**** BEGIN LOGGING AT Tue Dec 21 02:59:56 2021 Dec 21 08:19:58 Morning Dec 21 08:28:46 morning Dec 21 09:04:44 JaMa: morning! Dec 21 09:49:15 Morning! Dec 21 09:49:24 Good morning! I got qtwebengine compiled, I think. Now, vim is refusing to do_configure: https://pastebin.ubuntu.com/p/BF6NNhmfDR/ Dec 21 09:50:15 HerrieTP: for power key, check if nyx.conf has the right path to /dev/input/eventX Dec 21 09:50:54 for rotation, the sensor backend might differ too, but we should be able to find what pmOS is using Dec 21 09:51:24 erikinkinenfi[m]: this is vim-native, most likely caused by your host distro, you're using some bleeding edge fedora, right? Dec 21 09:51:53 erikinkinenfi[m]: can you share the config.log file? Dec 21 09:53:52 it's Fedora iirc Dec 21 09:54:53 Tofe: How do I check again which eventX I need? There was some monitor right? It worked with Megi's RC2, but not with Mobian RC6 it seems Dec 21 09:55:16 just use "cat" and hit the keys :) Dec 21 09:55:50 there are also some evdev tools but it's not necessarily more useful Dec 21 09:55:56 /proc/bus/input/devices provides some hints Dec 21 09:56:29 JaMa: This was on Manjaro: https://pastebin.ubuntu.com/p/Y5Mdc8CzK3/ Dec 21 09:57:07 configure:15129: ./conftest Dec 21 09:57:07 ./conftest: symbol lookup error: /usr/lib/libdl.so.2: undefined symbol: _dl_vsym, version GLIBC_PRIVATE Dec 21 09:57:48 Herrie: using /dev/input/eventX, the X might vary depending on what hardware is found first. It's often quite stable for the same kernel, but using a more "formal" address should be preferred I guess. However these formal paths are quite long and hard to remember Dec 21 09:58:45 erikinkinenfi[m]: does bitbake says somewhere near beginning that it disables uninative due to glibc version? Dec 21 09:59:53 meta/classes/uninative.bbclass: raise RuntimeError("Your host glibc version (%s) is newer than that in uninative (%s). Disabling uninative so that sstate is not corrupted." % (glibcver, d.getVar("UNINATIVE_MAXGLIBCVERSION"))) Dec 21 10:02:56 strange that I haven't seen this on ubuntu 21.10 yet as it also has glibc 2.34 Dec 21 10:03:50 JaMa: it said that on fedora but not here on Manjaro. Dec 21 10:04:09 what glibc version is in your Manjaro? Dec 21 10:05:35 and are you building with hardknott or honister? Dec 21 10:06:03 Tofe/JaMa: These are my inputs: https://bpa.st/NJ2Q Dec 21 10:06:09 both should support 2.34 now, but I haven't built hardknott for a while Dec 21 10:06:34 Herrie: https://wiki.pine64.org/wiki/PinePhone_Pro_Software_Releases#LuneOS that was fast :) Dec 21 10:06:49 JaMa: glibc-2.33-5 and https://pastebin.ubuntu.com/p/9pN8PXQVGB/ Dec 21 10:06:55 Tofe: Yeah I just uploaded something ;) Dec 21 10:07:48 Since we don't have a build server yet anyway... So I decided to upload my local image to GitHub as pre-release ;) Dec 21 10:08:04 Herrie: /devices/platform/gpio-key-power/input/input4 looks good I think Dec 21 10:08:21 erikinkinenfi[m]: can you comment-out these 2 lines and try to rebuild vim-native? Dec 21 10:08:24 meta-luneos/conf/distro/include/luneos.inc:require conf/distro/include/yocto-uninative.inc Dec 21 10:08:27 meta-luneos/conf/distro/include/luneos.inc:INHERIT += "uninative" Dec 21 10:14:16 JaMa: it seems that it built vim-native now. Dec 21 10:16:25 Tofe: Yeah I guess so. Dec 21 10:16:31 Also vibrator I should be able to fix now Dec 21 10:21:47 Btw, did you know that libera.chat is bridged to Matrix? Dec 21 10:25:02 strange I've rebuilt vim-native on ubuntu 21.10 with uninative enabled, not sure why it fails on manjaro Dec 21 10:25:21 no symbols in 21.10's objdump -x /usr/lib/x86_64-linux-gnu/libdl.so.2 Dec 21 10:25:38 and no _dl_v*sym in