**** BEGIN LOGGING AT Fri May 22 02:59:57 2020 May 22 05:45:14 Hi guys. Just 3 files left for commits. Currectly focussing on meta-xiaomi/recipes-kernel/linux/linux-xiaomi-rosy_git.bb May 22 05:45:31 would need your guidance May 22 05:45:57 since I am not familiar with git May 22 05:46:29 The file includes lie git://github.com/Tofee/android_kernel_xiaomi_msm8953.git;branch=los14.1 \ May 22 05:46:51 so definitely pro's guidance is must for me at this level May 22 05:47:19 *lie=line May 22 05:48:55 seems need to give local reference to my kernel image instead of github. Kindly support May 22 05:50:58 https://pastebin.com/gyDN8hZb These are contents of file from rosy which needs to be reconfigured to land May 22 05:59:46 bt40: morning May 22 06:00:22 You'd need to point it to the kernel that's in the Halium xiaomi_land.XML file May 22 06:00:50 So rename it to land, on top of the file replace rosy with land May 22 06:01:11 goog morning May 22 06:02:46 bt40: Give me a few minutes, will paste something May 22 06:04:03 thanks May 22 06:06:17 https://pastebin.com/LLtzzeMe May 22 06:06:39 BASically I pointed it to https://github.com/kskarthik/android_kernel_xiaomi_msm8937/commit/aad1958e0b7ffd00ca2af58ad3f25578d241f233 May 22 06:08:08 Halium in general also produces the kernel, but we do it in our build with newer GCC (9 at this time), instead of the one Google uses (4.7-4.9 depending on target device) which offers performance improvements etc and it's fully integrated in our build system. For 3.4 and 3.10 kernels this requires quite some patches to get things to work, but for 3.18 kernels it's mostly OK May 22 06:11:21 Thank you very much Herrie May 22 06:13:01 Now only two files left where i need promer mount points. meta-xiaomi/conf/machine/rosy.conf May 22 06:13:17 and meta-xiaomi/recipes-core/initrdscripts/initramfs-boot-android/rosy/machine.conf May 22 06:13:57 How to find those? Initially Tofe advised to keep that pending. Now only that part is left May 22 06:14:22 I have installed TWRP recovery in my device if that could help May 22 06:14:42 also installed SuperSu through TWRP recovery and rooted the phone May 22 06:17:04 sdcard_partition=mmcblk0p49 May 22 06:17:04 system_partition=mmcblk0p24 and boot partition......This one ANDROID_BOARD_BOOTIMAGE_PARTITION = "/dev/block/mmcblk0p21" .........these three address i need May 22 06:19:00 What is your output from "mount" when you're in recovery and start adb? May 22 06:20:03 please give me couple of mins. Will live connect and give result May 22 06:22:49 https://pastebin.com/ZbqwVAX5 this is the output of mount command from adb May 22 06:52:41 Ah that one doesn't give the mmc ones... May 22 06:52:50 What was the right command again... May 22 06:53:56 sudo adb shell "su -c 'cat /etc/fstab'" May 22 06:53:57 sush: cat: /etc/fstab: No such file or directory May 22 06:54:21 sudo adb shell "su -c 'df -h'" May 22 06:54:21 no such file or directory' May 22 06:55:22 sudo adb shell "su -c 'ls /dev/block'" May 22 06:55:22 This gives only mount points............not association. I can get names but not things like which one is boot May 22 06:55:54 cat /etc/mtab also not work May 22 06:56:17 Lsblk? May 22 06:57:40 sudo adb shell "su -c 'lsblk'" May 22 06:57:40 sush: lsblk: not found May 22 06:58:28 sush: cat: /proc/mtd: No such file or directory May 22 06:59:03 df or df -h? May 22 06:59:18 not work May 22 06:59:21 Sorry not behind laptop now so doing by heart May 22 06:59:32 Have some notes somewhere May 22 06:59:32 still appreciated man May 22 07:00:05 Found this: for my device(Sony Xperia Z2), in directory /dev/block/platform/msm_sdcc.1/by-name , there are documents named system, cache, etc.every file is linked to a partition. and I got this table: May 22 07:00:24 Will be at laptop shortly to check May 22 07:00:44 From https://stackoverflow.com/questions/20218852/how-to-identify-names-of-the-partitions/37662197#37662197 May 22 07:04:05 This one has some commands to try as well: https://www.cnblogs.com/shangdawei/p/4514128.html May 22 07:05:03 i have tried stack overflow on.................... /dev/block/platform/msm_sdcc.1/by-name this gives permissions error May 22 07:06:15 re run by su, got following result /dev/block/platform/msm_sdcc.1/by-name: No such file or directory May 22 07:08:44 sdcard & system are 049 and 024 as per https://github.com/Halium/hybris-boot/blob/9b328155366e489cd6fca88862884f69adf7a16c/fixup-mountpoints#1696 May 22 07:09:00 Just the boot I don't