[03:53] hello :P [03:54] Anyone here tested ubuntu-sdk in kubuntu? [03:54] i like more plasma 5 but i wanna make a great scope [07:09] Hi all [07:09] alecu ping [08:02] alecu ping === davidcalle_ is now known as davidcalle === _salem is now known as salem_ [11:15] mrqtros: pong [11:33] alecu hi there! Seems that my scope causes dash crash (sounds funny) [11:34] alecu I want to discuss it with someone from dash's team dunno anyone from it [11:36] mrqtros: well done! 😃 [11:38] mrqtros: saw your email, it's about the Academy scope, right? [11:38] alecu yes [11:38] popey thanks :D [11:39] mrqtros: I think pawel is the right person to handle that [11:39] mrqtros: saw he responded to your email [11:39] mrqtros: let me ping him [11:40] alecu thanks!) [11:40] alecu I am worrying about this because of participation in contest (Scope Showdown) [11:44] mrqtros: here's pstolowski [11:44] pstolowski: thanks for your help [11:45] pstolowski hello :) [11:45] mrqtros, hi! [11:46] pstolowski first of all - thanks for your reply in email [11:47] pstolowski does anyone else have such issue? When a fix will be released? [11:48] pstolowski is there any workaround? I am worrying because of participation in scope showdown [11:48] i have just received a couple of reports this morning and still investigating [11:49] mrqtros, ^ [11:50] mrqtros, i've a test cases with a dozen of stress tests and cannot make it crash. still unsure about root cause. [11:50] Can you simply run my scope? [11:51] mrqtros, good idea. do you have a click package for the phone by any chance? [11:52] pstolowski yes, w8 a little [11:52] pstolowski https://yadi.sk/d/p4bmL_Bgos74C [11:53] mrqtros, thanks, trying [11:53] pstolowski scenario is - open it, w8 for full download of 4 sources, then find something (for example "Haskell", 2 results), then hit back (or clear search query) [11:54] In most cases this causes dash crash on phone [11:56] mrqtros, right. reproduced. [11:56] pstolowski do you have logs or something like that? [11:57] pstolowski is it bug in scope or in Dash itself? [12:04] mrqtros, question - do you call register_category for all the categories always, even if they will have no results (such as with Haskell example)? [12:04] pstolowski I think yes [12:13] pstolowski I shouldn't? [12:13] mrqtros, no, that's absolutely fine, just trying to recreate the scenario [12:14] pstolowski source code is open, feel free to use it: https://github.com/QtRoS/courseexplorer/ === chihchun is now known as chihchun_afk [13:04] pstolowski any news? :) [13:11] mrqtros, no, sorry, still investigating [13:12] mrqtros, your scope is almost certainly innocent [13:17] pstolowski nice to hear that) === bpierre_ is now known as bpierre === benoitc_ is now known as benoitc [15:03] pstolowski how is it going? Can I help? [15:06] mrqtros, had a lunch break, so not much new. my stress tests https://code.launchpad.net/~stolowski/unity-scopes-shell/model-update-crashfix/+merge/286069 don't trigger the problem [15:07] mrqtros, would it be much work for you to modify your scope to push all results into a single category all the time, and see if this still reproduces it? [15:09] pstolowski not so much, but I can build click in evening only [15:10] mrqtros, i see. i don't need this click, just let me know if single category makes a difference [15:10] Ah, ok, can test it in evening, ~ 19:00 utc, is it ok? [15:11] pstolowski what kind of error do you get? [15:11] pstolowski segfault? [15:14] mrqtros, yes [15:15] Seems that bug suggested by you (https://code.launchpad.net/~chihchun/unity-scopes-shell/lp1535377/+merge/285123) is really affects [15:15] pstolowski, ^ [15:15] mrqtros, yes, i see same stacktrace [15:15] pstolowski so what are you trying to find out? [15:15] mrqtros, but i don't understand the root cause and don't agree with proposed fix [15:16] I also do not like this fix (to be honest) [15:16] mrqtros, i'm trying to find a way for reproducing it programatically in