| === shuduo|afk is now known as shuduo | ||
| === JamesTait is now known as Guest84209 | ||
| === Guest84209 is now known as JamesTait | ||
| === john-mca` is now known as john-mcaleely | ||
| kalikiana_ | nik90: ping | 11:05 |
|---|---|---|
| nik90 | kalikiana_: pong | 11:53 |
| === _salem is now known as salem_ | ||
| kalikiana_ | nik90: ah there you are | 12:18 |
| kalikiana_ | I just replied to the list | 12:18 |
| kalikiana_ | maybe have a look at the reply first | 12:18 |
| nik90 | kalikiana_: thnx. looking at the email now | 12:19 |
| nik90 | kalikiana_: I have seen and used http://developer.ubuntu.com/api/qml/sdk-1.0/U1DB/ before. Infact I did a hangout with stuart during which michael hall added it to the API docs. It was quite useful to list the functions of database, document etc. | 12:25 |
| nik90 | kalikiana_: so with regards to index, while specifying the expression is there a specific order we must use? | 12:26 |
| nik90 | also does it expect the expressions to be of the same size? | 12:27 |
| kalikiana_ | nik90: in general the query needs to list the same expressions defined in the index | 12:28 |
| kalikiana_ | if you want to search foobar, it needs to be in the index | 12:28 |
| nik90 | okay, so in essence an index provides the data, while queries uses them to filter and pass it on as a model. | 12:29 |
| kalikiana_ | yep | 12:31 |
| kalikiana_ | which as you've found is limited to wildcards basically right now. but there's plans for more filters ie. transformation functions in u1 lingo | 12:32 |
| nik90 | kalikiana_: I looked at the bug report you mentioned in the email and I think providing the capability to add user function to use as filters to provide some freedom to the developer | 12:33 |
| nik90 | kalikiana_: and regarding sorting of results, what I meant was in python, I saw calls like sorted(modeldata) where it returns the results in an alphabetical order, or in ascending/descending order. Unfortunately QML doesn't offer this natively. | 12:35 |
| kalikiana_ | ah right. in qml you can only do it "by hand" like copying the model and using move() with your own algorithm | 12:39 |
| kalikiana_ | hmm do you think this would make sense as a u1db feature? | 12:40 |
| kalikiana_ | nik90: maybe indirectly this could be solved through transformation functions | 12:41 |
| kalikiana_ | at the bottom of this mock api https://launchpadlibrarian.net/149329873/queryfuncs.qml | 12:42 |
| kalikiana_ | that's pretty much a way of sorting | 12:42 |
| nik90 | kalikiana_: yeah me and andrew thought that u1db would provide a efficient and more lean sorting function instead of us having to create one. | 12:44 |
| nik90 | kalikiana_: I have branch which shows of the sample app I have been working on at https://code.launchpad.net/~nik90/+junk/u1db-sample-app | 12:45 |
| nik90 | can you take a look and let me know if I am doing stuff correctly in the listview? | 12:46 |
| nik90 | I feel like I am complicating matters by using a repeater inside a listivew | 12:46 |
| timp | nik90: why would you do that instead of using a ListModel? | 12:48 |
| nik90 | timp: because strangely enough, when I specified contents.tracks, it threw a undefined output. | 12:50 |
| nik90 | timp: Also looking at my document structure, it seems it is a list inside a list. The tracks is a list object inside the playlist document | 12:50 |
| nik90 | kalikiana_: I also noticed a bug with example 5 in u1db-qt examples located at http://bazaar.launchpad.net/~uonedb-qt/u1db-qt/trunk/view/head:/examples/u1db-qt-example-5/u1db-qt-example-5.qml. | 12:51 |
| nik90 | I ran it without any modifcation and the output is rather bizarre | 12:52 |
| nik90 | I expect it to show 7 outputs in the listview, but it only shows 3. | 12:52 |
| nik90 | can you confirm that pls? | 12:52 |
| nik90 | timp: I am open to suggestions to simplifying that. I feel like I am over complicating it | 12:53 |
| timp | nik90: do you have a link (http?) of your code on launchpad? I can have a quick look | 12:55 |
| nik90 | timp: http://bazaar.launchpad.net/~nik90/+junk/u1db-sample-app/view/head:/u1db.qml | 12:55 |
| timp | nik90: thanks | 12:57 |
| timp | nik90: perhaps it is possible to create a ListModel and put the repeater inside that | 12:57 |
