[08:21] <pstolowski> Saviq, ping
[08:21] <tsdgeos> _salem: where do crashes like https://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-mako/6138/ get uploaded to? errors.u.c?
[08:21] <Saviq> pstolowski, hey
[08:21] <Saviq> tsdgeos, yes
[08:22] <Saviq> tsdgeos, but truth be told they don't always end up in the reports there for some reason
[08:23] <pstolowski> Saviq, hi! do you know if the 'finished' flag (set by shell/plugin and sent to scope) for click installation is set when it actually finishes intallation, or just on download completion?
[08:24] <Saviq> pstolowski, I'd say just download
[08:24] <Saviq> pstolowski, since we're using the DownloadTracker, which doesn't know about the installation afaict
[08:25] <pstolowski> Saviq, ok.. I think that's causing race for me on click side :(
[08:25] <Saviq> pstolowski, very possible
[08:25] <Saviq> pstolowski, I didn't like this from the beginning, but we never had time to implement this proper
[08:25] <pstolowski> Saviq, ok, will investigate further, perhaps will need some workaround in click
[08:26] <Saviq> pstolowski, problem is there's nothing else running... just the download tracker
[08:26] <Saviq> s/tracker/manager/
[08:27] <Saviq> pstolowski, so after that's finished, we'd have to do another progress bar for the installer
[08:28] <pstolowski> Saviq, ah, I see. is there an 'installed' signal you could subscribe to instead when progress reaches 100%
[08:28] <pstolowski> ?
[08:28] <pstolowski> Saviq, and delay 'finished' until you get it?
[08:29] <Saviq> pstolowski, the download manager does not install it
[08:29] <tsdgeos> Saviq: do we have any proper howto of how to get the info of those crash files? I remember we spoke on how to do it but have sincerely forgotten about it
[08:29] <Saviq> tsdgeos, you mean on errors.u.c or locally?
[08:29] <tsdgeos> Saviq: locally
[08:30] <Saviq> tsdgeos, you wanna do on device or on host?
[08:30] <tsdgeos> Saviq: whichever is easier
[08:30] <Saviq> tsdgeos, then device, probably
[08:30] <Saviq> tsdgeos, install apport-retrace
[08:30] <Saviq> tsdgeos, and then `apport-retrace -g file.crash` will get you in a gdb session in that crash
[08:30] <pstolowski> Saviq, ack, what i'm after is if you could subscribe to some other service to get notified about installation status?
[08:31] <tsdgeos> Saviq: okidoki
[08:31] <Saviq> tsdgeos, installing more dbg symbols on the phone will then let you see more
[08:31] <tsdgeos> sure
[08:31] <Saviq> pstolowski, that's the thing, there isn't any
[08:31] <pstolowski> Saviq, ok..
[08:31] <Saviq> pstolowski, and that would be building knowledge into the shell/dash, we need to make it generic
[08:34] <pstolowski> Saviq, who triggers actual installation?
[08:35] <Saviq> pstolowski, I think there's a hack in the download manager
[08:35] <pstolowski> Saviq, ah, one more hack, ok :/
[08:36] <Saviq> pstolowski, it was discussed on the ubuntu-phone ML some time recently
[08:36] <tsdgeos> Saviq: getting "BFD: Warning: /tmp/apport_core_0gtdfi is truncated: expected core file size >= 232255488, found: 36765696.", is that usual?
[08:36] <pstolowski> Saviq, thanks
[08:36] <Saviq> tsdgeos, yeah, we need to up the kill timeout
[08:37] <Saviq> tsdgeos, means the app was killed by upstart before apport managed to collect the crash report :|
[08:37] <tsdgeos> :/
[08:37] <tsdgeos> so useless report
[08:37] <Saviq> yeah
[08:40] <Saviq> tsdgeos, https://code.launchpad.net/~saviq/unity8/increase-kill-timeout/+merge/213003
[08:41] <tsdgeos> done
[08:44]  * Saviq flushes unity8 queue, then
[08:45] <Saviq> tsdgeos, top-ack?
[08:45] <tsdgeos> oh
[08:45] <tsdgeos> sleept
[09:18] <tsdgeos> hmmm
[09:19] <tsdgeos> the carousel is a bit bad in regards to delegate creation on startup
[09:19] <Saviq> tsdgeos, yeah, that's what I expected
[09:21] <tsdgeos> http://paste.ubuntu.com/7161682/
[09:21] <tsdgeos> now to find out why :)
[09:24] <tsdgeos> dednick: https://code.launchpad.net/~nick-dedekind/unity8/indicators.hint-interval/+merge/202104
[09:26] <tsdgeos> sizing seems to kick too late
[09:26] <tsdgeos> cards are 0x0 and thus carousel seems to think it needs to create them all
[09:26] <tsdgeos> then they get resized
[09:44] <Cimi> Saviq, saw this http://dorey.github.io/JavaScript-Equality-Table/ \o/
[09:48] <Cimi> tsdgeos, what's the issue?
[09:49] <Cimi> dednick, https://code.launchpad.net/~nick-dedekind/ubuntu-settings-components/access-point-init/+merge/212913/comments/503359
[09:49] <tsdgeos> Cimi: still don't know looking at it :)
[09:50] <Cimi> do we have some sort of semaphores to avoid using boolean vars like this? https://code.launchpad.net/~nick-dedekind/ubuntu-settings-components/access-point-init/+merge/212913/comments/503359
[09:50] <Cimi> in qml
[09:51] <dednick> Cimi: responded
[09:53] <Cimi> dednick, component.oncompleted is the first call?
[09:54] <Cimi> dednick, well yes
[09:54] <Cimi> dednick, but are we sure we won't have races?
