[04:05] hey guys so this multi-faceted bug has been annoying me for months... but now I have sort of a better picture [04:05] it has to do with the Unity menu bar integrated at the top and maximize/unmaximize [04:06] What is happening is that XTranslateCoordinates lies to me if I don't call XSynchronize before it, yet I will never receive further ConfigureNotify events [04:07] I'm receiving a ConfigureNotify saying that the window is now at 65, 24 in event->x,y , but XTranslateCoordinates will say it's still at 640, 251 [04:08] Putting an XSynchronize before XTranslateCoordinates causes annoying unwanted redraws but solves that particular problem. [04:16] this might be a silly X behavior :S [07:25] Cimi, https://code.launchpad.net/~cimi/unity-mir/unity-mir.stop-server_wizard/+merge/214983/comments/517509 [07:25] and corresponding https://code.launchpad.net/~cimi/unity8/unity8.upstart_unity-mir_variables/+merge/215152/comments/517510 [07:26] otherwise you can install incompatible versions [08:06] biab, doctors' time [08:11] seb128, how do I update a package be [08:11] version? (debian changelog) [08:11] dch -v 1.2.3-0ubuntu1 [08:14] what mhr3 said [09:39] mhr3: have a sec? [09:39] tsdgeos, shoot [09:41] mhr3: is it possible that you guys are emitting the components/template changed too much? [09:41] on creation of a category i'm getting this [09:41] http://paste.ubuntu.com/7358313/ [09:42] which given that we do quite expensive stuff when components/template change [09:42] it's a bit unfortunate [09:42] since they don't seem to be really changing :D [09:42] hmm, not that i would know of [09:43] tsdgeos, that's a single search, right? [09:47] mhr3: what you mean single search? === davidcalle_ is now known as davidcalle [09:48] it's the apps scope when the "local apps" is "created" [09:49] tsdgeos, in that case that really shouldn't be happening [09:49] tsdgeos, you might get a few count changed notifications, but not template/components [09:50] tsdgeos, your pastebin is from the filtermodel, right? [09:51] correct [09:52] but ideally it should not be a problem [09:52] maybe it is [09:52] let me remove the filter [09:53] seems i'm getting the same [09:53] i'll compile my own shell [09:53] and see if i can find out where they come [10:08] Cimi, your version updates need to be UNRELEASED, not utopic [10:11] Saviq, ok [10:12] Saviq, pushed [10:12] Cimi, thanks [10:13] Saviq, I guess I can put the maths inside a plugin for the carousel [10:13] Cimi, FYI: train considers everything that has a distro codename in the changelog is supposed to be in distro [10:14] Cimi, so it would fail, trying to find that version [10:14] I see [10:19] Cimi, LP says there's a conflict in your unity-mir branch, care to merge trunk please? [10:19] sure [10:19] Saviq, merged fine here [10:19] Cimi, yeah I know, LP was confused [10:20] Cimi, but that's what I was trying to resolve [10:20] I will push merged [10:20] Cimi, FWIW, no need to put [ Andrea Cimitan ] in the changelog if it's the only entry [10:21] Cimi, basically, you should just do `dch -v 0.4-0ubuntu1 "Blah blah"` and dch does everything you need [10:22] hah [10:22] Saviq, will do next [10:28] Saviq, a CarouselMath or so in DashViews is ok [10:28] or we should take listview and patch? [10:28] Cimi, no patching [10:29] I meant, subclas [10:29] you can't [10:29] it's not exported [10:29] lvwph did for flickable, no? [10:29] flickable, yes, listview, not [10:29] ok so much better to just add the maths [10:30] an object that calculates those [10:30] Cimi, I'm just not sure that will really help, TBH [10:30] Saviq, it's the only thing that can improve maths [10:30] Saviq, JS is really at its minimum [10:31] Cimi, you'd need to start with writing benchmark tests [10:31] Saviq, I doubt it is possible to improve formulas [10:31] I already did lots of work [10:31] Cimi, but anyway, you sure there's no more pressing