know... I would say it's safe to assume it's 21, but well... May 22 07:13:43 yeah May 22 07:17:02 so i think i should proceed with these most probable mount points. then will see later on. atleast build it. May 22 07:19:09 Yes May 22 07:19:24 I suspect 21 will be fine May 22 07:20:34 just backed up boot partition with twrp. will see if i get some info in logs May 22 07:22:00 good news May 22 07:22:37 boot image: /dev/block/mmcblk0p21 May 22 07:22:57 i got when i reflash super su in twrp logs May 22 07:23:04 Ah that's fine then May 22 07:23:55 ok now proceeding for next step. All commits done May 22 07:35:08 Halium folder grown over 25GB :) May 22 07:37:41 bt40: Yeah it will grow quite a bit May 22 07:37:55 QtWebEngine which uses Chromium uses quite a bit May 22 07:40:23 JaMa: I probably need to re-setup my builder. You suggest I stick with 18.04 or go to 20.04? I saw you added 20.04 as sanity tested distro May 22 07:40:29 What is the next step now May 22 07:41:26 You did the meta-webos-ports one too right? May 22 07:42:08 yes May 22 07:43:02 THen you're ready to build :) May 22 07:44:13 and how to build :) May 22 07:44:21 any command? May 22 07:49:35 When you're in your main directory (so where you have meta-webos-ports and meta-smartphone): Do ". setup-env" May 22 07:49:59 Then "$ MACHINE=land bb luneos-dev-package" May 22 07:55:18 .setup-env: command not found May 22 07:56:37 after putting space between dot and setup-env May 22 07:56:45 Altered environment for @luneos development May 22 08:04:13 Yes May 22 08:04:27 That's good May 22 08:04:36 THen "MACHINE=land bb luneos-dev-package" May 22 08:06:04 Then be patient ;) Might take a while LOL May 22 08:06:12 What machine are you using? May 22 08:36:52 core i3 sandybridge 10 year old :) May 22 08:36:55 4gb ram May 22 08:36:57 Hdd May 22 08:37:22 ERROR: Unable to connect to bitbake server, or start one (server startup failures would be in bitbake-cookerdaemon.log). May 22 08:38:28 Log https://pastebin.com/ND6cSU8Z May 22 08:39:43 internet working, tested. Still getting that error. May 22 08:39:45 Ah ok May 22 08:40:00 That's our build server that has issues... We have some Chinese DDOSing lately :( May 22 08:40:07 You can do the following: May 22 08:40:56 in conf/local.conf: Remove the # in front of PRSERV_HOST = "localhost:0" May 22 08:42:29 4GB RAM will be hard I'd say, but time will tell May 22 08:43:26 Morning! May 22 08:43:38 There are two local.conf files, should i modify both? one is in common folder, other in webos-ports May 22 08:43:55 THe one in webos-ports should be OK May 22 08:44:08 both folders are in webos-ports-env as parent May 22 08:46:02 webos-ports/conf/local.conf should be enough May 22 08:46:09 TOfe: Morning! May 22 08:46:12 yes, it's the one I modified too May 22 08:46:20 Still same issue here with apps.... May 22 08:46:22 Stuck on splash May 22 08:46:45 And /var/log/messages doesn't tell me much May 22 08:46:49 Herrie: one modification I forgot is that I rebased your branch on zeus, but that shouldn't impact this May 22 08:47:09 I'll bump SRCREV, build here, flash, and come back to you May 22 08:49:16 Tofe: Thnx May 22 08:49:47 ERROR: /home/simran/halium/build/webos-ports-env/webos-ports/meta-smartphone/meta-xiaomi/recipes-kernel/linux/linux-xiaomi-land_git.bb: This recipe does not have the LICENSE field set (linux-xiaomi-land) May 22 08:49:48 ERROR: Failed to parse recipe: /home/simran/halium/build/webos-ports-env/webos-ports/meta-smartphone/meta-xiaomi/recipes-kernel/linux/linux-xiaomi-land_git.bb May 22 08:50:37 bt40: can you paste your .bb content somewhere ? (bpaste or such) May 22 08:50:52 the LICENSE field should be inherited by the line "require recipes-kernel/linux/linux.inc" May 22 08:51:59 (if your copied your .bb from the rosy one) May 22 08:52:37 https://pastebin.com/XnWPnpkh May 22 08:52:43 that line seems missing May 22 08:52:50 bt40: Just add https://github.com/shr-distribution/meta-smartphone/blob/zeus/meta-xiaomi/recipes-kernel/linux/linux-xiaomi-rosy_git.bb#L1-L4 May 22 08:52:56 Those 1st 4 lines May 22 08:53:12 yep May 22 08:53:21 Tofe: I pointed him to your initial Rosy commit, but some thing changed since then May 22 08:53:41 ah, ok, yes, we refactored some things since then May 22 08:53:56 change done, restarting build May 22 08:55:11 Herrie: I would recommend 20.04, if you really need to rebuild older zeus revision then there are only 2-3 patches needed to be backported to build with 20.04 (or you can use docker with 