[09:55] <Saviq> Cimi, shame it shows everything twice ;)
[09:56] <Saviq> i.e. it's symmetrical by the diagonal
[09:56] <Cimi> yep
[09:56] <Cimi> I'm gonna share with us :D
[10:00] <tsdgeos> at least the table is symmetrical
[10:00] <tsdgeos> could be worse :D
[10:04] <dednick> Cimi: the oldValue is only used because those functions can be called from each other. nothing can call Component.onCompleted other than qt event loop (or whatever system it uses)
[10:05] <Cimi> dednick, sure
[10:06] <Cimi> dednick, but what happens if it is executing component.oncompleted
[10:06] <Cimi> dednick, then menu.checked changes?
[10:06] <Cimi> it will skip
[10:15] <tsdgeos> Cimi: Saviq: https://code.launchpad.net/~aacid/unity8/carouselNoNoSizeDelegates/+merge/213013
[10:17] <Saviq> tsdgeos, cool
[10:17] <Cimi> tsdgeos, Saviq I might use > 0 instead != 0 ?
[10:18] <tsdgeos> sure, let me change that
[10:21] <dpm> tsdgeos, hi! A quick question about Qt locales: I'm looking at a MP for the Calendar app that adds the SDK's DatePicker component, and I've noticed that for 'ca' it shows the wrong "$DAYNR $DAYNAME" (it should be the other way around) http://i.imgur.com/JXt1w7N.png - zsombi tells me that comes straight from the locale definitions, in which case it seems that ordering would be wrong for Catalan and would need to be changed upstream?
[10:21] <dpm> it seems to get it from Qt.locale.dateFormat(Locale.ShortFormat)
[10:21] <dpm> let me try if that prints the right order
[10:22] <tsdgeos> dpm: you mean you want "Dissabte 26" and not "26 Dissabte"?
[10:22] <dpm> yes
[10:22] <tsdgeos> may very well be what zsombi says
[10:26] <dpm> tsdgeos, so where are those locale definitions upstream?
[10:26] <tsdgeos> dpm: are you doing a small testcase or want me to?
[10:26] <tsdgeos> oh there's a big magic file with magic numbers
[10:27] <tsdgeos> once you see it you won't be able to sleep anymore :D
[10:27] <dpm> hahaha
[10:27] <tsdgeos> let me look for it
[10:27] <dpm> cool, thanks!
[10:27] <tsdgeos> Saviq: Cimi: pushed the != -> >
[10:28] <tsdgeos> or may well be the posix locales too, give me a few mins
[10:29] <dpm> no rush, thanks
[10:34] <Cimi> dednick, ?
[10:43] <dednick> Cimi: it can't happen while Components.onCompleted is executing...
[10:43] <Cimi> dednick, ah ok
[10:43] <Cimi> dednick, so all connections are blocked until component.oncompleted is executed?
[10:43] <Cimi> I mean, everything?
[10:43] <Cimi> don't know how multithreading works here
[10:44] <dednick> Cimi: it's a single thread as far as i'm aware
[10:47] <Cimi> might worth asking
[10:47] <Cimi> Saviq, tsdgeos hey
[10:48] <tsdgeos> yes?
[10:48] <Cimi> see this http://bazaar.launchpad.net/~nick-dedekind/ubuntu-settings-components/access-point-init/view/head:/Ubuntu/Settings/Menus/AccessPointMenu.qml
[10:48] <Saviq> dednick, this never worked afaict
[10:49] <Saviq> dednick, i.e. there was never a Changed signal unless there was a binding (and so the evaluation of the binding delayed setting of its value)
[10:55] <Cimi> dednick, can we use an alias to start?
[10:56] <Cimi> dednick, aliasing menu.checked with checkbox.checked?
[11:04] <dednick> Cimi: there is a comment there saying why we can't use a binding
[11:04] <dednick> it worked because the object used to be completed before we started setting any values
[11:05] <dednick> Cimi: because we used a Loader.
[11:06] <Cimi> dednick, sorry I don't understand
[11:06] <dednick> Loader { Component.onCompleted: { menuModel = model } } kind of thing
[11:07] <Cimi> dednick, I trust you, I'm just trying to understand
[11:07] <Cimi> dednick, so if we do for id:menu (parent), property alias checked: checkbox.checked
[11:07] <Cimi> dednick, what will happen, example?
[11:08] <Cimi> dednick, otherwise we can have a second property like serverChecked?
[11:08] <dednick> Cimi: at the moment it doesnt work in your code, because we use a binding to set the value of checked in the listview delegate. Because you are directly using the AccessPoint as a delegate, the object is NOT COMPLETE when you set "checked: model.isToggled" (or whatever), so the Connection to menu.checked in AccessPointMenu is never triggered
[11:09] <dednick> in the settings page, or in indicator code: We use a Loader, which we wait unitl it is complete, before setting the values.
[11:11] <Cimi> dednick, I understand. So why an alias won't work?
[11:11] <Cimi> sorry for asking and asking, but all this binding stuff and connections and loaders is confusing me now
[11:11] <Cimi> maybe I need a second coffee
[11:11] <Cimi> maybe it's aging
[11:11] <Cimi> since yesterday was my bday
[11:11] <dednick> Cimi: we can't use an alias, because we dont want the menu.triggered(menu.checked) code to run when the server updates the value.
[11:12] <Cimi> dednick, can't we use onCheckedChanged and so for that?
[11:12] <dednick> we only want to send info to server when the user changes the value manually.
[11:12] <Cimi> and trigger only when there is an actual click on the checkbox?
[11:13] <Cimi> dednick, like with that mousearea?
[11:14] <Cimi> dednick, menu.triggered is emitted only when             target: menu.__mouseArea
[11:14] <Cimi>             onClicked: {
[11:14] <Cimi>                 checkbox.clicked();
[11:14] <Cimi>             }
[11:14] <dednick> Cimi: no
[11:14] <Cimi> ok
[11:15] <dednick> Cimi: signals are not emitted order. checked might or might not have changed when we get that clicked signal.