bugs to tackle? the carousel behaves fine, sure it uses CPU, but only when you move it, and sure we could think to improve it [10:32] Cimi, but I don't think that's really a good use of your time right now [10:32] Saviq, I thought I need to learn C++ [10:32] Saviq, I read C and I understand everything [10:32] Cimi, that's not really C++ if you just move some formulas from JS to a C++ object [10:32] I read C++ and I start googling [10:33] it's a start [10:33] anyway thought was a good hit [10:33] Cimi, if you can't find more interesting bugs, I have some tasks that need attending [10:33] those are my current bugs https://bugs.launchpad.net/unity8/+bugs?search=Search&field.assignee=cimi [10:33] Cimi, I just don't see it making a visible difference [10:36] mhr3: seems it's not the scopes-shell launching the data changed signals [10:36] wonder why the changes on the qml level :/ [10:37] tsdgeos, bug #1297308 - should probably just fix this with the dynamic cards you're doing [10:37] tsdgeos, wow, i didn't know qml can even mess with them [10:38] mhr3: don't know yet :D [10:38] mhr3: just know that you're not emitting the signals but i'm getting some onchanged [10:38] still looking who to blame for that :D [10:39] tsdgeos, maybe it would make sense to do the card specialization on the plugin level? [10:39] although i'm not sure that's still what you're doing :) [10:39] mhr3: it is [10:39] i'm just doing it dynamically now [10:39] basically i have this monster [10:39] http://paste.ubuntu.com/7358572/ [10:39] that gives me the specialized code [10:40] i can make it prettier [10:40] uuh [10:40] but first i need to make sure it is performant [10:40] you're generating code [10:40] correct [10:40] evil++ [10:40] how can that be performant [10:40] because i only do it once [10:41] hmm... [10:41] and after that i create much simpler cards than before [10:41] scary though [10:42] copy&pasted code is worse [10:42] maintaince hell [10:42] at least this is all contained in the same place [10:48] Saviq, so that leads me just to the carousel bug :D [10:50] mzanetti, actually see ↑ [10:50] mhr3, only one generation per category/query, so much faster than a huge delegate of invisible things [10:51] Cimi, there's plenty of bugs that are not assigned to you ;) [10:51] Cimi, really, if you can't find one, we have features to do, too :) [10:52] Saviq, features please! [10:52] features yay! [10:52] * mhr3 would like attributes [10:52] Saviq, Cimi ^ [10:52] mhr3, yeah, I know, just don't want to step on tsdgeos's toes [10:53] although with tsdgeos working on that code, probably not... [10:53] right [10:53] Saviq, annotation renderer then [10:53] Cimi, [cimi] implement shadow under carousel cards: TODO [10:53] in https://blueprints.launchpad.net/ubuntu/+spec/client-1410-unity-ui-dash [10:54] ok cool [10:54] Saviq, I think this requires ubuntushape support, though [10:54] Cimi, how? it's meant to be outside of the shape [10:55] Saviq, otherwise you snap a png around it [10:55] Cimi, yeah, and? [10:55] not around, under, but yeah [10:55] Saviq, antialiasing on corners might suck [10:55] performance too [10:56] I will ask jouni [10:56] who is not online [10:56] rosie [10:56] who does not work anymore with us [10:56] hah [10:56] Cimi, it's just a drop shadow https://docs.google.com/a/canonical.com/file/d/0B8T-xKCA6bikUm56SG1pVEktZUk/edit [10:57] and https://docs.google.com/a/canonical.com/file/d/0B8T-xKCA6bikdGdJOHNvRUJYNHc/edit [10:58] let me try with qt graphicaleffects [10:58] [cimi] implement buffering progress in audio preview widget: TODO === MacSlow is now known as MacSlow|lunch [10:59] [cimi] make progress widget half-wide: TODO [10:59] [cimi] add pushdown effect to dash cards: TODO [11:00] thanks! [11:01] what's progress widget half wide? [11:01] Saviq: could you give lp:~aacid/unity8/dynamic_specialized_cards a try versus lp:~aacid/unity8/specialized_cards [11:02] and give me your opinion performance wise? [11:02] tsdgeos, will do [11:12] Saviq: on the device, on the pc everything is too fast [11:13] tsdgeos, yeah, I know [11:21] Saviq: was thinking of doing a cache of components [11:21] so that if i ask for the same template, compoent i return an already created Component instead of doing all the work again === _salem is now known as salem_ [11:44] tsdgeos, yeah, I'd cache them per-query at least [11:45] is there anything I need to do to run unity-scopes-tool? [11:45] Cimi, initctl emit scope-ui-starting === alan_g is now known as alan_g|lunch [12:19] seb128, hey, packaging q: we're splitting unity8-common out of unity8, we're apparently missing a Something (Breaks / Depends) somewhere, which results in conflicts when downgrading (i.e. unity8-common isn't removed when downgrading unity8), what am I missing? [12:19] https://code.launchpad.net/~saviq/unity8/split-common/+merge/214907 btw [12:22] Saviq, https://wiki.debian.org/PackageTransition [12:22] Saviq, you are in case #8 [12:22] Saviq, I think [12:24] seb128, tx! [12:24] yw! [12:46] tsdgeos, so, to cache card components, you thinking deep-compare template/components? or maybe we should stringify + concat them and md5sum or so? [12:46] tsdgeos, maybe we could even precrate some known ones on startup [12:46] Cimi, !!! [12:46] lp:~cimi/unity8/unity8.upstart_unity-mir_variables: 370 tag [12:46] s [12:46] Cimi, PLEASE clean them wherever you have a unity8 branch [12:47] Cimi, it's even more important locally than remotely [12:54] Saviq, dunno how they got there [13:00] seb128, another q: we have unity8-autopilot Depends: unity8 (>= $source:Version), I wonder if that shouldn't be == instead, WDYT? [13:00] Cimi, please let me know when they're gone, need to rebuild unity8 in the silo with some packaging tweaks [13:00] Saviq, I am running the script [13:01] Cimi, tx, and please make sure to run it on all local branches, too :) === MacSlow|lunch is now known as MacSlow [13:02] Saviq, is there a fast version that performs removing tags locally and not remotely? [13:02] Saviq, or can we intergrate the script in jenkins? [13:02] Cimi, so that's why you have them still [13:02] Cimi, you need to run the script on _all_ branches separately [13:02] local and remote [13:03] Cimi, just pass a folder as argument [13:03] Saviq, either work, == can be annoying between and arch any and all, because all is built on i386, so it creates installability issues on other arches between the builds ... but I guess it's not much of an issue nowadays with the CI train === alan_g|lunch is now known as alan_g [13:04] seb128, yeah, but then I had a modified test in autopilot, upgrading unity8 didn't pull it... so it failed [13:04] * Saviq tweaks [13:04] Saviq, ok, so change for a == if you really require them to match [13:05] probably easiest [13:05] seb128, == vs. =? [13:05] Saviq, there is no = [13:05] Saviq, >= == <= [13:06] interesting... [13:06] unity8 (= ${binary:Version}), [13:06] we have that all over the place... [13:07] seb128, shall I fix those ↑? [13:07] Saviq, wait [13:07] Saviq: correct, some precaching would be interesting [13:07] Saviq: i was thinking stringify [13:07] seb128, http://bazaar.launchpad.net/~unity-team/unity8/trunk/view/head:/debian/control [13:07] Saviq, sorry, I'm not awake today, no that's fine [13:08] seb128, kk thanks :) [13:14] pstolowski: hi! [13:15] sil2100, hey [13:16] pstolowski: ogra pointed out something in your last unity-scopes-api release - could you mention in the commit message (which is then used for the changelog) if you add any dependencies? ;) [13:16] When I think about it now, it's really useful when that's being done [13:18] sil2100, you mean the introduction of libunity-scopes-cli package? [13:19] pstolowski: no, there were some new dependencies added, like apparmour etc. from what I remember seeing [13:19] sil2100, ah, that