=== _salem is now known as salem_ === salem_ is now known as _salem === Wig is now known as Wug [08:36] Saviq, before you needfix me [08:37] Saviq, in the infographics, I need to get the user id [08:37] uid [08:38] so I did this http://paste.ubuntu.com/7420297/ [08:39] is not super nice as using a model role, otherwise we have to patch lightdm [08:49] anyone know how to set the display to "always on" on device? can't seem to get powerd-cli to do it... [08:58] dednick, you get something like https://www.youtube.com/watch?v=Z86V_ICUCD4 [08:58] but instead it just taps on your phone regularly [08:59] mhr3: lol [09:06] Cimi, better let's patch lightdm [09:06] ahahahahahha [09:06] I know it!!! [09:06] Cimi, especially since we only use a local copy of the LightDM plugin still ;) [09:06] Saviq, but we need to patch real lightdm [09:06] Cimi, until the split greeter [09:06] Cimi, sure, I know, let's [09:07] Saviq, hey [09:07] Saviq, https://blueprints.launchpad.net/ubuntu/+spec/client-1410-unity8-desktop-iso just as fyi [09:07] "by 16.04", not "for 16.04" ;) [09:08] subscribed [09:13] dednick, "sync item progress" is different than "transfer" is it? [09:16] Saviq: hmm. not sure exactly. I know there is an "indicator-sync" [09:16] Saviq: https://launchpad.net/indicator-sync [09:16] and an https://launchpad.net/indicator-transfer [09:16] dednick, ok, I'll leave it be then [09:16] Saviq: but the sync one is pretty old. [09:17] dednick, kk [09:17] so they may have been merged [09:17] dednick, I'll leave it be then, we'll adapt when we know more [09:17] ok [09:19] Saviq, "by" indeed, thanks ;-) [09:19] (fixed) [09:30] Cimi, we're to tell Loïc what are our biggest beefs with the shape, apparently the new implementation is too far off still... [09:30] Saviq, far off? [09:31] Saviq, like we're not getting things? [09:31] Cimi, incomplete still [09:31] Saviq, what are we getting? [09:31] Cimi, we need to list things we need to see what can be backported [09:31] I talked to kevin [09:31] hold on, logs [09:31] Cimi, I know about: [09:32] 1) ability to add any item inside and let ubuntushape clip it [09:32] 2) enable/disable inner shadows and border [09:32] 3) add external shado [09:32] 4) peformance improvements and caching [09:32] overlay support [09:32] Saviq, overlay support should fall into 1 [09:32] Cimi, 1) is possible via a ShaderEffectSource [09:33] Cimi, we're actually using it here and there [09:33] ;) [09:33] Saviq, yes [09:34] Saviq, in Components [09:34] now that I think of it, we should use it more [09:34] Saviq, but we want it integrated [09:34] Cimi, yeah sure, but I don't think that's a prio then, since there's a workaround [09:34] Saviq, so 2 and 3 though [09:34] Saviq, or more generally, ability to change border source [09:35] sometyhing like that [09:35] Cimi, don't we have that already? [09:35] Cimi, borderSource? [09:35] having always the inner shadow sucks [09:35] Saviq, we had [09:35] Saviq, now doesn't work anymore iirc [09:35] since a year or a bit less [09:37] Cimi, think it's still working, we use it in launcher and running tile [09:37] Saviq, where? [09:37] Cimi, grep for borderSource [09:38] Saviq, those refers to prefixed border sources in the sdk [09:39] Saviq, you cannot set to null [09:39] or anything you like [09:39] Cimi, well, it's set to "none" in LauncherDelegate, but sure, we can't override [09:40] Saviq, http://paste.ubuntu.com/7420521/ [09:41] Saviq, I wrote that code [09:41] in launcherdelegate :) [09:42] Saviq, I remember when I did borderSource: "CimiDoesntWantBorder" [09:42] someone needsfix that :D [09:43] maybe I can use it for the carousel... let's see [09:52] any volunteers? https://code.launchpad.net/~gerboland/unity-mir/fix-upstart-closed-apps2/+merge/218721 [09:53] for functional testing at the least [10:02] greyback, I'll take it [10:02] Saviq: thanks [10:25] Saviq, came up with something like that for lightdm http://paste.ubuntu.com/7420699/ [10:26] Saviq, how to I cast from a uid_t to guint64 in qt? is my approach correct? [10:26] + user.uid = (quint64)lightdm_user_get_uid(ldmUser); [10:27] Cimi, looks legit [10:27] ok [10:27] Cimi, although mterry will obviously be better to review [10:28] Cimi, is there nowhere you can get the uid from lightdm instead of getpwnam? [10:28] Saviq, that is a mistake [10:28] Saviq, that function is not called [10:29] I am using lightdm for the model [10:29] indeed [10:29] Cimi, right [10:29] I discovered lightdm already stored the uid [10:29] after