[11:15] <dednick> Cimi: // FIXME : should use Checkbox.toggled signal
[11:15] <dednick> :)
[11:15] <Cimi> yes
[11:15] <Cimi> but nobody approves
[11:16] <dednick> Cimi: it's the same for the slider. there is no "user changed value" signal
[11:16] <Cimi> what's saviq's opinion on those widgets?
[11:16] <dednick> although florian says the behaviour is well defined
[11:18] <dednick> Cimi: i have no idea :)
[11:19] <dednick> Cimi: believe me, i didnt want it this way. which is why i tried to add the toggle signal
[11:19] <dednick> it's a friggin pain in the arse
[11:21] <dednick> Cimi: i'll take a look at the checkbox. maybe they've changed something, but when i tried it didnt work
[11:27]  * Saviq just got the polishItems crash :|
[11:31] <Saviq> dednick, Cimi, let me know if you need me
[11:31] <Cimi> Saviq, we have this ugly code here http://bazaar.launchpad.net/~nick-dedekind/ubuntu-settings-components/access-point-init/view/head:/Ubuntu/Settings/Menus/AccessPointMenu.qml
[11:31] <Cimi> Saviq, nick wrote this because apparently it's the only way, we are both unhappy about it
[11:32] <Cimi> Saviq, connections and re-setting values
[11:32] <Cimi> instead bindings or such
[11:32] <Saviq> Cimi, so two-way sync? I don't think there's a better way than that
[11:33] <dednick> Cimi: been testing the Checkbox. looks like clicked slots do com after internal ones. But I have no idea if it's a "well known" order
[11:33] <dednick> Saviq: do you know if there's an order for signal handlers to be triggered?
[11:33] <dednick> "slots" or connections
[11:33] <Saviq> dednick, nothing to rely on
[11:33] <Saviq> Cimi, bindings only work one way, so if you need to sync two places, connections to onChanged is the only solution
[11:33] <dednick> Saviq: right. that's what i thought
[11:34] <tsdgeos> Saviq: what were you doing? can repro?
[11:34] <Saviq> tsdgeos, of course I can't
[11:34] <Saviq> tsdgeos, well, will try in a bit
[11:34] <Saviq> tsdgeos, but what I was doing is I managed to crash the click scope
[11:35] <tsdgeos> :/
[11:35] <Saviq> tsdgeos, you can try: my steps → uninstall and install webbrowser, search again
[11:36] <Cimi> Saviq, but when we have those booleans
[11:36] <Saviq> tsdgeos, it will get stuck on "Install", 'cause you shouldn't be able to uninstall it anyway
[11:36] <Saviq> (that's when scope crashes)
[11:36] <Saviq> tsdgeos, it crashed when I tried searching again
[11:36] <Cimi> Saviq, property bool enableCheckConnection: true
[11:37] <Cimi> it's a rudimental semaphore
[11:37] <Cimi> is there anything smarter in qml?
[11:37] <Saviq> Cimi, not really, but I'm not sure it should really be necessary, since it will just feed back once
[11:37] <Saviq> Cimi, well, it might be worse for slider, for exampe
[11:37] <Saviq> example
[11:38] <Saviq> but for bools it would just come back as the same value that's there on the target property already
[11:38] <dednick> Cimi: there was some talk about creating an object to manage the 2 way sync so we didnt have to do it in different classes, but i never got around to it.
[11:39] <dednick> for now we just have to stick it out
[11:55] <Saviq> tsdgeos, got it again, different steps, but searching seems to trigger it for me
[11:55] <tsdgeos> 'nteresting
[11:56] <Saviq> tsdgeos, searching in the apps scope was what I was doing in both cases
[11:56] <tsdgeos> Saviq: phone?
[11:56] <Saviq> tsdgeos, yes
[11:56] <tsdgeos> so you're runnign with run_on_Device?
[11:56] <pete-woods> Saviq: was it the same as this crash? https://bugs.launchpad.net/ubuntu/+source/unity-scope-click/+bug/1297297
[11:56] <tsdgeos> or how do you detech the crash?
[11:57] <tsdgeos> detech! new word
[11:57] <Saviq> tsdgeos, it just stops responding :)
[11:57] <Saviq> tsdgeos, and apport hogs the cpu
[11:57] <Saviq> tsdgeos, and there's a .crash file in /var/crash
[11:57] <tsdgeos> Saviq: but how do you know it's the polishItems thing?
[11:57] <tsdgeos> ah because it has time to dump
[11:57] <Saviq> pete-woods, no, yours is different
[11:57] <pete-woods> okay
[11:57] <pete-woods> thanks for looking
[11:58] <Cimi> dednick, is there like a property on the unitymenumodel I can use to see if I'm connected to a network?
[11:58] <Saviq> pete-woods, but I was actually talking about a unity8 crash right now, got another one in the scope
[11:58] <Saviq> pete-woods, let me find it
[11:58] <pete-woods> Saviq: it's more to help the guys figure out what happned
[11:58] <pete-woods> my crash file didn't seem to help them much
[11:58] <Saviq> pete-woods, https://bugs.launchpad.net/ubuntu/+source/unity-scope-click/+bug/1298314
[11:59] <pete-woods> ah, okay, quite different then
[12:00] <Saviq> pete-woods, mine has easy steps to repro, too, it's trying to install a non-click, already installed app (e.g. webbrowser)
[12:00] <pete-woods> yeah, mine was just I searched, it crashed, ??
[12:00] <pete-woods> and it didn't happen the next time
[12:00] <Saviq> pete-woods, our "bigger" issue is not that the scope crashes, but unity8
[12:01] <pete-woods> yeah, that's totally fair, unity8 trumps almost any other crash
[12:01] <pete-woods> except maybe kernel :p
[12:07] <Saviq> tsdgeos, and again
[12:07] <Saviq> tsdgeos, this seems like a reliable way to repro, just search, clear, search, clear
[12:07] <Saviq> try with slow / fast typing, selecting an item from the history and such
[12:07] <tsdgeos> ok
[12:07]  * tsdgeos tries
[12:08] <tsdgeos> got a crash, let's see if i can see if it's polishitems
[12:09] <tsdgeos> nothing in /var/crash
[12:09] <tsdgeos> confusing
[12:09] <Cimi> dednick, ?