wasn't me [13:19] sil2100, but I can pass it to the team [13:19] pstolowski: oh :) Thanks, sorry to bother then! [13:20] sil2100, no worries [13:25] Saviq, that branch is fine [13:25] Saviq, running on others [13:32] Cimi, thanks [13:39] damnit my unity8 has stopped starting and i don't see what has changed :S [13:43] * tsdgeos is seriously confused by our dependencies [13:43] unity-scope-click needs indicator-bluetooth [13:44] * tsdgeos reboots to see if this thing autofixes itself, not even the binary is starting === pete-woods1 is now known as pete-woods-lunch [14:04] tsdgeos, better? [14:05] Saviq: yes [14:05] actually it was my code ^_^ [14:05] seems the engine doesn't like a qml singleton that is malformed [14:05] and crashes :D [14:06] :) [14:07] davidcalle: ping [14:08] cwayne1, pong [14:10] davidcalle: how does the photo scope on desktop work? does it still use the shotwell db? [14:12] cwayne1, yes. On that topic, I think there is a bug opened for mediascanner to add pictures indexing, you might want bump that :p [14:12] kgunn: ping [14:12] davidcalle: ooh, do you happen to have a link? [14:13] cwayne1, https://bugs.launchpad.net/mediascanner2/+bug/1256334 [14:16] Saviq, I have a feeling the listview delegates are clipped [14:16] Cimi, feeling? :) [14:16] tsdgeos, are listview delegate clipped? [14:17] Saviq, more than a feeling [14:17] Cimi: not by default [14:17] Cimi, they're not clipped unless explicit [14:17] weird [14:18] davidcalle: ah, that'd be nice huh, so i'm trying to get a local photo scope on phone going, thinking i might maybe just use the gallery-app's index as that's similar to what the desktop lens does then... thoughts? [14:20] my feeling for something clipping my delegates https://www.youtube.com/watch?v=SSR6ZzjDZ94 [14:20] cwayne1, I don't know how the gallery app indexes metadata and stuff, but yes, that sounds like a plan until mediascanner becomes cooler :) [14:21] davidcalle: i like it, thanks :) [14:22] cwayne1, or you can port Shotwell. :p [14:22] hah! [14:23] might be just dropshadow being very bad [14:24] Saviq: ok, card cache pushed [14:24] card cache -> card component cache [14:24] tsdgeos, oh cool, was just getting onto testing this stuff [14:24] Saviq: now it feels to me that the dynamic specialized cards is as fast as specialized cards is [14:24] but having a non biased opinion is important [14:25] tsdgeos, are we garbage collecting the cache or something? [14:25] nope [14:25] but tbh it's like 4 or 5 different components at the moment [14:25] so not sure it makes sense [14:25] tsdgeos, not if you start opening remote scopes [14:25] from the scopes scope (still might not make sense) [14:26] or start searching, when different components come into play [14:26] Saviq: right, not sure how much memory is still used by the component itself [14:26] after all it shouldn't be much [14:27] but yes we may want to remove stuff from the cache at some point [14:27] tsdgeos, should be easy to check? like a Repeater alone, and a Repeater with model: 1000 of some small Component with UbuntuShape + Label { } [14:28] but yeah, doubt it's gonna be bigger than a few kB [14:31] ok not that easy === jono is now known as Guest72975 === dandrader is now known as dandrader|afk === alan_g is now known as alan_g|tea [14:38] tsdgeos, ok, I really doubt we have to worry about the cache ;) [14:38] :) [14:38] tsdgeos, unless we'd leak it or something [14:44] Saviq, mirco suggests the new ubuntushape will have dynamic shadow, I'd postpone then [14:45] Saviq, instead of using pngs and such [14:45] Cimi, eh... [14:45] new ubuntushape, good 'un [14:45] Saviq, I know it was scheduled for dec 2013 [14:46] Saviq, no idea why is not out yet [14:46] you mean 2012? [14:46] hah [14:46] I'll do png then === alan_g|tea is now known as alan_g [14:48] tsdgeos, one weird artifact is that it looks like we're slower at actually