18.04) May 22 08:56:04 JaMa: OK, will get to that later today then maybe May 22 08:56:28 I'm still on 18.04, in my chroot; after the ACG migration and the Qt validation, I'll upgrade it to 20.04 May 22 08:56:36 ERROR: Nothing PROVIDES 'luneos-dev-packag'. Close matches: May 22 08:56:37 luneos-dev-package May 22 08:56:37 luneos-package May 22 08:56:37 luneos-dev-image May 22 08:56:49 bt40: yep, typo :) May 22 08:57:14 Need to reclaim some 500GB from my 1TB and GParted Live doesn't like it and doesn't want to boot... So will just bite the bullet and do a fresh install ;) May 22 08:57:30 Luckily documented everything last time, so should be less of a hassle ;) May 22 08:58:09 Tofe, sorry. couldn't get......what should i do now May 22 08:58:27 i mean in which file May 22 08:58:28 bt40: add the "e" at the end of luneos-dev-packag May 22 08:58:36 it is not giving file name May 22 08:58:51 ok sorry got it May 22 08:59:03 sorry for irritating May 22 08:59:20 Herrie: it does work on tissot for me May 22 08:59:26 bt40: No worries, bitbake is quite a learning curve to start with... May 22 08:59:34 Tofe: Hmmz I have qemux86-64 here May 22 08:59:41 Though it shouldn't really differ May 22 08:59:50 Let me build for Mido here May 22 09:00:07 I agree... Can you rebase on zeus, just to really have the same thing? May 22 09:00:26 ah, damn, zeus changed in the meantime May 22 09:00:48 Tofe: Can you push your meta-wop to a new branch? May 22 09:00:51 Ill take that one then May 22 09:01:03 Hpe 16gb swap might turn magin for a tiny 4gb ram and process succeed somehow.... May 22 09:01:08 JaMa merged the Qt 5.15 upgrade I guess May 22 09:01:41 ok, just a minute May 22 09:01:42 bt40: In case your machine doens't like to build, I can run it here when you push your meta-smartphone & meta-webos-ports May 22 09:02:40 How should i thank you both guys....... Phenomenal support. May 22 09:03:04 Herrie, my dignity doesnot allow that. Already i owe you much May 22 09:03:08 :) May 22 09:03:18 Herrie: done in tofe/acg-wip May 22 09:04:13 bt40: just be prepared for quite some hours of build then :) Should be around 5-6 hours I guess, as it is from scratch. May 22 09:04:38 hmm May 22 09:05:22 I'm around 4-5 hours, with a i5 6600 skylake, 16GB of ram and a SSD May 22 09:05:38 your newer i3 should match my old i5 I guess May 22 09:06:22 Tofe, mine is i3-2100....secong gen. Around half as much your processing power..... May 22 09:06:26 :) May 22 09:06:39 will take 8-10 hrs May 22 09:06:56 ah, mmh, that could be a bit long yes. Your call :) May 22 09:07:14 Tofe: Let me build qemux86-64 first with your meta-wop May 22 09:08:19 bt40: a good 25% of the build time is just qtwebengine May 22 09:08:45 okay May 22 09:09:24 Fetching uninative binary shim from http://downloads.yoctoproject.org/releases/uninative/2.8/x86_64-nativesdk-libc.tar.xz;sha256sum=a09922172c3a439105e0ae6b943daad2d83505b17da0aba97961ff433b8c21ab May 22 09:09:24 May 22 09:09:40 currently stuck slow at this stage May 22 09:09:50 seems server side down/slowness May 22 09:10:13 mine internet is 50Mbps...., working fine on other devices May 22 09:10:42 fetching speed 1-2kbps :) May 22 09:11:01 Got 1Gbit/s fiber here ;) Download via WiFi gets to 400-500Mbit/s May 22 09:11:14 Herrie, netherland May 22 09:11:16 ? May 22 09:11:42 bt40: Yes new house so we got 1Gbit/1Gbit fibre right away :) May 22 09:12:03 Well it was 750/750 but they upgraded it to 1Gbit/1Gbit about 6 months ago May 22 09:12:04 * bt40 thinks european countries are awesome May 22 09:12:26 But even fast connection doesn't help with a slow server ;) May 22 09:12:34 yupp May 22 09:12:52 I'm on a 15Mb/1Mb dsl link, in France, so, well, it's not everywhere yet in europe :p May 22 09:13:12 and time travels much slower when you are excited for your first OS build :) May 22 09:13:28 Tofe: That's your own choice ;) May 22 09:13:38 You could get fiber if I recall :D May 22 09:13:54 hahaha May 22 09:14:06 Tofe: This image should work without any modifications right? May 22 09:14:39 Herrie: yes, but first there were strikes, then this virus, so I rely more on my home connection, I can't switch to fiber right now :) May 22 09:14:45 Herrie: right May 22 09:15:02 I just flashed it, booted, and started the mail app May 22 09:15:12 OK qemu is ready to be booted May 22 09:16:21 Tofe: Yes understandable... You cannot have them in parallel? Here I can have DSL & fiber at the same time (and even Coax for TV as well), so 3 sources... Of course you