[12:10] <Cimi> dednick, I'm looking for a way to know if I'm connected to a wifi or not, from wizard
[12:11] <Saviq> tsdgeos, assign yourself to bug #1297240 please
[12:11] <dednick> Cimi: sorry. no, i dont believe there is
[12:11] <Cimi> dednick, but the indicators
[12:11] <Saviq> tsdgeos, hmm? and that was unity8 launched just as usual, nothing special?
[12:11] <Cimi> dednick, the wifi indicator says connected
[12:11] <tsdgeos> Saviq: think so, yes
[12:11] <Cimi> dednick, on the panel
[12:11] <Saviq> tsdgeos, you might want to up the kill timeout in /usr/share/upstart/sessions/unity8.conf
[12:11] <Cimi> dednick, is this something I can fetch?
[12:12] <dednick> Cimi: oh. connected to wifi. er
[12:12] <dednick> there's an action
[12:12] <Saviq> tsdgeos, but there should be a .crash file regardless
[12:12] <dednick> Cimi: er, but i think it's just the root item icon, not a "i am conencted"
[12:13] <tsdgeos> scope runner crashed
[12:13] <Cimi> dednick, deom unitymenumodel?
[12:13] <Cimi> *from
[12:13] <tsdgeos> is there anyway to recover from that other than reboot?
[12:26] <Saviq> tsdgeos, restart scope-registry
[12:35] <dednick> Cimi: what do you actually need to show?
[12:36] <Cimi> dednick, I want to change label of the continue button from skip to connect when it's connected
[12:36] <Cimi> opposite
[12:36] <Cimi> when it's connected continue
[12:36] <Cimi> when it's not skip
[12:37] <Cimi> dednick, I can use a global variable set by the delegates when isToggled is true for one of them
[12:37] <mzanetti> Saviq: did you already trigger the right-edge review somehow?
[12:38] <Cimi> dednick, but was looking for better solutions
[12:38] <Cimi> dednick, thinking that I can have the same logic of the indicator row
[12:38] <Saviq> mzanetti, yes
[12:38] <Saviq> mzanetti, let me find out what's the progress
[12:38] <mzanetti> awesome :)
[12:40] <dednick> Cimi: dont think there is a way. probably best to check with ted or Wellark
[12:40] <Cimi> dednick, I'll go for the global variable
[12:40] <Cimi> dednick, but one day we will need it
[12:41] <Cimi> dednick, we need an "you have network, you don't"
[12:41] <dednick> Cimi: then check with the indicator gues
[12:41] <dednick> guys
[12:42] <dednick> there used to be an action for it, which is what the wifisection was using
[12:43] <Cimi> Wellark, ? :)
[12:44] <Cimi> dednick, do you remember this action?
[12:44] <dednick> Cimi: it was in the chewie network indicator
[12:44] <Cimi> ah ok
[12:45] <Saviq> mzanetti, they have a device and your ppa on it, so that's good progress!
[12:48] <Cimi> dednick, he might be on holiday
[12:49] <Cimi> Wellark
[12:54] <mzanetti> Saviq: cool, thanks
[12:56] <dednick> Cimi: well for now, i dont think the action exists. but maybe you should add a bug
[12:56] <Cimi> dednick, I did a quick workaround
[12:57] <Cimi> dednick, http://paste.ubuntu.com/7162565/
[13:00] <Cimi> mterry, can we start with fullscreen or not?
[13:00] <dednick> Cimi: > 0 ?
[13:00] <Cimi> dednick, well, if it has 2 is not connected...
[13:00] <mterry> Cimi, in wizard?
[13:00] <dednick> Cimi: one of them may be :)
[13:00] <Cimi> dednick, that's why I put 1
[13:01] <Cimi> ahah
[13:01] <Cimi> mterry, yup
[13:01] <Cimi> mterry, makes debugging on desktop harder though
[13:01] <mterry> Cimi, I mean we can.. I don't remember the logic in its main.cpp off the top of my head
[13:01] <Cimi> mterry, maybe I should just set it to be taller?
[13:01] <Cimi> mterry, the wizard has black empty space at bottom
[13:02] <Cimi> @unity is there a way to have apps filling screen but not being fullscreen?
[13:02] <mzanetti> Cimi: ?
[13:02] <Saviq> ???
[13:02] <Cimi> like, if panel is shown, anchors down the panel, otherwise fullscreen?
[13:02] <mterry> Cimi, yeah, you could get the size of the screen and resize the view
[13:03] <Saviq> mterry, I'm not sure that's actually possible
[13:03] <Saviq> greyback, ↑ what are we doing with surfaces these days?
[13:03] <mzanetti> it should be as long as you don't call showFullscreen() but just show() or similar
[13:03] <Cimi> Saviq, wizard starts outside unity, so should be fullscreen
[13:03] <mzanetti> hmm, yeah... unity-mir might override it
[13:03] <Saviq> Cimi, yes, so why shouldn't it be fullscreen?
[13:04] <Cimi> but if you run on desktop or inside unity, should not be fullscreen
[13:04] <greyback> Saviq: you'll have to be more specific
[13:04] <mzanetti> ah... showMaximized() perhaps helps with that
[13:04] <Saviq> mzanetti, I don't think that'll work
[13:04] <Saviq> Cimi, when will it ever be run on desktop or inside unity?
[13:05] <Cimi> Saviq, maybe we want to
[13:05] <Cimi> run it after?
[13:05] <Saviq> Cimi, but how does "fullscreen, but not fullscreen" solve that?
