**** BEGIN LOGGING AT Tue Aug 15 03:00:02 2017 Aug 15 07:22:17 Morning Aug 15 07:43:12 Morning! Aug 15 08:01:05 Herrie|Pre3: I've played a bit with ubuntu-ui-toolkit, and I'm now able to start some Ubuntu Touch apps on my PC, using Qt 5.9 Aug 15 08:03:21 However their styling capabilities is far too complex, imho. They have a lot of dependencies on Qt's internal APIs... Aug 15 08:04:09 Tofe: Ooh nice! I should have some LuneOS Components ipk for you tonight... Took a bit longer because I was busy with 5.9 stuff Aug 15 08:04:17 Any apps in specific? Aug 15 08:04:59 ubuntu-ui-toolkit is open source now? I remember in the past some custom QML stuff wasn't open sourced (like Lipstick for Jolla/SFOS)? Aug 15 08:05:06 I just tried two, so far. Mastermind and AnimeNews. Aug 15 08:05:16 ubuntu-ui-toolkit is LGPL-3.0 Aug 15 08:06:31 but it's not straightforward to try an app, because I have to make it point to my own build of the toolkit, so I've just tried little qml apps so far Aug 15 08:07:28 Ideally, I'd like to take a "click" app and convert it to ipk Aug 15 08:14:28 Tofe: That would be nice Aug 15 08:14:56 5.9 upgrade is a bit challenging due to gyp->gn migration Aug 15 08:15:13 The patches itself I updated for webengine, now need to do chromium still Aug 15 08:15:22 For SFOS, their components are only partially opensource (only the QML part), the rest is closed. so it's not possible to rebuild it for LuneOS. However, I think they use much less tricks than UbuntuTouch, so a simple wrapper might work. Aug 15 08:16:51 Probably. Aug 15 08:18:52 So will do all the patch updates first and then try to tackle the gn bits Aug 15 08:19:17 Shouldn't be rocket science just need to figure out where to move the required bits. Aug 15 08:20:43 ok... good luck :) Aug 15 08:28:38 Tofe: When you're back from your trip :P ? Aug 15 08:30:28 Herrie|Pre3: not yet, one more week! Aug 15 08:31:50 :) Aug 15 08:32:06 I'll be gone for 2 weeks in mid Sept :) Aug 15 08:35:06 :) Aug 15 09:56:16 Tofe: out of curiousity, what are you working on? Some uniform api to wrap all these QML components? Aug 15 09:56:55 saidinesh5: no, that was already begun with another initiative, whose name I can't remember Aug 15 09:57:09 yeah, i remember seeing one too.. Aug 15 09:57:31 I just wanted to see if I could easily have the ubuntu touch apps running on LuneOS, without too much effort Aug 15 09:57:36 Ahhh Aug 15 09:57:58 the difficult point was Qt 5.8, again :) Aug 15 09:58:40 (ubuntu guys have the same issue, ironically http://ubuntu-discuss.cattt.com/2017/07/re-status-of-ubuntu-touch-packages_7.html ) Aug 15 09:59:35 Ahh .. didn't know they had to patch it so much... Aug 15 10:00:01 With the sacrifice of some components, and upgrading the code a little bit, I got it to run https://github.com/Tofee/ubuntu-ui-toolkit/commit/696f222107e9835d4e19d5dd76860db1f647a8ee Aug 15 10:00:57 even the bits like swipe gestures ? Aug 15 10:01:03 saidinesh5: well, they have a bad habit of using too much the internal Qt APIs, and here it backfired Aug 15 10:01:24 saidinesh5: I think so, though I didn't test it on device Aug 15 10:01:29 Ahh and how tightly are they tied to Mir? Aug 15 10:01:48 saidinesh5: not much hopefully, just to Qt for this code Aug 15 10:01:56 I see Aug 15 10:02:31 also out of curiosity, what's the way to port Lune OS to a new device? get halium up and running and simply install the rootfs? or is there an easier way? Aug 15 10:03:21 saidinesh5: that's what we'll recommend yes, but we didn't really integrate Halium yet Aug 15 10:03:58 Ah.. I see Aug 15 10:05:52 saidinesh5: I used Halium for the Touchpad upgrade recently, and got it working, but it was a device not yet supported by Halium, so it's not very representative of the effort needed for another deivce like N5X or so Aug 15 10:07:13 * saidinesh5 looks at halium's documentation Aug 15 10:09:59 saidinesh5: roughly, for Halium, we still need to patch init.rc a little bit (maybe not anymore, I have to verify that part) and to adapt a bit how we handle the ramdisk image. Maybe some little bits about the exact hybris version to use, and/or libcompat usage. Aug 15 10:10:52 the ramdisk image? i thought