**** BEGIN LOGGING AT Thu May 06 02:59:56 2021 May 06 06:48:21 Morning! May 06 06:48:47 Tofe: Yeah I guess you're right, everything is downloaded in .repo and then created from there, so if we keep .repo it should be quick I guess. May 06 06:48:54 I just never did that at my end May 06 06:49:07 I don't think hybris-patches creates a lot of files to be honest May 06 06:52:19 Ah seems there are a few indeed: https://paste.ubuntu.com/p/68Z6yTyRJG/ May 06 06:58:58 pieguy128: Did you have any chance recently to toy with touchpad mainline kernel? May 06 10:07:50 I'm going to trigger new testing and unstable builds, so I'm going to revert the android-system-image changes for now May 06 10:24:28 JaMa: I don't know the "git clean" command, so maybe May 06 10:27:13 https://github.com/shr-distribution/meta-smartphone/commit/c8d6c65f52fdc69c6ed73ebd8f706c71fd8a3253 revert merged May 06 10:37:15 I've updated Herrie's jenkins-job PR: https://github.com/webOS-ports/jenkins-jobs/pull/20/files May 06 10:37:30 I think this should be enough once the manifest is updated to use our hybris-patches May 06 10:38:55 didn't we already merge that one May 06 10:40:35 JaMa: yes, already merged: https://github.com/webOS-ports/android/pull/9 May 06 10:40:51 https://github.com/webOS-ports/android/blob/luneos-halium-9.0/halium/default.xml#L25 don't we need to change this? May 06 10:41:01 ooh right ! May 06 10:41:24 https://github.com/webOS-ports/android/pull/9 could actually be reverted May 06 10:55:15 Well we'd need to point to our own fork of hybris-patches then at least May 06 10:55:25 Seems upstream is not really hurrying up with feedback on the PR's May 06 10:59:21 Ah wait I need more coffee it seems May 06 10:59:31 We should be good I guess May 06 11:14:01 should be good without updating the manifest or do we both need more coffee? May 06 11:14:34 I've triggered hammerhead build just to test that it will fail now May 06 11:15:24 but maybe it won't now when the hybris-boot isn't removed with https://github.com/webOS-ports/android/pull/9 May 06 11:19:06 JaMa: Well hybris-patches would fail then May 06 11:19:26 Since we use upstream hybris-patches which would try to patch non-existing hybris-boot May 06 11:19:38 At least that's how I understood the changes in jenkins-jobs.sh May 06 11:19:48 Or it will only print an error and still continue patching? May 06 11:20:59 yes, it should fail now and that will be better than silently ignoring apply-patches.sh issues May 06 11:21:22 then we should revert https://github.com/webOS-ports/android/pull/9 and change https://github.com/webOS-ports/android/blob/luneos-halium-9.0/halium/default.xml#L25 to use our hybris-patches without a patch for hybris-boot (already merged) May 06 11:22:21 Or leave as is May 06 11:22:32 If we use upstream, it will patch hybris-boot simply May 06 11:22:35 Which is fine May 06 14:32:26 good http://jenkins.nas-admin.org/job/LuneOS/view/halium/job/halium-luneos-9.0-build_hammerhead/3/consoleFull didn't fail and doesn't show any errors from apply-patches.sh May 06 14:33:30 will try halium-luneos-9.0-20210506-3-hammerhead with local build first before bumping them all in meta-smartphone again May 06 14:35:43 JaMa: I will test here as well, just I have my Hammerhead not nearby, so need to pick it up from my friend first May 06 15:41:27 hammerhead booted now with new halium image and hardknott May 06 15:42:08 but adb isn' May 06 15:42:12 t running May 06 15:43:43 heh, how to re-enable keyboard in QMLterm? May 06 15:44:31 nvm, see