my tests [15:16] pstolowski no much clear [15:19] pstolowski these methods beginInsertRows, beginMoveRows and so on are quite hard to understand [15:20] pstolowski let me look myself at model === javiercrowsoft1 is now known as javiercrowsoft [15:25] mrqtros, ok. btw, your category renderers have 'subtitle' mappings, but there is no subtitle attribute in results (shouldn't do any harm, except for some warnings from the shell i guess) [15:27] pstolowski yes, got it, will fix [15:35] ahayzen: yo [15:35] faenil, hey what's up [15:35] ahayzen: any news about music app? [15:36] faenil, define news :-) .. i've nearly got the convergence stuff working [15:36] at Uni at the moment, should hopefully be able to get it in a state tonight to push [15:36] ahayzen: great to hear, keep me updated :) [15:36] faenil, will do :-) [15:37] faenil, managed to find a way of making the tabs not go fullwidth :-) [15:37] cool :) [15:39] pstolowski seems that you should add test where some item is changing his position [15:39] pstolowski its* [15:39] pstolowski just shuffle input [15:42] mrqtros, i've plenty of such tests [15:44] pstolowski what is the point of moving results? Why not just add new instead of old? [15:45] mrqtros, should be slightly more effective from qml point of view; less flickering [15:50] pstolowski what about efficiency and CPU cycles? I think that such optimization has a lot of bad effects [15:51] pstolowski I never seen such things before, but I am working with Qt 6-7 years [15:52] What is the idea of this line of code: m_results.move(oldPos, row); ? [15:53] What if m_results do not have such amount of items? [15:56] Sry, last question is wrong [15:57] pstolowski got it! This line is harmful anyway: m_results.move(oldPos, row); [15:57] pstolowski what if 'row' is bigger than previous m_result.count() [15:58] mrqtros, that's the point, it shouldn't be bigger. if it is bigger, then the problem is somewhere else i think [15:59] pstolowski in my case it is bigger - empty search returns ~400 results, while search with query can return a little, for example 2 [15:59] mrqtros, of course. please read my comment to the MP [16:00] mrqtros, that's totally fine. the outer loop shouldn't let it go out of bounds as it's processing the rows [16:01] mrqtros, plus i've tests that start with 2000 results and go down to 10 results on next search etc. [16:01] pstolowski try from 10 to 2000 [16:01] pstolowski with matching items in different positions [16:02] mrqtros, i've 10, then 200 [16:02] pstolowski you can try my results as example, check indexes of "Haskell" course in result sets [16:05] pstolowski my input files are available in cache directory [16:05] pstolowski you can reproduce it easily [16:05] pstolowski I will be online in evening [16:05] pstolowski thanks a lot! === chihchun_afk is now known as chihchun === salem_ is now known as _salem === _salem is now known as salem_ === salem_ is now known as _salem [20:42] When I write a simple hello world app in the SDK, it deploys just fine on my device, but when I use, almost, the same cmake file for my own app, I get this error http://pastebin.ubuntu.com/15084110/ [20:57] faenil, how much do you know about the headers ? or is that Tim's area? [20:57] ahayzen: yep, timp's [20:57] ok :-) [20:57] :) [20:58] * faenil should really get some experience as "user" of the sdk [20:58] yes [20:58] * faenil is however working at 9PM atm... [20:59] timp, I've used PageHeader in my Pages which are inside Tabs .. which works nicely. However when i rotate my device and the tabs resize... i get the original header from the Tabs appear with like a transparent background [20:59] timp, is there a way for force this to be hidden? [21:11] * faenil is not in the SDK team, fwiw :D [21:13] faenil, you are a QML/Qt prototyper ? [21:13] ahayzen: yes, that is my role, although nowadays I help the SDK team develop the convergent components [21:13] faenil, awesome :-) [21:14] nowadays == last 