[13:05] <Cimi> Saviq, should fill the screen
[13:06] <Saviq> Cimi, so fullscreen :D
[13:06] <Cimi> but stay under panel if panel is in
[13:06] <Saviq> Cimi, that's not the screen
[13:06] <Saviq> Cimi, that's available geometry
[13:06] <Saviq> Cimi, and that's what maximized is for
[13:06] <Cimi> Saviq, X11 had workarea
[13:06] <Cimi> ok
[13:06] <Saviq> but I don't think we're handling maximized at all
[13:07] <Cimi> ah ok
[13:07] <Saviq> Cimi, but I really think you're overdoing it
[13:07] <Cimi> Saviq, so answer is not for now
[13:07] <mzanetti> Saviq: on the phone. I think its just ignored and shown (non-fullscreen)
[13:07] <Saviq> mzanetti, well, it's just always maximized anyway ;)
[13:07] <mzanetti> yep
[13:07] <mzanetti> so yeah, showMaximized might do the trick. but I haven't tried
[13:07] <Saviq> mzanetti, it won't
[13:08] <mzanetti> ok. I give up :)
[13:08] <Saviq> mzanetti, we don't have struts and such
[13:08] <Saviq> mzanetti, so it will just always be under the panel, even if it's not there
[13:08] <Saviq> i.e. there's no way to tell what's maximized atm
[13:08] <mzanetti> ah... I see
[13:10] <mterry> Cimi, the wizard isn't going to be run as an app though, just fyi.  not like some random click app
[13:10] <Cimi> mterry, yeah I know
[13:10] <Cimi> mterry, I'll do fullscreen
[13:10] <mterry> k
[13:12] <Cimi> mterry, seb128 https://code.launchpad.net/~unity-team/ubuntu-system-settings/wizard.wifi/+merge/212675
[13:13] <Cimi> one thing I want to make sure is that the upgrade of the deb will remove the 20-wifi.disabled file in /usr/share/ubuntu/settings/wizard/Pages
[13:13] <Cimi> qml/paged
[13:13] <Cimi> pages
[13:50] <greyback> hey folks, anyone here a qmake guru? I'm having a problem generating a pkgconfig file using qmake
[14:10] <tsdgeos> greyback: what's your problem?
[14:11] <tsdgeos> and why are you still using qmake? :D
[14:11] <greyback> tsdgeos: existing project, no choice
[14:11] <greyback> tsdgeos: http://bazaar.launchpad.net/~gerboland/qtubuntu/server-custom-input-destination/view/head:/src/platforms/ubuntu/ubuntumir/ubuntumirserver/ubuntumirserver.pro
[14:12] <greyback> tsdgeos: it just doesn't create the file at all
[14:15] <tsdgeos> greyback: the interwebs say
[14:15] <tsdgeos>  CONFIG += create_pc
[14:16] <greyback> tsdgeos: I had that initially, which didn't help. Then had http://stackoverflow.com/questions/6823862/qmake-creating-a-pc-file
[14:16] <tsdgeos> greyback: don't know moar
[14:17] <tsdgeos> i'd go with reading qmake's code
[14:17] <tsdgeos> did it once, it's not that hard
[14:17] <greyback> I was hoping to avoid that :)
[14:17] <greyback> but yeah
[14:22] <tsdgeos> hmmmm
[14:22] <tsdgeos> i'm running the shell manually on the phone and i have no keyboard
[14:22] <tsdgeos> any idea why?
[14:23] <tsdgeos> is there an upstart job for it or somethig?
[14:24] <tsdgeos> yep
[14:24] <tsdgeos> maliit-server
[14:24] <tsdgeos> worst name ever
[14:33] <Cimi> seb128, mterry when can you review that branch?
[14:34] <seb128> Cimi, why did you set it as merged?
[14:34] <Cimi> seb128, I set needs review, bloody dropdown menu
[14:34] <Cimi> seb128, I redid immediately to needs review
[14:34] <seb128> Cimi, weird, I got an email saying merged
[14:34] <seb128> Cimi, but adding to my review queue, it might not be for today though
[14:35] <Cimi> seb128, I'll rebase my location branch on top of that then
[14:35] <Cimi> kgunn, do we needs tests for the wizard?
[14:36] <Cimi> usually answer is we need tests for everything
[14:36] <Cimi> but this interaction things with funcionality might be more tricky and time consuming
[14:37] <kgunn> Cimi: how might it be more tricky ?
[14:38] <kgunn> Cimi: but yes, i would say let's create a nice AP test for that
[14:38] <Cimi> kgunn, because what you test if not the settings are correctly done?
[14:38] <Cimi> it's not about testing UI here
[14:39] <Cimi> kgunn, nice AP is utopia :P
[14:39] <Cimi> I can do an AP
[14:39] <Cimi> :D
[14:41] <mterry> kgunn, if you have some seconds, rebuilding unity-system-compositor and unity8 in silo 004 would be neat.  But not urgent
[14:41] <kgunn> Cimi: :D
[14:41] <kgunn> Cimi: i will take something less than utopia
[14:42] <greyback> Saviq: QSG_RENDERER_DEBUG=render,build,change,upload,roots,dump,noalpha,noopaque,noclip <- pick & choose what ones you want. "render" handiest
[14:42] <Cimi> kgunn, can I do this later?
[14:42] <Cimi> at a later time
[14:43] <kgunn> Cimi: can you do something super minimal at least ?
[14:43] <Cimi> kgunn, I'll have  alok
[14:44] <Cimi> kgunn, code is ready, just missing tests
[14:45] <kgunn> thanks Cimi
[14:46] <kgunn> mterry: kicked
[14:46] <mterry> kgunn, thanks!
[15:13] <Cimi> tsdgeos, Saviq an item's height does not take into account margins?
[15:16] <tsdgeos> hmmm
[15:16] <tsdgeos> what od you you mean with that?