| timp | I'm not sure if that is actually better than what you have now | 12:57 |
| timp | hmm.. a Repeater in a ListModel may not work, but some functional way of generating the model would be useful here. | 13:00 |
| timp | wait a sec.. what does aQuery return? if it is the list of tracks (and not list of playlists) then you don't need a repeater | 13:02 |
| nik90 | timp: I tried using a Multivalue list Item from the SDK which accepts a list. However when I used that I got a output stating TypeError: Cannot read property 'length' of undefined | 13:02 |
| nik90 | http://paste.ubuntu.com/6703066/ | 13:03 |
| nik90 | timp: would you expect two queries, one for playlist list and the other for playlist track info? | 13:05 |
| nik90 | Each document stores the playlist name and its associated tracks. As an analogy, think of a task and its tags in a document. I wish to display both the task info like name and its associated tags in the listview | 13:07 |
| nik90 | Which is why I have one query which retrieves the playlist name and its tracks | 13:07 |
| timp | I'm not sure what your query returns, but if it has all the tracks (ordered by playlist), you can use that as a model. | 13:11 |
| timp | then the tracks are the items in the listview, and playlist names the sections | 13:11 |
| timp | (it may be more complicated than what I'm saying, depending on what the queries do exactly. I didn't check that in detail) | 13:12 |
| nik90 | timp: oh yeah I totally forget to use section headers. But yeah in essence that's what I want to achieve. Have section headers as playlist names, and results and track info | 13:17 |
| nik90 | let me see if I can implement that | 13:18 |
| kalikiana_ | nik90: diff to avoid the repeater http://paste.ubuntu.com/6703112/ for testing I used JSON.stringify which is very good for debugging - I think there's either a bug causing name to be missing or I'm overlooking a typo in my changes | 13:18 |
| kalikiana_ | but that would be what you intended functionally? | 13:18 |
| kalikiana_ | minus the missing name, that is | 13:18 |
| nik90 | kalikiana_: yup. | 13:19 |
| nik90 | My intention was to specify contents.something and be able to display that easily | 13:20 |
| nik90 | without having to resort to repeater in a listview | 13:20 |
| kalikiana_ | my guess is you encountered an 'undefined' for a similar reason as I stumbled on, somehow a missing value, which in javascript becomes a catch-all undefined | 13:20 |
| nik90 | kalikiana_: I don't thinkn there is a typo causing name to be missing. I have that issue all day yesterday. | 13:21 |
| nik90 | the moment I change that to just playlist.tracks in the index expression, things work out | 13:21 |
| kalikiana_ | you mean not using name? | 13:22 |
| kalikiana_ | I think something might be wrong in Query as it doesn't include name in the result/ contents | 13:23 |
| nik90 | I mean when the index expression is ["playlist.name", "playlist.tracks"], the name is displayed. But if I change that to ["playlist.name", "playlist.tracks.file", "playlist.tracks.title"], then name becomes undefined | 13:24 |
| kalikiana_ | interesting. if I do that name is there but nothing else | 13:25 |
| nik90 | exactly | 13:26 |
| kalikiana_ | so somehow they kick each other out. even if I change the order same reltion | 13:27 |
| kalikiana_ | *relation | 13:27 |
| nik90 | yup..but this only happens with a multiple level list. Since if you have other properties like playlist description in the same level as the playnist name, they are both available | 13:29 |
| timp | nik90: cool. Are you writing a full app, or is it intended to demo the u1 functionality only? | 13:34 |
| timp | the u1db bindings look very interesting, but I haven't had time yet to check it out. | 13:34 |
| nik90 | timp: only demo purposes..but my full intention is to use u1db in clock-app. Hence this is a preparation for it :) | 13:35 |
| timp | nik90: okay. always good to have some demo programs :) Personally I prefer that over long tutorials when learning something new. | 13:36 |
| timp | it teaches you 90% of what you need to know in 10% of the time :) | 13:37 |