it now May 06 15:46:14 JaMa: no adb but you get UI and so on ? that's a bit strange May 06 15:46:23 correct May 06 15:46:26 often it's all-or-nothing May 06 15:46:43 wanted to check ps in QMLterm, but the keyb defeated me to type | :) May 06 15:48:08 we also ship fingerterm May 06 15:48:24 but QMLTerm isn't so nice to use, I remarked May 06 15:48:54 buttons are too big, UI gets cluttered, font isn't as effective as it could be May 06 15:49:40 yes, but managed to type it in the end :) shouldn't adbd be running? May 06 15:50:13 it should yes May 06 15:51:08 when started manually in fingerterm, it fails with "Failed to get adbd socker: No such file or directory" then Could not start mdnsd; Cloud not register mDNS service" May 06 15:51:36 and now it got stuck in half rotated screen :) May 06 15:52:28 will try to get usb-net working before continuing to debug in fingerterm, but first need to finish something for work May 06 15:52:55 I've triggered new hallium builds in parallel with new testing as it seems to help at least a bit May 06 15:53:18 next we should try new android image with testing to narrow adbd issue a bit May 06 15:59:51 mdnsd ? maybe hardknott introduce a new service by default, which is conflicting with our connman or something like that May 06 16:03:14 there were also some changes to adbd in meta-oe May 06 16:03:40 see https://github.com/openembedded/meta-openembedded/issues/147#issuecomment-832919328 May 06 16:03:51 so it's quite likely to be cased by hardknott May 06 16:04:10 and maybe we just need to set PREFERRED_PROVIDER_android-tools-conf = "android-tools-conf-configfs" May 06 16:12:27 there is still quite a few differences in the android image built on bonaire and uploaded from your local builds: https://paste.ubuntu.com/p/pxRMP7n4Wk/ May 06 16:15:30 or maybe not that many, somehow I got 2 versions of older 20210405 image and one of them is relatively close (at least with the list of included files: https://paste.ubuntu.com/p/J4N98Mdysf/ May 06 16:18:26 and the -b variant is the right one it seems, so there is few extra files: May 06 16:18:28 20210506-3-r0-new-rfs/res/images/charger/ May 06 16:18:29 20210506-3-r0-new-rfs/res/images/charger/battery_fail.png May 06 16:18:29 20210506-3-r0-new-rfs/res/images/charger/battery_scale.png May 06 16:18:51 $ cat 20210506-3-r0-new-rfs/system/etc/init/hybris_extras.rc May 06 16:18:52 on boot setprop ro.hybris.battery.capacity 2300 May 06 16:19:45 and missing 2 symlinks: May 06 16:19:46 lrwxrwxrwx. 1 martin martin 3 dub 5 14:31 20210405-1-r0-old-rfs-b/system/lib64/lib -> lib May 06 16:19:50 lrwxrwxrwx. 1 martin martin 3 dub 5 14:31 20210405-1-r0-old-rfs-b/system/lib/lib -> lib May 06 16:19:53 and May 06 16:19:54 -rw-r--r--. 1 martin martin 2,0M b?e 6 08:57 20210405-1-r0-old-rfs-b/system/etc/recovery-resource.dat May 06 16:20:07 symlinks is what Tofe added I think May 06 16:21:38 Yes: https://github.com/Tofee/android_device_lge_hammerhead/commit/5a7086365fcc14722938ff69ea1b803fd4efa7bf May 06 16:21:55 And they are in Halium's device manifest: https://github.com/Halium/halium-devices/blob/halium-9.0/manifests/lge_hammerhead.xml#L3 May 06 16:22:45 We should be using upstream halium-devices repo again as per https://github.com/webOS-ports/android/commit/060e72b3cdeeabfb5d892a1ef611f334046fa3c2#diff-d9b8e4a48f8e111ec5d21480d9d33a893b365dfa7f8550bbc0577e4d42afeac8 May 06 16:23:19 So the manifest (lge_hammerhead.xml) in halium-devices/devices on builder should be correct May 06 16:23:48 You might want to check the contents of lge_hammerhead.xml on the builder, just to be sure May 06 16:24:36 I don't think I have access to that at my end, or at least I don't know how ;) May 06 16:24:45 that's this symlink lrwxrwxrwx. 