) [09:23] yw! [09:36] idle-delay on the screensaver capplet seems to reset itself to 'never' after closing the capplet, but dconf-editor still shows the value I want [09:36] and screensaver doesn't kick in [09:38] tjaalton, what is "the screensaver capplet"? [09:38] shrug [09:38] * seb128 cleans old kernel and get 3Gb of disk space back [09:38] seb128: brightness & locking [09:38] it's quite ridiculous how those stack until filling your disk :/ [09:39] tjaalton, talk to darkxst or jbicha, likely another fallout of g-s-d 3.8 [09:39] ok, thanks [09:40] tjaalton, you check the value of org.gnome.desktop.session idle-delay, right? [09:40] yep, it's still whatever I set in the capplet [09:40] then when I reopen the capplet it's set as 'never' [09:41] but the gsettings config is still correct? [09:41] yrd [09:41] uh [09:41] yes [09:41] k, weird [09:42] could be a bug in the g-c-c Ui then [09:44] it's still at 3.6.3 [09:45] that's normal [09:45] why? [09:46] probably not guaranteed to work with g-s-d 3.8? [09:46] darkxst and jbicha spent weeks to get the g-s-d 3.8 update in shape and make sure it works with g-c-c 3.6 and they said they would deal with the fallout of updating g-s-d [09:47] so please open a bug report describing the issue [09:47] ok, sure [09:49] thanks === vrruiz_ is now known as rvr [10:02] seb128, thanks for uploading Cairo. [10:03] tkamppeter, thank you for backporting those fixes [10:09] seb128, do you know how do i figure out why is dee stuck in proposed? [10:10] mhr3, http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html [10:10] mhr3, basically blocked by beta2 freeze [10:10] Read ubuntu-devel-announce ;-) [10:10] but it's specifically for touch [10:11] The package is on desktop too though [10:12] so it won't get in until the freeze passes? [10:12] which is...? [10:17] hi all, can i ask a question? i've looked in the forum and google and can't find a similar case... [10:17] mhr3, tomorrow if everything goes ok [10:18] EtgarDizz, hey, try #ubuntu for user questions [10:18] mhr3: Tomorrow; but feel free to go argue why this is safe to let through in #ubuntu-release if you like :-) [10:18] is this channel only for dev? [10:18] yes [10:18] ok thanks :) [10:19] well, there is this thing that dee-qt was already released, even though it needed to be built with the not-yet-released dee [10:20] fortunately it "only" causes some runtime warnings and works otherwise... [10:23] seb128, btw the unity i386 problem suddenly and magically disappeared over the night [10:23] normally i'd go wtf... but starting to consider those things usual [10:24] :-/ [10:24] good that it went away at least I guess === smartboyhw_ is now known as smartboyhw [10:49] so we'll miss logind again with systemd > v205? === MacSlow is now known as MacSlow|lunch [10:54] reading the blog post by ovitters === greyback is now known as greyback|lunch [11:10] mhr3: right, dee is still in proposed since it affects desktop [11:15] tjaalton, let's see how that plays out... [11:15] tjaalton, I'm sure we can figure it out (or stay on our version for a while, priority is on mobile atm and what we have is working fine enough) [11:16] it's causing issues all over the place though [11:16] now sssd depends on logind [11:16] guess i'll just cripple the build on ubuntu, debian will have it all [11:16] tjaalton, ? [11:16] playing catch-up all the time [11:16] tjaalton, we use logind, do you mean they require > 204? [11:17] right now it should work [11:17] so why do you need to "cripple the build on ubuntu"? [11:17] talking about the next release [11:17] tjaalton, note that debian didn't go systemd yet afaik [11:18] so I'm not sure why you think you are better off there [11:18] That part of his blog post isn't right [11:18] i don't run debian, just maintain the package on both :) [11:18] They didn't go for systemd [11:18] the debian gnome team? [11:19] I know debian didn't choose anything yet [11:20] I don't think the GNOME team gets to choose the default init system [11:20] They've been doing as we are, making stuff require logind only [11:25] right [11:28] * Laney uploads a glib-networking with autopkgtests to debian [11:28] Thanks Upstream™ [11:32] Laney, great ;-) [11:46] seb128, tjaalton, I will look into it === greyback|lunch is now known as greyback [11:48] darkxst: thanks [11:48] darkxst, thanks [11:56] Does anyone knows why there is no network indicator icon in gnome-session-flashback? https://bugs.launchpad.net/indicator-network/+bug/1229294 [11:59] alberts, yup I know why [12:01] Bug 1189309 [12:01] wrong patch was uploaded, and I haven't got around to sorting that out === MacSlow|lunch is now known as MacSlow [12:11] darkxst: there is no icon, because nm-applet is not started? when correct patch will be uploaded? [12:21] sil2100: do you mind joining #ubuntu-ci-eng? [12:21] alberts, I will get to it soon [12:22] darkxst: ok. thanks for info! for now i just edited nm-applet.desktop file. [12:24] didrocks: ok, joining [12:24] Lunching now ;) [12:26] sil2100: thanks! [12:44] Laney: awake. [12:44] Laney: can you share the review url with me again? [12:47] desrt: Got to lunch now :( [12:47] But it is https://code.launchpad.net/~laney/ubuntu-system-settings/measure-real-directory-size/+merge/186844 [12:47] leave comments here or there and I'll pick them up when back [12:48] Laney: k [12:49] as i way saying before, my main issue is your weird use of pointers in 'struct MeasureData' [12:49] + uint *finished; [12:50] there should be no '*' here. it should just be a uint... and it should count down rather than up [12:50] ditto 'size' [12:54] okay. i see why you did it with 'size', but this is dangerous because you don't bind the scope of the object to the scope of the async call -- so if the async call completes after you've closed the window, you're gonna have a bad time [12:55] might make sense to use a cancellable here to deal with that case (you