tmp-glibc/sysroots-uninative/x86_64-linux/lib/.debug/libdl.so.2 Dec 21 10:31:20 Power key works with /dev/input/event4. With full path doesn't work Dec 21 10:34:09 At least I can now connect to the phone with terminal and SCP, so that's easier to debug and push quick fixes Dec 21 11:00:35 ok, good :) Dec 21 11:01:55 How would I go about identifying those sensors? Dec 21 11:04:29 good question... Dec 21 11:04:58 udevadm might give some hints, but I'm not sure how to look for it Dec 21 11:05:22 we're mainly looking for an accelerometer Dec 21 11:05:50 I didn't find any clue in pmaports for ppp, though they say they support it Dec 21 11:07:01 Tofe: yeah Dec 21 11:07:31 Rest doesn't work yet Dec 21 11:09:02 pmaports seems outdated a bit Dec 21 11:09:22 Seems Mobian and Manjaro are more active for the moment Dec 21 11:09:35 Hard to keep track of all projects and their progress Dec 21 12:14:51 99% built... Sound promising. Dec 21 12:15:00 *sounds Dec 21 12:18:15 Oh, it finished successfully now. How do I install it? Dec 21 12:23:49 You can just flash the .zip with adb sideload Dec 21 12:24:04 luneos-dev-package-sargo.zip or whatever it's called Dec 21 12:24:14 Or flash via TWRP or other recovery Dec 21 12:27:33 Herrie: ok, I'll try sideloading. Dec 21 12:35:03 I doubt it will work at first try, but who knows Dec 21 12:35:16 Mobian defconfig for PPP works fine on LuneOS ;) Dec 21 12:39:48 Herrie: it booted my old Manjaro installation. I backed it up and removed it from my phone and it doesn Dec 21 12:39:55 't boot anything Dec 21 12:40:24 Does it boot loop? Dec 21 12:40:43 Herrie: no, stays on vendor logo Dec 21 12:41:47 is LuneOS installed on system or userdata? Dec 21 12:48:25 Did you include the Halium Android bits? Dec 21 12:50:05 Herrie: Wdym by Halium Android bits? Dec 21 12:50:47 Well you need to include the Halium system.img in the build Dec 21 12:51:17 I,e,: https://github.com/erikinkinen/meta-smartphone/blob/hardknott/meta-google/recipes-core/android-system-image/android-system-image-sargo.bb Dec 21 12:52:38 Herrie: I am still not getting it. What should I do? Dec 21 12:52:59 Herrie: also, is LuneOS installed on system or userdata? Dec 21 12:54:13 Since we don't support GSI yet, each device needs it's own system.img (and vendor.img) from Halium for now. Dec 21 12:54:48 If I recall we install to userdata, but need to double check Dec 21 12:54:56 It's hidden somewhere in a package Dec 21 12:55:47 I do have a android-rootfs.img for Halium. Where should I put it. Dec 21 12:55:50 ? Dec 21 12:58:10 Ok, the luneos rootfs is under userdata. Dec 21 13:04:23 I'm surprised your build completed without the android-rootfs.img ;) Dec 21 13:04:49 Which means you might miss a RDEPENDS somewhere Dec 21 13:06:22 You need to add sargo here: https://github.com/erikinkinen/meta-webos-ports/blob/65265e8af8ce1c0b2796c85905ad2d9c1c0c23af/meta-luneos/recipes-core/packagegroups/packagegroup-luneos-extended.bb#L162 Dec 21 13:06:38 RDEPENDS:${PN}:append:sargo = " ${LIBHYBRIS_RDEPENDS} bluebinder" Dec 21 13:10:57 Herrie: How can I use a locally built android-rootfs.img? Dec 21 13:13:06 I'm checking, give me a minute Dec 21 13:13:25 Doing 5 things at the same time Dec 21 13:13:55 What we normally have is: halium-luneos-9.0-20201220-1-mido.tar.bz2 with system.img and vendor.img inside Dec 21 13:15:42 But I guess you could directly put the img file in meta-smartphone/meta-google/recipes-core/android-system-image/android-system-image-sargo and change the reference in: https://github.com/erikinkinen/meta-smartphone/blob/hardknott/meta-google/recipes-core/android-system-image/android-system-image-sargo.bb#L7 Dec 21 13:15:55 To "file://android-rootfs.img" Dec 21 13:16:24 Then run bitbake again, it will complain about the md5/sha256 sum and you can update them Dec 21 13:25:19 I'm wondering, have you guys made any progress yet on a proper website/front page rather than just a wiki on https://webos-ports.org? Dec 21 13:28:07 Not yet, should really look into that shortly, just real life gets into the way all the time Dec 21 13:28:37 We have it setup on GitHub, just didn't put anything yet Dec 21 13:29:08 I'm fighting a bit with alsa-ucm on PPP, not really sure what I'm doing wrong... Any thoughts on how to debug? Dec 21 13:29:18 Seems it somehow cannot find the files with the config Dec 21 13:29:38 I *THINK* I have stuff together in terms of the files and where they're installed, but I might be missing something somehow Dec 21 13:31:08 What alsa-lib version are you running? Dec 21 13:31:56 `strace alsaucm -c hw:0 set _verb HiFi 2>&1|grep -e "alsa/ucm.*ENOENT"` tells you what files it is looking for that it can't find Dec 21 13:32:58 These are the versions I updated to: https://github.com/openembedded/openembedded-core/tree/master/meta/recipes-multimedia/alsa Dec 21 13:35:52 And I have these config files: https://github.com/webOS-ports/meta-webos-ports/commit/8255ed16ea05880372f7ca30015d0bbc57ce19db Dec 21 13:40:48 I should get some more microSD's Dec 21 13:40:58 Having 1 between PP and PPP not very good Dec 21 13:51:49 Herrie: ok, no ui still but I got a mtp-device. My computer is also unable to open the mtp device. Dec 21 13:58:31 Well that's something at least Dec 21 13:59:20 That strace alsaucm doesn't give anything Dec 21 13:59:35 alsaucm listcards tells met hte list is empty Dec 21 14:03:47 Ah OK, seems I need to create another symbolic link Dec 21 14:04:19 Herrie: adb shell connects to luneos? Dec 21 14:04:27 It should Dec 21 14:04:35 It does! Dec 21 14:05:08 Well then you can just output "journalctl > /media/internal/mylog.txt" and use "adb pull /media/internal/mylog.txt" Dec 21 14:05:13 That's what I do usually Dec 21 14:05:54 Herrie: is this normal? https://pastebin.ubuntu.com/p/Dd5rBYyvtj/ Dec 21 14:06:12 Ehm no Dec 21 14:06:18 Don't think so Dec 21 14:07:11 https://pastebin.ubuntu.com/p/dRw5YcTJWv/ Dec 21 14:07:21 This explains the error... Dec 21 14:07:53 But why are they symlinks to each other? Dec 21 14:08:58 Normally, on other distros /vendor is not a symlink but a directory, if I remember correctly. Dec 21 14:09:21 s/directory/mountpoint/ Dec 21 14:10:59 What do you have in /android/vendor? Dec 21 14:12:14 I guess it assumes a vendor.img is present somehow: https://github.com/shr-distribution/meta-smartphone/blob/master/meta-android/recipes-core/android-system-image/android-system-image.inc Dec 21 14:13:30 empty dir. Dec 21 14:17:16 Tofe: ^ Dec 21 14:17:25 But getting this far isn't bad for a first try :) Dec 21 14:17:39 I could build a vendor.img for you just in case Dec 21 14:18:46 Doing that now Dec 21 14:19:33 Herrie: pushed my vendor.img after doing simg2img to it. Then, I changed the symlink and rebooted. Vendor seems to be mounted correclty now. Dec 21 14:20:41 OK nice Dec 21 14:21:24 Then your log should be a bit better Dec 21 14:21:44 Just post a full journalctl and I might have some thoughts Dec 21 14:22:51 https://pastebin.ubuntu.com/p/wt2kTmCCwV/ Dec 21 14:24:22 Seems you're missing a config file simply ? Dec 21 14:24:58 Herrie: which one? Dec 21 14:25:02 https://github.com/erikinkinen/meta-webos-ports/tree/hardknott/meta-luneui/recipes-luneui/luna-next/luna-next-conf Dec 21 14:27:48 Needs to be in https://github.com/erikinkinen/meta-webos-ports/blob/hardknott/meta-luneui/recipes-luneui/luna-next/luna-next-conf.bb#L13 