=== jamesh_ is now known as jamesh [07:53] mhr3, hey, we should drop qtdeclarative5-dee-plugin from unity8 deps, should we not? [07:55] Saviq, aren't we using it anywhere? [07:55] mhr3, will check [07:55] Saviq, but yea, from my pov, it isn't needed [08:00] Saviq: hud? [08:12] tsdgeos, do you know where can i find sailfish-specific qt patches? [08:13] mhr3: maybe in mer? [08:13] i was hoping for a link with directory listing of bunch of .patch files :) [08:15] tsdgeos, HUD uses the C++ APIs internally, not the plugin anywhere? [08:16] tsdgeos, i.e. grepping for Dee does not come up with anything QML [08:17] mhr3: i have no idea tbh [08:18] Saviq: but [08:19] ./Hud.qml:431: model: hudClient.results [08:19] ./hudclient.h:37: DeeListModel *results() const; [08:19] we need dee, no? [08:21] ok, but maybe not the model since we don't use DeeListModel {} in qml [08:21] s/model/plugin === davidcalle_ is now known as davidcalle [08:23] tsdgeos, yeah exactly, the HUD plugin exports everything itself [08:26] seb128, ping [08:27] pstolowski, hey [08:29] seb128, hi! where do I find source files for translations of unity-lens-applications (trusty), for let's say - italian? i'm looking at https://bugs.launchpad.net/bugs/1312401 (which I can repro with it, but not pl), seems like there is a messup in translation [08:29] Launchpad bug 1312401 in unity-lens-applications "Filter type names incorrect after upgrade to 14.04 and results don't correspond to the selected types" [Undecided,New] [08:30] pstolowski, https://translations.launchpad.net/ubuntu/trusty/+source/unity-lens-applications/+pots/unity-lens-applications/it/+translate === shiznix is now known as Guest98474 [08:31] seb128, thanks! [08:31] yw [08:32] pstolowski, if it timeout, try again, I had it to load after a few tries [08:35] not surprised that translations cause bugs if rosetta keeps timing out :P [08:35] seb128, mhr3 hmm, no luck, timing out... [08:36] lol [08:36] pstolowski, what string are you looking for? [08:36] pstolowski, https://translations.launchpad.net/ubuntu/trusty/+source/unity-lens-applications/+pots/unity-lens-applications/it/9/+translate [08:36] that's "Games" [08:37] pstolowski, you can also use https://translations.launchpad.net/ubuntu/trusty/+source/unity-lens-applications/+pots/unity-lens-applications/it/+translate?batch=10&show=all&search=Games and replace the search= [08:38] seb128, yeah, that url works just fine (I can browse the strings and they look fine I think), but I cannot download a .po, not entire branch with bzr [08:38] pstolowski, why do you need a po? [08:39] seb128, i'm not sure yet, just want to understand while PL locale works while IT is broken :/ [08:39] pstolowski, you can download https://launchpad.net/ubuntu/+source/language-pack-gnome-it-base/1:14.04+20140410/+build/5900514/+files/language-pack-gnome-it-base_14.04%2B20140410_all.deb [08:39] that's what has the .mo [08:40] you can use msgunfmt the .mo to see the content [08:41] seb128, cool, thanks for the hint. yeah, i've that installed on my system and I can reproduce the problem with IT [08:49] * greyback waves hello [08:49] * tsdgeos waves back to greyback [08:52] tsdgeos, believe it or not it still conflicts... bzr is stupid :| [08:53] greyback, o/ [08:54] * Saviq leaves killqt51 merge for later :| [08:55] Saviq: ?Ç¿ [08:55] ok :/ [08:55] Saviq: are we landing this for utopic? [08:56] paulliu: you there? [08:56] tsdgeos, yes [08:58] tsdgeos: yes [08:59] tsdgeos, FYI: http://anonscm.debian.org/gitweb/?p=pkg-kde/qt/qtdeclarative.git;a=shortlog [08:59] paulliu: about the zooming branch, it is much better now [08:59] but it still feels a bit weird [08:59] tsdgeos, all the → lines are package renames [08:59] tsdgeos: what's the weird thing? [08:59] that we're inheriting from debian [08:59] Saviq: damn debianites :D [09:00] pete-woods: your comment makes me wonder whether we should just limit the amount of items you can put into a gmenumodel [09:00] paulliu: let me try to take a photo, probably