could fire the cancellable from ~StorageAbout) [13:15] seb128: hello :) [13:17] sil2100, hey, how are you? feeling better? [13:18] seb128: a little bit, but I guessed it's better to try and do something productive instead of laying on bed [13:18] seb128: I have a question regarding the Bluetooth settings in u-s-s [13:18] that's the spirit ;-) I hope you get better soon! [13:18] sure [13:18] seb128: do you know if after it landed in trunk, it should work? [13:18] it should [13:18] seb128: i.e. switching bluetooth on and off [13:19] seb128: does it have some specific requirements in other packages? [13:19] it's reusing the indicator backend [13:19] indicator-bluetooth and qmenumodel [13:19] which are both already on the image === bcurtiswx is now known as Guest93821 [13:20] hmmm [13:20] sil2100, it's not working for you? [13:20] i seem to be hitting something similar to bug 871667 [13:21] Guest93821, hi, try #ubuntu for user questions [13:22] seb128: it seems not to work on my device, I have the latest bluetooth, indicator-bluetooth and ubuntu-system-settings (from daily-release), and switching bluetooth on/off in the settings doesn't change the state on the indicator [13:22] They're out of sync [13:23] sil2100, is bluetooth working at all on that device? if you close system-settings and reopen it, what status does it have? [13:24] seb128: it seems to be enabled all the time - when I disable it in the settings, close the settings and open again - it's ON again [13:24] charles, tedg: ^ can you help to debug that? [13:25] sil2100, what device do you have to test? [13:25] seb128: I'm testing on mako - nexus 4 [13:26] sil2100, can you adb shell it and check in /home/phablet/.cache/upstart/*.log if there is any error [13:26] could be unity8.log [13:26] or maybe the system logs [13:27] Looking [13:27] hah [13:27] Can't open /dev/rfkill for use as a killswitch backend: Permission denied [13:27] rfkill issues it seems [13:27] Couldn't write rfkill event: Bad file descriptor [13:28] :/ [13:28] well, I guess that's an issue that affects the indicator as well [13:28] can you open a bug and assign it to charles? [13:28] charles, tedg: ^ did you test bluetooth on a device/mako? [13:29] seb128: doing [13:29] sil2100, thanks [13:29] I've tested connecting on the Galaxy Nexus... I don't know that I shut it off. [13:29] sil2100, open it against indicator-bluetooth, that's where the backend code is [13:29] is the .Xauthority file desktop related? [13:29] or more the X team ? [13:29] bcurtiswx_, it's from lightdm/xorg [13:29] bcurtiswx_, why? [13:30] its not letting me login to my desktop session.. :( [13:30] rm it [13:30] yeah, it's a workaround to a larger issue (2nd time this has happened) [13:30] idk what the larger issue is [13:30] sometime there is a leftover with wrong permission that can create login issues [13:30] seb128, is there a reason for indicator-bluetooth to not have permissions for /dev/rfkill? [13:30] seb128: whats the perms on your .Xauthority file ? [13:30] charles, I've no idea how are the acl to that handled [13:30] pitti, ^ do you know? [13:31] bcurtiswx_, 664 user:user [13:31] seb128: mine got dropped to 600 user:user [13:32] bcurtiswx_, well, if it's not root owner it should be ok... [13:33] well, removing it lets me login again [13:33] odd [13:34] jdstrand: hm, aren't we past ui freeze? A red sound indicator for 13.10 is a bit late... [13:34] jdstrand: I don't understand why this is such a big issue now - we've been living with that on the desktop forever [13:35] seb128, sil2100, it looks like that's been read/writable by i-bluetooth in the past [13:35] larsu: it is a big issue because appstore apps receive no peer review, no code review, no nothing. as such a malicious app can record audio and send it off over the internet [13:36] seb128, sil2100, for example, that's how gnome-bluetooth does soft disables too [13:36] charles, well, the touch system images are new world, they are ro to start [13:36] charles, also currently polkit is not working because logind registration is buggy [13:36] jdstrand: same on the desktop when using non-free apps [13:36] charles, could either of those have an impact? [13:36] larsu: the appstore provides a direct channel for 3rd party developers to users, with Ubuntu/Canonical not making any guarantees or promises on the apps quality or behavior [13:37] seb128, probably so [13:38] larsu: yes, but on the desktop we have the Ubuntu archive. it is more difficult to get things into the archive. the archive is considered trusted and receives many eyes along the way. an app developer can upload an app into the store in minutes with no vetting or code review [13:38] jdstrand: and these apps have direct access to pulse? I remember people talking about a higher-level API for outputting music, which would run everything through a daemon. If we have that, then apps can't access the mic either, can they? [13:38] didrocks: one more thing, something very fishy with libfriends on armhf, build hangs eternally and blocks cu2d [13:38] jdstrand: yeah, fair enough. [13:39] Mirv: can you ask to the webops channel? [13:39] Mirv: they can ssh and sees I guess [13:39] jdstrand: don't get me wrong, implementing this will be fairly simple. I just don't see the point of it yet :) [13:39] seb128, i just received this message from launchpad: http://paste.ubuntu.com/6154476/ [13:40] larsu: apps typically use a higher api, but that api is available in libraries that apps use, as such, the process uses pulse. a malicious app can do whatever it wants and access pulse directly. but that doesn't matter-- right now, apps can use standard apis to record audio and there is no visual cue [13:40] I canceled the jenkins job to let others continue, but sil2100 didrocks next tick in 20 minutes won't start probably since apps+media only just started, unless you stop them [13:40] Mirv: ok, let's get that one building and