Dec 21 14:30:17 THen "systemctl restart luna-next" or reboot Dec 21 14:32:26 Herrie: that start-up sound was loud and scary. Dec 21 14:33:32 bt works, wifi doesn't Dec 21 14:34:08 sensor and brightness wor Dec 21 14:34:19 audio as well Dec 21 14:44:34 Wifi needs building as module and separate loading usually Dec 21 14:49:03 I'm outside now but can have a look later together with you Dec 21 14:49:33 There are some defconfig options + script Dec 21 14:50:07 CONFIG_WLAN=m Dec 21 14:51:50 And this: https://github.com/erikinkinen/meta-smartphone/blob/hardknott/meta-xiaomi/conf/machine/mido.conf#L35 Dec 21 14:52:42 and https://github.com/erikinkinen/meta-smartphone/blob/hardknott/meta-xiaomi/recipes-core/systemd/systemd-machine-units/wifi-module-load.service Dec 21 14:52:49 HerrieTP_: The driver is loaded correctly but /vendor/firmware_mnt/ isn't mounted properly Dec 21 14:53:29 Halium on sargo doesn't use any modules actually Dec 21 14:53:43 Even WiFi is inbuilt Dec 21 14:59:26 I, can you read me? my box' fiber might be disconnected at home Dec 21 14:59:30 HerrieTP_: what part should mount /vendor/firmware_mnt/ normally? Dec 21 14:59:41 ok, looks ike I'm still there, good Dec 21 14:59:56 Tofe: I can see your messages. Dec 21 15:00:10 erikinkinenfi[m]: thanks, perfect Dec 21 15:01:16 erikinkinenfi[m]: iirc, for firmware_mnt it's done by our lxc config at startup Dec 21 15:01:25 and there is a piece I think in the cmdline Dec 21 15:01:37 can you give me your complete kernel cmdline? Dec 21 15:03:06 Tofe: https://pastebin.ubuntu.com/p/BMfNRffMFR/ Dec 21 15:04:16 ok you have firmware_class.path=/vendor/firmware and maybe that's fine and I thought of something else Dec 21 15:07:05 on tissot we have firmware_class.path=/vendor/firmware_mnt/image Dec 21 15:07:53 but, well, it could be that /vendor/firmware should be mounted by the initramfs, depending on your android's fstab Dec 21 15:08:12 what is your halium's vendor repo? Dec 21 15:09:47 Ok, sargo has bt firmware on /vendor/firmware and wifi firmware on /vendor/firmware_mnt Dec 21 15:10:09 Every Halium distro for sargo uses stock vendor. Dec 21 15:14:26 https://pastebin.ubuntu.com/p/hmzjrx9P3B/ Dec 21 15:14:34 Tofe: it is in vendor's fstab Dec 21 15:15:39 mmmh but not mounted... could be paste me your dmesg ? mainly the part after Halium/LuneOS starts Dec 21 15:15:51 (and before systemd starts) Dec 21 15:16:41 I'm thinking, could it be the fstab hasn't been "cleaned up" and the selinux options are still here, preventing the mount to work correctly Dec 21 15:17:38 ah but that's your fstab *on the device*...ok... yes then it's probably the issue here Dec 21 15:17:54 I'm surprised that the other distros don't have any issue with mounts Dec 21 15:19:10 Tofe: wdym? Dec 21 15:20:08 on luneos, usually, we don't support the android selinux options in fstab, like "context=u:object_r:vendor_firmware_file:s0" Dec 21 15:21:22 and it's also the case for most halium distro, so when we build an Halium image, the Halium built script will remove these options from the fstab in the source code, and generate a new vendor.img without it Dec 21 15:22:33 you should see a complaint in the dmesg about Halium trying to mount this folder, without success Dec 21 15:23:54 If I understood you correctly, you took this vendor.img directly from Google? Dec 21 15:24:25 Yes, like does every other Halium distro on sargo Dec 21 15:26:24 ah, ok, maybe I'm missing something then Dec 21 15:33:05 Well Erik manually mounted it Dec 21 15:33:09 Maybe that's the issue somehow Dec 21 15:33:45 Herrie: should be done by Halium script in initramfs, usually Dec 21 15:34:00 Herrie: manually mounted what? Dec 