I added the function [10:29] Cimi, yeah, looks legit, MP please and get mterry to review [10:30] http://paste.ubuntu.com/7420719/ [10:30] ok [10:30] I'll wait a jenkins build [10:32] Saviq, just saw https://blueprints.launchpad.net/ubuntu/+spec/client-1410-suru-icon-switch [10:32] Saviq, I've https://code.launchpad.net/~seb128/ubuntu-system-settings/use-theme-icons/+merge/214950 up but it's blocked on your uitk change to land [10:32] Saviq, just as a fyi [10:33] cyphermox, ^ (since you have the W.I on that blueprint) [10:33] seb128, link them? [10:34] Saviq, done [10:34] cyphermox, took over your WI as well, set it as done === MacSlow is now known as MacSlow|lunch [11:13] Saviq, https://code.launchpad.net/~cimi/unity8/preview-text-summary-bottom-padding/+merge/217903 [11:13] Cimi, yup, will do it soon [11:18] pete-woods, hey, are those SVGs animated? [11:18] Cimi: nope, just boring ol SVGs [11:18] Cimi, animated SVGs are security nightmare [11:18] Cimi, since you need to support javascript [11:19] Saviq, so why we switched to use SVG? [11:19] Cimi, because we want to adapt the colours [11:19] Saviq, and you can do it other ways with properties [11:19] Cimi, and we switched to SVG to allow easier visualizers than QML [11:19] Cimi, and again, custom QML → security nightmare [11:20] Saviq, properties [11:20] anyway ok [11:20] Cimi, properties what? [11:20] I will tell designers anymations are gone [11:20] Saviq, you can expose colors [11:20] Saviq, like we do with the ubuntu themes [11:21] Cimi, expose them to what? to SVGs? yes, that's what we want to do [11:22] Cimi, but we didn't want to run any javascript (so animated SVGs and/or QML are not good) [11:24] Saviq, we could have exposed other properties from a text file [11:24] or anything really [11:24] and use animated qml [11:24] if we want just to change colours... [11:25] if we want to change shapes, it's fine [11:26] Cimi, NO ANIMATED QML [11:26] Cimi, security BAD [11:26] Cimi, and we didn't even want animated anyway [11:26] why bad? [11:26] Cimi, security [11:26] our shell is animated [11:27] our apps are animated [11:27] now is insecure? [11:27] Cimi, 3rd party infographics in the shell [11:27] ah 3rd party ok [11:27] Cimi, and anyway they're meant to be really as non-intrusive as possible [11:27] thought we were drawing the infographics [11:27] Cimi, so animation is not wanted [11:27] and give develelopers a couple of values to change [11:27] Cimi, nope, we want them to be installable as clicks and such === _salem is now known as salem_ === MacSlow|lunch is now known as MacSlow [12:18] Saviq, seen the design on departments behaviour yet? [12:19] mhr3, you mean https://sites.google.com/a/canonical.com/unity8dash/toolkit/2-navigation ? [12:19] mhr3, see bottom for my comments [12:20] indeed [12:21] Saviq, i'm interested in that from the model pov, i'll have to expose some kind of partial lazily loaded tree [12:21] the widget for that will be... interesting i guess [12:21] aaaargh foooking gsites [12:21] limited characters in comments... what is this, 1990s? [12:23] mhr3, since any change is meant to be applied instantly (it is, isn't it? [12:23] ) [12:24] mhr3, or maybe it isn't... [12:25] Saviq, well, yes, basically the widget isn't supposed to close by itself, unless you tap a leaf dep [12:25] mhr3, right, so nothing until you explicitly select a thing [12:25] and at the same time we don't have the whole tree, so loading children's children requires a query === alan_g is now known as alan_g|lunch [12:26] mhr3, yeah, sounds like lazy tree is what we need... [12:26] children exposed as a ListModel [12:26] FUN [12:26] and did i mention that we'll be completely relying on the scope giving us proper data? [12:26] kmn [12:27] but well.. doable === alan_g|lunch is now known as alan_g [12:27] mhr3, and how do we not rely on scope giving proper data now? :0 [12:28] Saviq, well, this is different, we still expect stateless queries, but them having overall hierarchy [12:29] mhr3, ah in that sense, yeah [12:29] Saviq, anyway, i take it the widget will not be an issue :) [12:29] Saviq, still, i'll need details on how you want the model(s) for it exposed [12:29] mhr3, as long as you prepare the model ;) === jhodapp|afk is now known as jhodapp [12:30] mhr3, definitely a ListModel for the results, and probably the two links separately [12:31] mhr3, in a stack? [12:32] so... a dynamic ListModel of department pages, filled as we descend in the hierarchy [12:32] mhr3, the