we will see if we manually trigger another tick [13:40] thanks for the head's up [13:40] charles: https://bugs.launchpad.net/indicator-bluetooth/+bug/1230275 [13:40] larsu: well, it is a hard requirement that apps not be able to eavesdrop in the background [13:40] Mirv: thanks! [13:41] didrocks: ponged webops with https://launchpad.net/~ubuntu-unity/+archive/daily-build/+build/5048682 [13:41] larsu: s/apps/confined appstore apps/ [13:41] jdstrand: this is in direct opposition to what I've heard before... Oh well, if mpt is fine with the red icon, I can do it this week. [13:41] mpt: ^^ [13:41] good :) [13:41] didrocks: lol, I just thought you meant me [13:42] larsu: webops? ;) [13:42] sil2100, thanks [13:42] didrocks: no, just you typing "good" after me talking to jdstrand [13:43] :) [13:43] attente, right, feel free to ignore it, or to check if the issue it points at is fixed in upstream git ... it's basically a buggy .po [13:43] larsu: part of the consideration is also that in the Ubuntu trust model, apps in the app store are untrusted. users don't have a context-less permissions prompt at install time (like on andriod), and therefore we need to be tasteful and appropriately, contextually cue the user for access to various apis [13:44] jdstrand, we shouldn't let background app do audio recording to start... [13:44] jdstrand: I know and I fully agree. I've just heard that apps cannot access pulse *at all* [13:44] jdstrand, how can that even happen btw, there is not "background app", right? [13:44] larsu: some apis, like online accounts or location, will have a runtime prompt that asks the user if the app can access the api (and cache the result so it is only once) [13:45] larsu: we could do the same with pulseaudio recording, but it seems simpler and more appropriate to give the visual cue [13:45] larsu: apps can access pulseaudio [13:45] jdstrand: yeah, so I've learned two minutes ago :) [13:45] Laney, desrt: I don't know what that get dir space code is doing but it made my system load to 6 with kswap0 hitting lot of resources [13:45] larsu: :) [13:46] jdstrand: I think we should prompt tbh. The red icon doesn't prevent anything, it only notifies. What if I'm not looking at my phone? [13:47] seb128: I think this also about foreground apps [13:47] seb128, yes, seems it's a problem with the upstream .po [13:47] seb128: well background vs not background doesn't really matter. technically it is currently possible for an app to laucnh something from its install directory in the background. in the future, we want to support that anyway with application lifecycle. even if that was not the case, a foreground app (eg, an addictive game) could be trojaned to listen to the user [13:48] jdstrand: wait, aren't games full screen? [13:49] larsu: on touch they don't seem to be-- I see all the indicators [13:49] asking the user "do you wish this recorder app to access the microphone" is a pretty dumb thing to do. [13:50] if the game goes full screen without indicators and records, there needs to be some visual cue for that too [13:50] mdeslaur: I don't see another way to prevent an app from recording audio if we don't have up-front permission settings [13:50] do we really care if the foreground app is accessing the microphone? [13:50] there is possibly an argument to both notify and prompt, but notify seems like enough [13:50] larsu: easy: don't allow background apps to access the microphone [13:51] mdeslaur: ya, this is what I'm arguing here [13:51] larsu: this isn't about enforcement-- this is about the user knowing what is happening [13:52] jdstrand: do we do something similar for the camera? [13:52] mdeslaur: I brought that up in the bug. I don't think there is a service for the camera [13:52] ok [13:52] jdstrand: you just said it's "a hard requirement that apps not eavesdrop in the background" [13:52] it is also something to consider [13:53] jdstrand: if we're only notifying the user, it might already be too late [13:53] larsu: I didn't phrase that correctly. it is a hard requirement that apps not record without the user knowing they are being recorded [13:54] recording without knowing is eavesdropping [13:55] desrt: how can it just be a uint? I make n copies of the struct and want to use the same counter for all of them [13:55] I think a visual cue is worthwhile on desktop and everywhere else. it is a meaningful improvement. a user needs to know if an untrusted app is recording them. I think the visual notification is enough in the short term. maybe a ontextual prompt is also warranted [13:56] jdstrand: fair enough. [13:56] but that can be added later [13:56] jdstrand: this leaves two issues for me: (1) I think the red icon is not clear enough. (2) full screen apps. [13:56] both of these can be solved by mpt [13:56] yes [13:57] I am not advocating what the visual cue is [13:57] who will most likely also be able to tell me wether I should use "fullscreen" or "full screen" [13:57] turning the icon red is just an easy way to describe it [13:57] it also happens to be very simple to implement :) [13:57] larsu, https://wiki.ubuntu.com/AccountPrivileges?action=diff&rev2=13&rev1=12 [13:58] mpt: hm, jdstrand just argued against that [13:58] larsu: as for UI freeze-- I will navigate that. if someone block it for 13.10, then we still need it for 14.04, so the time isn't wasted-- but if the changes and the cue is small enough, there shouldn't be an issue [13:59] jdstrand: ok [14:00] mpt: when an app requests it, the system would pop up a dialog, right? [14:00] larsu, right, just like it does on iOS [14:00] (I assume Android asks at install-time) [14:01] mpt: (a) this is probably not implementable before 13.10. (b) jdstrand and mdeslaur seem to strongly dislike it. [14:01] larsu: well, I'm not a designer...if design is fine with it, fine [14:02] * mpt reads backlog [14:04] adjusting pulseaudio to prompt for access is possible, but not