| === _salem is now known as salem_ | ||
| === salem_ is now known as _salem | ||
| === diddledan_ is now known as diddledan | ||
| === chihchun is now known as chihchun_afk | ||
| Gerowen | I'm thinking about trying Ubuntu Touch on my Nexus 7 tablet, but I was wondering if there's some way to preview it before-hand to see how it runs, what apps are available, etc. | 02:56 |
|---|---|---|
| DonkeyHotei | there are emulator images | 02:59 |
| genii | Gerowen: You might want to check out https://wiki.ubuntu.com/Touch/Emulator ...this runs on the computer, not on the tablet, but you get a feel for what it's like | 03:00 |
| === zhxt_afk is now known as zhxt | ||
| damja | Hi friends.. | 04:56 |
| damja | Anyone here now online? | 04:56 |
| damja | Can I install ubuntu touch in my firefox os based flame device? | 04:59 |
| DonkeyHotei | damja: is it one of the devices for which a port already exists? | 05:10 |
| === chihchun_afk is now known as chihchun | ||
| ric96 | any guesses why i am getting this ? | 05:30 |
| ric96 | make: *** No rule to make target `/home/ric/phablet/out/target/product/condor/obj/STATIC_LIBRARIES/libmake_f2fs_intermediates/export_includes', needed by `/home/ric/phablet/out/target/product/condor/obj/EXECUTABLES/recovery_intermediates/import_includes'. Stop. | 05:30 |
| peat-psuwit | ric96: You need to add f2fs-tools in local manifest | 05:38 |
| peat-psuwit | <remote name="github" fetch="https://github.com/" /> | 05:38 |
| peat-psuwit | <project path="external/f2fs-tools" name="CyanogenMod/android_external_f2fs-tools" remote="github" revision="cm-11.0" /> | 05:39 |
| ric96 | I did that and also added it to main.mk | 05:46 |
| ric96 | didnt help | 05:46 |
| ric96 | peat-psuwit: ^ | 05:48 |
| peat-psuwit | did you run "repo sync" ? | 05:53 |
| === duflu_ is now known as duflu | ||
| ric96 | i did not add it to mainfest but manually git cloned it to external | 05:56 |
| ric96 | and of course renamed it to f2fs-tools | 05:57 |
| ric96 | peat-psuwit: ^ | 05:58 |
| peat-psuwit | ric96: May I see your build/core/main.mk? Put it on pastebin or so. | 06:03 |
| ric96 | peat-psuwit: http://pastebin.com/cnSGAcfA | 06:17 |
| ubuntuser13 | where i can find ubuntu touch keyboard layout tutorials? | 06:19 |
| peat-psuwit | ric96: That's strange. Try running "make clean" and try again. | 06:29 |
| ric96 | peat-psuwit: didnt help | 06:35 |
| peat-psuwit | ric96: Then I'm out of idea. | 06:38 |
| ric96 | peat-psuwit: hmmm.... anyways thnx for you help | 06:44 |
| === chriadam is now known as chriadam|away | ||
| larlar | I have started to try to port ubuntu to lg g2 (d802 version). | 07:15 |
| larlar | I have gotten a working recovery built. | 07:16 |
| === marcusto_ is now known as marcustomlinson | ||
| dholbach | good morning | 07:47 |
| elimisteve | http://www.htxt.co.za/2015/03/02/mwc15-hands-on-with-the-ubuntu-mx4-a-better-class-of-phone/ | 08:38 |
| elimisteve | that's the only link I can find about Ubuntu Touch running on the MX4 at MWC! Hopefully there's more press than this? | 08:38 |
| === greyback_ is now known as greyback | ||
| === ecloud is now known as ecloud_wfh | ||
| elimisteve | Nevermind: https://www.google.com/?q=mwc+2015+ubuntu | 10:06 |
| JamesTait | Good morning all; happy Grammar Day! :-D | 10:10 |
| === iahmad_ is now known as iahmad | ||
| popey | http://www.reddit.com/r/Ubuntu/comments/2xto6a/i_was_at_the_canonical_stand_and_spent_5_minutes/ | 10:21 |
| popey | that made me chuckle | 10:21 |
| ogra_ | popey, heh, funny, mark *never* was CEO | 10:44 |
| oSoMoN | anyone knows who should handle requests to add a package to the touch seed ? (re: bug #1378184) | 10:55 |
| ubot5 | bug 1378184 in ubuntu-touch-meta (Ubuntu RTM) "RTM r3 why is Nepali Unicode texts not displaying in browser app" [Undecided,Confirmed] https://launchpad.net/bugs/1378184 | 10:55 |
| ogra_ | oSoMoN, rsalveti or me ... | 10:56 |
| oSoMoN | ogra_, so, can we add fonts-lohit-deva to the touch seed (both vivid and RTM) ? | 11:02 |
| ogra_ | oSoMoN, yeah, i need to talk to jibel if/how he wants to pre-test that ... seed changes dont go though silos | 11:03 |
| ogra_ | i guess its prety easy to test by just installin the font and switching to the language | 11:03 |
| oSoMoN | ogra_, no need to even switch to the language, just point the browser to http://www.ashesh.com.np/nepali-unicode.php and verify that nepali characters are correctly displayed | 11:04 |
| oSoMoN | ogra_, would you mind assigning the bug to yourself (to show progress, and to avoid it falling off the radar again)? | 11:05 |
| ogra_ | done | 11:05 |
| oSoMoN | thanks | 11:06 |
| === MacSlow is now known as MacSlow|lunch | ||
| === uuhimhere is now known as illusione | ||
| === _salem is now known as salem_ | ||
| ogra_ | pitti, i'm just wonderin about bug 1378184 ... shouldnt we perhaps make fonts a dep of the specific langpacks | 12:50 |
| ubot5 | bug 1378184 in ubuntu-touch-meta (Ubuntu RTM) "RTM r3 why is Nepali Unicode texts not displaying in browser app" [Undecided,Confirmed] https://launchpad.net/bugs/1378184 | 12:50 |
| === illusione is now known as desert_mirage | ||
| === alan_g is now known as alan_g|lunch | ||
| popey | http://m.bbc.co.uk/news/technology-31723029 | 13:22 |
| === desert_mirage is now known as great_facade | ||
| === great_facade is now known as keepItUp | ||
