**** BEGIN LOGGING AT Sat Mar 03 03:00:01 2018 Mar 03 07:51:39 Morning! Mar 03 07:52:03 Morning. Mar 03 08:22:29 Morning Mar 03 08:23:40 Tofe: I have a half working libglibutil recipe... I.e. my errors for missing .h files are gone with this, however oFono cannot find the functions that should be provided by the .c files of libglibutil. Mar 03 08:23:46 I put my changes on https://github.com/webOS-ports/meta-webos-ports/commit/3e3b3cca2c27d1ab2ba28ab57bf48169a2a79f03 Mar 03 08:23:59 Well actually here: https://github.com/webOS-ports/meta-webos-ports/commits/herrie/ofono Mar 03 08:24:41 I'm getting test-sms-filter.c:1147: error: undefined reference to 'gutil_log_timestamp' Mar 03 08:24:55 Herrie|Laptop: your bb is a bit weird: the .h aren't in the -dev package ? Mar 03 08:25:09 Tofe: I'm not sure :P Mar 03 08:25:20 I was just playing around Mar 03 08:25:30 I guess I could just add them to the -dev as well yes Mar 03 08:25:38 Now I get a -dev ipk where they are Mar 03 08:25:42 they should be put there quite automagically Mar 03 08:26:16 oh but wait... you don't actually build the code ? :p Mar 03 08:26:49 I don't see any autotools or cmake or anything Mar 03 08:28:07 you should inherit something more, to have a classic make && make install Mar 03 08:28:44 but... I'm not sure what ! Mar 03 08:29:31 what I would suggest is: do_compile() { make } and do_install() { make install-dev } and see how it goes Mar 03 08:30:51 maybe you'll need to change directory Mar 03 08:33:02 Let me try Mar 03 09:45:07 Hmmz tells me it cannot find Makefile Mar 03 09:54:31 ok needed to remove autotools Mar 03 09:55:09 it works? Mar 03 09:57:32 Not fully yet Mar 03 10:12:28 I forgot, did you migrate the CFLAGS for android-headers for the aarch64 targets ? Mar 03 10:12:50 I only did it for tenderloin, didn't want to break your build Mar 03 10:15:25 Tofe: Yeah think so Mar 03 10:15:41 ok good Mar 03 10:16:02 This one you mean? https://github.com/shr-distribution/meta-smartphone/commit/b6c20ab76fcfc1605d8d9ae6927bb4e3938418d4 Mar 03 10:16:51 Yes exactly; I'm changing quite some stuff around here, so I'm just checking what's our status Mar 03 10:22:47 Tofe: All shoudl've been PR-ed already Mar 03 10:22:55 good Mar 03 10:28:28 Herrie: what's the status for crespo and i9300 ? Mar 03 10:28:36 noone builds it ? Mar 03 10:28:57 and it wouldn't even build anymore? Mar 03 10:32:11 i9300 = Galaxy S3 it was never one of our targets Mar 03 10:32:54 Crespo = Nexus S also not our target ever Mar 03 10:33:26 We had Maguro which was i9250 if I'm correct and we dropped it due to lack of 3.4 kernel and long in the tooth Mar 03 10:33:35 I.e. slow Mar 03 10:34:20 Because I'm modifying their kernel recipe a bit, i.e. they won't inherit kernel_android anymore, to avoid dependency loops Mar 03 10:34:21 What we should migrate = meta-hp, meta-lg, meta-motorola, meta-xiaomi, meta-huawei, meta-oneplus Mar 03 10:34:49 Why woukpld you touch it at all? Mar 03 10:35:10 s/woukpld/would Mar 03 10:35:38 Or you modified kernel_android so you need to modify theirs? Mar 03 10:36:08 kernel_android is the one creating a loop Mar 03 10:36:36 it extends the kernel, but includes itself in boot.img which includes the iniramfs Mar 03 10:36:41 Tofe: So you're modifying it? Mar 03 10:36:47 Or.... ? Mar 03 10:37:12 I've created a new recipe, android-kernel-bootimg, which explicitely creates and deploys boot.img Mar 03 10:37:13 Just don't see why you'd touch a target we don't use? Mar 03 10:37:22 initramfs now includes udev Mar 03 10:37:35 that's common to all, a-priori Mar 03 10:38:03 Well Tenderloin might be exception. Mar 03 10:38:35 I'll need a bit more yes, but now including lvm won't be an issue at all Mar 03 10:38:40 it* Mar 03 10:39:17 Though keeping the current state, where it takes lvm from webOS, should be fine too Mar 03 10:39:30 I mean if we don't build i9300+crespo and those still inherit kernel_android that should be fine unless you changed kernel_android Mar 