for 13.10. it would need to interface with trust-store [14:05] (which is not implemented) [14:05] right, that's what I assumed [14:05] also, I've seen on other platforms that they give a visual cue. I know i've seen it on android, but I forget the circumstances when it is seen [14:05] sil2100, do you want to work on https://bugs.launchpad.net/ubuntu-system-settings/+bug/1230295 ? (hiding the welcome screen image selector from the background panel) [14:07] seb128: yessir! Assigning that to me :) [14:07] sil2100, excellent, thanks ;-) [14:09] jdstrand: really? I've never seen anything like that. There is a permission for using the microphone, though. [14:09] jdstrand, does "it is a hard requirement that apps not be able to eavesdrop in the background" equal "it is a hard requirement that Skype should stop working whenever you switch to another app"? If not, how? [14:10] larsu: like I said, I don't know the circumstances of where I saw it [14:10] mpt: so I said later that I didn't phrase that correctly [14:11] mpt: 08:53 < jdstrand> larsu: I didn't phrase that correctly. it is a hard [14:11] requirement that apps not record without the user knowing [14:11] they are being recorded [14:11] hi, on edubuntu (saucy) the dash isn't returning any results. any suggestions on how I should troubleshoot that? [14:11] mpt: that doesn't necessarily mean a visual cue, but it could. it doesn't mean we can't prompt, but it might [14:11] mpt: I too am not a designer [14:12] jdstrand, okay. I just made my first ever audio call with Skype on my phone, and got the "Skype wants to access your microphone" prompt. Once I accepted, there was no extra notification *until* I switched to another app. [14:12] mpt: what I don't want is for someone to install an app, and because we don't do any context-less installation permissions prompting, the user installs something and has no idea it is or can record them [14:13] Laney: oh [14:13] mpt: oh, you got a microphone prompt? what platform was this on? [14:13] Laney: that's weird :) [14:13] mpt: what was the extra notification once you switched to another app? [14:13] mdeslaur, iOS. [14:14] mpt: so you got two prompts? [14:14] Laney: so i guess i just have my original comment about the counting down rather than up [14:14] mdeslaur, larsu: It's the same in Skype, Heard, and the built-in phone app: a pulsing ribbon along the top [14:14] each time you dispatch an op add one to the int [14:14] random strange question, when you click restart on your machine you're presented wiht two large squares. One square is pre highlighted which is shut down, which is confusing as you clicked restart. Anyone else seeing this on saucy ? [14:14] each time one comes back, subtract one [14:14] finish when zero [14:14] So I shall find out what our current UI for returning to calls is [14:14] that way if someone comes along and adds another directory to check, they don't need to find the magic 4 hidden elsewhere [14:14] desrt: that could result in the event firing too early [14:15] if one finishes really fast [14:15] Laney: async doesn't work that way [14:15] the result comes back to the same thread [14:15] yeah? [14:15] and you're sending them all out at once [14:15] mpt: so you get the prompt, and if you accept, you get a ribbon? [14:15] fair enough [14:15] so they'll all go out before the first one has a chance to come back [14:15] I'll do that [14:15] my comments about the lifecycle of your object still stand, though [14:15] if you get destructed while the measure operations are in progress (which is somewhat likely since they're slow) you're gonna crash [14:15] mpt: you only get the ribbon if a _background_ app is accessing the microphone, right? [14:16] desrt: Ah, yes, I can cancel it in the destructor though presumably [14:16] yes [14:16] use a GCancellable [14:16] wilco [14:16] and from the async result function, first check the error for having CANCELLED set [14:16] if it was cancelled, drop it on the floor and avoid touching your (now freed) data [14:17] mdeslaur: the way I've understood it: he gets a prompt when an app wants to use the mic for the first time. When backgrounding an app that is recording, he gets the visual feedback [14:17] seb128: is this a bad thing? [14:17] larsu: right [14:17] interesting [14:17] * larsu is trying to find a video of that feedback online [14:17] seb128: the kernel is more efficient at IO when it has a lot of it to do because it can make better scheduling decisions and hide latencies better [14:18] desrt, well, I guess it's usually "linux sucks an handling io load" [14:18] seb128: so Laney thought (and i agreed) that it would be best to dispatch all of the measure operations at the same time [14:18] desrt, my mouse cursor hangs for seconds [14:18] hrmph. [14:18] that does sound worrying [14:18] load hits 6 [14:18] load hitting 6 is fine [14:18] with the cpu usage for system settings being 130% in top [14:19] maybe i should see if i can figure out a way to set the IO priority of the thread to be a bit lower [14:19] well, in any case it's going to be fine on the device [14:19] I got a bug on my laptop anyway [14:19] what's creating the issue on my laptop is that it's trying to get the space of ~ [14:19] doesn't count / [14:19] as a drive to calculate the sapce of [14:19] it takes a few seconds to take the xdg dirs [14:19] i hope you're not using measure() on / [14:20] there are far better ways to determine this [14:20] no, but I am for the home directory [14:20] homedir is tricky [14:20] does the phone have separate /home by default? [14:20] Laney, I think you should get the space of /data on the phone (write so on the merge request btw) [14:21] mpt: ah, apparently this has been added in ios7 [14:21] what is /data? [14:21] the rw part of the fs layout I think [14:21] but better to check with stgraber [14:21] mpt: not sure if you saw it in the bug, but I think we are going to need