| pitti | ogra_: or just seed them? that's what we do on desktop, as displaying foreign documents or websites shouldn't really require having the langpack installed | 13:33 |
| ogra_ | pitti, thats what i was questionin | 13:33 |
| ogra_ | g | 13:33 |
| ogra_ | but yeah, the webpage issue is indeed true | 13:33 |
| pitti | or emails and other docs | 13:34 |
| ogra_ | right | 13:34 |
| pitti | on desktop we have a more dynamic way of installing extra packages by language (check-language-suport), but obviously that won't make sense on touch | 13:34 |
| pitti | but I don't see a Nepalese specific font on desktop either | 13:35 |
| === dandrader is now known as dandrader|afk | ||
| === zhxt_ is now known as zhxt_home | ||
| === keepItUp is now known as disengaging_NOW | ||
| === MacSlow|lunch is now known as MacSlow | ||
| === alan_g|lunch is now known as alan_g | ||
| === dandrader|afk is now known as dandrader | ||
| Verc | HI :) Is there possibility to connect to WPA Enterprise network in Ubuntu Touch 14.10 ? | 14:06 |
| popey | Verc: i think we have an open bug for that. | 14:08 |
| Verc | ok | 14:08 |
| Verc | and when I create wifi network in my Desktop (14.04 LTS) the telephone cannot connect to it | 14:08 |
| Verc | if it is locked | 14:08 |
| Verc | if the network is not protected the telephone connects to it | 14:09 |
| robjh | if you were to plug a nexus 7 with ubuntu into a usb phone modem, would it work as a land line phone?! :O | 14:10 |
| popey | that makes no sense to me. | 14:12 |
| popey | Verc: you might be able to copy the network config over with "phablet-network" | 14:13 |
| robjh | heh :3 | 14:13 |
| Verc | any tutorial? I'm totally green | 14:14 |
| Verc | I don't know if there is that idiom in English :D | 14:15 |
| popey | plug phone into computer which is on the wifi you want to connect to | 14:15 |
| popey | run phablet-network on pc | 14:15 |
| popey | it copies the network-manager config over to the device | 14:16 |
| popey | might work | 14:16 |
| Verc | i'm trying :) | 14:18 |
| Verc | "pROVISIONING NETWORK ON DEVICE" | 14:18 |
| Verc | sorry | 14:18 |
| Verc | It doesn't work | 14:20 |
| Verc | Provisioning network on device Network setup complete | 14:20 |
| Verc | and on the telephone the name of the network is green but I cannot browse Internet | 14:21 |
| popey | Verc: if you "adb shell" to the phone and run ifconfig or ip addr, do you see an IP address assigned? | 14:22 |
| popey | (It could be this wont work, I know it was broken in the past) | 14:22 |
| Verc | phablet@ubuntu-phablet:~$ ifconfig lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:10276 errors:0 dropped:0 overruns:0 frame:0 TX packets:10276 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:770232 (770.2 KB) TX bytes:770232 | 14:25 |
| Verc | I think I cannot | 14:25 |
| Verc | inet6 is the IP ? | 14:26 |
| saidinesh5 | Elleo: is libskeyer working well for you? | 14:32 |
| Elleo | saidinesh5: the results it gives are really good, but the performance suffers a lot with anything but very short words unfortunately | 14:33 |
| Elleo | saidinesh5: I might try just having skeyer produce 1 candidate for use as the primary one, then presage do all the others, see what the delay is like then | 14:33 |
| saidinesh5 | Elleo: could you give me the offending words? | 14:33 |
| saidinesh5 | Elleo: also if I have deliberately disabled caching some data for the sake of speed when it comes to typed alphabets | 14:34 |
| saidinesh5 | *for the sake of memory | 14:34 |
| saidinesh5 | as opposed to swiping | 14:35 |
| kenvandine | popey, so you're into construct 2 right? | 14:35 |
| kenvandine | popey, http://people.canonical.com/~kenvandine/space-blaster.ken-vandine_0.1_all.click | 14:35 |
| kenvandine | popey, one of the construct 2 demo games | 14:35 |
| saidinesh5 | *if/I have | 14:35 |
| Elleo | saidinesh5: anything over 5 or 6 characters seemed to get too slow to be reasonably usable (since we need to assume the user is going to tap space very quickly after the last character they entered) | 14:35 |
| saidinesh5 | O_o | 14:35 |
| saidinesh5 | umm that doesn't sound right... | 14:36 |
| kenvandine | popey, took me a whole 5 minutes to build a click package for a game i exported from construct 2 :) | 14:36 |
| Elleo | saidinesh5: I can test some specific words a bit later, currently having to reflash my phone after an upgrade to Qt 5.4.1 went horribly wrong ;) | 14:36 |
| saidinesh5 | Elleo: can you give me the list of words you are using to benchmark? | 14:36 |
| saidinesh5 | Ahh hehhe | 14:36 |
| saidinesh5 | you can add the unigramFrequencyDistance check in case of typed words too in BruteForceMatcher | 14:37 |
| saidinesh5 | but that means 5MB of extra memory | 14:37 |
| saidinesh5 | but definitely a huge improvement in speed | 14:37 |
| saidinesh5 | hold on adding it so you can check | 14:37 |
| popey | kenvandine: hahah, i also made a click for that exact demo too :) | 14:38 |
| Elleo | saidinesh5: okay, cool, will do some proper tests with and without that | 14:38 |
| Elleo | saidinesh5: if I can revive my phone :P | 14:38 |
| saidinesh5 | hehh lets see who gets it done first: you reviving your phone or me fixing this issue | 14:38 |
| Elleo | saidinesh5: it just popped up a brand new boot failure message I've never seen before | 14:38 |
