=== duflu_ is now known as duflu === 21WACCEWU is now known as Wug === Wug is now known as Wug[bonk] === Wug[bonk] is now known as 21WACCEWU [08:43] tsdgeos, ping? [08:44] mhr3_: hi [08:44] tsdgeos, hey, last week saviq had an issue where the scope results were not showing up cause there was something broken in the renderer [08:44] tsdgeos, and we're seeing the same issue in new-scopes branch [08:45] tsdgeos, wonder if you know what the issue was [08:45] no :/ [08:45] mhr3_: is it white text vs white background? [08:45] or is it something more "broken"? [08:45] tsdgeos, no, cause there's no icons either [08:46] no idea then [08:46] tsdgeos, only the category header shows up and nothing else [08:46] but if you give me a way to reproduce it i can have a look [08:46] tsdgeos, just branch new-scopes and run the unity-scope-tool [08:46] perhaps run `restart scope-registry` in between [08:47] tsdgeos, oh plus the onlinemusic scope from the demo ppa [08:49] ok [08:51] mhr3_: "unity-scope-onlinemusic is already the newest version." ? [08:51] 0.1+14.04.20131029.1-0ubuntu1 ¿? [08:51] ah [08:51] silly me [08:51] missing an update [08:51] tsdgeos, you need this one https://launchpad.net/~unity-team/+archive/demo-stuff/+packages [08:51] yes yes [08:51] i always forget to update after add-apt-repository [08:57] tsdgeos, if you don't even see the category header then something else is wrong :) [08:57] so i see "Popular online" [08:57] and that's all [08:57] should i see more? [08:57] cool, yea that's the problem [08:57] you should also see the results [08:59] ok fix is easy [08:59] don't know if it's the correct fix yet though [08:59] http://paste.ubuntu.com/6908000/ [08:59] mhr3_: ↑↑↑ if it's blocking you [08:59] tsdgeos, yea, we mostly need to fix new-scopes [09:00] tsdgeos, can you explain why that messes stuff up? [09:00] because you are assigning to something that is readonly [09:00] and qml says [09:00] no [09:00] won't do that [09:00] and then all breaks [09:01] need to find out if it should not be readonly [09:01] or the assignment isn't neede [09:01] d [09:01] tsdgeos, so why is it there in the first place? isn't trunk's cardtool broken too then? [09:01] sure [09:01] i mean [09:01] no [09:01] because trunk isn't merged since ten years ago [09:02] because we have unreproducible autopilot errors in a phone we are not supposed to support anymore blocking stuff [09:02] yeah \o/ [09:02] mhr3_: but yes it is a mistake in some branch that we want to eventually merge to trunk [09:02] i'll hunt it down [09:03] tsdgeos, thx [09:05] karni: you there? [09:07] mhr3_: i'm going to propose a branch for merging to new-scopes, can you review it? Do you know if we have automerging in there or we manually merge stuff in it? [09:07] tsdgeos, manual merging afaik [09:14] mhr3_: https://code.launchpad.net/~aacid/unity8/no_readonly_headerheight/+merge/205538 [09:14] if you approve it i'll manually merge it [09:15] tsdgeos, acked [09:18] ok, merged [09:21] tsdgeos, thx [09:21] tsdgeos, do you know which category layouts are properly supported these days? [09:21] mhr3_: grid, carousel [09:21] i have coded all the rest, but they are not wired [09:21] tsdgeos, i'm looking at the designs and wondering if we can even do them :) [09:22] i.e. lots of qml glue is missing [09:22] mhr3_: Saviq told me about the addSpecialCategory thing you're working on, is it done, what's your timeframe if not [09:22] tsdgeos, can we do all of this for example? https://docs.google.com/a/canonical.com/file/d/0B-a_7E3tDxOgZDJvd3l4U2tKYnM/edit [09:22] tsdgeos, yea, there's mp for it [09:23] tsdgeos, feel free to review :) [09:23] we can't do the seemore/less [09:23] i mean we can [09:23] but it's the old expansion pattern [09:23] not that one [09:23] tsdgeos, https://code.launchpad.net/~mhr3/unity-scopes-shell/special-categories/+merge/205427 [09:23] besides that, it's just grid and carousel no? [09:23] tsdgeos, i think cimi is working on the more/less [09:23] tsdgeos, the carousel is overlaid though [09:23] just for the text summary [09:24] we do support overlay [09:24] ah, good [09:28] mhr3_: make tryCard and select the overlaid one (the last) [09:32] tsdgeos, missing title when using { [09:32] "template": { [09:32] "overlay": true, [09:32] "category-layout": "carousel" [09:32] }, [09:32] "components": { [09:32] "title": "title", [09:32] "art": "art", [09:32] "subtitle": "subtitle" [09:32] } [09:32] } [09:32] (and