shell might maintain the stack, though, maybe that's better [12:33] since we want to request a department page and only then push it onto the stack, to have the first items added [12:33] Saviq, you mean the model always being flat and updating (clear+re-populate) when navigating? [12:34] mhr3, hangout? [12:35] Saviq, want to talk with pstolowski about it too, so let's do one in a bit [12:35] mhr3, ok, ready when you are [12:35] pstolowski, or you available now? [12:35] yay diffs from gsites!!! [12:36] mikenagle, you get a diff if you subscribe to a site!!! [12:36] ah, he's on lunch === alan_g is now known as alan_g|lunch === jhodapp is now known as jhodapp|brb [12:53] AAARRGH [12:53] it even discarded part of my reply over email [12:53] like WTH [12:54] mhr3, i'm now [12:55] Saviq, pstolowski, k, ho in 5? [12:55] hooo! [12:55] k [12:55] too much sugar... or maybe too little? [12:57] Saviq: take more pain kilers [12:57] oh jeez this is the most half-baked comment support EVER [12:57] greyback, ah, the blue pill [12:57] wait [12:57] that explains things [12:57] * Saviq looks at the bottle [12:57] Saviq: they're made for comments, not essays :) [12:58] greyback, yeah, check that: only two-level threads, i.e. you can only reply to the top-level one [12:59] top-level threads have
after the name, sub-level ones don't! [12:59] Mike added attachments, I even got email notifications about them, and can open them, but there's no links on the site! [13:01] comments + email snippets = super easy to follow everything [13:02] sure, I get that, but otherwise they'd end up satisfying Zawinski's law [13:02] gotta draw a feature line somewhere [13:03] greyback, well, in docs you can reply to sub-comments can't you? [13:03] maybe it's the same, just here it suggests you can do more [13:04] * greyback decides not to get into this argument, goes to do something productive ;) === jhodapp|brb is now known as jhodapp [13:19] you guys talking about comment system on launchpad merge proposals? [13:20] * dandrader curious === alan_g|lunch is now known as alan_g [13:48] mterry, ping :) [13:49] Cimi, hello! [13:49] mterry, ciao! [13:49] mterry, I did this https://code.launchpad.net/~cimi/lightdm/uid-bindings/+merge/218958 [13:50] Cimi, ah cute. That's nice [13:50] Cimi, I'll let robert_ancell review it though. He's usually master-of-lightdm [13:50] mterry, my only concern is tests, I have no clue where to start [13:51] Cimi, lightdm tests are interesting yeah. Are there similar tests for other propertise? [13:51] not sure [13:53] Cimi, there are several other tests that deal with user properties (tests/scripts/user-*.conf [13:54] Cimi, but uid isn't something that can be changed like the other properties [13:54] Cimi, so testing is a little odder there [13:54] that's why my concern [13:55] Cimi, here's what you do. [13:57] who can trigger a build btw [13:57] Cimi, tests/src/test-qt-greeter.cpp and tests/src/test-gobjecte-greeter.c (and the python greeter if you're feeling funky) have support for LOG-USER to query a specific user property [13:57] would be nice to play with a arm build [13:58] Cimi, so add support for the UID field. Then add a new baby test script that queries the field [13:58] Cimi, then add two "runner" scripts in ./tests (you'll see all the others there) that calls your script with the gobject greeter and the qt greeter (to test both interfaces) [13:59] Cimi, your baby test script can look like ./tests/scripts/user-name.conf but simpler and just for UID [13:59] hey hey hey [13:59] one thing at a time :D [13:59] it's a complete new way of writing tests for me, I need to read code first :) [14:00] I'll start with reading those two files and add UID [14:00] Cimi, well go back and read above later :) [14:11] mterry, I don't see tests for username etc etc [14:11] Cimi, tests/scripts/user-name.conf [14:13] mterry, uid is not tested [14:13] Cimi, right... but you just added it, eh? [14:13] mterry, no [14:13] mterry, it was already in common etc [14:13] Cimi, you added the public api [14:13] like gid [14:14] Cimi, that's what this is testing [14:20] Cimi, https://code.launchpad.net/~cimi/unity8/preview-text-summary-bottom-padding/+merge/217903/comments/522179 [14:27] seb128, overlay scrollbars are so slow on my core i7 with 300$ GPU that I think it's the case to trash them [14:28] Cimi, you should fix them to not be slow? [14:29] Cimi, I'm not the one deciding, talk to Mark ;-) [14:29] seb128, cannot fix a hack [14:30] seb128, they