| saidinesh5 | oh | 14:38 |
| Elleo | which is... exciting :/ | 14:38 |
| saidinesh5 | did you use apt-get to update Qt? | 14:38 |
| kenvandine | popey, actually http://people.canonical.com/~kenvandine/space-blaster.ken-vandine_0.1.1_all.click | 14:39 |
| kenvandine | with sound :) | 14:39 |
| saidinesh5 | i only compile things in a chroot on device now... | 14:39 |
| Elleo | saidinesh5: we have a system for updating things from testing package silos (which uses apt-get under the hood, but only ugprades the specific packages in the silo, so as to avoid breaking other stuff) | 14:39 |
| popey | kenvandine: http://popey.mooo.com/screenshots/2014/device-2014-09-11-195619.png | 14:39 |
| popey | :) | 14:40 |
| saidinesh5 | Ohh link? its not that .... nice to currently do everything in a chroot... | 14:40 |
| popey | kenvandine: I didnt put it in the store because I wasn't sure it was allowed | 14:40 |
| Elleo | saidinesh5: the aim is that we want to shift to Qt 5.4.1 soon, but apparently maliit is segfaulting under Qt 5.4.1 (but is fine on Qt 5.4.0...) so I'm having to poke around a bit | 14:40 |
| saidinesh5 | but also i am using multirom so things break and make the whole deviec unbootable | 14:40 |
| saidinesh5 | Ahh | 14:40 |
| popey | kenvandine: have you seen they're making construct3 which runs on Linux? | 14:40 |
| Elleo | saidinesh5: it's all based on our internal landing system stuff, I don't think its possible for anyone outside of canonical to add anything to a silo unfortunately | 14:40 |
| kenvandine | popey, ah... you grabbed the same demo :) | 14:41 |
| popey | ya | 14:41 |
| saidinesh5 | i am on 5.4 too, but i am guessign the segfaults are from the new QVariant behavior | 14:41 |
| saidinesh5 | Ohh | 14:41 |
| kenvandine | it's super easy to package them... so anyone can build a construct game for ubuntu | 14:41 |
| popey | kenvandine: i made a skeleton around it for any construct2 app | 14:41 |
| popey | yeah | 14:41 |
| kenvandine | i just copied it into www and built a click | 14:41 |
| kenvandine | done | 14:41 |
| popey | yeah | 14:41 |
| Elleo | saidinesh5: well I think the QVariant stuff changed in 5.4.0, (which maliit runs fine on) not certain though | 14:42 |
| saidinesh5 | oh | 14:42 |
| kenvandine | tedg, adding my mir-connection-demangler to the apparmor profile just opened a can of worms... because the APP_EXEC uses aa-exec-click | 14:43 |
| kenvandine | tedg, which is a shell script, and runs other commands | 14:43 |
| kenvandine | tedg, so i guess i need to mimic what aa-exec-click does and do it myself? | 14:43 |
| Elleo | saidinesh5: apparently the crash happens in QV4::ExecutionContext::setProperty() which is interesting, as that's part of the internal QML JS runtime; so it sounds like it might be an "interesting" one to debug ;) | 14:44 |
| Elleo | saidinesh5: assuming I can ever get my phone working again :P | 14:44 |
| saidinesh5 | Ahh | 14:44 |
| saidinesh5 | looks like that QJsonValue | 14:44 |
| saidinesh5 | JSValue | 14:44 |
| saidinesh5 | or whatever that blog post said lol | 14:44 |
| Elleo | saidinesh5: could you link me to that blog post? sounds like it might be handy to read up on | 14:45 |
| kenvandine | popey, so if there are lots of developers writing construct 2 games, we should get them to publish them for ubuntu :) | 14:45 |
| saidinesh5 | yeah already searching | 14:45 |
| saidinesh5 | was on planet kde i think | 14:45 |
| Elleo | saidinesh5: oh, that rings a bell actually | 14:45 |
| kenvandine | popey, i still think QML is more fun to develop games with :) | 14:45 |
| Elleo | saidinesh5: did it discuss the differences between QV4 + QV8 for compiling stuff for performance? | 14:46 |
| === uuhimhere is now known as engaging_facade | ||
| tedg | kenvandine, No, you shouldn't need aa-exec-click | 14:46 |
| tedg | kenvandine, What are you using that for? | 14:46 |
| Elleo | saidinesh5: I might be thinking of a different article though | 14:46 |
| saidinesh5 | Okay my bad its for 5.4 | 14:46 |
| kenvandine | tedg, i'm not trying to :) | 14:46 |
| saidinesh5 | http://tsdgeos.blogspot.in/2015/02/qt-54-qml-c-qvariant-issues.html | 14:46 |
| tedg | kenvandine, Upstart will setup the apparmor profile for you. | 14:46 |
| kenvandine | tedg, it's what's used for exec | 14:46 |
| kenvandine | my exec-tool (modified version of what's in pay-service) | 14:46 |
| saidinesh5 | with 5.4.1 i think they bring back support for qt types | 14:46 |
| tedg | kenvandine, I'm not sure what you mean there. | 14:46 |
| kenvandine | builds the APP_EXEC | 14:46 |
| kenvandine | which is aa-exec-click | 14:47 |
| Elleo | saidinesh5: ah, okay | 14:47 |
| tedg | kenvandine, Don't do that. | 14:47 |
| tedg | kenvandine, :-) | 14:47 |
| kenvandine | so i should create my own Exec line? | 14:47 |
| tedg | kenvandine, There's no reason for you to be using aa-exec-click. | 14:47 |
| kenvandine | not what click has in the desktop file? | 14:48 |
| tedg | kenvandine, No, it should be the one given by the app. | 14:48 |
| saidinesh5 | Ooo this looks interesting: "Other important fixes in Qt 5.4.1 include enabling JIT by default in the Qt Quick Engine (QTBUG-43171)," | 14:48 |
| tedg | kenvandine, The app most certainly doesn't have that in it's desktop file. | 14:48 |
| saidinesh5 | => much better performance../ | 14:48 |