[15:16] <Cimi> tsdgeos, if I set margin 100px, height is not calculated
[15:16] <Cimi> nevermind, did in a different way
[15:17] <tsdgeos> ok
[15:28] <Saviq> davidcalle, glad you're liking the tool :D
[15:29] <Saviq> davidcalle, we actually plan(ned) to build some more logic into it, but it's definitely a good first step
[15:30] <davidcalle> Saviq, since I don't have a device, I've been using it for a couple of months now, it's my favorite 14.04 feature :p
[15:30] <Saviq> :D
[15:39] <paulliu> Saviq: https://code.launchpad.net/~paulliu/unity8/fix-lazyimage1/+merge/211728
[15:40] <paulliu> Saviq: BTW, I'm too unfamiliar with jenkins. Is there any newbie docs for that?
[15:41] <Saviq> paulliu, it's rather unstable recently
[15:41] <Saviq> paulliu, so don't worry
[15:41] <paulliu> Saviq: I access it through VPN but feel nothing to do there.
[15:41] <Saviq> mterry, hey, on the auto background
[15:41] <Saviq> mterry, should we not dpkg-depend on whatever provides that schema?
[15:41] <Saviq> mterry, I think it's a relatively new key?
[15:42] <Saviq> mterry, or maybe check that the schema/key exists, so we don't crash if it doesn't?
[15:43] <mterry> Saviq, yes, we should.  I was under the impression we already depended on it, but I see we don't
[15:43] <Saviq> mterry, make sure to put it in ./build, too
[15:44] <mterry> Saviq, I thought that was automatic now
[15:44] <Saviq> mterry, only for build deps
[15:44] <Saviq> mterry, not for runtime ones (somewhat impossible to be automatic due to some of them being generated)
[15:45] <tsdgeos> Saviq: are you adding the small optimization stuff i did in this landing?
[15:45] <Saviq> tsdgeos, I will
[15:45] <tsdgeos> ok
[15:45] <mterry> Saviq, ah...  it is already a dep, just with the old package name
[15:45] <tsdgeos> there's 3 small ones yet to be reviewed
[15:46] <mterry> or maybe I'm wrong
[15:46] <Saviq> tsdgeos, not really testable the carousel one is it...
[15:46] <tsdgeos> Saviq: i guess i could test it yes
[15:47] <tsdgeos> storing how many delegates are created
[15:47] <tsdgeos> ?
[15:47] <Saviq> tsdgeos, count them
[15:47] <tsdgeos> you can't check how many there will be at the end
[15:47] <tsdgeos> because it'll be correct anyway, but the total should be doable
[15:47] <mterry> yup, guess desktop-schemas and ubuntu-schemas are different
[15:47] <tsdgeos> let's see if i see how to do it though :D
[15:47] <Saviq> tsdgeos, well, yeah, only on creation
[15:47] <Saviq> mterry, gsettings-ubuntu-schemas ships that schema I think
[15:48] <Saviq> tsdgeos, I'll land it separately then, want to unblock the silos asap, we're running low
[15:48] <mterry> Saviq, yeah I got my aliases mixed up.  We were depending on gsettings-desktop-schemas.  Which of course is a different thing  :)
[15:48] <tsdgeos> Saviq: land separately the test or the whole thing?
[15:48] <mterry> Saviq, branch updated
[15:48] <Saviq> tsdgeos, the test I think, that's an important, and somewhat obviously correct, thing
[15:48] <Saviq> mterry, thanks
[15:48] <tsdgeos> Saviq: ok
[15:54] <Cimi> dednick_, the access point menu is basically an extension to the checkmenu, can we merge them?
[15:55] <dednick_> Cimi: an extension? not really. It has a checkbox in it, but it's not the same
[15:56] <dednick_> as far as i remember
[15:56] <Cimi> dednick_, almost like
[15:56] <Cimi> dednick_, it has some extra stuff
[15:57] <Saviq> tsdgeos, actually, I'll land that with the right edge, don't want to wait for rebuilding
[15:57] <Saviq> and right edge I'll need to rebuild anyway
[15:58] <Cimi> dednick_, I find it confusing we have checkbox on left here
[15:58] <Cimi> dednick_, while in optionmenu the tick is on the right
[15:59] <tsdgeos> Saviq: ok
[16:00] <dednick_> Cimi: the AccessPoint alignment is different, it also has an icon between text and checkbox, as well as an icon on the right.
[16:01] <Cimi> ok
[16:02] <Cimi> dednick_, reusing that checkbox with the logic for the connection would be nice though
[16:03] <dednick_> Cimi: yeah. that's the point of having a bi-directional syncing object
[16:04] <dednick_> just havent dont it yet
[16:09] <Cimi> dednick_, where is checkable menu used?
[16:11] <dednick_> Cimi: um. not sure it is anymore. used to be on bluetooth or something
[16:11] <dednick_> but i think it's using switch now
[16:11] <Cimi> dednick_, it is
[16:11] <Cimi> dednick_, let's drop it?
[16:50] <tsdgeos> Saviq: Cimi: https://code.launchpad.net/~aacid/unity8/test_carouselNoNoSizeDelegates/+merge/213109
[16:50] <tsdgeos> looks huge
[16:50] <tsdgeos> but it's only a few lines + indentation of the test
[16:50] <tsdgeos> test->rest
[16:51] <Cimi> tsdgeos, does this work?
[16:51] <tsdgeos> what work?
[16:51] <Cimi> tsdgeos, thought the prevopus was unit test
[16:51] <Cimi> previous
[16:52] <tsdgeos> ?¿
[16:52] <Cimi> tsdgeos, this one requires the window
[16:52] <tsdgeos> yes
[16:52] <Cimi> tsdgeos, from early times, I remember having unit tests
[16:52] <Cimi> and qmltests
[16:52] <tsdgeos> ah
[16:52] <tsdgeos> see what you men
[16:52] <tsdgeos> a
[16:52] <Cimi> this looks like it's a qmltests so requires cmake change
[16:53] <tsdgeos> right need to move it
[16:53] <tsdgeos> yes
[16:53] <tsdgeos> well spotted
[16:53] <Cimi> cool
[16:53] <Saviq> tsdgeos, also, could we move TestCase outside of Carousel?