will stop working on mir anyway :) [14:30] Cimi, write a non hack implementation then? [14:30] seb128, can't [14:30] seb128, they require separate windows [14:30] build support in Mir :p [14:30] I drop them [14:31] or I change design [14:31] I want them internal like facebook [14:31] and osx (prefer facebook impl) [14:32] Cimi, that will be supported on Mir, btw ;) [14:32] Cimi, separate surface of a certain type [14:32] Saviq, that move and send positions back and forth in sync with the app? [14:32] Saviq, workarea hints etc? [14:33] it's shitload of work [14:33] let's just make them internal [14:33] Cimi, still doable, if done properly in the toolkit :) [14:33] superfast, not hackish [14:33] Saviq, i doubt [14:47] ping Saviq, kgunn: we need to meet to talk about testability requirements for this cycle and goals for the malta sprint. [14:47] can we do it next week? [14:48] elopio, sure, I'm off Monday, available rest of the week [14:48] Saviq, kgunn: what about tuesday, 10 minutes from now? [14:49] elopio use my calendar [14:49] elopio, OMG Tuesday is in 10 minutes?! [14:49] right. [14:49] elopio, and yeah, just find a slot in our calendars [14:49] Saviq: yes, sorry you didn't get to enjoy your free weekend :) [14:49] * elopio dives into google calendar. [14:50] elopio, ;) [14:55] kgunn: wow, you have all your mornings colored in the calendar. [14:55] * elopio hugs kgunn. [14:56] I've sent the invitation for wednesday [14:58] elopio, that time doesn't work for me ;| [14:59] * Saviq looks for a better time [15:13] greyback: what's the purpose of the ProcInfo class in unity-mir? why is it passed to app manager? [15:15] dednick: it's used to support the --desktop_file_hint hack mainly. It fetches the command line for a PID and let's us parse it === dandrader is now known as dandrader|afk [15:16] greyback: sure. but why is an empty one passed to the App manager? [15:17] greyback: does ProcInfo mean "all process info" or "a single processes info" [15:17] dednick: empty? [15:17] it's used here: std::unique_ptr info = m_procInfo->commandLine(pid); [15:19] we set the PID for ProcInfo to fetch, then use it to parse the PID [15:19] greyback: yeah, i know. But why is it passed to the app manager as a parameter? the class has no context. [15:19] dednick: ah, for each of testing [15:19] ease [15:19] so it can be easily mocked [15:19] greyback: ah. ok [15:20] greyback: so it's supposed to identify "all process info" [15:20] not just a single one [15:20] seb128, Saviq I think we should change designs and have scrollbars with the internal overlay [15:21] Cimi, that works for me, but I'm not a designer [15:21] is massively easier to implement, and lot faster [15:21] Cimi, you want to present that to Mark? ;) [15:21] Saviq, I could, no problem [15:21] Cimi, TBH I don't care much, doesn't involve my work either way ;) [15:22] Saviq, neither mine [15:22] Cimi, sure, but you care, I don't [15:22] but will for who in mir side will have to implement this [15:22] and qt side [15:22] and gtk [15:22] Cimi, toolkits only, Mir will support it anyway [15:23] Saviq, yes but if it's internal to the window, mir needs to do 0 [15:23] Saviq, otherwise we need features from the server [15:23] Cimi, they need to do 0 anyway, those features are planned regardless of scrollbars, so they will get implemented [15:24] whether they'll be used for scrollbars or not is a completely different topic === gatox is now known as gatox_lunch [15:24] Cimi, and I believe you're overreacting, if you can have a child surface and enforce/know its position in relation to the parent surface, it's really not that difficult to do [15:25] Saviq, it is not child surface [15:25] Cimi, it would be in Mir [15:25] Saviq, children stay within the parent dimensions [15:25] Cimi, said who? [15:25] Saviq, cannot float around the screen outside the parent [15:25] Saviq, X [15:25] Cimi, yeah, and? [15:26] since when do we care about what X implements? [15:26] Cimi, really, please trust people that thought about this [15:26] Saviq, not sure they thought about this [15:27] Cimi, believe me, they did [15:27] about how to implement overlay scrollbars in qt? [15:27] and gtk? [15:27] in Foo [15:27] cool then [15:27] I'll let them work on them [15:27] Cimi, doesn't matter in what [15:27] Cimi, what matters to Mir is making it possible [15:28] in a sane way [15:30] Saviq, performance will still suck though :) [15:30] on scrolling [15:30] Cimi, why? [15:30] Cimi, what's different with overlay