1 martin martin 3 dub 5 14:31 20210405-1-r0-old-rfs-b/system/lib64 -> lib May 06 16:25:08 I guess the other level was when the ln was re-executed in the same directory May 06 16:25:12 as the ln call is missing -n May 06 16:25:57 change that to "ln -snf" to prevent creating 20210405-1-r0-old-rfs-b/system/lib64/lib when re-executed May 06 16:26:17 Tofe: ^ May 06 16:26:44 but doesn't harm now when we cleanup between builds and we don't need them May 06 16:26:58 any idea what happened with recovery-resource.dat May 06 16:28:44 JaMa: ah, yes, I didn't anticipate that May 06 16:45:51 adbd runs in gatesgarth image, so it's most likely this meta-oe commit: May 06 16:46:01 d00a1f363d android-tools: split adbd to the separate package May 06 16:46:06 or the android-tools-conf changes May 06 16:47:28 hardknott doesn't have android-tools-adbd installed, but there was adbd from somewhere, strange May 06 16:47:56 https://paste.ubuntu.com/p/gvvy4R7mTk/ May 06 16:50:02 JaMa: I don't see adbd in hardknott in your paste May 06 16:50:50 correct May 06 16:51:07 that's why I'm adding android-tools-adbd to the image May 06 16:51:22 but I was starting some adbd in the fingerterm before May 06 16:51:31 so something was there May 06 16:57:35 sh-5.1# cat /etc/luneos-release May 06 16:57:35 Rockhopper release 1.0-jansa/hardknott-hardknott-jama (luneos-Eiskaffee) May 06 16:57:38 fixed May 06 16:59:00 /android/system/bin/adbd /usr/bin/adbd May 06 16:59:43 /android/system/bin/adbd is the one which was failing in fingerterm I think May 06 17:00:31 sh-5.1# /android/system/bin/adbd --version May 06 17:00:32 Android Debug Bridge Daemon version 1.0.40 May 06 17:00:32 sh-5.1# /usr/bin/adbd --version May 06 17:00:32 Segmentation fault May 06 17:00:45 but good enough for me :) May 06 17:24:43 oh, /android/system/bin was in PATH ? or you typed the full path ? May 06 17:25:16 so we need android-tools-adbd in our image RDEPENDS for hardknott only May 06 18:22:47 yes, that's what I did in https://github.com/webOS-ports/meta-webos-ports/commit/e6afec205107d8fd791f26e8fc3d7242c5799e63 May 06 18:24:19 /system/bin is in PATH in fingerterm May 06 18:24:36 sh-5.1# echo $PATH May 06 18:24:36 /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin May 06 18:24:47 but not in "adb shell" May 06 18:28:39 https://paste.ubuntu.com/p/8MB98YX4Jv/ not sure where fingerterm gets that May 06 18:29:23 ./profile:PATH=$PATH:/system/bin May 06 18:29:27 I guess here :) May 06 18:29:56 ./meta-android/recipes-core/base-files/base-files_3.0.14.bbappend: echo "PATH=\$PATH:/system/bin" >> ${D}${sysconfdir}/profile May 06 18:40:25 JaMa: fingerterm is mainly used for sailfishos, where they run android alongside the rest of the system, so it would make sense for them May 06 18:40:57 ah, no, I should have read further before writing a mistake :D May 06 18:45:37 * JaMa hands cup of coffee May 06 18:47:18 https://github.com/shr-distribution/meta-smartphone/commit/369b6a8bff2bb5c2c33300dd2e7fc419c2f037f1 May 06 18:50:29 JaMa, mmm...coffee May 06 22:08:06 meta-smartphone change to use new halium builds from today merged, will build new testing, then unstable images May 06 22:17:10 JaMa: thanks ! **** BEGIN LOGGING AT Thu May 06 23:07:26 2021 **** BEGIN LOGGING AT Thu May 06 23:39:43 2021 **** ENDING LOGGING AT Fri May 07 02:59:56 2021