03 10:40:04 They'll have the dependency loop because of initramfs Mar 03 10:40:20 which impacts them too Mar 03 10:40:48 I guess I need to see what you propose to understand Mar 03 10:41:32 I'll do some PR'ing during the day, we'll be able to discuss based on that Mar 03 10:42:12 The change is pretty extensive, and I wouldn't want to break someone else's work who used meta-smartphone too Mar 03 10:42:17 uses* Mar 03 10:43:30 (though we are the only users of that layer currently, for android kernels, that I know of) Mar 03 10:44:52 Tofe: Well AsteroidOS too Mar 03 10:45:11 Ah yes right... Hmm I really have to be careful :) Mar 03 10:45:56 But Florent seems to be pretty OK with changes normally Mar 03 10:46:23 For this one it'll be worth discussing with him a bit, to be sure Mar 03 10:46:47 But I'll begin a WIP PR, and we'll see Mar 03 11:07:04 Tofe: https://github.com/FlorentRevest/meta-sm-t210-hybris/blob/c932624e1b5a74d7afc05c3d3d5ae7d42ba64dd8/recipes-kernel/linux/linux-sm-t210_git.bb and https://github.com/Openwide-Ingenierie/meta-boot2efl/blob/master/recipes-kernel/linux/linux-hammerhead_git.bb seems to be the only ones inheriting kernel_android Mar 03 11:07:42 And GranPC but he's from oru community I think Mar 03 11:08:29 Well I've made my initramfs modification more flexible, so if we don't add udev in the machine conf then there won't be any loop dependency Mar 03 11:08:54 (maybe it should be in the distro conf, now that I think of it) Mar 03 11:09:11 but you'll see, I'm beginning to commit Mar 03 11:20:11 Tofe: So the end goal here is that we can use Halium's Android repo and we don't need to drop their hybris/halium-boot anymore right? Mar 03 11:21:02 Not yet :) We will use their init script, but we still build the iniramfs and boot.img Mar 03 11:21:59 Take brings nearly all the simplifications of using Halium-boot Mar 03 11:22:49 i.e. we don't manage android partitions mounting, ramdisk unpacking... Mar 03 11:23:17 And we use directly the system.img created by the Halium android build Mar 03 11:24:49 For the moment I'm not yet ready to give up on our toolchain for boot.img; our way of doing thing is still pretty elegant, and well integrated Mar 03 11:35:47 Here you go! https://github.com/shr-distribution/meta-smartphone/compare/pyro...Tofee:pyro-halium?expand=1 and https://github.com/webOS-ports/meta-webos-ports/compare/pyro...Tofee:pyro-halium?expand=1 Mar 03 11:36:18 it's still WIP, mainly the system.img images are still to be built on jenkins :p Mar 03 11:37:12 I migrated also tuna; I know it's useless, but it really wasn't much work. Mar 03 11:47:44 Tofe: Yeah our toolchain is "better". Newer GCC v.s. ancient 4.8/4.9 from Google will give quite some performance improvements Mar 03 11:48:36 Tofe: Well for Tuna there might be a 3.4 kernel from unlegacy guys at some point that's workable Mar 03 11:50:26 Herrie|Pre3: for me, our toolchain is better simply because it's integrated; I don't want to rely on a "meta-toolchain" like jenkins to build a deployable image Mar 03 11:52:06 Tofe: Also agree there. The whole Android repo stuff feels wrong to me... Mar 03 11:52:16 Far from elegant Mar 03 11:53:30 I'm looking into a solution that could help us get rid of the patches to a large extend Mar 03 11:53:36 For the Android bits Mar 03 11:53:56 Not sure it'll work but have good hopes Mar 03 11:53:58 that would be welcome :) Mar 03 11:54:26 At least for all the .jar and .apk which should be 95% Mar 03 11:57:29 PR-ed stuff to Halium to make newer ports a bit easier. Ie 41/52 builds used qcom common and sepolicy but they were specified in device manifests Mar 03 11:57:41 So moved all to central one Mar 03 11:58:14 Same for other repos that were quite common and only 20kb Mar 03 11:59:03 In the whole repo sync mass download made little sense to me not to sync those centrally Mar 03 12:56:19 good :) Mar 03 13:21:25 I guess we could somehow take the Halium device manifests going forward Mar 03 13:21:30 