| nik90 | timp: +1 but only if the demo program has comments. otherwise it becomes rather unclear what the code does. | 13:37 |
| nik90 | timp: in the u1db-qt examples, examply 5 and 6 have no comments making it guesswork to see what they are for. | 13:38 |
| nik90 | timp, kalikiana_: I tried using a section header as suggested, however it outputs that contents is not defined. -> http://paste.ubuntu.com/6703235/ | 13:42 |
| kalikiana_ | actually I've never used section stuff - nik90, on a guess try model.contents.name | 13:47 |
| kalikiana_ | but not sure if it is actually a delegate or how it behaves | 13:47 |
| nik90 | kalikiana_: it behaves exactly like a delegate. so passing contents.name should work like normal. | 13:48 |
| kalikiana_ | nik90: oh, actually could you be hitting the same bug as before with name missing? | 13:49 |
| nik90 | kalikiana_: I thought so. But after changing the index expression back to the orignal where contents.name should work but it doesnt with a section header. | 13:50 |
| nik90 | but that's fine. | 13:51 |
| nik90 | Should I report a bug about the name being undefined issue? | 13:51 |
| kalikiana_ | yes please file a bug for the name. if possible mention the weird relation with the tracks, that has to be a hint at the underlying bug | 13:52 |
| kalikiana_ | I'm trying the section stuff now | 13:52 |
| kalikiana_ | nik90: so… what exactly should the section header use? why is there text: as well as the section.property? | 13:57 |
| kalikiana_ | not sure if I'm using it right | 13:58 |
| nik90 | kalikiana_: section property defines the text that will be shown in the section header. In our case that would be the playlist name. | 13:59 |
| kalikiana_ | but it doesn't work if I put "foo" in there | 13:59 |
| kalikiana_ | by which I mean I don't see anything besides the list | 14:00 |
| nik90 | http://qt-project.org/doc/qt-5.0/qtquick/qml-qtquick2-listview.html#section.criteria-prop | 14:01 |
| nik90 | or better here is an example http://qt-project.org/doc/qt-5.0/qtquick/views-listview-sections-qml.html | 14:02 |
| === alex_abreu is now known as alex-abreu | ||
| jdstrand | so, I wrote a small qml app for myself that uses a column and a few rows that each has a Label and a TextField. it works fine except that in the last TextField, when I click in it, the OSK covers it up. It seems that the app should adjust to slide up over the OSK, but it doesn't automatically. I've tried wrapping it in a Flickable, but it doesn't work. | 14:05 |
| jdstrand | I've seen this in a number of other apps too. what is the correct technique to have the app adjust to when the osk pops up? | 14:06 |
| nik90 | jdstrand: http://nik90.com/ubuntu-touch-keyboard-tricks/ | 14:07 |
| timp | nik90: section.property: "name" works? | 14:07 |
| nik90 | jdstrand: short summary, you need to set anchorToKeyboard: true | 14:07 |
| nik90 | timp: no :/ | 14:07 |
| nik90 | timp: I tried model.contents.name, name, ccontents.name | 14:08 |
| nik90 | reallys strange | 14:08 |
| timp | nik90: the example http://qt-project.org/doc/qt-5.0/qtquick/views-listview-sections-qml.html uses a string to define the section.property | 14:08 |
| jdstrand | nik90: cool, thanks! | 14:09 |
| * jdstrand tries | 14:09 | |
| nik90 | timp: I think they made a mistake. Since I use the section feature in the clock app -> http://bazaar.launchpad.net/~ubuntu-clock-dev/ubuntu-clock-app/trunk/view/head:/clock/WorldClock.qml | 14:11 |
| nik90 | timp: you essentially give it a model property so that you have multiple section headers | 14:11 |
| nik90 | timp: actually you are correct! My mistake | 14:13 |
| nik90 | in the clock app, I also use a string | 14:13 |
| kalikiana_ | nik90: this doesn't show up at all and the delegate isn't even evaluated if I put a console.log in there http://paste.ubuntu.com/6703381/ | 14:17 |
| kalikiana_ | by the example it should be used like that | 14:17 |
| nik90 | kalikiana_: I realise that.. no worries, I will try to figure this out later | 14:20 |
| kalikiana_ | okay | 14:22 |
| timp | kalikiana_: can you see what's wrong here? https://jenkins.qa.ubuntu.com/job/ubuntu-ui-toolkit-trusty-amd64-ci/458/console | 14:24 |
| timp | poor jp getting all the failure logs ;p | 14:24 |