that only has basic telnet + little script to actually start the systemd right? Aug 15 10:13:37 saidinesh5: the ramdisk image of halium, for android's init - doesn't do much, but it contains the init,c files and the name is slightly different from what our scripts expect currently Aug 15 10:14:31 I see Aug 15 10:14:53 saidinesh5: also I'd like to keep the build of the kernel as part of our build toolchain, instead of putting it into android's build toolchain Aug 15 10:15:38 That shouldn't be a problem, but we might have some surprises with our gcc 6 compiler Aug 15 10:15:54 gcc 5 is the one that broke compatibility right? Aug 15 10:16:02 mostly yes Aug 15 10:17:20 but it's just a couple of patches, we know how to solve that issue now Aug 15 10:19:59 Ah I see Aug 15 10:20:21 from the looks of it,halium is more or less similar to porting sfos sdk.. infact theyr'e using hybris boot etc.. Aug 15 10:20:31 so Will try start the port maybe this week/end Aug 15 10:20:44 any preference on halium 7.1 or 5.1? Aug 15 10:31:27 saidinesh5: We don't have anything on 7.1 yet, but should be fine too. As long as libhybris is fine with 7.1 :P Aug 15 10:32:02 I see.. will give it a shot then Aug 15 10:33:26 saidinesh5: which device? Aug 15 10:33:33 Which device you're looking at? Aug 15 10:33:42 Xiaomi Mi3/Mi4.. (cancro) .. msm8974,... Aug 15 10:33:49 very close to Nexus 5 i think Aug 15 10:34:25 You should be fine with it. Aug 15 10:34:37 We have a hard kernel 3.4 requirement now Aug 15 10:34:43 So that's good to keep in mind. Aug 15 10:34:48 I was supposed to get anbox up and running on it, but they say this device is about to get a kernel upgrade to 3.18 soon.. So waiting on that till then Aug 15 10:34:50 Ahh Aug 15 10:35:00 yeah this one has 3.4.. runs sailfish perfectly well as of now Aug 15 10:35:07 OK :) Aug 15 10:35:23 We need 3.4 for recent glibc and systemd in Yocto Pyro. Aug 15 10:35:38 3.10+ for Anbox would be nice :) Aug 15 10:36:36 Ahh..Yes. I think the newer systemd even requires a couple of patches to my 3.4 kernel.. Aug 15 10:36:57 saidinesh5: should be no big problems then; if you need to build the kernel using luneos's gcc, you can have a look at the latest commits in our mako/hammerhead kernels https://github.com/Tofee/android_kernel_lge_hammerhead/commits/hammerhead/3.4/cm-12.1 Aug 15 10:37:23 Shouldn't be too hard to get it working on LuneOS, Anbox that is when you have a proper kernel Aug 15 10:37:59 Yeah.. so far so many issues i have had with anbox was just kernel issues Aug 15 10:37:59 We'd like to have it too, but 3.4 seems to be close to be impossible to get Anbox working. Aug 15 10:38:28 probably.. Aug 15 10:38:38 i think thereis a 3.10 kernel for Nexus 5 too right? Aug 15 10:38:51 plus someone was even working ona mainline kernel for nexus 5 btw. Aug 15 10:38:57 maybe worth looking into .. Aug 15 10:39:13 if you are fine with just serial console... sure... :P Aug 15 10:39:41 <- this guy was looking at mainline kernel :P Aug 15 10:39:48 bshah: hehe :) Aug 15 10:40:40 bshah: heh there is also a mainline kernel for Nexus 7 flo.. so that's some place i can look at Aug 15 10:41:56 saidinesh5: mainline kernels are fine... as long as the rest comes along, i.e. graphics, sound etc. If it's just the mainboard code, the interest is limited for us Aug 15 10:42:07 for Nexus 7, they do I think Aug 15 10:42:45 https://wiki.linaro.org/LMG/Kernel/FormFactorEnablement Aug 15 10:45:15 saidinesh5: that's a good initiative, but it's still maintained outside of mainline Aug 15 10:46:11 still, it's good, let's be positive :) Aug 15 10:46:12 Yep Aug 15 10:46:33 i mean they can't really merge all the android specific bits that were rejected by the mainline right? Aug 15 10:47:55 right; and there are dependencies to the android firmware, which must be a nightmare to maintain :) Aug 15 10:49:45 Yep Aug 15 18:08:45 JaMa: Were you able to find any clues for the VBox/Qemu issues? Aug 15 18:59:23 Herrie: not yet, I wanted to test it in fresh build with latest master branches and there are too many failures all over the place due to switch to openss11 Aug 15 18:59:26 openssl11 **** ENDING LOGGING AT Wed Aug 16 03:00:00 2017