pay 3 times as well :P May 22 09:16:29 (which means the container file didn't help for this issue, btw) May 22 09:17:47 Herrie: I'm not that desperate :) May 22 09:18:03 Hmmz same issue May 22 09:18:07 Damn May 22 09:18:16 QML apps are fine, Enyo 1/2 stuck on splash May 22 09:18:26 I did see something in the logs previously which I didn't pay much attention to May 22 09:18:28 Let me pull it May 22 09:19:37 2020-05-22T09:17:25.642933Z [49.221147908] user.warning luna-next [] LS_REQUIRES_SECURITY {"SERVICE":"com.palm.applicationManager","CATEGORY":"/","METHOD":"focusApplication"} Service security groups don't allow method call. May 22 09:21:20 Building Mido now, just to be sure May 22 09:21:35 Herrie: I had that one too May 22 09:21:38 Note: Currently running 4 tasks....Seems 4 are selected based on processor cores May 22 09:22:01 bt40: you can tweak that in the conf/local.conf file May 22 09:22:49 4 tasks is the default, and there is also the number of parallel builds per task (it's 4 for me, I think I tweaked it a bit) May 22 09:22:55 Tofe, aren't already best? 4 is the max cores (2 physical, 2 virtual with hyperthreading) May 22 09:23:16 atleast for me 4 is best:) May 22 09:23:20 bt40: depends, some of the tasks just fetch the source code, so the cpu is free May 22 09:24:18 but as long as your cpu is always at 100% and you don't max out your ram, it's fine May 22 09:26:50 bt40: the BB_NUMBER_THREADS depends quite on your IO as well, unless you have fast NVME drives, then staying on 4 or 2 is probably better (especially after the very first build) May 22 09:26:58 lots of errors and warnings with failure///// May 22 09:27:00 https://pastebin.com/mKZPaScx May 22 09:27:03 bt40: how much ram+swap you have? May 22 09:27:56 JaMa, 4GB ddr3 ram, 16gb swap on 7200 rpm desktop hdd-sata, May 22 09:28:13 bt40: you can comment out SSTATE_MIRROR now when our fileserver isn't up2date because of disk array issues May 22 09:28:42 bt40: oh :/ that will be much more than 8 hours for you very likely May 22 09:28:52 okay thanks for guidance man....appreciated May 22 09:30:00 you should lower the BB_NUMBER_THREADS to just 2 May 22 09:30:19 it sucks when fetching stuff, but with that amount of ram you really don't want anything else building when qtwebengine is May 22 09:30:34 bt40: SSTATE_MIRROR is also in conf/local.conf, I also have it commented out May 22 09:30:55 processor is just sitting at 10 percent. tofes idea seems good. more process could have boosted speed May 22 09:31:17 see "individual builds" sheet in https://docs.google.com/spreadsheets/d/1qtpACcUO0T1bhHPZxDQFvtYEesNTTfxuvFYFdU8kyjI/edit#gid=1288574922 to know what to expect May 22 09:31:33 Tofe: mido is ready, flashing May 22 09:33:29 JaMa: is it possible to unfliter the sheet ? We only see threadrippers there :p May 22 09:33:52 JaMa, which column tells time? cells are not width changing as i dont have google account, text is hidden May 22 09:34:59 Tofe: not on this sheet (all tests are from my threadripper), but you can see similar information on top20.builds sheet May 22 09:35:02 bt40: you can download it as .xlsx or .ods May 22 09:35:48 longest qtwebengine was 5030s on my 1600AF with 16G ram, now we have contender to at least double that :) May 22 09:36:20 bt40: "e" as elapsed, it's using the /usr/bin/time formating string for header May 22 09:37:21 * bt40 wonders how these guys find awesome names like HAMMERHEAD, THREADRIPPER, May 22 09:37:27 JaMa: 5000s is the kind of time I get for qtwebengine too May 22 09:38:06 Tofe: if you select qtbase or qtdeclarative then there is also ryzen 1600AF and athlon 3000G, and for qtwebengine 1600AF (I forgot to time it on 3000G and really didn't want to repeat that test) :) May 22 09:38:41 on Athlon 3000G it was 3600 for qtbase.do_compile May 22 09:39:23 Qt is really the big piece May 22 09:39:27 and tb40 will unfortunately be much slower than that May 22 09:39:35 Currently 4 running tasks (0 of 0) 100% is 0 here percentage of tasks completed? How can i find percentage of total work completed any trick? May 22 09:40:30 bt40: I guess it's computing which tasks are to be done, there are about 6500 or so May 22 09:41:15 there is no percentage and number of task is quite misleading as well (because some tasks like qtwebengine.do_compile take many times more then other) May 22 09:41:24 hahaha......it means its like preparing to copy files when we move big files. May 22 09:41:27 or fetching the sstate from fileserver May 22 09:41:30 