21 15:34:07 The vendor.img, no ? Dec 21 15:34:12 no, Dec 21 15:34:32 I said I pushed it because I was missing one. Dec 21 15:34:49 erikinkinenfi[m]: what halium script does droidian use? in case it's patched to support these selinux options Dec 21 15:35:19 I took a look at other distros. They do mounting vendor fstab as systemd service. Dec 21 15:35:34 At least droidian and manjaro Dec 21 15:36:07 ah, ok, interesting Dec 21 15:36:41 though I'm not sure how they do that :p Dec 21 15:36:41 https://pastebin.ubuntu.com/p/9YBhSC6c4q/ I did test changing the lxc pre start to this and I have wifi! Dec 21 15:37:56 Tofe: take a look at the script I sent. Dec 21 15:38:13 erikinkinenfi[m]: yes, I see it uses parse_mount_flags so remove the spurious mount options Dec 21 15:38:18 quite nice Dec 21 15:38:34 it's the one from droidian ? Dec 21 15:39:01 it looks like an improved version of the one we currently use Dec 21 15:39:37 No, that is a one I wrote based on the droidian where it is located in separate script run as a systemd service before lxc@android Dec 21 15:39:55 ah ok Dec 21 15:40:42 well, still, I'm tempted to propose that in a PR for our lxc prestart script Dec 21 15:41:01 should check it a bit more just in case, but looks good Dec 21 15:42:07 Tofe: it also fixed lte Dec 21 15:43:20 :) nice Dec 21 15:46:01 https://github.com/webOS-ports/lxc-android/blob/tofe/fixes/var/lib/lxc/android/pre-start.sh ok, it's saved somewhere, will check when back home Dec 21 15:54:59 Tofe: should power key turn off the screen? Dec 21 15:55:39 it should, but could be you need to tweak the value Dec 21 15:55:58 Herrie: the value? Dec 21 15:56:01 I was having the same issue on PPP Dec 21 15:56:51 Also, the screen turns off automatically after a while but doesn't turn on? Dec 21 15:57:11 nyx.conf: https://github.com/webOS-ports/meta-webos-ports/blob/hardknott/meta-luneos/recipes-webos/nyx-conf/nyx-conf/s2/nyx.conf Dec 21 15:57:36 You can do "cat /dev/input/event0" event1, event2, event3 and event4 Dec 21 15:57:47 See which event one will show output when you press the power key Dec 21 15:57:51 Add add it to the nyx.conf Dec 21 15:57:59 File should be in /etc/nyx.conf on device Dec 21 15:58:11 Just fixed this on PPP today, so is fresh in memory Dec 21 15:58:21 The screen off is a bug in our state machine somewhere we need to tackle I guess Dec 21 15:58:25 I see that as well Dec 21 15:58:34 Just like brightness sometimes acting up Dec 21 16:06:07 Is it also a known bug that SMS don't simply work? And calls have no audio routing? Dec 21 16:07:28 Setting the event device in nyx.conf worked1 Dec 21 16:10:17 Some of these things are device specific, hard to say if they're bugs Dec 21 16:10:30 We recently upgraded to Halium 9, could be we need to tweak a few things still Dec 21 16:19:03 Herrie: so, sms work on other devices? Dec 21 16:22:44 To be honest I haven't tried it for a while Dec 21 16:22:49 When I did last time it did Dec 21 16:23:06 But we've gone through quite some rework and upgrades, could be we broke something along the way Dec 21 16:23:34 Do outgoing calls work in general? I mean does the call go through? That would indicate that RIL works at least Dec 21 16:24:05 Yeah, calls work without audio. Dec 21 16:24:30 OK so RIL should be fine, then SMS should work too normally Dec 21 16:24:35 and audio works in other places but not in calls. Dec 21 16:26:09 Herrie: service org.webosports.service.messaging does not exist. Dec 21 16:28:58 Hmm it should let me check Dec 21 16:29:50 It should be in /usr/palm/services/org.webosport.service.messaging Dec 21 16:30:58 What's the exact error you're getting in logs? Dec 21 16:32:43 Herrie: What log? systemd? Dec 21 16:33:15 journalctl Dec 21 16:33:25 journalctl | grep messaging would do Dec 21 16:37:49 Herrie: adb shell stopped working Dec 21 16:38:33 Ah I had that happening sometimes too Dec 21 16:39:04 audio is always a mystery when it goes through the PulseAudio black box anyway ;) Dec 21 16:41:15 Herrie: Service is not listed in service files. Dec 21 16:41:42 That is from journalctl run on fingerterm Dec 21 16:41:49 Can you paste a full journalctl again? Dec 21 16:42:28 Herrie: I couldn't make adb work even after rebooting. Dec 21 16:43:37 Sometimes adb on PC dies ;) Dec 21 16:43:41 adb kill-server Dec 21 16:43:46 And then try again Dec 21 16:44:05 adb: no devices/emulators found Dec 21 16:45:15 mtp is also gone Dec 21 16:45:29 Well you might be able to SSH in via WiFi ;) Dec 21 16:45:33 Or SCP Dec 21 16:45:38 That works on my PPP Dec 21 16:48:43 https://pastebin.ubuntu.com/p/sHFdpXJgZW/ Dec 21 16:48:51 Herrie: ssh worked Dec 21 16:49:54 OK seems something related to LS2 security I guess Dec 21 16:50:08 I see the following related bits: Dec 21 18:34:33 sargo activitymanager[1862]: [] [pmlog] DB8 MOJ_SERVICE_WARNING {"sender":"com.palm.telephony","method":"complete","payload":{"activityName":"telephony-send-outgoing-sms","restart":true},"error":"No such file or directory","reqErr":2} Dec 21 16:50:27 Dec 21 18:34:33 sargo ls-hubd[1281]: [] [pmlog] ls-hubd LSHUB_NOT_LSTED {"SERVICE_NAME":"org.webosports.service.messaging","EXE":"/usr/sbin/activitymanager","APP_ID":"(null)","PID":1862} Service not listed in service files (cmdline: /usr/sbin/activitymanager -c {"log":{"appender":{"type":"syslog"},"levels":{"default":"notice"}}}) Dec 21 16:50:48 I'll have a look into it in a bit Dec 21 16:51:21 webOS/LuneOS uses something called activitymanager to schedule things etc. And it uses it's own LS2 (Luna Service 2) bus with security features Dec 21 16:51:46 These security features weren't there before (just public/private) and we needed to migrate the whole OS without any documentation basically ;) Dec 21 16:51:59 Managed to do most, but could be something was forgotten somewhere Dec 21 16:52:33 I think I know why calld don't have audio. Dec 21 16:54:04 Herrie: pulseaudio is missing a module. Dec 21 16:55:03 It is not even installed Dec 21 16:56:34 Which module is missing? Dec 21 16:56:49 module-droid-hidl Dec 21 16:57:09 https://github.com/mer-hybris/pulseaudio-modules-droid-hidl Dec 21 16:57:38 There isn't even a recipe for that? Dec 21 16:58:46 That could be, we'd need to write one then :) Dec 21 17:02:56 That's something for tonight then Dec 21 17:12:34 * erikinkinenfi[m] uploaded an image: (62KiB) < https://libera.ems.host/_matrix/media/r0/download/erikinkinen.fi/vxvMjOtDcDyQBKPKDeuekVsV/68db4c58-1425-4541-9e3b-55818c5220148380819822028370619_thumbnail_1200.jpg > Dec 21 17:13:28 How can I adjust the scaling? Some things are quite small. Dec 21 17:22:45 Yes you can, you need to change some value again let me find the link Dec 21 17:24:17 These values: https://github.com/webOS-ports/meta-webos-ports/blob/hardknott/meta-luneos/recipes-webos/luna-sysmgr-conf/luna-sysmgr-conf/sagit/luna-platform.conf#L21-L22 and https://github.com/webOS-ports/meta-webos-ports/blob/hardknott/meta-luneos/recipes-webos/luna-sysmgr-conf/luna-sysmgr-conf/sagit/luna-platform.conf#L33-L36 Dec 21 17:35:55 Tofe: Found some regression in the PP with the uBoot changes, will push a fix Dec 21 17:36:15 Also I managed to break sound on PP, I guess same issue as on the PPP with alsa-ucm Dec 21 17:52:49 Herrie: changing those values worked well. Dec 21 17:53:17 These are fine tuning usual Dec 21 17:53:37 Not critical to get them