to consider the camera as well [14:21] do all cameras have hardware leds? [14:22] I love how all of this is coming up a couple of weeks before the release... [14:22] jdstrand, mdeslaur: Yes, the ribbon is only for a background app using the mic. Android apparently does the same thing. [14:22] mdeslaur: the nexus 4 doesn't show an led in ubuntu [14:23] mdeslaur: I guess it could be made to do so [14:23] seb128: I think /data is android stuff [14:23] I guess you mean /userdata [14:24] jdstrand, mdeslaur: I was hoping we could just reuse whatever the built-in phone app did for calls when you switched to another app. But the design for that isn't finished. X-/ [14:24] Laney, yes, sorry [14:24] but /userdata/user-data is the interesting bit which is the same as the home directory anyway [14:24] the rest of it is "Used by Ubuntu" [14:24] jdstrand: actually, I don't think the front facing camera even has a led [14:24] Laney, used by Ubuntu is a separate mount/partition no? [14:25] Laney, e.g that's the size of the base image [14:25] Laney, can't we get the ~ as being "disk - - ubuntu image"? [14:25] - free space [14:26] I am trying to avoid saying things which only make sense on touch devices [14:26] jdstrand, I don't see a mention of camera in the bug report, but it makes a lot of sense that it would have the same prompt and/or reminder that the mic does [14:26] Laney, well, anyway I doubt devices are going to have enough datas to be an issue as it is on a desktop config [14:26] so Used by Ubuntu is total space - free space - home directory - click packages [14:27] seb128, Laney: /data is the Android data partition (/data in the android container), /userdata is the physical partition [14:27] mpt: I mentioned it in a comment down below. that bug was about pulseaudio, so I didn't mention more about the camera [14:27] seb128, Laney: Android partitions are usually symlinks to /android/, if that helps [14:27] mpt: I need some more info about the camera though, which I am gathering now [14:27] stgraber, thanks ... is there any partition that correspond to the user dir? [14:27] I think we could declare it OK for devices and try to ionice the priority down [14:28] in the future [14:28] Laney, wfm [14:28] and also maybe use the progress reporting to show you that it is working [14:28] at least it's good enough for v1 [14:28] make the bar fill up or something [14:28] on the phone, if there is only one user and we have a separate home partition, ... [14:28] seb128: the home directory is stored on the userdata partition, so if you need to check how much space is left for the user, that'd be the one [14:28] measuring /home is a silly idea [14:29] er. measuring ~/ i mean [14:31] ubiquity-dm is not loading the following indicators: sound, keyboard, system. How should I go about troubleshooting that? as far as I remember there was suppose to be a "ubiquity" profile to load those indicators... or something =) [14:31] larsu: ^ [14:32] xnox: yep, add the profile to the indicator files [14:32] xnox: they're in /usr/share/unity/indicators/*, or in data/ in most indicator source trees [14:33] mpt: thought (which you probably are already thinking of)> when considering design-- there are three cases-- mic only, camera only, mic+camera [14:34] larsu: hm, ok. [14:34] ChrisTownsend: I not sure I understand your compiz fix...I can no longer switch workspaces by clicking on an app in the launcher...is that intended? [14:34] xnox: "hm" doesn't sound good. I can whip up patches for them in a bit if you want [14:35] jdstrand, yeah, and it seems to me they should have the same UI. Maybe that's what triggered my instinct against using the sound indicator: there's no camera indicator. :-) [14:35] yeah [14:35] (Or video indicator, for a more precise analogy.) [14:35] mdeslaur: That is a regression in my "fix":-(. I have a fix upstream to take care of the regression, but we have to wait on distro to release a new Compiz package. [14:36] desrt: I'd like to avoid making assumptions about the partition layout if possible [14:36] I could over report "used by Ubuntu" and just count the XDG directories [14:36] mdeslaur: https://bugs.launchpad.net/compiz/+bug/1228352 [14:37] ChrisTownsend: ah, great, thanks :) [14:37] mdeslaur: I'm really sorry about this. [14:37] kinda silly that we don't have a good way to know how big a directory is... [14:37] i guess it's a bad idea for the same reason that noatime is a good idea [14:37] ChrisTownsend: a bug is fine...I was afraid it was a conscious decision :) [14:38] mdeslaur: [14:38] mdeslaur: Oops. I was wayyy too agressive on fixing something else. [14:39] jdstrand, so the *draft* returning-to-calls design I just saw uses a special temporary indicator. That doesn't solve the screencast problem (perhaps an app needs reviewer blessing to hide it), but it would mean that Skype for example would be consistent for background voice call vs. video call. [14:40] And its menu could list all the apps that are currently using the camera or mic, e.g. if you were running Skype and Heard at the same time. [14:44] mpt: this very much sounds like the counterpart to the sound menu [14:47] larsu, how so? [14:47] mpt: telling me about all the things on the system that are outputting sound (even though the sound menu doesn't _really_ do that) [14:47] mpt: s/outputting sound/using some subsystem/ [14:48] I'd argue that the sound menu should be doing that btw, but I'm sure you'll disagree ;) [14:48] having designed the sound menu and all… [14:50] Laney: seems gst 1.2 is not yet in debian, will test once the packages are available in there [14:50] Laney: do you have a FFe for it already? [14:50] rsalveti: yes it is [14:50] and no I don't [14:51] larsu, the spec implies (perhaps wrongly) that only one Touch app will be playing audio at a time. By "audio" I meant music/video/podcasts/streaming. [14:51] rsalveti: I just started the packaging but