will make it easier to show what i mean [09:00] pete-woods: the added parameter is a gvariant list of the new items to add [09:00] tsdgeos: btw, I'm running unity8 on Galaxy Nexus, it is a bit slow now. [09:00] larsu: ahh, okay, didn't realise, I was just thinking about all the same checks I added to HUD [09:00] paulliu: it's not about speed :) [09:00] tsdgeos: ok [09:00] pete-woods: but I guess if I restrict section size to something like 1000 you'll have problems with the hud? how many items do you usually have in there? [09:00] would you be okay with any limit=? [09:00] larsu: I think you're spot on with the idea of there being some sort of safety valve, though [09:01] pete-woods: right. Normal menus will be totally fine with 1000 as a max, but will the HUD? [09:01] larsu: the biggest app is Gimp, at least as far as I've discovered so far [09:01] larsu: HUD is only dealing with normal menus [09:01] on the phone they're even smaller [09:01] like 10 entries [09:02] pete-woods: so total amount of menu items for the whole menu bar is the max for the hud [09:02] makes sense now that I think about it :) [09:03] seb128, ok, I found where the problem is, it's in the translation file of home scope; it's also shipped with language-pack-*-base - what should be the project to report the bug against? [09:03] larsu: I'd be much happier with someone complaining their app needs a bit more space in HUD, than for someone to just blow HUD / unity-panel-service's memory usage through the roof [09:04] pstolowski, ubuntu-translations [09:04] seb128, ty [09:04] yw [09:05] pete-woods: I agree on the memory usage part, but what do you mean by "their app needs a bit more space in HUD"= [09:05] ? [09:05] larsu: I mean, say their menu had been truncated somehow [09:05] and it only indexed the first 1000 items.. [09:05] something like that [09:05] pstolowski, that bug already has a component for that, reassign the ubuntu part ot language-pack-gnome-it-base I guess [09:06] pete-woods: cool, I'll do that than. Thanks [09:06] *then [09:07] larsu: thanks :) [09:10] paulliu: so this is the original situation http://imgur.com/txGuDoa and this after i do some zoom+pan http://imgur.com/P0NyuLo [09:10] paulliu: i don't think it makes sense i can pan "past" the image [09:10] and end up with 70% of empty space [09:10] paulliu: am i making sense? [09:19] pstolowski, do you have specifics about what is wrong in those translations? [09:19] tsdgeos: ok. I got it. [09:23] tsdgeos, there's transitional packages in place fortunately, but we still might get bitten here and there (I already have a branch to switch unity8's deps to the new names) [09:31] good :) [09:32] * greyback back in 5, hoping reboot solves bluetooth bugs [09:33] seb128, yes, i've added it to bug report [09:33] pstolowski, saw that, thanks [09:58] Cimi, re bug #1210458 - it's irrelevant now I'd say [09:58] bug 1210458 in unity8 (Ubuntu) "Dash should scroll down before previewing if the previewed item is not fully on screen" [Undecided,New] https://launchpad.net/bugs/1210458 [09:58] Cimi, it was about the "see-through" part of the preview, which we don't have any more [09:59] Saviq, didn't realise we don't have it anymore :) [09:59] yep, irrelevant [09:59] Cimi, already marked as such [09:59] /food [10:15] interesting https://bugs.launchpad.net/unity8/+bug/1288548 [10:15] Launchpad bug 1288548 in Unity 8 "can drag from all four edges simultaneously" [Undecided,New] [10:19] brrr, i got some cold tonight, feeling a bit sick-ish :/ [10:30] Trevinho: hello! I would like to ask some questions regarding LIM [10:31] sil2100: hi [10:31] Trevinho: actually, I need to know if a bug someone reported to appmenu-qt5 is a bug in appmenu or LIM itself [10:31] Trevinho: so, we got this bug reported: https://bugs.launchpad.net/unity/+bug/1302084 [10:31] Launchpad bug 1302084 in appmenu-qt5 (Ubuntu) "Menu in the global title bar is missing after a window child in Qt5 applications" [Critical,In progress] [10:32] Trevinho: it seems the bug only happens with LIM enabled, with global menu all works ok and the menu is correctly exported - with LIM it seems the system gets confused on where the menu should be located for Qt5 apps [10:33] Trevinho: I checked from my appmenu-qt5 side, and it seems we're not doing anything specific in this case - when the child window appears no re-parenting happens, there's actually no action from the appmenu [10:33] sil2100: yeah, I know about that, but it's not an unity issue [10:33] sil2100: indicator-appmenu just doesn't give us the menus back in that case [10:33] Trevinho: oh? [10:34] sil2100: yaeh, just try to call [10:34] gdbus call --session --dest com.canonical.Unity.Panel.Service.Desktop --object-path /com/canonical/Unity/Panel/Service --method com.canonical.Unity.Panel.Service.SyncOne 'libappmenu.so' [10:34] after that you've opened a child iwndow [10:34] Trevinho: I'm a bit out-of-date in how LIM works, but is indicator-appmenu still used when LIM is enabled? [10:34] sil2100: indeed [10:35] Trevinho: ok, so it's a known indicator-appmenu issue then - I'll try what you mentioned to double check :) THanks - do you know if we have a bug for that already? [10:35] (besides this one?) [10:38] sil2100: mh, wait.. I've done another check now and it seems that the mnenus are there... IT's another issue then... We don't get the signal of the menus being updated, and that's probably the case, then... I was wrong about that [10:38] sil2100: I checked that long time ago, and it had all the symptoms of a weird indicator issue, but it seems only partially a problem [10:39] sil2100: I mean, the appmenu-qt should not, imho, expose menus for dialog windows === MacSlow is now known as MacSlow|lunch [10:39] Trevinho: hmmm [10:41] Trevinho: from appmenu-qt side of things, it so happens that we're not actually exporting anything new - it's like, we're exporting the main menu of the parent application when it's starting and then even when the dialog window is opened, we simply do not do anything - the previous menu is left exported and no additional action is performed [10:41] Trevinho: is that not correct? [10:41] Since theoretically, the child dialog window will never have a menu of its own, so Qt5 apps expect the main menu to be still used [10:42] sil2100: imho appmenu-qt should notify that menus changed anyway, and not exporting menus for child windows [10:42] I'll try looking if that's possible, since hm, it's Qt5 that's deciding things like that most probably [10:42] I think by design dialog windows should never expose menus (unless they don't provide one) [10:43] Trevinho: ok, thanks for the info, I'll dig into that a bit more and poke you later if I have anything :) [10:50] mhr3: any hits what could be the reason? [10:50] SmartScopesClient.get_search_results(): Failed to retrieve search results for query 12 [10:50] SSQueryObject::run(): unity::LogicException: Node does not contain a string value [10:51] mhr3: I can see the message, I just don't know what "Node" conceptually maps to in SSQueryObject [10:51] mhr3: result set seems fine to me http://paste.ubuntu.com/7351649/ [10:51] karni, it will be json node [10:51] oh [10:51] so trying to read a string from the json, but can't [10:52] karni, perhaps missing dnd_uri? [10:52] not sure [10:52] the error msg isn't overly helpful :) [10:53] result 'uri' is there, not sure what you mean by dnd_uri. also, found this https://code.launchpad.net/~stolowski/unity-scopes-api/optional-dnd-uri/+merge/205002 [10:56] karni, doesn't mean the json parser finds it optional too ;) [10:57] of course if that's the issue, we need to fix it [10:57] mhr3: this used to work on Friday, so I think it's not that :D must be my mistake somewhere. [10:57] Ignore me for now. [10:57] karni, ah, hm [10:58] mhr3: ah, I was supposed to bug you about scope:// uri support ;D Although it's not as important as it was for me already. So, low prio. [11:00] karni, :) i want to get to that today, but doing reviews