to work Dec 21 17:53:50 The UI gives me nostalgia even though I haven't used WebOS ever. Dec 21 17:54:34 All multitasking in iOS and Android is poorly borrowed from here Dec 21 17:55:19 That's so true Dec 21 17:55:57 Does the UI support any kind of theming? Dec 21 17:56:38 Theming not for now, it was there in legacy but this UI was rewritten from scratch Dec 21 17:56:55 Though it's probably not rocket science to add it Dec 21 17:57:24 Was not high on priority list as you can imagine Dec 21 17:59:14 Yeah, I can imagine it wasn't but hopefully at some point. I like the UI but it would be better with a more modern theme. Dec 21 18:09:47 Does incoming SMS work? That should be separate compared to outgoing one Dec 21 18:10:25 HerrieTP: No, incoming sms didn't work Dec 21 18:12:27 OK need to do some more end-2-end testing at my end Dec 21 18:14:41 Is there a way to install a single package without reinstalling the whole system? Dec 21 18:28:04 Yup simply put it in /media/internakl Dec 21 18:28:11 media/internal that is Dec 21 18:28:12 The .ipk Dec 21 18:28:30 Then opkg install --force-reinstall --force-downgrade mypackage.ipk Dec 21 18:29:24 That's via command line, there are ways via UI as well, but I haven't used those in ages to be honest, so not sure to what extend they work Dec 21 18:33:47 OK let's look at those recipes to be added Dec 21 19:13:57 OK first issue: | /home/herrie/LuneOS/honister/webos-ports/tmp-glibc/hosttools/ld: cannot find -lgbinder Dec 21 19:14:14 While I do have it in my DEPENDS: https://bpa.st/A52A Dec 21 19:14:29 Same error for -lglibutil Dec 21 19:18:35 Herrie: where should this file get installed? https://github.com/webOS-ports/org.webosports.messaging/blob/master/files/sysbus/org.webosports.service.messaging.role.json Dec 21 19:22:30 I found a bunch of similar files for other services but messaging isn't there? Dec 21 19:22:57 In /usr/share/luna-service2/roles.d I think Dec 21 19:23:30 This was one of the migration things... We had to specify this for each and every service... Quite a PIA to do without proper documentation and tools ;) Dec 21 19:23:51 So could be it slipped through somehow Dec 21 19:24:34 ls: /usr/share/luna-service2/roles.d/org.webosports.service.messaging.role.json: No such file or directory Dec 21 19:26:19 Yeah seems it's not there, so that's not helping Dec 21 19:31:21 I think I see the problem, let me try something Dec 21 19:45:38 OK I seem to know where it comes from, will address it in a bit Dec 21 20:03:56 OK fixed it: Try this: https://github.com/webOS-ports/meta-webos-ports/commit/ede6131e962fb7ebcd2d472bad0617ad9b10ebed Dec 21 20:04:28 That should pick them up and put them in the output IPK file in tmp-glibs/deploy/ipk/all Dec 21 20:04:41 This might solve it, or at least partially Dec 21 20:11:34 For sound we are still using https://github.com/webOS-ports/meta-webos-ports/blob/hardknott/meta-luneos/recipes-multimedia/pulseaudio/pulseaudio-modules-droid_git.bb Dec 21 20:14:34 I know we should have migrated to something else, but so far it worked very well Dec 21 20:14:48 Tofe: Yeah but that doesn't work for everything on 9.0 it seems ;) Dec 21 20:17:05 We'll need a recipe for https://github.com/mer-hybris/audiosystem-passthrough too Dec 21 20:18:27 Tofe: Yeah I'm trying to write that Dec 21 20:18:32 But getting the above errors Dec 21 20:18:44 My recipe: https://bpa.st/A52A Dec 21 20:19:11 But LD complaining about ld: cannot find -lgbinder and -lglibutil Dec 21 20:21:15 Herrie: maybe pkgconfig didn't find them? Dec 21 20:22:16 Herrie: check your inherit pkgconfig Dec 21 20:22:42 and take inspiration from