| jdstrand | yes, to reiterate what ted said, don't do that :) | 14:48 |
| tedg | kenvandine, You can't use the desktop file in ~/.local/share/applications, that's a cache that should go away. | 14:48 |
| kenvandine | Exec=aa-exec-click -p com.ubuntu.developer.ken-vandine.hub-exporter_hub-exporter_0.2.1 -- qmlscene $@ hub-exporter.qml | 14:48 |
| tedg | kenvandine, You need to use the desktop file that the application ships. | 14:48 |
| jdstrand | kenvandine: don't you just need to fork/execvp() or similar? | 14:48 |
| kenvandine | oh... so that's why you always copy them around :) | 14:48 |
| Elleo | saidinesh5: looking at the bug it appears JIT was just disabled on windows | 14:49 |
| kenvandine | jdstrand, i am using execvp | 14:49 |
| kenvandine | but i was using the wrong APP_EXEC :) | 14:49 |
| tedg | kenvandine, symlink, but yes :-) | 14:49 |
| Elleo | saidinesh5: we have some nifty caching stuff for the JIT code in RTM at the moment, but that still has to be ported to 5.4 I think | 14:49 |
| kenvandine | tedg, i was wondering why you always make copies :) | 14:49 |
| kenvandine | ok... link | 14:49 |
| saidinesh5 | Ohh | 14:49 |
| Elleo | saidinesh5: which is why app start-up is much quicker on RTM than on vivid | 14:49 |
| jdstrand | right, so, just exec your executable | 14:49 |
| popey | kenvandine: i actually bought the paid version of construct2 :) | 14:49 |
| popey | kenvandine: made a couple of games already | 14:49 |
| kenvandine | popey, it's just not as much fun as QML :) | 14:50 |
| saidinesh5 | Elleo: i really thought you guys were using that qtquick compiler heh | 14:50 |
| jdstrand | alright, seems ted has it under control, I just wanted to reiterate what he was saying :) | 14:50 |
| kenvandine | writing games should be fun :) | 14:50 |
| kenvandine | jdstrand, thanks | 14:50 |
| popey | kenvandine: it is for me, who doesn't know qml very well | 14:50 |
| Elleo | saidinesh5: nah, I think we wanted a solution that worked as well as possible for all developers, rather than just those that could license stuff from digia | 14:50 |
| kenvandine | tedg, so indeed... i guess i need my hook to make a symlink too | 14:50 |
| saidinesh5 | oh.. wasnt there a fossy version on github? | 14:50 |
| tedg | kenvandine, Just add a click hook, and click will do it for you. | 14:50 |
| jdstrand | (you won't have access to aa-exec-click, can't aa_change_profile(), etc, etc) | 14:50 |
| kenvandine | tedg, yes... | 14:51 |
| tedg | kenvandine, You don't even need to give click an exec | 14:51 |
| Elleo | saidinesh5: yeah, it's not usable yet though | 14:51 |
| saidinesh5 | ohh thanks for the warning | 14:51 |
| tedg | kenvandine, Just have it maintain a directory of symlinks | 14:51 |
| kenvandine | jdstrand, yeah, that's what i hit... and i was like "no way i want to use aa-exec-click" :) | 14:51 |
| Elleo | saidinesh5: or at least it wasn't when we last looked at it, iirc it fails on anything using Loaders | 14:51 |
| jdstrand | hehe | 14:51 |
| kenvandine | tedg, yeah, i was trying to be slick and just reuse what i knew would be there | 14:51 |
| Elleo | saidinesh5: and didn't have arm support at the time (although I think they were working on that) | 14:51 |
| kenvandine | popey, i played with construct 2 for a bit last night, i pretty quickly felt constrained, just couldn't do what i wanted | 14:52 |
| jdstrand | kenvandine: iirc, isn't this just an executable down in /usr/lib somewhere? | 14:52 |
| saidinesh5 | ohh i thought it compiled things to C++? | 14:52 |
| kenvandine | it's a nice tool though | 14:52 |
| saidinesh5 | whats so big about arm support? | 14:52 |
| kenvandine | jdstrand, well... my mir-connection-demangler is | 14:52 |
| kenvandine | which uses execvp on the app's exec | 14:52 |
| popey | kenvandine: interesting. | 14:52 |
| kenvandine | which would most likely be qmlscene | 14:52 |
| jdstrand | I see, I was missing context | 14:53 |
| Elleo | saidinesh5: looks like they have arm support now: https://github.com/qmlc/qmlc | 14:53 |
| kenvandine | i need to be able to launch any app that provides a picker | 14:53 |
| Elleo | saidinesh5: it's basically just doing the normal Qt JIT stuff in advance | 14:53 |
| kenvandine | popey, it's fun to play with though | 14:53 |
| popey | yeah | 14:53 |
| kenvandine | popey, so are there a lot of games created with it? | 14:53 |
| jdstrand | kenvandine: oh, interesting, I thought it went the other way around. I thought the app needed to execvp mir-connection-demangler | 14:54 |
| Elleo | saidinesh5: whereas we let Qt do its JIT at app start-up and then save the JIT code then (so the developer doesn't have to worry about it) | 14:54 |
| kenvandine | jdstrand, ah... no other way :) | 14:54 |
| Elleo | (and it keeps pure QML apps arch independent) | 14:54 |
| saidinesh5 | Ahh | 14:54 |
| Elleo | saidinesh5: whereas I think the qml compiler from Digia does a lot more advanced stuff | 14:54 |
| saidinesh5 | neat | 14:54 |
| saidinesh5 | like? | 14:54 |
| kenvandine | mir-connection-demangler is just used to get the handle for the mir connection | 14:54 |
| kenvandine | so the app can use it | 14:54 |
| kenvandine | so it needs to be the middle man | 14:55 |
| popey | kenvandine: yeah, I rummage around in their forums a bit but seems mostly hobby devs | 14:55 |