I'll have a look into that Mar 03 13:21:44 Doesn't make much sense we maintain our own Mar 03 13:22:03 Also they have a by device sync and not a one massive one currently Mar 03 13:22:06 I'll have a look into that Mar 03 13:22:12 Will save some repo sync size for sure Mar 03 14:04:16 how do i sideload apps from the FTP i mirrored yesterday? Mar 03 14:50:12 vanyasem[m]: You can adb push them to /media/internal/ and use Preware to install them Mar 03 14:53:44 vanyasem[m]: Just keep in mind that there are 3 types of apps for legacy webOS: Mojo JS, not supported due to not being available open source Mar 03 14:53:58 EnyoJS, supported and I stress tested quite some and they worked mostly fine Mar 03 14:54:11 And PDK/SDK apps, which won't work due to missing bindings Mar 03 14:57:53 Unfortunately there's not really a way to tell them apart based on name Mar 03 15:08:45 Tofe: Changes look good in general! Mar 03 15:47:17 Aaawww... Now my boot.img images don't boot ! What did I change ! Mar 03 16:54:53 Tofe: Maybe typo somewhere? Mar 03 16:55:48 Could be, but the kernel is just crashing, it's weird Mar 03 17:11:26 how's it going guys? Mar 03 17:26:50 HaDAk: A little slow in terms of releases, but lots of stuff behind the screens. Halium is proving to be very useful! Mar 03 17:27:00 You have any device laying around that you can tinker with? Mar 03 19:03:46 elvispre: How's the OnePlus revival going? Mar 03 19:32:05 Herrie: Tantalisingly! I found a recovery method that flashes ColorOS onto it using a COM port USB driver method, but it's a "cut down" version that only puts the recovery image on. Mar 03 19:33:01 That just leaves me with a recovery tool that, like fastboot, can't flash anything because the phone is still locked. Mar 03 19:34:08 What I need is the "full version" that actually installs ColorOS so I can switch that into Developer Mode and unlock the device. Mar 03 19:35:26 I do have *a* full version, but it doesn't boot so I guess it for the OnePlus One or something. Mar 03 19:48:16 elvispre: Hmmmz Mar 03 19:48:26 I was having a hard time with my Redmi Note 4x too.... Mar 03 19:48:52 I needed to short some pins, but these were somehow changed between different ROMs... So most guides were useless :P Mar 03 19:50:00 Guessing which pins to short must have been fun! Mar 03 19:50:54 Well I found some other guide eventually and it was still fiddly but well Mar 03 19:51:09 This is an EUR 175 phone, so getting the back off was easy :) Mar 03 19:51:19 Just something between casing and screen and it just comes off :D Mar 03 19:51:25 No glue, nothing Mar 03 19:54:35 My main problem is the amount of "advice noise". Lots of people ask "My phone is locked and I have no OS. How can I unlock it?" Almost all of the answers explain how to access dev mode from the OS or tell you to fastboot oem unlock. That is, they ignore the actual question. Mar 03 19:56:34 You've seen? https://forums.oneplus.net/threads/guide-mega-unbrick-guide-for-a-hard-bricked-oneplus-x.417648/ Mar 03 19:56:45 I had to get mine in the qdloader9008 mode to reflash Mar 03 19:57:26 Yes. That's what got me out of my machine gun boot loop to where I am. Mar 03 20:01:07 I assume you saw this one too? https://forums.oneplus.net/threads/solution-recover-from-hard-bricked-oneplus-one.104943/ Mar 03 20:02:37 Yes. That's the one that flashes but does not boot :-P Mar 03 20:05:52 I can probably do something like get all / some of those components from a LineageOS build and construct my own Msm8974DownloadTool "kit", but I haven't tried that very hard yet. (My first quick attempt failed.) Mar 03 20:07:32 how does one make a preware feed? Mar 03 20:07:37 is there a documentation somewhere? Mar 03 20:22:38 vanyasem[m]: This should help I guess: https://forums.webosnation.com/canuck-coding/282255-how-easily-setup-your-own-ipkg-feed-dropbox.html Mar 03 20:35:43 also, how do i copy stuff from the web browser? **** ENDING LOGGING AT Sun Mar 04 03:00:01 2018