| zubozrout | Hello. Could someone please help me out with my QML project. This is one of the files: http://paste.ubuntu.com/14456672/ and for some reason this code: connection_detail.console_out += "→ " + type + "(" + num + ")" + "\n"; is called more times than how many items are present in the connection_trains_detail_info_model. More precisely if there are n items in the model then the components in the delegate are called n fac | 09:29 |
|---|---|---|
| zubozrout | So if I mark model items as "1","2","3" ... then this is what is called in the delegate: "1","2","3","2","3","3" - and although I've been wondering about that for the whole day now I haven't figured it out. | 09:29 |
| zubozrout | The odd thing is that it renders correctly, there are only connection_trains_detail_info_model.count items in the listview visible - but it destroys my text output + it is not much efficient if it has to call some items multiple times. | 09:33 |
| zubozrout | Huh, ok, using Repeater helped ... not that I like it, but it works :) | 10:10 |
| bartbes | popey: didn't you look at running the ubuntu sdk on arch a while back? | 15:16 |
| popey | ya | 15:18 |
| bartbes | did you get it running? | 15:18 |
| popey | not quite | 15:18 |
| popey | it's in progress | 15:18 |
| bartbes | well, at least adding a second core to my ubuntu vm seems to have helped | 15:18 |
| bartbes | I finally have time to look at the terminal emulator again | 15:20 |
| bartbes | so let's see if I can find a codepath that is run both on a physical and an on-screen keyboard | 15:20 |
| bartbes | actually, can I run the osk on the desktop? | 15:26 |
| popey | bartbes, been playing with sdk in docker, DanChapman has a dockerfile which works on ubuntu, not tried on arch yet | 15:29 |
| bartbes | docker should work, got a link? | 15:29 |
| popey | lemme push what i have | 15:29 |
| bartbes | this is fun, the desktop version's saying it can't load the xcb qt platform plugin | 15:30 |
| popey | https://github.com/popey/ubuntu-sdk-docker | 15:30 |
| popey | i just took what DanChapman had done and stripped it down a bit, as it had lots of extra bits for dekko | 15:30 |
| popey | i just wanted a bare docker container | 15:30 |
| popey | that might be ubuntu specific (I mean, the host) so if it doesn't work on arch, let me know, not tried it there yet | 15:31 |
| popey | will do later when kids in bed | 15:31 |
| popey | should be lighter weight than a full vm | 15:31 |
| bartbes | what would be ubuntu-specific about it? | 15:32 |
| popey | dunno, user ids and stuff in the launch line | 15:32 |
| bartbes | you can get the xauthority path from $XAUTHORITY, I'm fairly sure, but I've never seen a system that didn't have it in $HOME | 15:34 |
| bartbes | having qt creator running on my own x session will probably make for a much nicer experience, let's see | 15:36 |
| popey | http://imgur.com/yMY1wHl | 15:37 |
| popey | works on my local machine - as you can see I don't have ubuntu-sdk / qtcreator installed at all, so it's running from docker image | 15:37 |
| bartbes | it's building now | 15:37 |
| popey | patches welcome :) | 15:39 |
| popey | be nice to be able to mount some part of your host filesystem (like ~/source) inside the docker container, not looked at that | 15:39 |
| bartbes | you mount your home | 15:42 |
| bartbes | just look at /home/popey ;) | 15:42 |
| bartbes | or /home/alan, I guess | 15:42 |
| popey | well yeah. | 15:43 |
| popey | but I didn't know how to do that :) | 15:43 |
| bartbes | your current script does mount it | 15:43 |
| popey | oh, doesn't for me | 15:43 |
| bartbes | well, it claims to, I haven't actually run it yet | 15:43 |
| popey | oh, i lie | 15:44 |
| popey | it does | 15:44 |
| popey | :D | 15:44 |
| * popey marks fixed ;) | 15:44 | |
| bartbes | so many deps! | 15:45 |
| bartbes | I guess the next step is having an armhf click thing ready inside of it | 15:47 |
| popey | that would be handy | 15:47 |
| popey | but I'm find with providing the minimal necessary to get started, the script _could_ also do the "click chroot create" bit too | 15:48 |
| bartbes | especially since you pass --rm, so you'd have to create a new one every time | 15:48 |
| popey | so when you first launch the container, all the stuff is just there | 15:48 |
| popey | good point | 15:48 |
| bartbes | yeah, it wouldn't be too difficult to have a second Dockerfile which uses the first as a base | 15:48 |
| bartbes | or you could of course mount the chroot in | 15:49 |
| popey | I think I'd rather keep it all inside the container.. | 15:50 |
| popey | so it's basically 1) install docker, 2) run script, 3) wait, 4) win | 15:50 |