| kenvandine | popey, yeah, that's the impression i got, clearly there are more construct2 games than there are ubuntu games... | 14:55 |
| popey | well, yes. | 14:55 |
| kenvandine | but it's nothing like libgdx or cocos2d-x :) | 14:55 |
| * jdstrand wonders if there is another problem lurking somewhere in there | 14:56 | |
| Elleo | saidinesh5: I'm not sure exactly what it does, but iirc someone did some benchmarks and got some pretty impressive improvements on some things (particularly large listviews iirc) | 14:56 |
| popey | kenvandine: or unity3d | 14:56 |
| kenvandine | jdstrand, oh i'm sure i'll hit another wall | 14:56 |
| kenvandine | popey, indeed | 14:56 |
| Elleo | saidinesh5: but that's proprietary, and has to be licensed from Digia unforunately, so I haven't looked at it in detail | 14:56 |
| kenvandine | but it is so easy to package a construct2 game, we should promote that | 14:56 |
| saidinesh5 | Ahh hmm | 14:56 |
| popey | kenvandine: blog it :) | 14:56 |
| saidinesh5 | on android i can certainly see how slow the app startup is though | 14:56 |
| kenvandine | popey, i will then! | 14:57 |
| saidinesh5 | and especially horrible if you use their qtquick controls! | 14:57 |
| mpt | What’s the Touch equivalent of the ubuntu-meta package? i.e. deciding exactly what packages get shipped on the touch image | 14:57 |
| jdstrand | I just wonder about inserting a new parent for qmlscene, etc | 14:57 |
| kenvandine | mpt, ubuntu-touch | 14:57 |
| Elleo | saidinesh5: personally I quite like Jolla's approach (and Nokia's old approach) of having all the common qt+qml libs in the app launcher's memory and just fork that to spped up start up time, but apparently that causes some security issues I believe | 14:58 |
| kenvandine | jdstrand, well it still runs under the same profile it normally would | 14:58 |
| kenvandine | i suspect it'll be fine | 14:58 |
| Elleo | speed* | 14:58 |
| jdstrand | if there are things that are going to be expecting qmlscene or similar in /proc | 14:58 |
| kenvandine | but then again... the deeper i get in this... the more layers i find :) | 14:58 |
| mpt | kenvandine, <https://launchpad.net/ubuntu/+source/ubuntu-touch> and <https://launchpad.net/ubuntu-touch> both 404 | 14:58 |
| jdstrand | I think the apparmo rbits will be ok | 14:58 |
| jdstrand | I wonder about other parts of the system | 14:58 |
| jdstrand | the sdk | 14:59 |
| jdstrand | etc | 14:59 |
| saidinesh5 | hmm..... iirc N9 used to use screenshots to make the startup look faster too | 14:59 |
| kenvandine | mpt, oh... ubuntu-touch-meta | 14:59 |
| kenvandine | sorry :) | 14:59 |
| ogra_ | mpt, ubuntu-tuoch-meta | 14:59 |
| Elleo | saidinesh5: yeah, that's super cheating :P | 14:59 |
| saidinesh5 | hehh but worked nicely | 14:59 |
| ogra_ | (without the typo :P ) | 14:59 |
| jdstrand | I'm not saying there is a problem. I'm merely saying I wouldn't be surprised if there was | 14:59 |
| saidinesh5 | even apple does that i think | 14:59 |
| kenvandine | jdstrand, yeah... lets hope not :) | 14:59 |
| mpt | kenvandine, thank you kindly :-) | 14:59 |
| kenvandine | np | 14:59 |
| mpt | I’ll add it to the Avengers page | 14:59 |
| ogra_ | does anyone still use that ? | 15:00 |
| jdstrand | I can say that one won't be able to use the current Exec line that is created in the desktop file in ~/.local/share/applications the the desktop hook provides | 15:00 |
| shashankmc | Hey, anyone has any clue if ubuntu touch can be flashed on one touch idol x plus? | 15:00 |
| jdstrand | cause it does: Exec=aa-exec-click -p ... -- <original Exec=> | 15:00 |
| kenvandine | jdstrand, yeah... that's what i just learned :) | 15:01 |
| mpt | ogra_, the Design team does … Unless you’re steeped in the code it’s not at all obvious which package controls which UI | 15:02 |
| kenvandine | i was trying to avoid my hook from keeping it's own symlink... but that didn't work out | 15:02 |
| jdstrand | but that probably doesn't matter cause I think these days ual is always preferred with unity8, and the .desktop file in ~/.local/share/applications is for non-mir environments | 15:02 |
| mpt | ogra_, for example, today I discovered that a bug report being filed on “ciborium” was not actually an accident | 15:02 |
| ogra_ | mpt, ah, good then ... i just wasnt sure if it is still used | 15:02 |
| ogra_ | heh, yeah | 15:02 |
| ogra_ | we also have nuntium :) | 15:03 |
| jdstrand | alright, well, I suspect there may be an issue lurking somewhere, but I hope not | 15:03 |
| saidinesh5 | Elleo: git pull | 15:33 |
| saidinesh5 | filtering even words out by caching some data. So increased memory usage by 4MB. but you should have a much higher speed now... | 15:34 |
| Elleo | saidinesh5: okay, cool; still trying to revive my phone | 15:35 |
| saidinesh5 | was doing this only for swipes before because words there are really big.. 30 characcters on average | 15:35 |
| saidinesh5 | Ahh sure np | 15:35 |
| saidinesh5 | Also another interesting bug was increasing the UI responsiveness of skeyer itself | 15:35 |
| saidinesh5 | i was using onMouseMoved: displayPoints.push(Qt.point(mouse.x, mouse.y)) | 15:36 |
| saidinesh5 | which was working well on my desktop but absolutely sucking on friends' shiny android phones | 15:36 |
| saidinesh5 | to draw the swipeTail | 15:36 |
| === dholbach_ is now known as dholbach | ||