[16:53] <tsdgeos> Saviq: why?
[16:53] <tsdgeos> Saviq: we have all the tests like this one
[16:53] <tsdgeos> Saviq: wait, it's outside the carousel, no?
[16:54] <tsdgeos> oh, it' not :D
[16:54] <tsdgeos> Cimi: Saviq: pushed
[16:55] <Saviq> tsdgeos, I generally put TestCase next to the tested item, not inside it
[16:55] <tsdgeos> yes yes
[16:55] <tsdgeos> done
[16:55] <Saviq> ok
[16:56] <Saviq> biab
[17:06] <kgunn> tsdgeos: ...who would be best to look at this ?
[17:06] <kgunn> https://bugs.launchpad.net/ubuntu/+source/qtubuntu/+bug/1297900
[17:07] <tsdgeos> kgunn: gerry :D
[17:07] <tsdgeos> or ricmm
[17:07] <tsdgeos> maybe
[17:07] <tsdgeos> i can have a look i guess too
[17:07] <kgunn> tsdgeos: yep...i was thinking of begging for ricmm
[17:07] <kgunn> mmm
[17:07] <tsdgeos> but i'm not the optimal guy
[17:07] <kgunn> i think ricmm is in the process of moving
[17:08] <kgunn> mzanetti: ^ maybe ?
[17:08] <kgunn> gerry's kinda busy
[17:08] <greyback> he is. But lemme take a glance
[17:08] <kgunn> tsdgeos: i know you're working on the other crasher...
[17:08] <kgunn> if you're done, then yes pls
[17:08] <tsdgeos> i am working in the other crasher yes
[17:08] <greyback> that crash /always/ happens when unity8 rejects the client's connection
[17:09] <greyback> so the output of unity8.log would help a lot
[17:09] <greyback> usually due to using the --desktop_file_hint badly
[17:12] <greyback> kgunn: commented on bug, not with answers, but with where someone should look
[17:19] <mzanetti> strange... the link in that bug shows a 100% pass rate
[17:19] <mzanetti> ah, there's the crash
[17:21] <mzanetti> kgunn: greyback: yep, this is in the log: ApplicationManager REJECTED connection from app with pid 4983 as no desktop_file_hint specified
[17:22] <kgunn> mzanetti: so dialer app is not special...gotta have a hint just like everyone?
[17:23] <mzanetti> I think so, yes
[17:23] <greyback> you should check how the apps are being launched - via command line with --desktop_file_hint, or via upstart.
[17:23] <greyback> well upstart-app-launch
[17:42] <josharenson> kgunn: I'm in contact with fginther re: ci tests, but he is currently in a meeting. Do you know where the tests live so I can start poking around?
[17:44] <kgunn> mmm
[17:45] <kgunn> mzanetti: if josharenson wanted to modify jenkins ci job to add a whole new test (e.g. glmark2 on mir)...where would he look ?
[17:47] <mzanetti> kgunn: hmm. I could definitely help him get started. I think we should talk to the ci guys first, to ask for permission etc
[17:47] <mzanetti> fginther: hello :)
[17:48] <mzanetti> fginther: are you ok with us creating some jenkins jobs to try in getting set up something new? And me giving permissions to josharenson to do that?
[17:49] <mzanetti> josharenson: your vpn connection and everything set up already?
[17:50] <fginther> mzanetti, josharenson, this all depends on what your trying to do
[17:50] <josharenson> I don't think I have vpn...
[17:51] <mzanetti> indeed... josharenson: what would the requirements be for such a job? any special packages needed or hardware requirements?
[17:51] <kgunn> fginther: mzanetti josharenson... so this is at my request :) basically adding glmark2
[17:51] <josharenson> ^^
[17:51] <kgunn> which is a seperate project to be part of the ci run
[17:51] <kgunn> for mir
[17:51] <fginther> I can setup you up with a sandbox job to get something started if you need to experiment, but if you already have a test, there is no need for that
[17:52] <fginther> kgunn, glmark on touch? on desktop?
[17:52] <josharenson> fginther: I need to experiment. I am new here and have never seen the tests.
[17:52] <kgunn> fginther: glmark2 on touch/armhf
[17:52] <kgunn> ultimately we'd like to get a report out as well....in order to track trend
[17:53] <kgunn> you may remember the qa report dashboard that robotfuel put together
[17:53] <kgunn> then we hope to expand that into more ui tests that will measure perf as well
[17:53] <kgunn> so baby step 1 out of about 6 here
[17:53] <fginther> kgunn, ok so this is part of a big picture :-)
[17:54] <fginther> but it's good to know where this needs to go
[17:54] <kgunn> yeah step1...add glmark2 as a test for arm mir ci on mir/devel
[17:54] <kgunn> step2 reporting
[17:54] <fginther> josharenson, do you have a phone device to test on?
[17:54] <josharenson> yes
[17:54] <kgunn> step3 (semi-optional) - create a warning on jenkins....(in addition to failure)
[17:55] <josharenson> fginther: I have nex7 flashed w/ ubuntu currently
[17:55] <fginther> josharenson, that'll work.
[17:56] <fginther> josharenson, we have the ability to add tests to mir ci as is being requested. the first step is to get the test into something that can be provisioned and executed in a somewhat generic environment
[17:57] <fginther> josharenson, i.e. if all that is needed is to add some packages and execute a single command, we'll be in good shape to add this
[17:58] <josharenson> fginther: how about parsing the results from the execution?