and stuff atm [11:00] kewl [11:01] besides having multiple JSON root elements, json formatter claims it's correct. what is up. [11:08] karni, well, the only thing that isn't string is the use_background [11:08] mhr3: it wasn't it. preparing an interesting paste for you. [11:10] mhr3: sent you PM === alan_g is now known as alan_g|lunch [11:36] https://bugs.launchpad.net/unity-scopes-api/+bug/1313673 [11:36] Launchpad bug 1313673 in unity-scopes-api "Category JSON parser expects string in render_template" [Undecided,New] [11:39] https://bugs.launchpad.net/unity8/+bug/1234730 ? [11:39] Launchpad bug 1234730 in Unity 8 "Scopes still visible with all plugins disabled" [Undecided,Triaged] [11:42] Cimi, invalid now === MacSlow|lunch is now known as MacSlow === alan_g|lunch is now known as alan_g [12:36] Saviq, so after all https://code.launchpad.net/~macslow/unity-api/version-bump-to-0.1.3/+merge/216122 is not really needed, if the package-bump is happening with lp:~mhr3/unity-api/localized-strings [12:36] MacSlow, read the second sentence in https://code.launchpad.net/~macslow/unity-api/version-bump-to-0.1.3/+merge/216122/comments/516349 [12:37] or third, actually [12:37] Saviq, we really need to split up the shell interfaces [12:37] they shouldn't be in lp:unity-api [12:37] mhr3, it does get confusing a bit indeed [12:37] mhr3, maybe they should just live with lp:unity8... [12:38] Saviq, there's no unity-notifications.pc.in [12:38] Saviq, hm.. perhaps, didn't think where to put it [12:38] mhr3, still valid? https://bugs.launchpad.net/unity8/+bug/1263246 [12:38] Launchpad bug 1263246 in Unity 8 "[Dash] clicking on search result causes click scope to receive two activations" [Undecided,New] [12:40] MacSlow, sorry, it's set(Version...) in include/unity/shell/notifications/CMakeLists.txt [12:40] Saviq, or did you imply "unity-notifications-impl-3" with "unity-notifications.pc.in"? [12:40] MacSlow, it ends up as unity-notifications.pc in the package [12:40] Saviq, oh... I didn't know that. [12:41] Saviq, on of those magic-behind-the-doors things of cMake?! [12:41] MacSlow, not really, we're configuring the file explicitly, only using a single template instead of per-component ones [12:41] Cimi, well, none of the code it points to gets executed these days [12:41] Cimi, so i'd say no [12:42] mhr3, incomplete then [12:42] thx [12:42] Cimi, is incomplete code for "let's hope it auto-expires in 60days"? [12:43] mhr3, ideally yes [12:43] mhr3, the point is to make sure valid bugs are not marked as invalid [12:45] Saviq, going to eat some kielbasa francuska === _salem is now known as salem_ [12:46] Cimi, french don't do sausage :P [12:46] Cimi, it's good you did that in polish, english would sound dirty :P [12:46] sounds equally dirty in polish ;) [12:47] Saviq, bought in a polish shop [12:47] I just read the label :D [12:48] they are pretty good [12:48] right, so it's polish "french sausage", that makes (almost) more sense :) [12:49] Saviq, stop trolling the frenches! [12:49] "french don't do sausage" [12:49] where did you get that?! [12:50] seb128, might be just french living near german border :) [12:50] shame that didrocks is not there to defend himself [12:50] he's surely busy eating snails :) [12:51] Saviq, the version used in include/unity/shell/notifications/CMakeLists.txt needs to be 3 I assume [12:51] MacSlow, yes [12:57] Saviq, ok... I hope now everything is finally in order [12:57] MacSlow, k, will look === jhodapp|afk is now known as jhodapp [13:17] Saviq, was this due to openEffect? https://bugs.launchpad.net/unity8/+bug/1283865 [13:17] Launchpad bug 1283865 in Unity 8 "Opening the dash preview causes high CPU usage" [Undecided,New] [13:17] Cimi, I doubt that [13:18] Cimi, this one would need evaluation still, but I don't think this is the case any more === ken__ is now known as kenvandine [13:45] Saviq: hey man, howzit going :)? wanted to ask if we have already any support built in for departments or other fancy scope features, or