| Elleo | interesting, do you get a lot more mouse movements reported on touch perhaps? | 15:36 |
| saidinesh5 | was surprised but i figured out pushing so many points on such tiny screen but such high screen resolution => lots of reallocs() | 15:37 |
| Elleo | ah | 15:37 |
| saidinesh5 | so using a prealllocated ringBuffer now in the UI | 15:37 |
| Elleo | cool | 15:37 |
| saidinesh5 | heh another cool trick was in the bruteforcematcher | 15:39 |
| saidinesh5 | previously i was using QMap<qchar, int> ish data type to store the frequencies of characters in a word | 15:39 |
| saidinesh5 | and i needed their "diffs" | 15:39 |
| saidinesh5 | so allocating/deallocating a new QMap<qchar,int> and copying stuff there to diff things was just waaaaaay too many reallocations | 15:40 |
| saidinesh5 | and QMap<qchar,int> is not exactly O(1) either | 15:40 |
| saidinesh5 | so now using QVector< QPair<QChar, int> > to store frequency vectors | 15:41 |
| saidinesh5 | and doing things in O(n) as opposed to the O(nlog n) as previously done | 15:41 |
| Elleo | sweet :) | 15:41 |
| saidinesh5 | basically given 2 sets , finding intersection... | 15:42 |
| saidinesh5 | is much easier with "sorted sets" | 15:42 |
| mariogrip_ | ogra_ will qcom cne work with ubuntu touch? | 15:42 |
| saidinesh5 | now the one place that can use some speed up is the editDistance method itself.. but holding it back until you tell me it i really needed :P | 15:43 |
| Elleo | heh | 15:43 |
| saidinesh5 | kcachegrind output btw http://imgur.com/a/iuePf#0 | 15:44 |
| saidinesh5 | before/after | 15:44 |
| Elleo | great :) | 15:48 |
| Elleo | I suspect I'm going to have to resort to a --wipe now, reflashing doesn't seem to be going anywhere | 15:49 |
| ogra_ | mariogrip_, hmm, that would likely require some support in network-manager ... not sure if there is anything yet ... i guess cyphermox could tell you | 15:49 |
| mariogrip_ | ogra_ is there any way of disable it? | 15:50 |
| ogra_ | no idea, i didnt know about cne 10min ago :) | 15:51 |
| ogra_ | but i assume if your driver supports it there might be a switch in the driver to turn it on7off | 15:52 |
| mariogrip_ | I might use the init.rc setup of mako. | 15:53 |
| saidinesh5 | Elleo: next time use a chroot :P | 15:53 |
| Elleo | saidinesh5: next time let other people worry about Qt upgrade issues :P | 15:54 |
| saidinesh5 | lol the last time i had qt upgrade issues the only option for me was chroot | 15:54 |
| saidinesh5 | cuz apt get was pulling in things that messed with /boot and breaking the whole device instead of just that multirom | 15:55 |
| ogra_ | yeah, better use chroots | 15:55 |
| dobey | mardy: did you see my comment on your MP? also, i'm not sure how you're testing it exactly, but it seems to me like it would just not work, as well. | 15:55 |
| Elleo | saidinesh5: yeah, I need to be testing issues on the phone though to be sure they're fixed | 15:55 |
| Elleo | saidinesh5: plus the silo system *should* avoid the sort of issue you were running into | 15:56 |
| * ogra_ thinks we should probably provide a "lxc-developer container" | 15:56 | |
| Elleo | (assuming the silo doesn't contain packages that break stuff) | 15:56 |
| ogra_ | Elleo, the base system does | 15:56 |
| ogra_ | our images are not designed as RW images | 15:56 |
| saidinesh5 | Elleo: i thought multirom isnt supported by utouch? | 15:56 |
| mardy | dobey: yes, you are right that queryinfo fails, but the write succeeds (for some reason signond allows that); but you are right, a script would be better | 15:57 |
| ogra_ | eventually you will hit a package that tries to regenerate the initrd or one that has files in the writable space | 15:57 |
| Elleo | ogra_: yeah, I realise that; but I'm trying to test Qt 5.4.1 upgrades to our images from silo 12 | 15:57 |
| ogra_ | either case will break your upgrade | 15:57 |
| dobey | mardy: that's odd that write is allowed. wouldn't that be a problem that any app could exploit by simply adding itself to the acl? | 15:58 |
| === chihchun is now known as chihchun_afk | ||
| === dandrader is now known as dandrader|lunch | ||
| mardy | dobey: no, it's not like that; it's that because of a workaround we added, "unconfined" can add itself to any ACL | 16:11 |
| dobey | mardy: ok. could we maybe just change signond instead to always add unconfined to any accounts that don't have it when something tries to access them? | 16:15 |
| dobey | mardy: that would at least move it to a single place for all accounts, and we could avoid duplicating the code to add unconfined to the ACL across all the plug-ins | 16:16 |
| mardy | dobey: no, modifying the ACL on the fly doesn't seem correct; we could add a debian patch to let "unconfined" always pass, but it's a pity since ATM we don't have any upstream patches | 16:22 |
| ryan_evos | Is the Ubuntu Touch Email client (part of Core Apps) still a thing? | 16:25 |
| dobey | mardy: why doesn't it seem correct? and i'm not sure why that would need to be a patch. requiring all plug-ins to duplicate code to add the same ACL seems incorrect to me | 16:25 |
| ogra_ | ryan_evos, dekko is ... | 16:26 |
| dobey | ryan_evos: a) there is no "Touch." it's just Ubuntu. b) it's not part of core apps no, but dekko does exist yes | 16:26 |
| dobey | (at least, i don't think it's part of the core apps project) | 16:26 |