loading the images? [14:49] than in trunk, that is [14:50] or maybe we're just blocking now, and not blocking in dynamics... [14:51] Cimi, had to resubmit based on your branch: https://code.launchpad.net/~saviq/unity8/split-common/+merge/217622 [14:51] Saviq: yes, it does indeed seem that making the stuff async seems like it takes mroe to load for some reason [14:51] Saviq: that happens with the manually specialized cards too, right? [14:51] Cimi, there are small changes to packaging, that will be reviewed by the landing team [14:52] tsdgeos, yeah [14:52] tsdgeos, caching still enabled is it? [14:52] yes [14:52] should [14:55] tsdgeos, it feels like the only way to make it real nice would be to actually load everything in an expanded category... [14:55] Saviq: not sure i get what you mean [14:56] tsdgeos, I mean that if we wanted it to be buttery smooth, we'd have to load all items into memory when expanding [14:56] ah [14:56] yes [14:56] tsdgeos, basically not destroy delegates... [14:57] or limit the number of stuff [14:57] unity7 here we go [14:57] what's the point of 200 items in there, is people really going to dig on that many? [14:58] tsdgeos, I agree, just *USE THE SEARCH, LUKE* [15:00] Saviq: can't seem to run "make test[testName]" anymore on unity8. anything we need to do to make it work? [15:00] Saviq: but still we want this since it makes delegate creation faster anyway [15:00] tsdgeos, yeah [15:01] tsdgeos, but yeah, I can't see no difference between dynamic and "static" components === dandrader|afk is now known as dandrader [15:01] Saviq: plugin paths dont seem to be working anymore. [15:01] we're still blocking on moving between scopes [15:01] dednick, example? [15:01] make testDash [15:01] Saviq: ^ [15:02] dednick, I think we missed one commit from Albert [15:02] dednick, https://code.launchpad.net/~aacid/unity8/card_optimizations/+merge/213660 [15:02] dednick, it's landing today [15:02] Saviq: so you don't appreciate a speed difference between trunk and the specialized cards? [15:02] tsdgeos, huh? [15:03] tsdgeos, "static" as in specialized [15:03] ah [15:03] ok [15:03] got me confused there :D [15:03] :D [15:03] Saviq: so you mean "dynamic looking good" [15:03] tsdgeos, definitely [15:03] ok, need to polish the code now [15:03] and make it look like the other ones [15:03] since there's some definite regressions [15:03] Saviq: http://pastebin.ubuntu.com/7359929/ [15:03] indeed [15:04] Saviq: doesnt seem to work with the patch [15:04] unless it didnt rebuild the cmake files [15:04] rebuilding now. [15:08] Saviq: uh, no. that didnt work either [15:08] dednick, hmm [15:09] Saviq: make qmltests doesnt work for me either [15:10] dednick, try with the branch from http://people.canonical.com/~didrocks/citrain/silos/landing-007/unity8/ ? [15:10] worked fine here [15:12] mhr3: SearchReplyProxyFwd.h belongs to unity-scopes-api? (figuring which source package I want) [15:12] karni, yep [15:13] karni, you should just #include [15:13] mhr3: I want to know what exactly 'auto' is in auto cat = reply->register_category("cat1", "Category 1", ""); [15:13] be it Category.h or something else [15:14] karni, http://developer.ubuntu.com/api/devel/ubuntu-14.04/cplusplus/unity-scopes/classunity_1_1scopes_1_1_search_reply.html#aaa061806a96f50ff66abc6184135ea66 [15:14] mhr3: yay! :) === pete-woods-lunch is now known as pete-woods [15:18] tsdgeos, so... is it possible we were doing better with old scopes? or was it just that we didn't pay attention enough? (/me will flash some old image soon to check) [15:18] Saviq: what do you mean "better"? [15:18] as in faster? [15:19] tsdgeos, yeah [15:19] i'd say it was similar to what we have now [15:19] but old memories are treacherous [15:19] tsdgeos, yeah, I'll try with some old image in a bit [15:27] Saviq: the silo one works [15:28] dednick, so yeah, will be there in before tomorrow [15:29] dpm, your request