**** BEGIN LOGGING AT Mon May 18 02:59:57 2020 May 18 08:44:51 Morning May 18 09:11:57 o/ May 18 09:23:16 Tofe: Seems my luna-appmanager ACG now at least boots... Still some minor things in logs, but well... And still the other issues (no background, no webapps, but progress) May 18 09:56:13 well, I guess luna-appmanager is far from obvious May 18 09:56:46 Tofe: Well it's the most tricky one probably in the end May 18 09:57:04 Since most of it in OSE is now done by SAM, which is a stripped down version of what we use it seems May 18 09:57:09 Quite some API calls are no longer there May 18 11:20:11 Tofe: The things I saw was in common for all the bits having issues with LS permissions is that they seem to have an empty SERVICE identifier: May 18 11:17:42 qemux86-64 user.warn LunaAppManager: [] [pmlog] LS_REQUIRES_SECURITY {"SERVICE":"(null)","CATEGORY":"/","METHOD":"registerApplication"} Service security groups don't allow method call. May 18 11:30:00 "SERVICE":"(null)" ... is that expected ? May 18 11:37:38 Tofe: Not sure, but that seems to cause the problems May 18 12:00:54 Hi greats May 18 12:01:04 meta-luneos/recipes-webos/nyx-modules/nyx-modules/rosy.cmake May 18 12:01:58 contents of this file in build directory are different than provided on github reference ;link for commits for same rosy devices. Which one are latest? May 18 12:03:03 I am adding these files to port OS to my xiaomi device codenamed land snapdragon 430. T have copied rosy folder and changed its name to land. May 18 12:03:54 I am now confused whether should i keep default build file from my build directory or replace it with online one? May 18 12:04:23 https://github.com/webOS-ports/meta-webos-ports/commit/2e3da5ceb0f10b02ffc8195d610dfc8deea7fe92?diff=unified This is online commit link May 18 12:05:11 second last file when viewed in combined commits mode. May 18 12:06:47 contents in my computer are these: https://pastebin.com/9kFFAxUT May 18 12:14:34 Bt40 use the contents from Zeus branch May 18 12:15:02 Some things were migrated to webOS OSE version of nyx-modules May 18 12:15:18 Which differs from what we had from Open webOS a bit May 18 12:16:44 Herrie, Thank you for supporting as always May 18 12:16:52 zeus mean online one? May 18 12:17:06 no keyword found in local file May 18 12:17:42 Zeus branch of meta-webos-ports repo May 18 12:17:52 The file variant from there May 18 12:18:20 any commands or referal procedure for that? May 18 12:18:35 how to connect with that branch and fetch May 18 12:21:58 ok got it. sorry for inconvonvenience May 18 12:28:03 set(NYXMOD_OW_TOUCHPANEL FALSE) Does it needs to be false? Will touch work if false? May 18 12:28:12 Need to put my little one to sleep, will be back in 10-15 and send you the link May 18 12:28:41 set(NYXMOD_OW_HAPTICS FALSE) will haptics work with false? May 18 12:28:59 ok sure May 18 12:41:42 bt40: Basically put this content in it: https://raw.githubusercontent.com/webOS-ports/meta-webos-ports/zeus/meta-luneos/recipes-webos/nyx-modules/nyx-modules/rosy.cmake May 18 12:42:14 NYXMOD_OW_TOUCHPANEL_MTDEV = touchpanel we use May 18 12:42:53 The ones mentioned under: # provided by nyx-modules-hybris are provided by libhybris via Android drivers May 18 12:44:02 nyx = HP's (webOS former owner) and LG's own way of accessing hardware components in a uniform way. Since we're using Android devices these things (haptics, led are done via libhybris) May 18 12:46:50 thanks for useful info May 18 12:51:39 I have replaced PREFERRED_PROVIDER_virtual/kernel ?= "linux-xiaomi-rosy' to land. and VIRTUAL-RUNTIME_android-system-image = "android-system-image-land" from rosy. May 18 12:51:48 Hope i am going right May 18 12:52:24 meta-xiaomi/conf/machine/land.conf This is file in which i made changes May 18 12:53:33 I am also concerned for these two things; Should i change? ANDROID_BOARD_SYSTEMIMAGE_PARTITION_SIZE = "3221225472" May 18 12:53:33 ANDROID_BOARD_FLASH_BLOCK_SIZE = "131072" May 18 12:53:45 My storage is 16gb internal May 18 12:54:07 IMAGE_ROOTFS_EXTRA_SPACE = "200000" May 18 12:54:42 So i wish shortage of storage does not give error during flashing and turn my device to brick. May 18 12:55:35 bt40: Those values should normally be OK, let me double check May 18 12:57:03 FLASH_BLOCK_SIZE = standard as per https://github.com/SRIDHARDV/android_device_xiaomi_land/blob/halium-7.1/board/filesystem.mk#L8 May 18 12:57:25 ANDROID_BOARD_SYSTEMIMAGE_PARTITION_SIZE is fine too May 18 12:57:40 What you might need to change in land.conf