| ogra_ | ryan_evos, https://appstore.bhdouglass.com/app/dekko.dekkoproject | 16:26 |
| ryan_evos | Was looking at this: https://wiki.ubuntu.com/Touch/CoreApps/EmailClient | 16:27 |
| dobey | ryan_evos: that is dekko | 16:28 |
| ryan_evos | Ah | 16:28 |
| ryan_evos | Got it | 16:28 |
| dobey | that wiki page is way outdated and i don't think it's actually part of core apps, though | 16:28 |
| dobey | mhall119, popey: ^^ is that wiki page still relevant wrt dekko? can you delete/update it as appropriate? | 16:29 |
| dobey | ok i need to get lunch | 16:33 |
| === dpm_ is now known as dpm | ||
| DonkeyHotei | is rsalveti here? | 17:01 |
| ogra_ | pitti, juts FYI, the image build looks fine | 17:01 |
| ogra_ | Building dependency tree... | 17:01 |
| ogra_ | Package 'systemd-sysv' is not installed, so not removed | 17:01 |
| ogra_ | ubuntu-minimal is already the newest version. | 17:01 |
| ogra_ | thats all i get new with your changes | 17:02 |
| mhall119 | Mirv: are you the guy to talk to about getting QtOrganizer API docs? | 17:03 |
| pitti | ogra_: thanks! | 17:03 |
| === dandrader|lunch is now known as dandrader | ||
| === danielg4 is now known as DonkeyHotei | ||
| DonkeyHotei | is rsalveti here? | 17:15 |
| OerHeks | DonkeyHotei, likely AFK | 17:20 |
| seb128 | hum | 17:21 |
| davmor2 | mhall119: he will likely be offline now but I would hazzard if Mirv isn't he can point you in the right direction maybe email him | 17:21 |
| seb128 | my rtm bq displays a wifi signal icon in the indicator but the aps list shows none as connected to | 17:21 |
| seb128 | is that a known issue? | 17:22 |
| davmor2 | mhall119: he is on holiday too till tuesday-ish iirc too | 17:22 |
| seb128 | kenvandine, mandel, bah, didn't see that for a while, but I just saw a case of duplicated updates on the settings panel on rtm 249 :/ (http://people.canonical.com/~seb128/updates.jpg) | 17:28 |
| kenvandine | seb128, ugh | 17:28 |
| kenvandine | i wonder how that happened... we clear that model | 17:29 |
| attente_ | bfiller: hi, would you mind looking at https://code.launchpad.net/~attente/maliit/1245925/+merge/250311 | 17:29 |
| kenvandine | seb128, did you maybe backout of the updates page before it was done checking then go back in? | 17:29 |
| seb128 | kenvandine, I had n-m in a weird state, see just before | 17:29 |
| kenvandine | it should still be getting cleared, but maybe there's a race | 17:30 |
| seb128 | kenvandine, it would display the icon as being connected to wifi but no ap selected | 17:30 |
| seb128 | kenvandine, I might have done that as well yes | 17:30 |
| kenvandine | oh... so maybe the check was hung | 17:30 |
| seb128 | could be | 17:30 |
| kenvandine | and it resumed after a second check started | 17:30 |
| seb128 | yeah | 17:30 |
| kenvandine | which shouldn't be possible, but if the connectivity API reported as connected when it was, perhaps | 17:31 |
| kenvandine | s/was/wasn't/ | 17:31 |
| seb128 | right | 17:31 |
| seb128 | well, first time I saw it in a while | 17:31 |
| bfiller | attente_: going to pass that on to Elleo to have a look | 17:31 |
| kenvandine | seb128, it probably couldn't happen if it wasn't for the nm bug :) | 17:32 |
| attente_ | bfiller: ok, thanks | 17:32 |
| kenvandine | i haven't seen it in ages either | 17:32 |
| seb128 | kenvandine, could be :-) | 17:32 |
| Elleo | attente_: will take a look tomorrow, just about to finish for the day :) | 17:39 |
| Elleo | saidinesh5: performance is *much* better, that's really cool | 17:39 |
| attente_ | Elleo: sure, thanks :) | 17:39 |
| saidinesh5 | Elleo: :D | 17:40 |
| saidinesh5 | swipes are still a little slower though | 17:40 |
| saidinesh5 | but will have to fix that later | 17:40 |
| saidinesh5 | typing is quite harder for me since the last 20 days cuz of what looks like RSI | 17:40 |
| saidinesh5 | Elleo: and the memory usage? | 17:40 |
| Elleo | saidinesh5: ouch, one of my friends had pretty bad RSI a few years back; really important to get it looked at as soon as possible and don't do much typing if its hurting | 17:41 |
| saidinesh5 | oh :\ | 17:41 |
| Elleo | saidinesh5: iirc he found some pads that elevated his wrists as he typed helped a lot | 17:41 |
| saidinesh5 | hmm i should try that | 17:42 |
| Elleo | saidinesh5: but best to get some proper advice from a doctor, it got a lot worse for him because he ignored it for ages and carried on working ridiculously long hours despite it | 17:42 |
| saidinesh5 | oh define ages | 17:43 |
| saidinesh5 | its been 20 days for me | 17:43 |
| Elleo | saidinesh5: memory usage is a bit high (78MB), but I think we can probably completely replace hunspell with libskeyer, which will bring us back down by 10-15MB (depending on the dictionary) | 17:43 |
| Elleo | saidinesh5: but that's an acceptable trade off as far as I'm concerned for the much better performance :) | 17:43 |
| saidinesh5 | interesting that skeyer is only 71MB | 17:43 |
| saidinesh5 | but wait thats a debug build? | 17:44 |
| Elleo | yeah | 17:45 |
| saidinesh5 | Ahh | 17:45 |
| saidinesh5 | yeah so everything should be around 50MBish then | 17:45 |
| saidinesh5 | in release build | 17:45 |
| Elleo | yeah, and getting rid of hunspell with be a big benefit | 17:46 |