| ilhami | hey | 15:51 |
| ilhami | :D | 15:51 |
| popey | bartbes, have added the click chroot create step to mine which is running now | 15:54 |
| bartbes | oh, btw, is there a way to run the osk on the desktop? | 15:59 |
| popey | possibly | 16:01 |
| popey | Elleo may know | 16:01 |
| popey | creating the click chroot barfs... | 16:02 |
| popey | http://paste.ubuntu.com/14459138/ | 16:03 |
| popey | aha http://stackoverflow.com/questions/33235395/run-chroot-within-docker | 16:03 |
| bartbes | or, since you seem to have 10 minutes, just port click to docker :P | 16:07 |
| popey | hah | 16:07 |
| popey | ITYM snappy | 16:07 |
| bartbes | how does the chroot work anyway, does it just come with a statically linked qemu? | 16:08 |
| popey | yeah | 16:08 |
| popey | the --privileged seems to only work with docker run, not docker build | 16:08 |
| popey | which is somewhat less useful | 16:08 |
| popey | might have to just let the user do the click chroot create on first run as usual | 16:08 |
| popey | and remove the --rm? | 16:08 |
| popey | rather than try and second guess it in the dockerfile | 16:09 |
| bartbes | does running take a while for you, or is it failing? | 16:11 |
| bartbes | I'm guessing it might be because I use nvidia drivers | 16:13 |
| bartbes | considering the gl errors I'm getting | 16:13 |
| popey | which bit? | 16:13 |
| popey | the run click window which builds the armhf chroot seems to hang | 16:14 |
| bartbes | no, just launching the container | 16:14 |
| popey | libGL error: failed to open drm device: Permission denied | 16:14 |
| popey | libGL error: failed to load driver: i965 | 16:14 |
| popey | couple of them in the output | 16:14 |
| popey | intel laptop here | 16:14 |
| bartbes | http://hastebin.com/tirosefete.txt | 16:14 |
| popey | oof | 16:16 |
| bartbes | welp, back to the vm I go | 16:21 |
| farad | has anybody gotten an error similar to this? http://pastebin.ubuntu.com/14459067/ | 16:22 |
| farad | only occurs when building for the phone, not for desktop builds | 16:23 |
| farad | it is the output of QtCreator, when building a "QML app with simple UI (cmake)" | 16:24 |
| farad | I am experimenting with the CurrencyConverter demo project from here: https://developer.ubuntu.com/en/apps/qml/tutorials/building-your-first-qml-app/ | 16:24 |
| dobey | farad: hmm, seems to be a bug. the arm moc binary can't be run anyway for cross compiling, because it's not building under qemu. i'm not sure why whatever is pointing to that specific binary is doing so. it should be using /usr/bin/moc instead | 16:26 |
| farad | it built fine exactly once, all consecutive builds failed | 16:27 |
| bartbes | so, back to this qt xcb platform plugin issue | 16:27 |
| farad | I tried to link this moc path to /var/lib/schroot/chroots/click-ubuntu-sdk-15.04-armhf/usr/bin/moc but the error message persists, which is quite strange, as the file is definitely existing in that case | 16:31 |
| dobey | farad: how did you link it exactly? | 16:38 |
| farad | well, I did not really change anything specific in the default configuration | 16:39 |
| farad | I created a Kit for armhf, and connected the phone | 16:39 |
| farad | the only thing I changed after the first (successful) build was the project description | 16:40 |
| dobey | i mean, i presume you ran ln -s to create a symlink yes? | 16:40 |
| dobey | what was the exact command you ran? | 16:40 |
| farad | yes | 16:40 |
| farad | I first tried to link against the path I gave some messages before, and now I also tried to create a symbolic link to /usr/bin/moc | 16:41 |
| farad | same result | 16:41 |
| dobey | ln -s /usr/bin/moc /var/lib/schroot/chroots/click-ubuntu-sdk-15.04-armhf/usr/lib/arm-linux-gnueabihf/qt5/bin/moc | 16:41 |
| dobey | is what should have been run | 16:41 |
| dobey | if you were creating the link from the host OS | 16:41 |
| dobey | if you are creating the link inside the chroot using the "Maintain" button in the settings page in qtcreator, then you need "ln -s /usr/bin/moc /usr/lib/arm-linux-gnueabihf/qt5/bin/moc" | 16:42 |
| farad | I did it using the terminal on the host and used the first command | 16:43 |
| dobey | then at least you should not get the same error | 16:43 |
| dobey | beyond that, i'm not sure why it would want to use that file in the first place | 16:44 |
| dobey | only thing further i can really recommend for you, is to file a bug against the sdk at https://bugs.launchpad.net/ubuntu-sdk-ide | 16:45 |
| farad | OK, thank you dobey! | 16:47 |
| bartbes | is there a way I can start an app from the command line and/or get its output? | 16:49 |