I'm now fixing another MP so I'll get back to it later [14:51] Laney: cool, sync happened in launchpad a few hours ago [14:51] cool [14:54] mpt: right, this was my understanding as well (only one app plays something at a time) [14:56] larsu, oh, so a list of one. Right. :-) You might want more than one app using the mic at a time though. [14:56] mpt: really? [14:56] larsu, e.g. recording a call. [14:57] mpt: good point, that might be a separate app. [14:58] mpt: anyway, do we have a decision regarding that make-the-sound-indicator-red bug? [14:58] * larsu would like to know whether he should get coding [14:59] jdstrand: actually, I guess this ^^ is more a question for you. Matthew amended the spec, but that can't be implemented before 13.10. [15:02] attente, hey [15:05] larsu, mpt: I'm not clear on the changes. I understand we want a prompt. do we plan to have a visual cue too? [15:06] jdstrand, what was your conclusion from the camera investigation? === m_conley_away is now known as m_conley [15:08] mpt: I am about to file a bug on it. right now the camera is happening on the android side. however, for 14.04 the plan is to have it on the Ubuntu side. rsalveti and I feel that there will likely need to be a service that apps would contact. therefore, that service is, like pulseaudio, in a position to prompt [15:10] mpt: I don't have any insight on an "indicator-video", but I'm guessing it is not useful in and of itself (ie, I don't know what it would do besides give a visual cue) [15:10] right [15:12] jdstrand, if there was a prompt would you be happy with a visual cue only for background untrusted apps, not for foreground apps (or future trusted screencast apps)? [15:12] larsu: so I guess the answer to your question is, if design says the indicator change is not the proper implementation, I'll release note the lack of notification/prompting for mic. we can adjust the bug to say it should interface with trust-store, and target that for early 14.04 [15:13] mpt: yeah-- that makes sense to me [15:13] mdeslaur: ^ [15:14] jdstrand, I think the background visual cue for background apps using the mic or camera should be the same as for a phone call using the mic or camera. In part because many of those apps will be phone app substitutes (Voip clients). [15:14] makes sense [15:14] jdstrand, mpt: yes, that makes sense...I really like the notification for background apps. The prompt is a bit annoying to me, but I don't see a better solution. [15:14] jdstrand, unfortunately that design isn't finalized, but it looks like it's going to be a separate temporary indicator that contains items naming the app(s) in question. larsu, how much work would that be? [15:15] jdstrand: I think a release note is ok for now...we don't have much of a choice anyway :) [15:16] I find the prompt useful fwiw :) it is first time only and can be cached. it even allows for the possibility of preseeding the prompt [15:16] s/preseeding the prompt/preseeding the cache/ [15:16] mpt: I don't know about camera, but mic is fairly easy. Indicator-sound already detects that to show the mic volume. Would you want this for 13.10? Sounds very much like a feature to me ;) (also, I have tons of bugs to fix) [15:16] jdstrand: it's useful until you get 4 of them in a row: 1- microphone, 2- accounts, etc. [15:17] mdeslaur: if everything is using trust-store, perhaps there is a way to aggregate the prompts [15:17] yes! [15:17] seb128, hey [15:17] jdstrand: hrm, perhaps an uber-helper that an app could call that would display them all at once pre-emptively? [15:18] anyway, future thought [15:18] mdeslaur: I don't know what that would look like implementation-wise, but possibly [15:18] jdstrand, I was just going to say the same thing. :-) The buttons should be insensitive for a second after the prompt appears anyway, to prevent race conditions, and that would often be enough time to aggregate other requests. [15:19] yeah-- aggregating the prompts would be pretty slick [15:19] attente, I'm looking at the current phone settings ... language is going to work once polkit is there, what about the other ones? Could you hide the widgets that have no working backend with some comments, we can add them back when the backends are there [15:20] jdstrand: the good thing about this being in a library is it will be easy to add that in the future [15:20] attente, doing the way Laney did in https://code.launchpad.net/~laney/ubuntu-system-settings/hide-greeter-messages/+merge/186061 [15:20] yep [15:20] larsu, whether it's for 13.10 is between you and security and release management. ;-) [15:20] attente, do you know what options they plan to support for this release? [15:20] desrt: do I need to free the cancellable? [15:21] seb128, sure... but it's going to be a very empty page... [15:21] I also didn't finish my thought-- with trust-store, it should be possible for us to also do things like have Skype in a preinstall not prompt [15:21] Laney: yes. a good place to do that is right after you cancel it. [15:21] sure [15:21] Laney: all of the operations that are tied to the cancellable will take their own refs [15:21] so g_cancellable_cancel(c); g_object_unref(c); is a common sight [15:21] attente, well, we want to list only thing that work, most pages are getting empty ... but better than having non working UI (and we can get the other back by setting USS_SHOW_ALL_UI=1 [15:22] seb128, last i spoke to tmoenicke, he said he'd be working on the ubuntu-keyboard gsettings migration this week [15:22] seb128, ok, no problem [15:22] attente, the schemas landed today, let's see what we can add back, I guess meanwhile it's going to be the language selector only... [15:22] attente, thanks [15:23] mpt: re larsu, security team and release-- that for larsu somes done to if you think indicator-sound should change its icon in the future, or if we'll implement the background visual cue in another manner [15:23] mpt: it sounds like you don't want the visual cue in the indicator [15:23] mpt: is that accurate? [15:27] thinking out loud: I wonder if a notification is enough instead of a prompt. Nobody will ever click on "No" anyway. ie: "Skype is using the microphone." the first time it uses it. [15:27] "Fart app is using the microphone." [15:27] jdstrand, yes, for two reasons. First, I think the background cue should be the same as for phone calls, and the current design for that does not involve the sound indicator. Second, a red indicator would suggest a problem that you could open the menu to fix, which wouldn't be the case here. [15:28] mpt: right. I'll update the bugs and followup on this [15:28] thanks [15:28] larsu: I'll release note it [15:29] jdstrand: cool, thanks. [15:29] mdeslaur, might be a counterexample. :-) [15:29] larsu: are you responsible for indicator-sound and not pulseaudio? [15:29] jdstrand: yes. [15:29] ok [15:30] mpt: fair enough :) === dednick is now known as dednick|lunch [15:44] seb128: can you pull and re-check the measure branch please? [15:45] it won't be any faster but I wonder if the actual bugs are fixed [15:45] building to see how fast it is on the nexus 4 too [15:46] Laney, ok === dednick|lunch is now known as dednick [15:59] yeah it took like one second there === psivaa is now known as psivaa-afk-bbl [16:11] charles, hey, do you have any idea about https://bugs.launchpad.net/ubuntu/+source/indicator-session/+bug/1230381 ? czajkowski is having the issue [16:11] seb128, I'll take a look in a minute [16:12] charles, thanks [16:12] czajkowski, ^ [16:13] Thank you :) [16:13] * czajkowski hugs seb128 [16:13] seb128: are you in Paris by any chance on October 10th ? [16:13] * seb128 hugs czajkowski back [16:13] I owe you pints! [16:14] czajkowski, no, and I'm not living close from Paris ... going to be for another time I guess ;-) [16:15] don't stand for that [16:15] he can definitely just hop on the tgv [16:15] :P [16:15] lol [16:15] heh [16:16] I'm over for mongoDB paris and trying to meet people when I travel, this week was Dublin and meet people from the mozilla community. Hoping to meet people in Munich as well. [16:16] kill 2 birds with one stone :) [16:16] czajkowski: we're importing seb to canada [16:17] I'm there also :) [16:17] but only before it snows! too cold after that! [16:18] I heard -40 isn't actually that bad [16:18] we'll be in montreal until the 18th, then in toronto [16:18] until the 26th [16:19] desrt: http://www.lczajkowski.com/2013/09/21/travelling-over-the-coming-weeks-and-catching-up-with-people/ [16:20] czajkowski: friends in toronto? [16:20] czajkowski, nice list :D [16:20] jealous of your air miles [16:21] Laney: air miles are a scant reward given to partially compensate someone for sitting with their knees at their neck in an aluminium tube where they are subjected to viruses and conditions that increase their chances of catching them [16:23] Laney, do you know if the "stats on the welcome screen" control has a working side on unity? [16:23] sure - but if they can make this terrible process any less terrible then i'm going to grab the chance [16:23] seb128: i think so but check with mterry [16:24] Laney, ok, thanks [16:24] needs pk anyway [16:24] oh, right, one of those :/ [16:26] desrt: in Guelp and Toronto :) [16:26] czajkowski: cool. attente is here too. let's get a crepe some night. [16:26] charles: that takes me up to November 14th when I get my new puppy Bash :) [16:27] desrt: excellent! [16:27] I'll also be in NYC every 3 months for team building :) and puerto rico in feburary for company all hands :D [16:31] could have a worse place for an allhands [16:37] this is true :) this week it's Dublin for the EMEA side of the company, which is nice as I get to see family here. [16:38] mpt: fyi, you probably saw I revamped bug #1224756, but I also filed bug #1230391. That needs design and I wasn't sure how to get it on your radar, so I am mentioning it here [16:40] jdstrand, thanks, assigned. [16:41] Laney, we are not going to have automatic tz for v1 right? [16:41] no [16:41] that toggle is ntp [16:41] Laney, tz as well? [16:41] ? [16:42] Laney, https://wiki.ubuntu.com/TimeAndDate?action=AttachFile&do=get&target=phone-settings-time-and-date.png [16:42] ^ that's ntp right? [16:42] Laney, what about [16:42] https://wiki.ubuntu.com/TimeAndDate?action=AttachFile&do=get&target=phone-settings-time-zone-automatic.png [16:42] mpt: I also mentioned bug #1230391 in the bug I just filed on the camera service (1230366). guessing the design is the same, but fyi [16:42] the second one is some clever stuff but it doesn't exist [16:43] first one is ntp, yes [16:43] it asks timedated to turn that on/off, didn't actually check if it works on the device yet [16:43] Laney, ok, so the second one should be change to only have manual ... I'm going to hide the selector etc [16:43] Laney, seb128: NTP and/or carrier-provided info [16:43] yes, we don't have carrier stuff though [16:43] mpt, we don't have the backend for carrier-provided info [16:43] so it is just ntp now [16:43] ok [16:43] that's why we can't do automatic timezone [16:43] that's not going to happen for v1 [16:44] I spoke to someone (tvoss?) last week about that and he said they haven't even thought about it yet [16:44] mpt, we can only do manual in https://wiki.ubuntu.com/TimeAndDate?action=AttachFile&do=get&target=phone-settings-time-zone-manual.png [16:44] that's "timed" [16:44] mpt, what do you recommend doing? hidding the selector, e.g just keeping the entry and the list? [16:44] mpt, should we have an heading before, like "enter your location"? [16:44] seb128, that's exactly what I was typing, yes. :-) [16:45] seb128: that "No matching place" looks really really tiny on the phone here [16:45] can you fix that to be bigger when you mp this? [16:45] seb128: ...I meant the hiding the selector bit. No need for a separate