Unfortunately the progress will not be linear: it will be stuck a long time with a couple of tasks (qtbase.do_compile and qtwebengine.do_compile) May 22 09:43:24 bt40: sstate is the cache from a previous build; if you have a slow cpu but a good internet link it can be a lifesaver :) May 22 09:43:58 great May 22 09:44:05 the only trick is that the current sstate cache on our server has to match your version too May 22 09:44:06 bt40: and I'm sorry if it sounds mean to make fun from your slow pc, but it's really not suited for building this, me or herrie can build it for you or of course you can build it, just expect it will be quite bad experience and it might take whole day to finish, if you're fine with that, than OK May 22 09:44:59 bt40: and it's not just this first build, next time we do some change like small qt upgrade or this ACG migration is merged, it will be again many hours for your PC to rebuild with new changes May 22 09:45:44 JaMa, I am a mechanical engineer by profession. Computer is my hobby. Until u dont face hickups, try to understand errors and learn, it is a bit of unsatisfactory to an enthusiast May 22 09:46:20 so learned a lot of stuff while porting webos. will learn more using this process to May 22 09:46:45 bt40: I think we understand very well what you mean :) May 22 09:46:52 ok, as long as you understand what you're going into :) May 22 09:47:00 foor example, sstate concept told by tofe. new for me. I really enjoyed learning and knowing this thing May 22 09:47:47 I just mean, that your PC is probably better utilized by you learning/debugging/writting than being hammered to trying to build it :) May 22 09:47:51 Sure i know my machine is not suitable, stuck due to coronavirus pandemic. May 22 09:48:02 hahaha May 22 09:48:07 JaMa, hit the nail May 22 09:48:11 hahaha May 22 09:49:22 I've spent enough time working on overloaded PC running too many builds on background, so I just felt sad for you to be in this possition now :) May 22 09:49:28 bt40: still 0 out of 0 ? May 22 09:49:40 my pc is like 98 years old man with arthritis trying to run marathon. May 22 09:49:47 yes :) May 22 09:50:30 if you do "ps -axf" what is it doing behind the scene, a "wget" or so ? May 22 09:50:45 as long as it doesn't catch fire or wents very cold... :) May 22 09:51:03 bt40: or my favourite htop with tree view May 22 09:51:29 JaMa: building luneos is a very good test after changing the cpu fan, but it's also a bit stressing I must say May 22 09:51:39 ps axf frm another terminal? May 22 09:51:51 bt40: yes May 22 09:52:20 Tofe: I've at least +5 degree C in my room since the upgrade to threadripper May 22 09:53:12 it's really "cooking a recipe", with that cpu May 22 09:53:33 https://pastebin.com/RG2QuHE1 May 22 09:53:39 Tofe: and the radiator mounted on top of the case is really effective, so from 25 ambient temperature, it exhausts 35-40 air under my desk (not far from me :/) May 22 09:53:46 Tofe, that is the aoutput of commmand May 22 09:53:47 Tofe: I really need an AC now May 22 09:54:03 but I'm the recipe :) May 22 09:54:04 bt40: ok, it's fetching things May 22 09:54:30 bt40: maybe sstate ? but I thought you disabled the sstate mirrors ? May 22 09:54:53 yes, sstate and failing in many cases based on previous paste https://pastebin.com/mKZPaScx May 22 09:55:09 i didn't did anything to config regarding sstate May 22 09:55:32 Tofe: mido is OK here too May 22 09:55:41 you should comment out SSTATE_MIRROR in conf/local.conf May 22 09:55:50 bt40: then you should probably stop the build (double ctrl+C) and -- what JaMa said May 22 09:55:52 So weird qemux86-64 act up May 22 09:56:10 JaMa, +5 degrees in cold climate is good. In hot climate that might be unwanted May 22 09:56:24 it would be useful for you to fetch at least some prebuilt sstate from our fileserver, but right now it's not in sync, so it wouldn't reuse much (if anything at all) May 22 09:57:06 Herrie: at least I'm glad you can reproduce the working scenario May 22 09:57:14 bt40: yes, summer time in europe will be nice next to this heater :) May 22 09:57:34 Me too May 22 09:57:39 summer time in europe is equivalent to winters in india May 22 09:57:41 Herrie: can you upload your qemu built image ? I could have a look too May 22 09:57:52 ok stopping the built May 22 09:57:54 Tofe: It's on the regular fTP location ;) May 22 09:57:58 ok thanks May 22 09:58:02 but 99% of my builds are still done remotely in server room with AC luckily May 22 09:58:24 JaMa: Did