[18:00] <mzanetti> josharenson: you need to produce a certain xml file
[18:00] <mzanetti> can't reacall the name. the same as autopilot or qmltestrunner produce
[18:01] <mzanetti> other than that, you can collect whatever files you want
[18:01] <fginther> josharenson, that's less straight forward. From what I know of glmark, these results don't really fit th junitxml format
[18:01] <fginther> correct, we can collect any artifacts when the job runs. The post processing could be done elsewhere (which is what the dashboard does)
[18:02] <fginther> getting the performance regression feedback kgunn mentions in step 3 isn't something we've implemented yet, but are working on
[18:03] <kgunn> fginther: ack...i'll live with whatever....i just want to get this in place :)
[18:03] <mzanetti> josharenson: you can browse other jenkins jobs to see what they produce, and to learn how jenkins works (unless you already know that)
[18:03] <mzanetti> I'll send you some links
[18:04] <josharenson> I'll run the benchmark and see if it leaves behind any parsable artifacts
[18:04] <josharenson> I'm sure there is _someting_
[18:05] <fginther> josharenson, kgunn, there's already a bug this https://bugs.launchpad.net/ubuntu-ci-services-itself/+bug/1252933
[18:06] <Cimi> kgunn, the two relevant branches for the welcome wizard are here, but they might need more refinements with the review: lp:~unity-team/ubuntu-system-settings/wizard.wifi and lp:~cimi/ubuntu-system-settings/wizard.privacy
[18:06] <Cimi> as well as AP test
[18:06] <fginther> it was originally from an email from thomi. it's a broader focus than what you're asking for here, so we can probably chunk this into smaller pieces
[18:07] <Cimi> (and I'm waiting for the build for the phone)
[18:08] <kgunn> Cimi: thanks for the ap test!
[18:08] <Cimi> kgunn, HAVEN'T DONE YET :p
[18:08] <Cimi> ops caps lock
[18:57] <Cimi> seb128, in the diagnostics's PageComponent, there's this /* Confirm the setting stuck and reflect it in the UI. */
[18:58] <Cimi> seb128, what's that?
[19:41] <kgunn> mterry: gonna add this mp for the flicker bug
[19:41] <kgunn> https://wiki.canonical.com/InformationInfrastructure/IS/VPN
[19:41] <kgunn> whoops
[19:41] <kgunn> or that
[19:41] <kgunn> https://code.launchpad.net/~andreas-pokorny/mir/no-initial-display-configuration-sent-to-hosting-server/+merge/213126
[19:44] <mterry> kgunn, ok nice
[19:44] <kgunn> mterry: also...they think osk bug is fixed on tip of mir-devel...so we should get that on rebuild
[19:44] <mterry> kgunn, I'm investigating why notifications in the greeter appear too tall  :-/  may have a unity8 rebuild request when I figure that out
[19:54] <seb128> Cimi, dunno, ev wrote that setting page, ask him for details
[19:55] <mzanetti> kgunn: https://bugs.launchpad.net/ubuntu/+source/qtubuntu/+bug/1297900/comments/4
[19:56]  * mzanetti EOD
[19:57] <vini_> hi
[19:57] <vini_> i get an error of not enough disk space on /boot while trying to run system-update
[19:58] <seb128> Saviq, ^ not sure if the issue mzanetti just listed is https://errors.ubuntu.com/problem/85c403c4a05cd32a48a73b226340850faa45e785 but it could be
[19:59] <vini_> sudo apt-get clean doesn't help, any other soultion?
[20:00] <mzanetti> vini_: check how much free disk space there is and what is occupying it
[20:01] <mzanetti> vini_: this might help: http://askubuntu.com/questions/2793/how-do-i-remove-or-hide-old-kernel-versions-to-clean-up-the-boot-menu-or-free-sp
[20:01] <Cimi> seb128, that code is quite weird
[20:01] <Cimi> seb128, who approved?
[20:02] <seb128> Cimi, Laney iirc, but he's eod for today
[20:02] <Cimi> seb128, ok
[20:03] <seb128> Cimi, if you see issues feel free to do changes/send patches, ev has only been sending that merge request and moved to other things then
[20:03] <Saviq> seb128, I think it'd say SIGABRT at least
[20:03] <Cimi> seb128, might do
[20:03] <Saviq> seb128, as that above would be a raise() + abort()
[20:03] <Cimi> seb128, this is the code, you might agree with me http://paste.ubuntu.com/7164686/
[20:03] <Saviq> seb128, and there's completely nothing in the trace
[20:03] <Saviq> ah waity
[20:04] <Cimi> seb128, first sets a variable, then check for it to be different
[20:04] <seb128> Saviq, it does?
[20:04] <Saviq> seb128, yeah, it's a raise() + abort()
[20:04] <Saviq> seb128, so yeah, it could be that, we'd have to see stderr to know for sure
[20:05] <seb128> Cimi, I don't understand the logic of that code, so I'm not sure why it's this way ... there could be a reason or it could be buggy, better to check with ev tomorrow
[20:05] <Cimi> I will
[20:06] <seb128> Saviq, right, in any case we need to add some hook so qmlscene issues log extra info
[20:06] <Saviq> seb128, yeah
[21:28] <Saviq> mterry, hmm indicator-sound segfault :/
[21:29] <mterry> Saviq, :(  during build?
[21:29] <Saviq> mterry, yeah, tests
[21:29] <Saviq> https://launchpad.net/~ci-train-ppa-service/+archive/landing-004/+sourcepub/4046039/+listing-archive-extra
[21:29] <mterry> Saviq, I think ted_ has a branch for that.  let me see
[21:29] <mterry> Saviq, lp:~ted/indicator-sound/test-failures
[21:29] <mterry> Saviq, while there, maybe add lp:~ted/indicator-sound/phone-greeter-mute
[21:30] <ted_> They're dependent on each other :-)
[21:30] <mterry> ted_, ah yes.  lp:~ted/indicator-sound/phone-greeter-mute should be sufficient
[21:31] <Saviq> mterry, kicked
[21:32] <mterry> thanks!