https://github.com/webOS-ports/meta-webos-ports/blob/hardknott/meta-luneos/recipes-support/libgbinder/libgbinder.bb , because the Makefile they use has the same structure Dec 21 20:22:47 and thus the same issues too Dec 21 20:28:02 Well even when I add inherit pkgconfig it still doesn't work, same error Dec 21 20:31:41 Herrie: SMS work! Dec 21 20:32:27 Restarted builder, because bitbake was acting up Dec 21 20:32:59 Let me try again Dec 21 20:35:20 Same error Dec 21 21:11:19 erikinkinenfi[m]: Nice! Sometimes it are these small things ;) Dec 21 21:12:33 HP sold webOS to LG, who then decided to not contribute anything to open source for 5 years and then did a code dump without any history in 2018 or so... So we basically had to go through all the components, code and migrate... A real headache as you can imagine Dec 21 21:28:25 Herrie: do wired headphones work for you? Dec 21 21:30:13 I haven't tried those in forever to be honest Dec 21 21:30:25 Pretty sure they won't work on my PPP since I'm still fighting alsa there Dec 21 21:30:51 I don't have much else at hand at the moment I could play with I think Dec 21 21:31:53 Herrie: the Halium pa module package was too old for Halium 9 to have working headphones. I think I fixed it. Dec 21 21:32:35 Ah OK Dec 21 21:32:48 Yeah sometimes a bump does the trick ;) Dec 21 21:32:59 But often the Mer guys break stuff too Dec 21 21:33:53 I think hardly anything we got from them worked out of the box without patches ;) Dec 21 21:33:55 Well, I actually had to change the upsteam to be another Halium distro. Mer does thing significantly different here. Dec 21 21:34:15 OK Mobian? Dec 21 21:34:19 Or ,,, ? Dec 21 21:34:28 We're a bit in between with some things Dec 21 21:34:37 We can re-use a lot of Mer though usually ;) Dec 21 21:34:46 Droidian but Ut would have worked fine probably. Dec 21 21:34:48 https://pastebin.ubuntu.com/p/BYqnMrh5M6/ Dec 21 21:35:05 I had to add pulse user to input group as well... Dec 21 21:35:44 And, I added this to webos-system.pa: load-module module-switch-on-port-available Dec 21 21:36:23 I think we had some nasty patches previously for headphone detection Dec 21 21:36:25 Let me check Dec 21 21:37:09 Yep, the Droidian sources have those applied and some more patches as well like the evdev headset detection. Dec 21 21:38:34 Well we had this patch too in our PA-droid: https://github.com/webOS-ports/meta-webos-ports/blob/hardknott/meta-luneos/recipes-multimedia/pulseaudio/pulseaudio-modules-droid/0001-Add-support-for-detected-external-connection-changes.patch Dec 21 21:38:52 That is for older Halium versions. Dec 21 21:39:28 It works via udev and newer android versions do not provide that way to check for headset. Dec 21 21:39:39 We need the evdev check. Dec 21 21:40:00 https://github.com/droidian/pulseaudio-modules-droid/commit/5cab2af137a87fd017567d06a9ec2e802593d90c Dec 21 21:41:20 Ah OK that makes sense Dec 21 21:41:54 Good you right away test some of these things, some get missed at our end when updating. It's usually 2-3 people doing everything and as you can imagine we forget something sometimes Dec 21 21:44:00 The fix for Messaging probably also fixes IM issues I had earlier :) Dec 21 21:44:06 With Skype etc not arriving Dec 21 21:49:54 Yeah, I like testing things I haven't tried before. So, I've enjoyed the process of getting LuneOS running on my sargo. Dec 21 21:53:54 Same here, not a great programmer, but good at bit fiddling, trial and error and bug hunting/fixing as long as it doesn't touch c(++) I can survive LOL Dec 21 21:54:33 Learning new things along the way. PPP was also a whole new thing for me, first time with serial console this way **** ENDING LOGGING AT Wed Dec 22 02:59:56 2021