8months :D [21:15] but I'm in the Design team, one of the few developers who act as the bridge between designers and the engineering teams, basically :) [21:15] (spoiler: we need more bridges :P) [21:17] faenil, hah :-) [21:17] :) [22:09] ahayzen: ohh, I haven't been testing the PageHeader combined with Tabs a lot, because for the Tabs we'll have a new pattern [22:09] ahayzen: what do you mean with the tabs resize? [22:09] timp, i'm workong on the convergence stuff [22:09] the idea is that whenever you have a Page with a PageHeader in it, it automatically hides the old header. Even in landscape. [22:10] timp, so what i've managed todo is make the tabs take up say 2/3 width, by giving the page a PageHeader [22:10] so maybe you ran into a UITK bug [22:10] timp, but when i rotate and then go back the header appears with no bg [22:10] then if i push anything to the stack and go back it is fine again [22:10] I don't understand it. Do you have a screenshot? [22:10] timp, yeah give me a minute [22:11] ok [22:11] timp, also i've been mapping the PageHeader stuff to things like page.head.backAction (to support legacy things) as i'm migrating things over to the new PageHeader [22:11] timp, where is the action for the tab drawer? should it be in .backAction ? i cannot see it [22:12] ahayzen: in PageHeader? For that, you can configure the whole PageHeader.leadingActionBar [22:13] timp, i've managed to get them working on the PageHeader, what i mean is, how can i hijack the action from the tabs ? [22:13] ahayzen: so set PageHeader.leadingActionBar.actions: [your, tab, actions] [22:13] (as in the navigation drawer thing) [22:15] timp, this is the bit of the 'old' header that can reappear after rotations https://drive.google.com/file/d/0B3XynHVKfrvMMWNMUmI0TFhxWVk/view?usp=sharing [22:15] ahayzen: maybe I'm a bit slow today because I was sick last week ;) can you explicitly say what you want to accomplish? [22:15] you want to trigger an action when the selected tab is changeD? [22:16] timp, i'm using the PageHeader thing, because i need the tabs not to be full width [22:18] timp, i have all the actions from the pages mapped correctly. Just have two issues. 1) I need to either make the navigation-menu action and overlay thing to switch tabs, or somehow extract the one from the tabs component ... 2) when i rotate i can get it to the state where the old header appears [22:18] ahayzen: the old header appearing after going to landscape and back to portrait looks like a bug to me. Can you report it? If there is a test program to reproduce it, that would be great. [22:18] timp, yup i can do that [22:18] thanks [22:19] timp, and the tab navigation thing (in the leading actions position on the old header) can i access that through the Tabs {} component? or should it be the Page.head.backAction ? [22:19] so the bug will be for 2). [22:19] yup [22:20] you cannot access the actions in the old header navigation overflow [22:20] but you can see it is there when the bug occurs :-) hehe [22:21] timp, where is the component that does that in the SDK ? I've found bits of it in the TabBarStyle, maybe i could just emulate it for now [22:22] or just make my own [22:22] I still don't understand the problem [22:22] ahayzen: it is in Themes/Ambiance/1.3/PageHeadStyle.qml [22:22] ahayzen: basically it is this code http://pastebin.ubuntu.com/15086428/ [22:22] i need the tabs navigation button to switch tabs, to be visible when using the PageHeader [22:23] ah cool! [22:23] ahayzen: that OverflowPanel is the same as is used in ActionBar for the overflow button [22:23] hmm [22:23] ahayzen: so if you give a list of actions to switch tabs to leadingActionBar.actions, that should give the same result [22:23] timp, sweet! i'll try that thanks :-) [22:25] ahayzen: just set the Actions to set the selectedTabIndex [22:25] if that is possible [22:25] i'll have a play about [22:25] timp, thanks for the pointers :-) [22:26] no problem :) [22:26] * timp off