you convince the powers higher up to invest in some new builders? May 22 09:59:06 Herrie: no, I've only consolidated my builds to just a handful of the more powerfull one and let them decommission the weaker ones May 22 10:00:04 Herrie: wendell might be able to benchmark 32c Epyc, if it's faster than 3970x thanks to 8 memory channels then it might be very nice replacement even for the "faster" remaining xeons May 22 10:00:44 JaMa: I'm always surprised at the penny wise, pound foolish mentality of big corporates ;) May 22 10:00:49 Herrie: the jobs I've triggered last Friday were still running yesterday (at least until one of the builders run out of disk space and I had to requeue the rest after cleanup) May 22 10:01:34 started at 3.30pm IST from scratch. Lets see how much time it takes May 22 10:02:39 yes, pity that our IT doesn't understand what test-oe-build-time is showing, because I understand that explaining that these "powerfull" blades are junk is a bit difficult :) May 22 10:02:39 Now all cpus at 100% load May 22 10:03:37 bt40: it'll start with building the build toolchain, and then get to the real stuff May 22 10:04:17 Setscene task (/home/simran/halium/build/webos-ports-env/webos-ports/meta-webos-ports/meta-luneos/recipes-webos/event-monitor-network/event-monitor-network.bb:do_package_setscene) failed with exit code '1' - real task will be run instead May 22 10:04:30 and on the other hand, if I get 5 dual epyc 7745 servers, then I wouldn't be able to even go through the build logs fast enough to keep them utilized :) May 22 10:04:43 bt40: you can ignore this one May 22 10:04:49 bt40: probably a consequence of the failed sstate fetch May 22 10:05:20 bt40: it means that it was trying to fetch sstate archive which seemd to be on fileserver (over http), but then it failed to fetch it and will need to actually build it locally instead May 22 10:05:36 Currently 4 running tasks (105 of 6507) 1% May 22 10:09:24 So glat that after 7 years I got this wonderful irc ommunity once again. Earlier best community was linux mint. Once i left using mint around 2012, all other communities are not welcoming. Ubuntu, and lot of others. May 22 10:09:50 This one is similar to mint in experience, so much helping guys May 22 10:10:30 * JaMa thinks we're desperate to get more developers :) May 22 10:11:07 * bt40 will surely devote his time to webos guys May 22 10:11:37 Specially these 3 guys whome i am porting. May 22 10:11:54 with whom* sorry for loose english May 22 10:49:03 10% in 50 mins approx 9-10h projected time May 22 10:49:42 swap playing nice role here: 700mb utilized from 0mb normal level. Ram 2.5gb used, 1gb free May 22 11:14:37 bt40: "Slow" PC here is the bottleneck really.... May 22 11:16:31 I had 2010 machine as well previously (for building) and a Win10 desktop, I replaced those 2 by very powerfull Threadripper 2990WX with 128GB RAM about 1,5 years ago. LuneOS now builds in about 50-55 minutes from scratch. JaMa got an even newer Threadripper 39xx which makes mine look "slow" ;) May 22 11:17:05 Anyway if you have a working port, we can always decide to add it to our builder for example ;) May 22 11:17:59 128GB ram.....wow May 22 11:18:18 bigger than ny OS partitioon 100gb hahaha May 22 11:20:04 bt40: Yeah it's a bit overkill but I expect this machine to last me 5-10 years at least so was worth the investment May 22 11:20:49 would have been costed you nearly a second hand car May 22 11:20:50 Saving energy by having 1 vs 2 PC's etc ;) May 22 11:21:02 saving energy.....hahaha May 22 11:21:15 bt40: Yes... I think it was somewhere between 4000 and 5000 EUR :S May 22 11:21:45 Also due to 1TB NVM SSD, 2* 10TB disks for storage etc May 22 11:21:54 This amount is my 1.5 year earning man May 22 11:22:00 hahaha May 22 11:43:05 bt40: Yeah I know, also here a big amount of money, but OK ;) Sold an appartment with good profit, so this was present for myself ;) May 22 12:21:43 OK builder installed with Kubuntu 20.04 May 22 12:22:13 Now need to configure kernel bits, ZNC but SSH, TeamViewer already working ;) May 22 13:26:15 can u guys share how webos look on your phones. Hope you use webos also alongside other OSes May 22 13:26:28 screenshots May 22 13:27:05 * Hex87 also has nick bt40 May 22 13:30:01 Hex87: https://www.youtube.com/watch?v=G9lbXE_wWpQ did that some months ago, about the progress of the port to pinephone May 22 13:30:21 ok May 22 13:33:36 I hope i'd be able to make calls and send text mesages atleast on 