heading, compare for example. [16:45] Laney, sure, but shouldn't we fix the toolkit to not have text being tiny by default? [16:46] mpt, makes sense [16:46] probably [16:47] seb128: mpt: You'd get a text field and some text saying "No matching place" [16:47] is that clear enough? the field could have some placeholder text [16:47] Laney, I'm submitting a bunch of trivial changes and then reviewing your directories work again [16:47] There is a "Set the time zone" sub heading too [16:47] Laney, mpt: I found weird to have "No matching place" when you open the dialog and didn't enter anything btw [16:47] seb128: ok === slomo_ is now known as slomo [16:53] seb128, hmm, it shouldn't have been showing that regardless of whether there's anything else in the rest of the screen... [16:53] Ah, my fault. [17:01] mpt, that text should only be there is there is some text entered and no match I guess? [17:01] seb128, exactly. How's that? [17:02] seb128: how did your lightdm testing go? did anything just work? [17:03] Laney, no, I ran into several issues due to ro and started opening bugs for other components :/ [17:03] Laney, second round tomorrow [17:04] haha [17:04] mpt, looks good ... the 2 texts are inconsistent in their use of a "." at the end [17:04] ro is a right pain [17:04] it is :/ [17:06] seb128, you don't miss a thing, do you. ;-) That's tentatively deliberate: one is an instruction, the other is an empty-state error. If the empty state error was followed by an instruction on how to fix the error, *that* would end in a period. [17:06] e.g. "No matches" "Try different spelling, or a larger city nearby." [17:07] mpt, haha [17:07] mpt, I see, thank for the explanation ;-) [17:07] rsalveti: I synced gstreamer into ppa:ubuntu-desktop/gstreamer-1.1 [17:07] should have gst/base/bad built now, good/ugly arriving soon [17:08] will re-apply the ubuntu diffs tomorrow morning [17:10] going out for a ride now, see you! [17:11] Laney, have fun! [17:13] mdeslaur: want me to look at those unit tests for the i-session MP? [17:14] I'm in the neighborhood anyway for another change === psivaa-afk-bbl is now known as psivaa [17:30] charles: ah, sure! Sorry about that, I was still building it when I did the bzr upload [17:31] mdeslaur: no worries [17:48] mdeslaur: it all looks innocent enough in the unit tests [17:48] mdeslaur: here's a diff that includes the fixed tests [17:48] mdeslaur: http://paste.ubuntu.com/6155459/ [17:48] that should make jenkins happy [17:49] hrm, odd...I wonder why my original patch didn't trigger a test failure [17:49] besides... unit tests in indicators? what a novel idea :D [17:49] mdeslaur: because the original patch only changed behavior if n_users==1 [17:49] oh, and it's not testing that [17:49] I see [17:49] mdeslaur: probably I should add a test to cover that logic branch [17:49] I didn't realize it wasn't being covered until this ticket [17:50] so, in addition to the screensaver fix, thanks for that too :) [17:50] hehe [17:50] charles: want me to update my merge request, or are you good? [17:52] mdeslaur: it's not a big deal, but less work overall if you update your mr [17:58] charles: ok, resubmitted, thanks [18:00] mdeslaur: cool, approved [18:04] charles: thanks! [18:39] hey attente, did all your language settings and keyboard MPs get in? [18:40] jasoncwarner, the ubuntu-keyboard MP is merged, but language settings not yet though [18:40] attente do you know what is holding up the language settings? can we get those in today for EOW image? [18:41] jasoncwarner, i just disabled the maliit settings because they won't work on the backend, so the page is looking pretty sparse [18:42] it just needs someone to review and approve [18:43] and while the ubuntu-keyboard MP is merged, i don't think it's landed yet... [18:43] attente ok. can you poke a few people to get your stuff merged? if no one does, ping me with various MPs and I'll poke as well [18:43] Laney, seb128, are either of you around? [18:43] attente, yes [18:44] jasoncwarner, attente: I can review that in a bit sure [18:44] seb128, i don't think the ubuntu-keyboard schema is landed yet though [18:45] thanks, seb128 [18:45] attente, https://launchpad.net/ubuntu/+source/ubuntu-keyboard/0.99.trunk.phablet2+13.10.20130925-0ubuntu1 [18:45] attente, it is [18:45] jasoncwarner, yw! [18:45] oh, i didn't realize, thanks seb128 [18:46] yw [19:05] Laney: thanks [19:07] hey robru, did you end up picking up cellular settings? make any headway? [20:09] jasoncwarner, no, sorry. that merge that ken showed me was *enormous* and I'm not familiar with the codebase. asked for a bit of help and never got a response. [20:09] figured he was too busy [20:10] attente, ping [20:10] bschaefer, hey [20:11] attente, hey, soo we are having some fun ibus issues [20:11] bschaefer, yep... [20:11] and it looks like the cause is due to the ibus-daemon always being started...even when its no suppose to? [20:11] attente, or do you already know about it :) [20:12] bschaefer, talking to bregma about it right now in #systems [20:12] maybe here is a better place to discuss it [20:12] attente, o well then ill join and listen :) [20:12] * bschaefer doesn't know what #systems is [20:12] bschaefer, it's on canonical's private irc server [20:13] attente, right, tried freenode and no one was there :) [20:13] bschaefer, anyways, short and long of it is that i don't think i-keyboard is starting ibus, it's actually g-s-d [20:14] i-keyboard is only ever talking to ibus when it switches to an actual ibus IM [20:14] attente, yes g-s-d sounds more like it cause its the package that places the dbus ibus service [20:14] and g-s-d has this line that makes me suspect [20:14] https://git.gnome.org/browse/gnome-settings-daemon/tree/plugins/keyboard/gsd-keyboard-manager.c#n951 === m_conley is now known as m_conley_away === thumper is now known as thumper-dogwalk === thumper-dogwalk is now known as thumper