| Elleo | bartbes: for using the OSK on the desktop just install ubuntu-keyboard and whichever languages you need (e.g. ubuntu-keyboard-english, ubuntu-keyboard-emoji, etc.), then run maliit-server and start a Qt application with the environment variable 'QT_IM_MODULE=maliitphablet' set, but you'll get double input on all keys unless you're running xenial as there was a bug in qt resulting in multiple presses on multi-touch areas from mice | 16:54 |
| bartbes | still beats not being able to find the touch control flow | 16:54 |
| bartbes | hmm, doesn't seem to be working | 16:56 |
| bartbes | oh, maliit-server does say '"libubuntu-keyboard-plugin.so" "en" is not enabled' | 16:57 |
| bartbes | Elleo: got it working, totally forgot the console doesn't open the osk on startup | 17:56 |
| Elleo | bartbes: great :) | 17:58 |
| bartbes | turns out, the osk not only sends the double keypresses, as you suggested, it also sends each keypress twice? | 18:00 |
| bartbes | that might explain the behaviour I was seeing | 18:00 |
| bartbes | I'd never have figured this out without the osk on the desktop | 18:23 |
| bartbes | popey: I think I finally got it working | 18:23 |
| popey | bartbes, docker? | 19:15 |
| farad | @dobey: my issue was fixed by rebuilding the armhf kit | 19:24 |
| bartbes | popey: no, a control key in the terminal emulator | 19:25 |
| popey | oh, yay! | 19:25 |
| bartbes | this time actually tested on my phone, too ;) | 19:26 |
| bartbes | shall I mail the patch again? | 19:26 |
| popey | could you propose a merge? | 19:33 |
| popey | oh, you mailed already, cool | 19:33 |
| popey | will take a look, thanks bartbes | 19:33 |
| bartbes | yeah, I never did figure bzr/launchpad out | 19:34 |
| popey | dobey, you use a chroot on a device? I found via https://lists.launchpad.net/ubuntu-phone/msg12190.html and mentions here... | 21:22 |
| popey | do you make the phone R/W to put debootstrap on it? | 21:22 |
| popey | I was hoping to run debootstrap without actually going RW and using apt | 21:22 |
| dobey | popey: i do have chroots on my device | 21:22 |
| dobey | no need for debootstrap | 21:23 |
| popey | oh, how did you set them up?> | 21:23 |
| dobey | https://askubuntu.com/questions/620740/recommended-way-to-install-regularcli-deb-packages-on-ubuntu-phone/623311#623311 | 21:23 |
| popey | i want to use one for building stuff | 21:23 |
| dobey | grab the daily-preinstalled image and unpack it | 21:23 |
| popey | magic, ta | 21:23 |
| popey | sweet! | 21:23 |
| * popey upvotes | 21:23 | |
| dobey | whee. 349 more of those and i'll have the last privilege tier unlocked :) | 21:25 |
| mcphail | popey: remember if you want to debootstrap for a device which doesn't have it, you can run it on a normal machine with the "--foreign" parameter, copy the files across to your chroot, chroot in and run the "debootstrap/debootstrap" script with the "--second-stage" parameter to complete the process. But a pre-rolled chroot is nicer ;) | 21:28 |
| bartbes | fancy | 21:32 |
| bartbes | I do like debootstrap | 21:33 |
| farad | is there a recommended way of rotating a camera preview, so that you see through the camera preview what you would see without the phone in front of your face? Setting "autoOrientation" to true leaves me with a preview, rotated by a quarter clockwise or anti-clockwise - depending on the orientation of the phone | 22:06 |
| farad | do I have to do this manually, using the orientation sensor, or is there a simpler way? | 22:07 |
| bartbes | isn't that just a case of not rotating? | 22:22 |
| bartbes | seeing as the camera has the same rotation as the phone itself | 22:22 |
| farad | yes, but the screen is rotating as well | 22:23 |
| bartbes | yes, so disable screen rotation | 22:23 |
| bartbes | unless you want to rotate your ui, in which case you have to do the inverse rotation on the picture | 22:24 |
| farad | hmmm, this will block the title bar from moving to the top, won't it? | 22:24 |
| farad | OK | 22:24 |
| bartbes | I just tried the default camera app and it seems to rotate its ui, I'm not sure if it manually "unrotates" the image, or manually rotates the ui | 22:25 |
| bartbes | but you basically want to make sure the image isn't rotated | 22:25 |
| farad | hmm, I guess it manually rotates its UI and fixes the preview | 22:27 |
| farad | good idea! I will proceed with that. | 22:27 |
| farad | thank you | 22:27 |
| farad | I thought setting "automaticOrientation" of the MainView to false would block the rotation, but I cannot see any changes in the behaviour | 22:54 |
| farad | the screen still rotates if the phone is rotated | 22:55 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!