4gVoLTE carrier. Having no other secondary phone. May 22 13:45:33 Need to save system resources, just came from a minor freeze. Will join later on. Thank you Tofe and JaMa for guidance. May 22 13:47:04 * Hex87 's porting progress: 26% in 3h 45 mins....Slow and steady wins the race May 22 14:30:51 JaMa: Seems that zeus still requires python 2.7 on 20.04? Can I just install it or you have another suggestion? May 22 14:33:27 herrie: just install it, I do the same for my zeus based builds, let me check Dockerfile.. May 22 14:33:39 # To build with Yocto 3.0 Zeus you need few more: May 22 14:33:39 # ERROR: The following required tools (as specified by HOSTTOOLS) appear to be unavailable in PATH, please install them in order to proceed: May 22 14:33:42 # makeinfo python2 python2.7 May 22 14:33:45 RUN apt-get install -y texinfo python2 May 22 14:34:22 check also libxml2, you probbly have it in full OS, but 20.04 docker doesn't have it by default May 22 14:34:25 # Ubuntu 20.04 docker image doesn't have libxml2 installed and webruntime and libmusicbrainz fail because of that May 22 14:36:07 and python3-distutils which you probably already have and it should be all http://paste.ubuntu.com/p/xPJwd6tJZG/ May 22 14:40:35 OK let me have a look May 22 14:40:44 Installed the 2 bits now seems to continue May 22 14:40:49 I got a full install in general May 22 14:41:01 Hmmz ZNC with SASL seems to be acting up a bit :S May 22 14:41:07 Last time it worked like a charm :S May 22 14:41:13 Or maybe it's Freenode's server May 22 15:59:32 JaMa: is it actually worth it, to update my chroot ? it's not my main system, and very little of that chroot is actually used for building May 22 15:59:52 Plus, I see it still uses python2 May 22 16:21:15 Tofe: no at least as long as our jenkins is using 18.04 May 22 16:22:09 it was worth only because Herrie was already reinstalling because of other reasons and in desktop there are some other useful improvements, but for OE build there is no benefit yet May 22 16:22:57 ok May 22 16:23:14 Except for ZNC it works well it seems May 22 16:23:36 And need to tweak my Broadcom dkms again... May 22 16:24:05 Ubuntu guys are slow updating packages, were done for Debian a month ago already May 22 16:24:24 In order to get the 5.6 kernel working May 22 17:08:22 Herrie: WARNING: 13:07:52.991: Attempted flush to non-raster surface QQuickView(0xb33050) of type QSurface::OpenGLSurface <-- what's that ? could it be it's related ? May 22 17:11:36 Tofe: Yeah that one is suspicious May 22 17:11:49 I just built new qemux86-64 with latest Zeus (so without ACG) May 22 17:11:52 Let me see what it does there May 22 17:11:58 Could be somehow qt related ? May 22 17:13:33 https://codereview.qt-project.org/c/qt/qtbase/+/269676 do we have this in our eglfs code ? It should, but it looks the same May 22 17:16:20 we do have this commit. Then maybe our eglfs usage is incorrect somehow May 22 17:23:56 Liking the 20.04 so far, quite some things got easier to setup :) May 22 17:24:21 The issue is probably coming from https://github.com/webOS-ports/luna-webappmanager/blob/herrie/ls2-fix/src/webapplicationwindow.cpp#L214 somehow May 22 17:25:00 Just got this latest Zeus image with 5.15rc starting here May 22 17:25:04 Let me see what that one doesr May 22 17:25:28 Will try first as is, then maybe rebase the ACG bits May 22 17:26:46 Herrie: I have a few ideas to try May 22 17:27:44 Ideally it should work for both hwcomposer and eglfs, but let's fix qemu first May 22 17:27:45 It works OK on the 5,15rc May 22 17:27:58 oh ? great ! May 22 17:28:03 Though I do get a crash randomly when closing app May 22 17:28:05 Let me check log May 22 17:28:08 me too May 22 17:28:15 so it's not the Qt upgrade May 22 17:29:57 mmmh I can't reproduce the crash now... May 22 17:30:13 maybe it was just the first boot of luna-webappmanager May 22 17:32:25 https://paste.ubuntu.com/p/TFmgsWnBb2/ May 22 17:33:55 Line 4164 and onwards should give some clues May 22 17:35:38 oh, the compositor itself died May 22 17:38:59 looks like it's because of the CardGroupListView binding loop... but that one isn't obvious, I'll have to investigate. May 22 17:41:56 Seems some of the QML code needs some minor tweaking here and there in preparation of Qt 6 by the looks of it May 22 18:20:56 Tofe: I think we're pretty OK for the ACG bits in general, I will start to prepare some PR's for the obvious stuff May 22 18:21:42 Herrie: so with Qt 5.15, the mail app and such do start normally ? May 