still "just" search and categories. [13:45] karni, we still don't have those fancy things *designed*, dude [13:45] or, to call it the other way, we have them designed 6 times over already [13:46] and now we're waiting for another ;P [13:46] Saviq: ah, kewl dude. was just wondering, cose I could make use of them now :D so, needed a small update. thanks :) [13:46] Saviq: hehe, gotcha! [13:46] but yeah, there's first visuals incoming, so we'll be jumping on it soon, first tsdgeos is fixing the performance issues [13:47] we'll probably tackle emblem + attributes first, after ↑ is done [13:47] but that'd be small, and then departments here we come [13:47] Saviq: very cool, thanks Michał! === dandrader is now known as dandrader|afk [13:56] mhr3: actually, I take that back - I could make use of scope:// uris ;D (new work assigned, hehehe) [13:57] karni, you might actually want the annotation results [13:58] karni, facundo was just talking with pstolowski about them [13:58] orly [13:58] * karni chases facundo [13:58] pstolowski: unless you want to tell me more about it? ↑ [13:59] karni, hi :) [13:59] pstolowski: hi! so, any work going on around annotations? [14:03] karni, scopes-api wise they're implemented, also scopes:// uris are there. take a look at the Annotation class. Annotation is meant to carry a canned query link(s) with label(s) and optional icon, with no ability to define their look. Once shell handles scopes:// uris, you will be able to put them in Result uris, and they will invoke search with arbitrary scope [14:04] karni, to make it clear (I think putting everything in one paragraph made it confusing): [14:04] pstolowski: lovely. so, we're basically waiting for ubuntu-scopes-shell work [14:04] pstolowski: no, it made sense :) [14:04] karni, there is Annotation class that you can use for simple links that are going to be displayed between categories [14:04] pstolowski: is there a design doc for this? [14:04] karni, and there are scopes:// uris which you'll be able to use as normal uris [14:05] karni, the one and only big UX spec [14:05] pstolowski: just wondering - so, if it's a remote scope, how would it make use of an Annotation class? has this been already figured out? [14:05] pstolowski: the future dash ux doc? [14:06] karni, yup [14:07] karni, no, we've just been discussing this with facundobatista today [14:07] aha [14:08] Saviq, https://bugs.launchpad.net/unity8/+bug/1201106 shall I write the JS math in C++? [14:08] Launchpad bug 1201106 in Unity 8 "horizontal swipe of Videos Carousel makes unity consume as much as 90% cpu" [Low,Triaged] [14:08] Cimi, first this would require profiling [14:09] Cimi, to see that this actually is where the CPU is being eaten [14:10] Saviq, ok let's do it === dandrader|afk is now known as dandrader [14:32] hey! Any unity8 autopilot tooling experts around? [14:33] Wellark, also have a look in #ubuntu-autopilot [14:34] I have specific questions about unity8 setup, but I will go there [14:39] dandrader: Can re-review? https://code.launchpad.net/~paulliu/unity-mir/logout/+merge/216336 [14:39] Saviq: mterry: hey, now that U is open, anything stopping the greeter-ux-fixes branch from landing? [14:40] cwayne, no. It's approved, it will be in next unity8 landing I believe [14:43] cwayne, Qt is stopping it [14:43] cwayne, as in syncing from Debian, should be sorted today / tomorrow [14:44] Saviq: ah, i gotcha. thanks! [14:46] paulliu, ok [14:56] Saviq: I've found a way to get the current call from the telephny service with all the info I need for the active call display. But I believe we only want it to display if we havent got the phone app in focus. Any way you can think to do that? [14:57] dednick, good question... [14:57] Saviq: i guess we could sniff the application manager for it? [14:57] dednick, the obvious, hackish, solution would be to hardcode dialer-app somewhere [14:57] Saviq: yeah [14:58] dednick, a potentially better way would be to route the info through the telephony service [14:58] dednick, this would have the advantage of anything that was allowed talking to the service inhibiting the in-call hint [14:59] Saviq: hm. yeah. possibly through some registration mech. [14:59] dednick, last, but not least, you interrogating telephony service, sounds limited - are we thinking to support 3rd party apps to use this hint? [14:59] Saviq: do apps know when they are "unfocused"? [15:00] dednick, they might not know right now, but mterry is working on that AFAIK [15:00] Saviq: the telephony serices uses telepathy i think [15:01] dednick, well, yeah, I'm just wondering what's the plan to support Skype, WhatsApp, G+ Hangouts etc. [15:01] dednick, if all of them are meant to bind into telepathy - we're good - if not, we might need to think more about it [15:01] Saviq, I'm not especially working on the app side -- but apps do get mir lifecycle events in the form of Qt ApplicationActive events [15:01] Saviq: i'm pretty sure we discussed this awhile ago with telephony peeps, and we decided that if we want integration, we need it all to go through telepathy [15:01] dednick, ^ [15:02] dednick, ok, then I think all the info for that hint needs to come from there [15:03] dednick, especially since we need to know what app to focus, and I think there will still be a separate Skype app, even if it will all go through telepathy [15:03] that's somewhat of a design / architecture question, too... [15:04] sounds like we need a Malta session for this [15:04] Saviq: ya. good idea. [15:04] kgunn, ↑ "support for 3rd-party voice / video calls (Skype, WhatsApp, Hangouts)" [15:05] kgunn, I'm worried we need to inject that into the first week, though, and join remotely [15:05] or maybe not, is phone "Core"? [15:06] nope, Bill's folks are there for the first week [15:08] kgunn: speaking of breakouts - I remember you saying there should be one for trust sessions? [15:10] Saviq, tsdgeos is there a way to profile the carousel from make tryCarousel? [15:11] Cimi, sure, only you need to run qmlscene yourself [15:11] Cimi, and pass -qmljsdebugger=port:BLAH [15:12] maybe we should enable ↑ by default in try* [15:12] yes please! [15:12] I was editing the cmake :) [15:12] Cimi, patches welcome! [15:14] Saviq, 3768 ? [15:14] Cimi, whatever [15:14] Cimi, that's the default probably [15:15] yes [15:17] Saviq, you sure is the right flag? http://paste.ubuntu.com/7353052/ [15:17] I see on the shell QML Debugger: Waiting for connection on port 3768... [15:17] Cimi, yeah, and? [15:17] but when I start remote profiling in the IDE doesn't seem to work [15:17] UbuntuSDK [15:17] Cimi, you need to stop it [15:18] Cimi, to get data [15:18] I did [15:19] no data [15:19] Cimi, did you switch the tab? [15:19] Cimi, I get data in Timeline [15:20] and Events [15:20] of course I did [15:20] Cimi, SOA#1 then [15:21] Saviq, First I run the command [15:21] then I enable external profiling in ubuntusdk [15:21] I go back to the qmlscene app and play with it [15:21] stop the profiling [15:21] no data... [15:21] http://paste.ubuntu.com/7353064/ [15:21] full make command line called [15:22] Cimi, works for me [15:22] but wow the rendering is b0rked :| [15:23] Saviq, hah like tsdgeos [15:23] yeah [15:23] Saviq, well I can propose the branch but doesn't work here :D [15:23] also tried running the command manually [15:23] Cimi, find out why it doesn't work, then :) [15:23] /home/cimi/Development/unity8/carousel-perf/builddir/tests/uqmlscene/uqmlscene -qmljsdebugger=port:3768 /home/cimi/Development/unity8/carousel-perf/tests/qmltests/Components/tst_Carousel.qml -I /home/cimi/Development/unity8/carousel-perf/builddir/tests/utils/modules -I /home/cimi/Development/unity8/carousel-perf/builddir/tests/mocks [15:24] maybe I am missing some qml modules? [15:24] not likely [15:24] Cimi, try telnetting into that port [15:24] does the timer run in QtCreator? [15:25] Saviq, http://paste.ubuntu.com/7353064/ [15:25] ops [15:25] cimi@draco:~/Development/unity8/carousel-perf/builddir$ telnet 