subtitle) [09:34] trueth [09:34] and if you leave the mascot there it looks out of position [09:35] but at least it's there :) [09:35] so i guess someone needs to look at it and that someone is me? [09:37] tsdgeos, well, i'm not your team lead :) but would be nice if you did [09:38] there is no team lead :D [09:38] someone hinted i was acting one this week [09:38] don't believe it [09:40] j/k [09:40] i'll have a look [09:41] sil2100: do you have any idea what happened to the unity8 silo/landing/train thing? still blocked by autopilot failures? [09:41] and any idea where i can check the status of that? [09:54] mhr3_: it works [09:54] it's just the tryCard thing that is not as good as it should [09:55] and doesn't set all the stuff needed for carousel to work [09:55] mhr3_: but if you use tryCardTool [09:55] oh? [09:55] you'll see it work [09:55] great then [09:55] even a bit crappy since it's just wireframes either [09:55] i can fix tryCard to work if you want [09:56] tsdgeos, working tests are always good, primary thing is that it will work [10:01] mhr3_: i pushed the fix of the test to new-scopes [10:03] cheers [10:09] tsdgeos: hi! Not sure if anything changed, on Friday Didier wanted the maguro failures on unity8 resolved, right? [10:10] sil2100, ping, ping, libunity landing? [10:10] sil2100: i guess, why are we still caring about galaxy nexus if we don't support it and almost noone has one anymore to try to reproduce the problems? [10:16] tsdgeos: the rationale is that maguro is slow, and we'll start using the emulator soon which is even slower - so making stuff work on maguro will help things working on the emulator [10:17] mhr3_: ping! Give us a few moments [10:17] sil2100, i did, i asked last week :P [10:18] sil2100, it's a unity7 pkg, so it better land before ff [10:18] Aaaah, libunity! [10:18] sil2100: the rationale is fine, but if it's just blocking our builds and noone can reproduce/fix it, not really helping the flow [10:18] anyway i know i'm not going to win [10:19] so i go to my cave [10:19] Mirv: ^ could you give it a shot maybe? [10:19] Mirv: could you see if you could use the good-old cu2d to land the new libunity? (stand-alone) [10:19] Mirv: I'll add a landing for that in the old spreadsheet [10:20] +1, there probably won't be anymore changes there, so no need to use train for it [10:30] sil2100: sure in theory, but it's known that the tests don't pass at the moment for unity stack so I'm not sure how to validate it [10:32] Mirv: I guess mhr3_ can help out with finding the right set of tests to run/do manually in the meantime :) [10:34] ok. and on device I guess unity8 AP etc. [10:35] my device is as usual in Qt 5.2 mode, but I'll see if I can do that still today or tomorrow morning (provided mhr3_ will instruct how to test on desktop) [10:35] the changelog is essentially the dropping of gee [10:38] Sure, thanks! [10:40] Mirv, indeed, it's just dropping gee, so what could break is scopes themselves (especially aggregation) and then sound indicator (the play/pause buttons in there)... if that works it should be good [10:44] tsdgeos, when will we have branches merged? [10:44] Cimi: when some tests we can't reproduce failing don't fail [10:45] mhr3_: ok, but no problems known? should it work fine with the unity itself from october? (since nux+compiz+unity haven't been updated in trusty at all) [10:45] Cimi: meanwhile, can you review https://code.launchpad.net/~aacid/unity8/preview_header [10:45] Mirv, sure [10:46] tsdgeos, in your code seems all of them are optional [10:47] tsdgeos, shall we have title required? [10:47] it's not required afaics [10:47] at least the tryCard doesn't make it mandatory either [10:47] mhr3_: hmm, I updated libunity (the packages that come from the source package) from daily-build and now my windows don't have decorations for some reason [10:47] tsdgeos, but if you set to use it, you might have the title no? [10:48] tsdgeos, what's the point of using the header if you have all the json null? [10:48] Mirv, ?? compiz weirdness? [10:49] Cimi: makes no sense [10:49] Cimi: but that's the guy that created the json fault [10:49] Cimi: what do you suggest us to do? [10:49] mhr3_: hmm ok I guess compiz weirdness, not seeing the problem on guest account [10:50] and yes I was testing