| kalikiana_ | timp: components::DatePickerAPI::test_1_changeDateToNextMonth() Uncaught exception: Cannot read property 'children' of null | 14:24 |
| kalikiana_ | Loc: [(0)] | 14:24 |
| timp | ah yes I see now. | 14:25 |
| timp | searching for "fail!" is a lot handier than going through all the "fail" instances in the file | 14:25 |
| mhall119 | jdstrand: did you get the help you needed? | 14:25 |
| timp | kalikiana_: weird. I got that from the headerInit MR... nothing to do with DatePicker. | 14:25 |
| jdstrand | mhall119: yes, thanks :) | 14:26 |
| nik90 | kalikiana_: https://bugs.launchpad.net/u1db-qt/+bug/1266478 | 14:29 |
| ubot2 | Launchpad bug 1266478 in U1DB Qt/ QML "Query model doesn't work properly with a multilevel document list" [Undecided,New] | 14:29 |
| nik90 | kalikiana_: can we expect bug fixes for u1db-qt since it seems like all contributors have moved to ui-toolkit? | 14:30 |
| kalikiana_ | nik90: I am trying to make some time. I think it will be worth a lot to have you actively giving feedback, in the past I sometimes had to "hunt down" people who had problems but didn't report | 14:33 |
| nik90 | kalikiana_: yes I can do that..well that's the least I can do :) | 14:34 |
| timp | nik90, kalikiana_ an example program (as small as possible) that shows the bug included in the bug report is always useful :) I guess for this one, it would need to include a tiny db | 14:35 |
| nik90 | timp: I will innclude a sample program where you just need to download the file and run it to see the bug. Good point | 14:35 |
| timp | nik90: awesome. that eliminates the need to hunt down people to figure out how to reproduce the bug :) | 14:36 |
| nik90 | kalikiana_: I hope those bug fixes propogate to a ppa for easy testing and I will rally people to do the bug reporting and testing to help you. | 14:37 |
| nik90 | timp, kalikiana_: added test program. | 14:48 |
| nik90 | to bug report | 14:48 |
| nik90 | mhall119: I have another tutorial which shows how to add content to the ubuntu welcome screen infographics. -> http://nik90.com/adding-infographics-to-ubuntu-phone/ | 14:52 |
| nik90 | mhall119: you think that would be good for the tutorial page? | 14:52 |
| mhall119 | nik90: I'll have to check with some people, last I heard we didn't want to recommend that API just yet | 14:56 |
| nik90 | mhall119: oh. I wasn't aware of that | 14:58 |
| mhall119 | nik90: it's been a while, so it may have changed | 14:59 |
| mhall119 | I don't even remember the reasons for it, tbh | 14:59 |
| jdstrand | nik90: it worked great. thanks! | 14:59 |
| nik90 | jdstrand: u r welcome :-) | 15:00 |
| nik90 | mhall119: no worries. Although the core apps like music are already using it now | 15:00 |
| nik90 | and clock will use it soon enough to indicate the alarms | 15:01 |
| mhall119 | nik90: that's fine, I think, I'm just not sure if non-core-apps should be using it yet | 15:01 |
| nik90 | ok | 15:02 |
| mhall119 | nik90: it may have been that we didn't have the apparmor policy back then, IIRC | 15:02 |
| nik90 | true | 15:04 |
| === matiasb1 is now known as matiasb | ||
| jdstrand | nik90, mhall119: fyi, I just learned that Qt.inputMethod.hide() will hide the osk. This might be a good addition to http://developer.ubuntu.com/apps/qml/tutorial/ubuntu-on-screen-keyboard-tricks/. Eg, I used nik90's method to always show the button. Then I use in onClicked for that button: Qt.inputMethod.hide() | 15:25 |
| mhall119 | thanks jdstrand | 15:28 |
| === charles_ is now known as charles | ||
| nik90 | jdstrand: ah thnx that is a useful method. | 16:01 |
| === sarnold_ is now known as sarnold | ||
| === gatox is now known as gatox_lunch | ||
| mhall119 | new app dev video: http://www.youtube.com/watch?v=8Hhtd2JUBZo&feature=youtu.be | 16:47 |
| === marcoceppi_ is now known as marcoceppi | ||
| === gatox_lunch is now known as gatox | ||
| === boiko_ is now known as boiko | ||
| ahayzen | nik90, thanks for the email :) | 20:23 |
| nik90 | ahayzen: :) | 20:24 |
| * ahayzen subscribes to the bugs | 20:24 | |
| nik90 | ahayzen: I think those bugs are not critical stoppers for the music-app | 20:24 |
| nik90 | since you wouldn't be anyway showing both the playlist name and tracks at the same time? | 20:25 |