127.0.0.1:3768 [15:25] telnet: could not resolve 127.0.0.1:3768/telnet: Name or service not known [15:25] anyway no [15:26] mzanetti, Saviq, if you guys could chime in... https://code.launchpad.net/~paulliu/unity-mir/logout/+merge/216336/comments/517190 [15:26] dandrader: ack [15:26] Cimi, telnet doesn't understand colon [15:27] Cimi, space instead [15:27] to me it feels like this logout api and how things are wired together deserves a bit more thought [15:27] Saviq, ok I can connect, but cannot see anything [15:30] Cimi, that's because QtCreator sends some commands to actually start the debugging [15:30] Cimi, it's some internal protocol [15:31] Saviq, but from telnet I see my commands like echoing in the app [15:31] on the terminal, output [15:31] so maybe I can write that command in telnet [15:31] dandrader: yeah. I don't think this belongs in the applicationmanager [15:32] dandrader: feels more like a session API. Don't we get this from lightdm? [15:35] mzanetti, no idea [15:35] dandrader: what exactly is the purpose of that? just to close down all apps on logout? [15:36] mzanetti, from AppManager's point of view: yes [15:36] * mzanetti doesn't feel like we should expose a D-Bus iface for that... rather connect the AppMan to something else and react to a logout signal of some sort [15:36] but I just jumped on the subject when I saw that MP. not aware of the grand plan (if any) [15:37] I guess knowing the grand plan would be helpful here... Saviq, seems we need you on this when you have a minute [15:38] dandrader, mzanetti, there indeed isn't a grand plan yet, I don't know enough about the legacy session management interfaces and mechanisms [15:38] something we are planning to get through on Malta [15:39] maybe mterry could add some insight? [15:39] maybe mterry can shed some light too, given he's our local lightdm expert [15:39] :) [15:39] ditto :) === jhodapp is now known as jhodapp|lunch [15:39] * mterry reads back [15:39] Trevinho, kudos! [15:39] davmor2, you too! [15:40] Saviq: thank you! :) === alan_g is now known as alan_g|afk [15:40] thanks [15:40] mzanetti, dandrader: are you guys talking about logging out? [15:40] mterry: yes [15:40] who triggers that, and how would re react in unity8 to it? [15:40] s/re/we/ [15:41] mzanetti, dandrader: lightdm doesn't actually manage that. I think in GNOME and unity7 there is a DBus name owned by the shell that handles that [15:41] mterry: so the whole session management thing is handled by the shell? [15:41] mzanetti, dandrader: and then lightdm notices that the process it started died and shows a greeter [15:41] yeah. org.gnome.SessionManager [15:41] ah... I see [15:41] and that's from gnome-session [15:42] mzanetti, dandrader: note that for locking or switching to a new user, lightdm does get involved [15:42] ok [15:42] but its the shell probably that calls LightDM::startNewSession() or similar [15:42] mterry, also, for power button, right? 'cause if there's no session manager, power button or ctrl+alt+del goes through to the vt and acts directly even from X [15:42] mzanetti, greeter asks lightdm to start a new session when user logs in, then lightdm starts the appropriate process [15:43] org.gnome.SessionManager is rather big [15:43] mzanetti, so lightdm launches shell, not the other way around [15:43] yeah sure, I meant when you're in a session and want to start a second one [15:43] but anyways... I got the idea [15:43] Saviq, shell handles power button today yeah [15:43] Saviq, and likely in future too [15:43] ok... so we need an interface in the shell. [15:43] mterry, sure, but it's not a simple key event, is it [15:44] mzanetti, ah yeah. If you want to switch to another user, there is a lightdm DBus interface for asking for that [15:44] mterry, since if you have an empty X11 server, power button and c+a+d do halt/reboot [15:44] mterry, so it's the session manager that prevents that from happening, seb128 mentioned some inhibition mechanism [15:45] mzanetti, dandrader, in any