compiz+nux+unity for fun before that, probably related [10:50] tsdgeos, fix the json [10:53] Cimi: right, so if the solution is "fix the json", what change do you want me to do in the qml code of that branch? [10:53] mhr3_: ok thanks, I should be able to test the desktop side [10:53] tsdgeos, 38 - title: root.widgetData["title"] || "" [10:53] + title: root.widgetData["title"] [10:54] Cimi: but that makes no sense [10:54] why? [10:54] why are you forcing me to have a title? [10:54] tsdgeos, because if I use a header, I must provide a title [10:54] where does it say that? [10:55] tsdgeos, it's like having a button without a label and an icon [10:55] it makes no sense [10:55] Mirv, and if desktop didn't break, unity8 won't either :) [10:55] Cimi: but i can have a header with an emblem and no title [10:55] tsdgeos, title is always important [10:56] it's the title of the app, no? [10:56] sure [10:56] it's the title of something [10:56] this is a preview [10:56] not a preview of an app [10:56] title is important [10:56] tsdgeos, well you got my point [10:56] tsdgeos, I'd kinda force developers to keep things right [10:56] but since we provide stuff for randomguy to do a preview, i don't see why we should force title on him [10:57] because otherwise he uses no title making it not consistent? [10:57] Cimi: honestly your change won't force anything [10:57] only will give us a console warning [10:57] and the developer will still don't care [10:57] tsdgeos, following your thought we should have those checks for empty strings all around unity [10:58] and we mostly do [10:58] and it's not a check for empty string [10:58] is a check for undefined data in a map [10:59] ./plugins/Unity/Indicators/qml/MenuItemFactory.qml:229: text: menuData && menuData.label || "" [10:59] ./qml/Dash/Card.qml:71: title: cardData && cardData["title"] || "" [11:02] mhr3_: about the addSpecialCategory thing, can you explain why are we doing that? [11:02] tsdgeos, because of the running apps [11:02] mhr3_: but it works as we have it now, no? [11:02] so, why are we doing it? [11:03] tsdgeos, i'm not sure i understand, the new plugin didn't allow shell to "override" a category [11:03] the old one did yes, so we're mostly bringing it back [11:04] in a slightly different shape [11:04] mhr3_: ok, so overrideResults is gone [11:04] right [11:05] hmmm [11:05] so we need addSpecialCategory for running apps and for the dashhome then [11:06] tsdgeos, there's no dashhome with new scope [11:06] s [11:06] ah right [11:06] forgot about that [11:07] tsdgeos, so far we don't have apps scope, so try to test it with music or something, but fwiw apps scope id will most probably be "clickscope" [11:12] tsdgeos, this is not always [11:12] tsdgeos, I'm scared of developers adding empty headers [11:12] Cimi: ok [11:13] Cimi: do you understand that your change won't fix anything? [11:13] I do [11:13] ok [11:13] because just gives a warning [11:13] that noone will see anyway [11:13] ok [11:13] if you want to force CardHeader to have a title, discuss with design and propose a code change that actually breaks the thing if it's not there [11:13] tsdgeos, but warning is better than nothing, no? [11:14] but later please [11:14] we need to get all this done by friday [11:23] if i do `system-image-cli -b 0` i'm getting the same thing as if i did a full reflash, right? [11:25] ie i can do it on dirty device? (one with ppa pkgs installed) [11:27] also, is it faster than phablet-flash? [11:28] tsdgeos: anyway, I know it's all a very risky topic, but did anyone of you have any success with those maguro unity8 failures in overall? I see one, test_hide_hud_click, seems to happen on almost every maguro test in smoketesting [11:29] tsdgeos: you think it would be easy to identify and/or fix? [11:29] sil2100: i don't have a maguro [11:29] i've been running the test for 1 hour in a loop [11:29] and nothing failed [11:29] in my nexus4 [11:30] if we are fine being blocked, fine [11:30] sil2100: is a test app crashing occasionally on maguro? [11:30] Do you know if unity8 tests are going fine on the emulator now? [11:30] greyback: yes, but this failure seems to be unrelared to the qmlscene-crash [11:31] sil2100: i have not touched the emulator yet, and honestly don't feel like doing [11:31] sil2100: we have one target which is nexus4, and stuff works there [11:31] if people feels