scrollbars? [15:30] Saviq, because there is a client that waits for signals coming from the window manager/server [15:31] and this is out of the rendering pipeline of the toolkit [15:31] Cimi, and why does that have to suck? [15:31] because it is slower than direct input [15:31] if you move a mouse inside a mousearea [15:31] Cimi, slower than 16ms? ;) [15:31] is direct input [15:32] Cimi, I really fail to see the bottleneck here, it's one real value to communicate every frame [15:32] otherwise you move a window, this window is moved on the server that generates an event with how much it moved [15:32] there is a callback connected in the toolkit that translates this into input [15:33] Cimi, _one real value every frame_ really? bottleneck? [15:33] Saviq, it is super slow here on compiz [15:33] well now [15:33] on my machine [15:33] faster than a 3000 retina macbook [15:33] why do you translate unity+compiz+X directly to Mir? [15:34] Cimi, yes, we know you have a fast notebook [15:34] Saviq, cannot imagine if it is slow here [15:34] how it could be on a slower hardware [15:34] Cimi, good you're not writing it then [15:34] end of message, sorry [15:34] anyway [15:34] indeed [15:34] was about to say [15:34] just let's try to change design so we can have something much faster [15:35] Cimi, it's a broken argument [15:35] Cimi, "let's change design because it's slow" [15:35] Cimi, it's not inherently slow === dandrader|afk is now known as dandrader [15:35] Cimi, it's slow in the current implementation, let's make the implementation better [15:35] Cimi, some humility please [15:35] Saviq, well I also don't like the design [15:36] Cimi, so start with that, not with "it's slow on my beefy machine" [15:36] Cimi, that argument won't get you anywhere [15:36] I don't like it and is slow because of implementation [15:36] Saviq: ok, you have permissions to move it, just find a better spot. [15:36] I much prefer other implementations of scrollbars [15:36] or we can meet with you, and then with kevin. [15:36] elopio, would be easier if everyone had their work hours marked in the calendar :| [15:38] yeah, I've just updated mine. [15:40] Cimi, sure, but that doesn't lead to "let's change design" unless "I don't like it" resonates with those responsible for design and/or unless "it's slow because of implementation" is analyzed and determined to be unfixable, which in this case I tell you it's crazy to think it will be slow in every implementation - again, it's just one real value to communicate every frame [15:41] Cimi, we'd be in a much worse situation if we'd had problems to communicate that [15:41] mzanetti, look at https://blueprints.launchpad.net/ubuntu/+spec/client-1410-unity-ui-appsplash please and comment / add anything you think important [15:59] Saviq: btw.! [16:00] Saviq: we noticed an unity8 crash in our smoketesting recently, could you have someone look at that by any chance? I can fill in a bug if needed, the crash can be found here for instance: http://ci.ubuntu.com/smokeng/utopic/touch/mako/19:20140509.1:20140506.1/7941/unity8/ [16:00] Saviq: (hi!) [16:06] sil2100, isn't it the same crash on unity8 exit? [16:07] grrrrr why no stacktrace :[ [16:08] why aren't those preprocessed again? :) [16:08] :( rather === gatox_lunch is now known as gatox [16:12] Saviq, best way for carousel shadow now is adding a property to Card, can I? [16:12] sil2100, same crash - bug #1315251 [16:12] bug 1315251 in Unity 8 "unity8 7.86+14.10.20140429.2-0ubuntu1 crashes on recent utopic images" [High,Triaged] https://launchpad.net/bugs/1315251 [16:12] Cimi, can't the shadow be added in CardCarousel to the delegate? [16:12] Cimi, since it's the only place where we'd use it? [16:13] Saviq, but the delegate is the Card [16:13] Cimi, yeah, and? [16:13] Saviq, and to add shadow I need to set borderSource of artShape to an invalid value [16:13] base off of https://code.launchpad.net/~aacid/unity8/dynamic_specialized_cards/+merge/218089 btw [16:14] Saviq, as well as placing under artShape [16:14] so cardShadow [16:14] bool [16:14] might be easy, light [16:15] Cimi, why do you need to remove the border, we don't want it in carousel? [16:15] Saviq: oh, so this one is the crash on stop as well? Thought it's something new as we had one unity8-crash on test-stop already 'fixed by itself' recently [16:15] Saviq, nope [16:15] Saviq: anyway, ok, thanks for having a quick look! [16:15] sil2100, apparently not any more [16:15] Saviq, the border is an inset [16:15]