to do something about the paths is not exactly easy [15:29] although i guess i could just run on the generated pot :) [15:29] run sed* [15:29] Saviq: ah. i'm being an idiot. was doing bzr merge http://.... rather than lp: :) [15:30] dednick, right, that might not have worked :D [15:30] heh. [15:30] hi mhr3. As I said, it was just a nitpick, not a big deal if it cannot be fixed [15:30] i just assumed it did [15:31] dpm, anything can be fixed... it's just tricky :) [15:35] Saviq, http://bazaar.launchpad.net/~saviq/unity8/split-common/revision/835 [15:35] Cimi, reduces any further diffs [15:35] Cimi, when you add items to the same list [15:35] Cimi, 'cause you don't need to add the , [15:35] sure [15:36] haven't thought of that [15:36] ok [15:37] Saviq, approved [15:37] Can't debug scopes in Ubuntu SDK. Known issue? &"warning: GDB: Failed to set controlling terminal: Inappropriate ioctl for device\n" [15:37] Cimi, kk [15:37] #ubuntu-sdk basically empty, I suppose this is the place to ask sdk related questions as well. === ChrisTownsend1 is now known as ChrisTownsend [15:51] elopio: Is there documentation/guidelines for writing tests for scopes? [16:01] karni: autopilot tests for the scopes? [16:02] elopio: actually I'd be more interested in unit tests. any examples out there you know of? [16:05] karni: on the unity8 branch you will find tests/qmltests/Dash/tst_GenericScopeView.qml [16:06] but mzanetti or tsdgeos are the ones that can help you with that. [16:07] elopio: yeah, I meant testing a c++ Scope, not the QML part of Dash toolkit. Anyways, will ask around. Thanks :) [16:07] karni: ok. alecu and his team has tests for the click scope. [16:08] elopio: thanks! [16:23] karni: our coverage is far from perfect, but we are aiming to improve it [16:24] karni: we are using gmock and gtest for testing our scope; it's not the best for testing qt stuff, but we are slowly getting rid of qtisms in the scope code [16:25] karni: what scope will you be working on, and what network library will you use? [16:25] alecu: confidential :D I'll talk to yoi when I'm back. [16:25] uhh [16:26] my canonical nick suggests I'm on lunch, not here ATM [16:26] alecu: but already thanks for the hints, I'll get back to you! [16:29] you always lunch late :-) === alan_g is now known as alan_g|EOD [17:06] mterry, I unassigned you from bug #1267623 since your branch doesn't actually take you to apps (it does unlock, but doesn't go to apps) [17:06] mterry, and that's all gonna change anyway when we move dash to be an app [17:06] Saviq, k === dandrader is now known as dandrader|lunch [18:34] 20:28 < asac> with bregma being on vacation who would know something about how the unity8 session with mir works? [18:34] 20:28 < asac> on deslktop? [18:34] 20:28 < asac> and lightdm etc. [18:34] Saviq: ? [18:38] asac, I would need to understand it more, is there any issue in particular? [18:44] Saviq: ok i connected stgraber to you [18:44] will be back in 1h === dandrader|lunch is now known as dandrader [19:03] Saviq: is shell rotation blocked by something still, or is it waiting the developer time to implement it? [19:03] mhall119, https://blueprints.launchpad.net/ubuntu/+spec/client-1410-unity-ui-rotation [19:50] thanks Saviq [19:56] Saviq: mind if I add thoughts to the whiteboard on that BP? [20:00] dednick, did you see I replied on that indicator-sound MP? [20:05] Saviq, all tracks are automatically buffered on playback of one item in the audio preview [20:26] mhall119, not at all [20:26] Cimi, huh? there's only one Audio element [20:27] Saviq, when I tapped play I saw progressbar buffering on all of them [20:27] probably because there's indeed only one [20:27] yeah [20:27] Cimi, yes, because you probably bound directly to the single Audio item, which was buffering the one it was playing [20:27] sorry [20:27] you need to check whether it's the one that's actually playing [20:28]