is the parition names May 18 12:59:35 THe ANDROID_BOARD_BOOTIMAGE_PARTITION May 18 13:00:47 and value should be? May 18 13:01:10 That's a good question ;) May 18 13:04:04 https://github.com/webOS-ports/meta-webos-ports/commit/2e3da5ceb0f10b02ffc8195d610dfc8deea7fe92?diff=unified I made no changes to these files. Just name change from rosy to land. Hope I do not miss anything.After changes required in bootimage partition name, not it has made me think again.:) May 18 13:04:26 *not=now May 18 13:06:47 Found a beautiful guide by Tofe . https://gist.github.com/Tofee/f3a65d5ea391da20817d3746f0031357 May 18 13:07:11 * bt40 thanks Tofe for this online guide May 18 13:08:45 As per tofe porting guide: ANDROID_BOARD_FLASH_BLOCK_SIZE. This can be found in the file "BoardConfig.mk" of the CM12.1 Android build (see step 1) May 18 13:09:11 this could be the clue which i need to scan to find name also May 18 13:13:50 bt40: well this is just the raw notes I took when porting rosy, I hope it's helpful :) It's meant to be merged into the main porting guide, but I never quite got the time to do that May 18 13:14:29 If you have complementary notes don't hesitate to share it too, we'll merge it too May 18 13:17:22 Tofe, can u please guide where we can find name of ANDROID_BOARD_BOOTIMAGE_PARTITION May 18 13:18:03 i downloaded lineage os, no file found as per guide. seems guide refers to build file and i have downloaded final image. thats why difference May 18 13:18:12 do i need to download source code? May 18 13:41:33 bt40: you have already the source code, it's Halium :) May 18 13:41:56 Some of the partitions are here: https://github.com/Halium/hybris-boot/blob/master/fixup-mountpoints#L1694 May 18 13:42:16 Just seems boot is not listed there May 18 13:42:29 you can find BoardConfig.mk in devices/xiaomi/land May 18 13:43:58 another way to have the boot partition is to list the partition table, from an adb shell, with "fdisk -l /dev/block/mmcblk0" May 18 13:45:19 Tofe: Boardconfig points to /board/ folder with various .mk files in there ;) May 18 13:46:54 ah, yes... weirdly enough, they've put most of the config in https://github.com/HyperTeam/android_device_xiaomi_msm8937-common/blob/lineage-15.1/BoardConfigCommon.mk May 18 13:48:00 Found a file https://github.com/AQSD/android_device_xiaomi_land/blob/cm-14.1/BoardConfig.mk May 18 13:48:12 more detailed May 18 13:48:57 looks correct yes May 18 13:49:45 but not found partition name May 18 13:49:58 seems need to access adb stuff via usb cable May 18 13:56:53 Getting the boot partition name correct is critical May 18 13:57:48 yes, I did erase the bootloader by mistake once, well, let's say it's a good thing there's a backup bootloader partition... May 18 14:08:32 -include vendor/xiaomi/land/BoardConfigVendor.mk where can i find this? many boardconfig.mk files point to this May 18 14:08:50 searching file name results bank May 18 14:16:32 that one should be in vendor/xiaomi/land in your Halium download May 18 14:17:57 this one is blank right now May 18 14:18:16 that's possible yes May 18 14:22:16 The mtdblock devices are the phone's internal flash storage. mmcblk0 is the phone's SD card. as per stack overflow.. May 18 14:22:42 so right now config file is pointing to mmc which is wrong. definitely need to alter that May 18 14:23:34 https://github.com/TeamWin/android_device_xiaomi_land/blob/android-9.0/recovery/root/etc/recovery.fstab tadaa May 18 14:24:15 ah wait no, there isn't the mmcblk path May 18 14:26:41 yes this is what i meant to say. May 18 14:27:04 mmcblk0 will be wrong mount May 18 14:27:59 u genius guys now can tell what should i write there referring to this May 18 14:29:53 bt40: mmcblk0 is definitely the internal storage, as per https://github.com/Halium/hybris-boot/blob/master/fixup-mountpoints#L1694 May 18 14:30:10 sd card would be something like mmcblk1 May 18 14:30:23 ok May 18 14:32:27 shall i proceed withsame rosy file address? May 18 14:32:44 o u guys suggest something else May 18 14:34:18 I think you can leave the mmcblk thing apart right now, and move on with the rest, and come back to this later May 18 14:37:34 ok May 18 14:38:05 thank you guys for todays session and support. Herrie Tofe May 18 14:51:50 adb shell fdisk -l /dev/block/mmcblk0 May 18 14:52:35 system/bin/sh: fdisk: not found May 18 14:54:54 adb shell cat proc/partitions command gives following mount points https://pastebin.com/5WB4vncf May 18 14:57:34 bt40 that one doesn't have the names May 18 14:57:48 I think