| saidinesh5 | so this means i should work on finishing off the API first. wrt language models. 2 things needed: last used term needs more priority, need to add tests for incrementProbabilityOnce | 17:46 |
| saidinesh5 | oh yes you were even using 2 threads before | 17:46 |
| saidinesh5 | for hunspell | 17:46 |
| Elleo | yeah, we'll still have it multithreaded as we can't really afford any delay between a key being pressed and it being input | 17:47 |
| Elleo | even slight latency there feels really awkward | 17:47 |
| saidinesh5 | afaik for normal types every result is within 16msec, so it shouldnt matter.. no? | 17:47 |
| Elleo | saidinesh5: we'll still have presage doing stuff as well though | 17:47 |
| saidinesh5 | i would really like nice benchmarks though | 17:47 |
| saidinesh5 | Ahhh i am actually looking for proper bigram models | 17:48 |
| saidinesh5 | so i dont have to include presage | 17:48 |
| saidinesh5 | google's dictionaries were supposed to have bigrams too | 17:48 |
| saidinesh5 | but i couldnt find them | 17:48 |
| Elleo | yeah, we're currently using presage with 3-grams but I haven't evaluated its abilities on different amounts of n-grams really | 17:49 |
| Elleo | I think we just went with 3-grams because its the default presage recommends | 17:49 |
| saidinesh5 | Ahh | 17:49 |
| saidinesh5 | at this point i am actually not sure what you mean by 3-grams for presage though | 17:50 |
| saidinesh5 | if it is q-grams or n-grams | 17:50 |
| saidinesh5 | that is | 17:50 |
| Elleo | saidinesh5: n-grams | 17:50 |
| Elleo | where n is 3 | 17:50 |
| saidinesh5 | i know but i meant what presage means by ngrams | 17:50 |
| Elleo | ah right | 17:50 |
| saidinesh5 | because 3grams means 100s of mb of data | 17:50 |
| saidinesh5 | proper 3grams | 17:50 |
| Elleo | yeah, I have dug into presage's internals in any great detail | 17:51 |
| Elleo | haven't* | 17:51 |
| saidinesh5 | Ah | 17:51 |
| saidinesh5 | like q-grams are for filtering words before computing edit distance | 17:51 |
| saidinesh5 | n-grams are more for predicting a word based on previous words | 17:51 |
| saidinesh5 | most research papers i ve seen are calling q-grams as n-grams too | 17:51 |
| Elleo | yeah, presage basically builds a database of 1, 2 and 3 word n-grams and their frequency | 17:52 |
| Elleo | but I'm not sure if it has a threshold below which it might not bother recording them for example | 17:52 |
| saidinesh5 | wow.. how big is that database | 17:52 |
| Elleo | (e.g. if you only see a 3-gram once it might not be worth reporting) | 17:52 |
| saidinesh5 | yeah but the google 2 gram database was heaavy | 17:53 |
| Elleo | saidinesh5: depends on the sample data, by default presage produces its databases from books | 17:53 |
| saidinesh5 | Ah | 17:53 |
| Elleo | saidinesh5: e.g. a picture of dorian grey ends up giving you a ~6mb database | 17:53 |
| Elleo | which is the default for english | 17:53 |
| saidinesh5 | Ahh interesting | 17:53 |
| saidinesh5 | 6MB doesnt sound much actually | 17:53 |
| saidinesh5 | since the words are already indexed | 17:53 |
| saidinesh5 | i can use integer arrays | 17:54 |
| saidinesh5 | to store bigrams | 17:54 |
| Elleo | I did try building a really big database once, but presage's performance gets pretty dodgy with large databases (getting on for 40mb iirc) | 17:54 |
| saidinesh5 | interesting | 17:54 |
| saidinesh5 | btw. google uses bigrams on android | 17:54 |
| Elleo | anyway, I have to go catch a train | 17:54 |
| saidinesh5 | Ah sure cya later | 17:54 |
| Elleo | but thanks so much for your skeyer work, I think it's going to be a big boost to our keyboard :) | 17:54 |
| saidinesh5 | hehh no mention :) | 17:55 |
| saidinesh5 | this way more people work on and fix skeyer :P | 17:55 |
| saidinesh5 | profit . win-win. etc.. :P | 17:55 |
| mhall119 | thanks davmor2 | 17:58 |
| === alan_g is now known as alan_g|EOD | ||
| rokki | Hi guys! I tried to compile reminders-app in my 15.04 desktop, but got the following error message: Can not use "Organizer" module which has not yet been found. | 18:07 |
| rokki | Any idea? | 18:07 |
| davmor2 | rokki: might be better asked on #ubuntu-app-devel | 18:07 |
| === dandrader is now known as dandrader|afk | ||
| rokki | davmor2: thanks, I saw this channel in the reminders-app wiki page as the IRC support channel | 18:08 |
| === dandrader|afk is now known as dandrader | ||
| ocrbanda | hello | 20:42 |
| brendand | barry, i could look at hooking the upgrade reboot to autopkgtest-reboot, can you give an idea of what's involved? | 21:07 |
| brendand | barry, somewhere to start from | 21:07 |
| barry | oh brendand, come back! :) | 21:47 |
| dobey | does anyone have a fresh rtm flashed onto a phone? | 22:01 |
| === salem_ is now known as _salem | ||
| dobey | i guess not :-/ | 22:08 |
| aLeSD_ | hi all | 23:51 |
| aLeSD_ | I started to port a device to ubuntu touch .. the documentation says that if I have a AOSP tree of the device it's easy | 23:51 |
| aLeSD_ | I have this : https://github.com/bq/aquaris-E5 | 23:52 |
| aLeSD_ | it's the kernel source of my device... | 23:52 |
| aLeSD_ | the documentation says : If you're repository is compatible with the AOSP tree, it's just a matter of adding the device specific git repositories | 23:53 |
| aLeSD_ | but my question is where ? | 23:53 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