| ahayzen | nik90, well the recent list is pretty much work other than the sorting | 20:25 |
| ahayzen | nik90, i need to filter the tracks by a playlist name :/ | 20:25 |
| nik90 | true, but why dont you use a index for playlist names alone where you can do the playlist name sorting | 20:26 |
| nik90 | with that you can get the docID and use that to retrieve the tracks separately? | 20:26 |
| nik90 | we have the issue when we try to combine both of them in a single query | 20:26 |
| ahayzen | nik90, doesn't it only give u the first track of each though IIRC? | 20:27 |
| nik90 | this is just a workaround. Obviously the fix would make life easier but until then a temp solution would suffice | 20:27 |
| nik90 | nope using a repeater, I was able to show all the tracks | 20:27 |
| ahayzen | nik90, yeah but then the swipeDelete/reordering is a nightmare ;) | 20:27 |
| === wendar_ is now known as wendar | ||
| nik90 | ahayzen: oh yeah :P | 20:28 |
| nik90 | btw were you able to achieve reordering in a listview? | 20:28 |
| nik90 | normally...(no reference to u1db) | 20:29 |
| ahayzen | nik90, hehe we have a custom implementation | 20:29 |
| ahayzen | nik90, i'm working on one which is portable for any listview :) | 20:29 |
| nik90 | nice! | 20:29 |
| ahayzen | nik90, thts the plan anyway :P | 20:29 |
| ahayzen | nik90, just need some patches to land in the UITK first | 20:30 |
| nik90 | good luck! | 20:30 |
| ahayzen | thanks lol | 20:31 |
| nik90 | I meant that sincerely :) it will benefit other devs as well | 20:31 |
| ahayzen | hopefully it will | 20:31 |
| labsin | What's the best way to package my libs with my click app? If I link my executable to shared libraries that aren't in the default images, do I need to copy and relink them, or can I use some CMake magic? | 20:38 |
| mhall119 | labsin: you'll need the libs in your .click package, yes | 20:39 |
| mhall119 | labsin: there's still some work underway to support multi-arch libs, so it may not fully work just yet | 20:39 |
| labsin | mhall119, I saw the commit just 2h ago | 20:39 |
| labsin | But I was wondering if I really need to copy all my libs to lest say ./lib and then link them, or do you know any convenient ways? | 20:40 |
| mhall119 | I believe the work being done will support a specific folder structure, and automatically allow linking to libs in those | 20:45 |
| mhall119 | click apps are all launched by an upstart process, which will be capable of adding the right library folder to use for linking at runtime | 20:46 |
| nik90 | balloons: ping | 20:54 |
| balloons | nik90, pong | 20:55 |
| balloons | I know I have a pending review for you.. it popped up on the list of leftovers :-) | 20:55 |
| nik90 | balloons: yup indeed :) | 20:55 |
| nik90 | balloons: I fixed the stuff we discussed before | 20:55 |
| balloons | I'll try and review today.. working on getting my devices updated | 20:57 |
| balloons | excellent.. so it should be quick and easy.. does it pass on your nexus 4? | 20:57 |
| nik90 | It passed on my desktop :P | 20:57 |
| nik90 | balloons: you are going to angry if I told you I have still not figured the exact process on testing on phone | 20:58 |
| * nik90 ducks | 20:58 | |
| balloons | nik90, ohh boy, lol | 20:59 |
| balloons | we should fix that | 20:59 |
| nik90 | balloons: btw is https://bugs.launchpad.net/ubuntu-clock-app/+bug/1256864 still valid? | 21:02 |
| ubot2 | Launchpad bug 1256864 in Ubuntu Clock App "test_alarm.TestAlarm.test_longpress_new_alarm fails on #40 on mako" [High,Confirmed] | 21:02 |
| nik90 | if yes, can we move it to the mir project since it isnt a clock app isue? | 21:02 |
| nik90 | issue* | 21:02 |
| balloons | nik90, ohh man.. boo on me for not putting the bug number in there | 21:05 |
| nik90 | yay one less clock bug :P | 21:08 |
| nik90 | balloons: the continous integration failed | 22:41 |
| nik90 | with a message Unapproved changes made after approval. | 22:41 |
| balloons | ahh funzies | 22:41 |
| nik90 | But I did not make any changes | 22:41 |
| balloons | I approved, should merge now | 22:43 |
| nik90 | thnx | 22:52 |
| === salem_ is now known as _salem | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!