22 18:21:51 Yes May 22 18:22:01 great May 22 18:22:05 Well I didn't test ACG yet though May 22 18:22:17 It's unusual that a second migration helps the first one :) May 22 18:22:38 But current Testing Zeus which Qt 5.15rc1+ mail app works in Qemux86-64 May 22 18:23:03 ok, I was already fine with upgrading Qt, now I really want it :) May 22 18:25:00 I can try to test 5.15 with ACG though May 22 18:25:11 Let me put children to bed shortly so I can focus a bit LOL May 22 18:28:35 Well I've begun a first beer (kwack), so my focus will probably regress May 22 18:28:41 -c May 22 18:30:37 LOL ;) May 22 18:30:51 No alcohol in the house here atm except for vodka LOL May 22 18:30:56 But that's too strong for me ;) May 22 18:33:14 here it's just beer. Or does wine count as alcohol also May 22 18:33:41 (frenchs disagree on that last point) May 22 18:39:25 LOL, like Russians don't count beer & wine as alcohol LOL May 22 19:18:31 Tofe: OK building @ my end now, so ACH with QT 5.15RC1+ May 22 19:38:14 Also a qtwebengine rebuild here jay May 22 19:50:27 Herrie: Tofe: For QML issues with 5.15 we might need fix similar to this in more places https://github.com/webOS-ports/meta-webos-ports/commit/f01c102207081a59efc160e8ef6671ed97bcc294 https://github.com/webOS-ports/meta-webos-ports/commit/b8c53ccc0a312992e93e78353aa75d7af7a905fd https://github.com/webOS-ports/meta-webos-ports/commit/2029354ff07aa713059e2f3402613ca6ecf3ed6e this is what was originally May 22 19:50:33 blocking home screen with 5.15 and I haven't noticed it in other places (yet) May 22 19:52:53 JaMa: Ah OK May 22 19:52:59 Good to know May 22 19:53:08 Seems my 20.04 builder is working pretty OK May 22 19:53:29 Got it up and running fairly quickly this time, including everything I want it to do.. Was fighting a bit with ZNC but seems OK now too May 22 19:53:48 Got latest 5.6 kernel running as well, so seems all is good May 22 19:54:31 Built a zeus + qt 5.15rc1+ testing for 2 targets, now doing zeus + qt 5.15rc1+ + acg testing May 22 19:59:14 fast builder :) May 22 19:59:20 Yes :) May 22 19:59:58 Lots of small things improved in 20.04, so setup was easier for KDE Wallet, Teamviewer, krfb, printer etc :) May 22 20:00:04 So that helps, less fiddling ;) May 22 20:00:13 Notes I kept from last time helped a lot too ;) May 22 20:06:35 I don't live in command line like you ;) May 22 20:06:44 Still prefer GUI for quite some things ;) May 22 20:11:45 yeah, browser is the only GUI thing I use regularly :) May 22 20:12:50 and for printer/scanner I've used only app in android phone for last couple years, so I no longer even bother to set it on desktop (even when that app is really annoying) May 22 20:15:44 JaMa: For some reason my printer only works properly when I connect to the same mesh device :S May 22 20:15:57 Never really bothered to look into it with too much detail though May 22 20:33:16 Tofe: On 5.15rc1 + ACG email works in qemux86-64 May 22 20:38:44 Tofe: Seems mostly OK, can add email, Google account, contacts sync. Just Testr gives me white screen when starting May 22 20:38:49 Will look into that tomorrow May 22 20:41:23 May 22 22:36:17 qemux86-64 LunaWebAppManager[778]: CRITICAL: 22:36:17.009: Uncaught TypeError: Illegal constructor May 22 20:41:23 May 22 22:36:17 qemux86-64 LunaWebAppManager[778]: [778:778:0522/223617.010016:INFO:CONSOLE(72)] "Uncaught TypeError: Illegal constructor", source: file:///usr/palm/applications/org.webosports.app.testr/build/enyo.js (72) May 22 20:45:47 Looks easy to fix, the error is precise May 22 20:47:49 Probably something new Chrome doesn't like May 22 20:49:27 Tofe: Yup http://webos-ports.github.io/org.webosports.app.testr/debug.html May 22 20:49:35 My Chrome shows it there too May 22 20:54:58 Seems it doesn't like the Geolocation kind May 22 20:55:08 Seems it might have been defined elsewhere May 22 20:55:11 Already May 22 20:57:24 If I change it to Geolocation2 it's OK May 22 20:58:32 So easy fix indeed May 22 20:58:44 ok good :) May 22 21:01:55 Will push that tomorrow May 23 02:32:46 Yesterday porting ended unexpectedly. Not because of ram or other error, but due to disk shortagee. 110GB partition got full at just 26% compiling May 23 02:33:45 Will attempt today by migrating halium folder to 1 dedicated 160GB partition for porting only. Hope it could be enough. May 23 02:34:20 Just wandering whole compiling woud start again from scratch or will resume from 26% where it left.. :) **** ENDING LOGGING AT Sat May 23 03:00:00 2020