**** BEGIN LOGGING AT Tue Jul 12 02:59:58 2016 Jul 12 03:54:34 I have an intriguing result on 1143, Herrie! Running luna-send -n 1 palm://com.palm.service.contacts.linker/dbUpdatedRelinkChanges '{}' Jul 12 03:54:49 seems to fix things. Jul 12 04:35:25 Is anyone here? Jul 12 04:35:34 Yes. Jul 12 04:35:52 :) Jul 12 04:36:38 There's a problem with the porting guide, for reliable porting using cm10.1 and cm11, you need oracle java 6 Jul 12 04:36:47 Not 7 Jul 12 04:37:06 Ah? That’s good to know. Jul 12 04:37:42 Unstable builds with java 7 can occur on cm11 and 10.1 Jul 12 04:37:56 Cm12 and above fixed that problem Jul 12 04:38:36 Hrm Jul 12 04:40:00 I didn't even know you could build luneOS with cm12 and above Jul 12 04:41:04 Do you work on porting? Jul 12 04:41:16 Or development? Jul 12 04:41:32 I work on apps and services. Jul 12 04:41:55 Ah, okay Jul 12 04:42:34 Is there dual sim support at all planned? Jul 12 04:43:39 Heh. I don’t know that any of our currently supported devices have 2 slots. Jul 12 04:44:00 Depends on the telephony lib, I suppose. Jul 12 04:45:17 I'm working on the Galaxy Star Pro, which has two sims Jul 12 04:46:09 * DougReeder nods Jul 12 04:46:52 Also, what about devices that don't have fastboot? Like most Samsung devices? Jul 12 04:47:02 I don’t know. Jul 12 04:47:19 You can't build a fastbootable kernel Jul 12 04:49:11 Have you guys tested luneOS on less than 1GB of RAM? Jul 12 04:51:58 I doubt it. Jul 12 04:53:10 :D Jul 12 04:55:25 How's keyboard development going? Jul 12 04:55:48 Is there full autocorrect support yet? Jul 12 04:56:16 I don’t think so Jul 12 04:58:21 Is there language support beyond english anywhere in the OS? Jul 12 05:01:29 The apps are partily internationalized. Jul 12 05:02:04 Finishing that and localizing should be straighforward. Jul 12 05:03:57 Xml strings? Or something else? Jul 12 05:04:23 A JSON file for each language. Jul 12 05:05:05 Ah, okay. Jul 12 05:05:21 Thanks Jul 12 06:31:50 Morning! Jul 12 06:32:22 ah, poor Zhongtiao1, his timezone hardly matches ours... Jul 12 06:33:12 I could answer some of his questions, if he was still onlin Jul 12 06:33:13 e Jul 12 06:33:32 Maybe he'll read the logs, so let's give it a try Jul 12 06:34:57 For dual sim, we currently don't support it, but our infrastructure (oFono) supports it afaik. But we have no UI to choose the active sim card, and we make assumptions that there is only one at the moment. Jul 12 06:53:28 For lack of fastboot, I don't really know; fastboot isn't needed in itself for LuneOS, but it is a convenient way to flash the new kernel. However, for example for Nexus 4, if we go through recovery, fastboot isn't involved at any time. Jul 12 13:44:32 JaMa: ping Jul 12 13:49:43 JaMa: if I want to duoplicate a finished build to another machine, but stripped from the heavy "work" content, would it still avoid rebuilding all the dependencies if I need to rebuild locally one package ? and can I simply delete the "work" subfolder or should I do something more subtil ? Jul 12 13:59:57 Tofe: you should copy just sstate-cache and metadata Jul 12 14:02:59 JaMa: well, I'd like to be able to rebuild one package if needed (let's say luna-next), and it will certainly need dependancies like sysroots, doesn't it? Jul 12 14:05:04 one typical scenario is I'm on vacation, with a light laptop, and I want to hack into some source code to test some fix, and rebuild just that one package (with bb -f -c compile toto && bb toto, typically) Jul 12 14:06:17 "light laptop" and "compile" don't sound like good combinations Jul 12 14:06:56 GodGinrai: I agree, thus the willingness to avoid compiling what doesn't need to Jul 12 14:07:22 But compiling a little package should be fine, even on a i3 Jul 12 14:10:52 The machine I'm talking about would be something like i3, 4Gb RAM, 120Gb SSD; vacation implies having time to spare, to compilation time isn't a big issue. But the small SSD here means I have to strip down a finalized build Jul 12 14:11:08 s/to/so/ Jul 12 15:48:06 Tofe: with sstate-cache moved all necessary dependencies will be unpacked to sysroot from sstate-cache Jul 12 15:49:11 coping e.g. whole sysroot to different path wont work at all (because of relocation issues) if you unpack sstate which isn't so slow, it will resolve relocation for you Jul 12 15:50:21 removing work directory while keeping sysroot and stamps also will fail, because bitbake will expect that configure was already executed (based on do_configure stamp) and compile step will fail, because there won't be any sources in work directory Jul 12 15:51:02 with rm_work it's a bit better, because work directory will be already smaller and stamp files for tasks which expect anyting in work directory are already removed - so it will unpack from sstate the steps it needs Jul 12 17:02:02 JaMa: oh, ok, yes, I didn't think about all this Jul 12 17:02:11 JaMa: thanks! Jul 12 17:08:04 JaMa: thanks for all these explanations! one last question: what is the "metadata" folder" you talked about ? Jul 12 17:17:55 Tofe: I meant the webos-ports directory without tmp-glibc Jul 12 17:18:12 like matching oe-core/meta-oe/meta-webos-ports repositories Jul 12 17:18:34 because if you build with different metadata it might change the sstate checksums and prevent you to reuse sstate Jul 12 17:29:33 ah, yes, of course :) Jul 12 17:55:05 DougReeder: OK good on the testing. Just how is this different from the force autolink? Jul 12 18:43:20 nizovn: ping Jul 12 18:44:10 Herrie|Pre3: pong Jul 12 18:44:38 I don't remember if we had a way in qml via plugin to check if a file exists? Jul 12 18:44:48 I somehow recall we did but not sure Jul 12 18:45:55 i think we don't have this Jul 12 18:46:28 nizovn: OK, I also couldn't find it... Jul 12 18:46:35 So I guess we don't Jul 12 18:48:25 yes, seems so Jul 12 18:49:17 Seems legacy lunasysmgr c++ code had some checks to see if a file existed before using it Jul 12 18:49:39 I.e. for audio playback of banners Jul 12 18:50:44 yes, sounds reasonable Jul 12 18:51:48 I guess to have some generic function for this would be good that we could call in QML through c++ plugin? Jul 12 18:52:26 yup Jul 12 18:55:53 Is that something you could write? Jul 12 18:56:37 yes Jul 12 18:57:28 Great :) Jul 12 20:26:12 Herrie, they are different methods on the service. I don't know how they differ, but I can guess the callback method only operates on the changed records, while force link would scan all records. **** ENDING LOGGING AT Wed Jul 13 02:59:58 2016