we need to get stalled fine [11:32] tsdgeos: I'm thinking that if the problem here is the slowness of maguro, maybe if the timeout could be bumped for the failing assert in that test? [11:32] but i'm not going to take any blame for stuff not working on stuff we don't support and stuff i can't reproduce [11:32] sil2100: ok [11:32] A hacky way around it, but hmmm [11:32] sil2100: hey, maybe, if i had a maguro, i would get a fix, but i don't [11:36] sil2100: i doubt very much the problem is slowness [11:36] i've run the test on nexus4 [11:36] while running "stress --cpu 8 --io 4 --vm 2 --vm-bytes 128M " also [11:36] and it's just fine [11:36] tsdgeos, if you want i could give you ssh tunnel to a maguro :) [11:37] mhr3_: i guess that could help [11:37] \o/ [11:38] tsdgeos, you sure? you won't see much :) [11:38] mhr3_: i don't need to see anything [11:38] mhr3_: just need to run the test [11:38] and make it succeed [11:38] well [11:38] see if it fails first [11:38] and then make it succeed [11:38] seeing would help [11:38] but it's not critical [11:39] i can just add billions of prints [11:39] mhr3_: can you phablet flash it with latest stuff? [11:39] ok, let me try to setup everything [11:39] first i need to charge it a bit though [11:39] mhr3_: thanks [11:40] mhr3_: thanks! === mhr3__ is now known as mhr3 [11:49] tsdgeos, :/ someone changed our network setup, can't open ports, so.. not good [11:49] :/ [11:49] tsdgeos, i can do it at the office tomorrow, if that's good enough for you [11:50] unity: anyone has a galaxy nexus ? [11:50] @unity: anyone has a galaxy nexus ? [11:50] i mean [11:50] tsdgeos, I have [11:50] * mzanetti too [11:51] Cimi: mzanetti: you have two options, fix the autopilot failures we have or give me ssh so i can try to [11:51] Cimi, mzanetti: can anyone of you flash it to the latest version and provide ssh access for tsdgeos ? :) [11:51] Cimi: mzanetti: you choose :D [11:51] (or that) [11:51] ;D [11:51] It seems to be something easily reproducible on maguro from what I see [11:52] ok... need to charge and flash tho [11:52] tsdgeos, I can do both [11:52] tsdgeos, I have to run to the doctor though now :) [11:53] tsdgeos, maybe i can setup ssh before [11:53] Cimi: that'd be great [11:53] oh yeah [11:53] tsdgeos, so what want me to do? [11:53] Cimi: setup ssh so i can login from here, if you have time to phablet-flash it first the better [11:54] if not i guess i can manage from here [11:54] tsdgeos, you want a linux pc or ssh to the phone? [11:54] Cimi: ssh to the phone [11:54] or ssh to a pc where i can adb to the phone [11:54] you choose [11:54] but i guess the first is easier [11:55] tsdgeos, ok if I set static ip [11:55] tsdgeos, battery is dead [11:55] tsdgeos, hah [11:55] tsdgeos, don't forget that flashing would remove the ssh server [11:55] tsdgeos, charging then flashing [11:55] tsdgeos, 30 mins+ [11:55] Cimi: ok [11:55] Bad luck! [11:55] :\ [11:56] yeah... mine won't turn on even with power attached [11:56] will take a little here too [11:56] Cimi: if you gotta run to the doctor I can set it up [11:57] mzanetti, I can do it [11:57] ok [11:57] mzanetti, I don't have appointment, just grab medicine [11:57] ah ok [11:58] mhr3: about the progress preview type [11:58] mhr3: i don't understand this [11:58] "id": null, // mandatory here! action $id triggered on completion/error(/cancel) with argument for status [11:59] tsdgeos, there should be just a regular widget_id [11:59] tsdgeos, when triggering actions, we're sending widget_id + action_id [12:00] is this something i care at all from the UI POV? [12:00] tsdgeos, for progress it should be the widget_id as specified by the scope author + action_id specified by us [12:00] i mean for progress i just show stuff, don't send anything back, no? [12:01] tsdgeos, the action should be auto-triggered when stuff happens [12:01] http://developer.ubuntu.com/api/devel/ubuntu-14.04/cplusplus/libunity-scopes/classunity_1_1scopes_1_1_preview_widget.html#progress [12:02] mhr3: what do you mean with auto-triggered? [12:02] tsdgeos, download finishes - shell triggers "finished" action on the progress widget [12:04] can i say that doesn't make any sense? [12:05] no [12:05] i knew it :) [12:06] tsdgeos, the entire progress widget really needs to tied to the progress provider service [12:06] mhr3: why is that on the frontend and not the