case, what Paul was doing was a quick'n'dirty thing to allow logging out from the unity8 preview session [15:45] Saviq, I'd have to confirm, but I think it's just that whichever session has input focus gets the press, eh? [15:45] Saviq, mterry: http://www.freedesktop.org/wiki/Software/systemd/inhibit/ [15:45] Saviq: ok... do we want the quick'n'dirty? [15:45] mzanetti, dandrader, looking as we don't have a grand plan on that yet, I'd rather land something that's good-enough [15:45] seb128, OK. Saviq ^ new and shiny way to do it [15:47] dandrader: Saviq: imo it still should be a own plugin, not in the ApplicationManager plugin === alan_g|afk is now known as alan_g [15:47] should still be quick enough [15:48] mzanetti, sure, I actually have no hard opinion on this... not until we come up with a grand plan ;) [15:48] so I'll let you guys compromise :) [15:49] dandrader, ↑ [15:49] yeah... so especially if this is quickndirty I don't think the already complex applicationmanager should get more quick'n'dirty stuff [15:50] just a Unity.Session plugin with a signal logoutRequested() and a slot logout() should do I guess [15:53] mzanetti, there are two kinds of logout calls coming through d-bus: one that should just logout immediately and another that will prompt the user for acknowledgement [15:56] dandrader: ah ok... so yes, slightly more then, but still a sepearate plugin for the dbus iface, no? [15:56] mzanetti, and if it's quick and dirty, we don't need to change AppMan at all. just but that loop to close all apps in unity8 and be done with it [15:56] and then just close all the apps through AppManager's unity-facing API [15:56] yep [15:57] s/just but/just put [15:57] right [15:57] well, I could see a use for ApplicationManager::closeAll() in the future too... but yeah, to not have to bump unity-api and everything we can just loop over them in qml [15:58] and add the closeAll in the next unity-api iteration [15:58] ok, so we have an agreement. will update the MP. [15:58] thanks [16:02] Saviq, tried still no luck [16:03] Cimi, dunno how to help you, didn't hear anyone having such issues - try with a small QML UI project started from QtCreator itself [16:06] Saviq, qmlprofiler -p 3768 -attach 127.0.0.1 works for me [16:06] Saviq, with qmlprofile you press r to start/stop recording [16:06] then it saves to a file I import in the sdk [16:07] Cimi, so sounds like qtcreator fails? [16:07] Saviq, yeah [16:07] Cimi, wonder if the ubuntu plugins cause that (I don't have them installed) [16:10] Saviq, https://code.launchpad.net/~cimi/unity8/qmljsdebugger/+merge/217474 === jhodapp|lunch is now known as jhodapp [16:41] Saviq, confirm is the ubuntu plugin that breaks qtcreator profiling === dandrader is now known as dandrader|lunch [17:41] * mterry upgrades to utopic [17:42] mterry: freals? [17:42] cwayne, felt wrong to be on a stable release [17:42] mterry: welp, i might as well do it too [17:43] cwayne, that's the spirit! ;) [17:43] i need some challenge, ya know [17:43] 14.04 is too boring :P [18:14] kgunn, what has to happen to transition a silo like 002 to utopic? === dandrader|lunch is now known as dandrader [18:47] mterry: i think they have to "take down the silo" and retarget it [18:47] mterry: do you want that to happen now ? [18:48] kgunn, sure? And add lp:~mterry/lightdm/resettable while there, if you can [18:48] mterry: ok...the main thing, you'll lose whatever packages are there right now [18:49] of course, you'll get new utopic ones soon after [18:50] mterry: btw, i already had that mp in there.... [18:53] kgunn, I know we talked about it last week, but I didn't see it in the PPA, so wasn't sure [18:54] maybe it didn;t build ?... i'll double check it this go round mterry [18:56] kgunn, I solved a merge conflict with the branch today, maybe that conflict happened last week before you built it [19:49] Cimi, filed bug? === dandrader is now known as dandrader|afk === salem_ is now known as _salem === dandrader|afk is now known as dandrader