backend? [12:06] imho the frontend all it should care is "show 50", "show 75", "done" [12:07] tsdgeos, it is, but shell is the bridge [12:07] ok, i don't think it makes sense [12:07] tsdgeos, scope might not be running to receive those notifications [12:07] but we're of course not going to change it [12:07] tsdgeos, all it wants is result of the operation [12:07] so let's pretend i think it makes sense [12:08] tsdgeos, and since shell does know both the progress and the result, it makes sense to notify the scope when the result is available [12:08] tsdgeos, plus the completion will change the preview [12:14] mhr3: the contract we have for previews right now is http://paste.ubuntu.com/6908739/ does it mean that besides triggered your model is also going to expose finished and failed? [12:14] does it do that already? [12:15] tsdgeos, no, that's the progress provider's job [12:15] (the thing that gives you progress == 60%) [12:16] what is its job? === MacSlow is now known as MacSlow|lunch [12:16] to notify you of the progress and whether the operation was successful/failed [12:17] you're killing me :D [12:17] you told me the shell has to notify the scope [12:17] wait, let me try to find something [12:17] yes? [12:18] yes, as in you're supposed to call .triggered(progress_widget_id, "finished", meta) [12:18] but you do that in response to progress indicator telling you that is what happened [12:18] *provider [12:19] ok [12:19] so that is what i was asking [12:19] if we use triggered [12:19] or something else [12:19] but the docu you pointed says [12:20] "Upon completion the scope will receive preview action activation with id "finished" or "failed" depending on the outcome of the operation." [12:20] will you take care of converting that .triggered(progress_widget_id, ) to that? [12:20] sure [12:20] that is what's happenning with any "triggered" method call [12:21] tsdgeos, this might help to see the entire flow https://docs.google.com/a/canonical.com/drawings/d/1AKbddm35Tw9IoxZ8T33HBO30qO8c35B_IwEzacmf3bI/edit [12:21] the 3rd part especially [12:21] mhr3: so can you tell me what triggered call exactly you want for success and failure? [12:22] mhr3: the unity-scope-scopes landing - will this have some impact on the current image? i.e. is this scope enabled by default now? [12:22] sil2100, no [12:22] .triggered(progress_widget_id, "finished", progress_widget_properties_map) [12:23] tsdgeos, ^ + same for "failed" [12:23] mh. my nexus 10 keeps showing device out of space when installing packages... [12:23] mhr3: but s/finished/failed [12:23] right [12:23] mhr3: what's in progress_widget_properties_map ? the same thing that you gave us? [12:24] tsdgeos, yep [12:24] mhr3: we don't do that on buttons, saviq told me not to, do you expect it too for buttons? [12:24] yea, it should be there [12:24] why? [12:25] tsdgeos, but needs to per-action [12:25] to be* [12:25] i mean you gave it to us, you sure know its contents, no? [12:25] tsdgeos, cause scope can push there extra data that you don't care about, but it does [12:26] ok [12:26] so that needs changing in the regular actions too [12:26] mhr3: did this package get pre-NEWed by an archive admin already? [12:27] tsdgeos, ie if actions = [{"id": "foo", "label": "Foo", "session-id": "0x1234ef"}, {...}], i expect triggered(widget_id, "foo", {"session-id": "0x1234ef", ...}) [12:27] sil2100, no idea, robru was dealing with that [12:27] mhr3: you'll get back id and label too [12:27] tsdgeos, yea, fine [12:28] ok, i added it to my todo [12:28] thing is growing too much :D [12:29] tsdgeos, btw i saw designs where there is an action combo + regular button on the right side, do we support that? [12:30] kind of [12:30] we don't support "combos" [12:30] tsdgeos, https://docs.google.com/a/canonical.com/file/d/0B-a_7E3tDxOgUUY0MjFTQThBZFE/edit [12:30] we support a button that when you click on it [12:30] it grows [12:30] so it's an "ugly combo" [12:30] but yes [12:30] if you give us 3 actions [12:30] you'll end up with the first action in the right and then on the left a More/Less buttons that shows the other 2 actions [12:31] ah, ok [12:31] so the right one is "primary" action? [12:31] so yes but if looks a bit ugly because there's no proper combo in the SDK yet and we didn't feel like coding one :D [12:31] yep [12:31] ok [12:31] also if you give us two actions [12:32]