mount might give you that as well May 18 14:58:51 adb shell ls -al /dev/block/platform/msm_sdcc.1/by-name May 18 14:59:00 this one i will try May 18 14:59:14 currently facing permission error, fixing that May 18 15:57:48 Herrie: doesn't give names, but gives sizes ! and we have BOARD_BOOTIMAGE_PARTITION_SIZE := 67108864 which means 65536kb, it gives use either mmcblk0p21, mmcblk0p22 or mmcblk0p30. It could be either boot or recovery, or a boot backup... not enough :) May 18 16:09:24 Herrie: if I want to help with the latest things, what do I take ? your latest ACG PR + I bump appmanager ? May 18 16:12:06 Tofe: Well PR doesn't really seems to work now :S May 18 16:12:17 And I updated a few things locally, though nothing too much May 18 16:12:21 luna-appmanager mainly May 18 16:12:25 So you can start from there May 18 16:12:46 Or I can push my meta-webos-ports to a new branch for now, might be easier May 18 16:13:03 Gimme a few mins ;) May 18 16:13:42 as you wish; though you probably know ACG much better than me now May 18 16:16:24 Tofe: Every time i think I understand how it works, I get new surprises LOL May 18 16:17:02 :p May 18 16:20:54 Tofe: https://github.com/webOS-ports/meta-webos-ports/commit/eb4b0777f78d0aff9aeed87525be4779bf63179f May 18 16:21:00 That's my current build May 18 16:21:06 ok ! May 18 16:21:27 We should be almost there though May 18 16:24:53 At least going by the logs May 18 16:25:01 Once we get webapps running we might see more issues of course May 18 16:25:25 And probably some more cleanup of legacy bits May 18 16:34:21 why do you have stuff like meta-luneos/recipes-qt/qmltermwidget/qmltermwidget.bb ? May 18 16:35:58 anyway, as long as qtwebengine isn't rebuilt, I'm ok May 18 16:36:05 Tofe: I think it won't May 18 16:37:00 Didn't you add QMLTermWidget as per https://github.com/webOS-ports/meta-webos-ports/commit/d5d150114aa0d644cd90418aecc175c8c2a12116#diff-7ef30bd82f781846329eb888bb28edce ? May 18 16:37:03 :P May 18 16:40:08 well yes, but I just don't understand why it appears in your commit May 18 16:40:41 anyway, nevermind May 18 16:41:17 Tofe: Well also not sure... I transferred from my Linux machine to Windows machine, could be due to line endings May 18 16:41:27 For most I thought I removed them May 18 16:41:33 But could be I missed this one May 18 16:42:18 I was just surprised I had so many recipes to rebuild, when only a few changes were done compared to the previous ACG changeset May 18 16:43:33 JaMa is not sitting still ;) May 18 16:43:57 https://github.com/webOS-ports/meta-webos-ports/commit/512d7f26aa24eb3a5cb5c5825670e1b3243f8707 May 18 16:44:07 I think I cloned mine on Friday before he made the changes ;) May 18 16:44:26 So it picked those up as well now when I took the new zeus from today and applied my changes on top May 18 16:49:42 If I were you, I'd rebase the wip branch on zop of your zeus-acg branch, to have a clear view of what changes are applied for appmanager May 18 16:50:14 But I'm not you, so you do the way you prefer May 18 16:51:17 Tofe: Yeah well I need to PR to Zeus again at some point ;) May 18 16:51:22 Differences are minimal anyway May 18 16:52:37 Anyway I added a few things that were missing, but nothing major May 18 16:52:41 Mainly some srcrev stuff May 18 16:52:51 rebuild finished, let's roll May 18 16:53:58 Yeah it's quick in general May 18 16:54:06 WHen there's no QT rebuild involved May 18 17:02:01 most of the remaining errors are like LunaAppManager [] LS_REQUIRES_SECURITY {"SERVICE":"(null)","CATEGORY":"/","METHOD":"registerApplication"} Service security groups don't allow method call. May 18 17:02:10 with AppManager, SysService and WebAppManager May 18 17:02:40 Tofe: In the logs I see getPreferences and getSystemTime that have issues. It seems both are called from https://github.com/webOS-ports/luna-systemui/blob/webOS-ports/master/data/System-Service.js#L56 May 18 17:03:14 you mean LunaSysService: [] [pmlog] LS_REQUIRES_SECURITY {"SERVICE":"(null)","CATEGORY":"/time","METHOD":"getSystemTime"} Service security groups don't allow method call. May 18 17:03:54 Yes May 18 17:04:28 in that error, is LunaSysService the caller ? the callee ? May 18 17:05:34 looks like it's the callee May 18 17:05:39 Yes May 18 17:06:20 Damn, couldn't ls2 include the caller in its error message ? "(null)" isn't very useful May 18 17:06:35 Well it could be it's not being passed somehow May 18 17:06:39 We had that before, remember? May 18 17:06:45 With some services were it was empty May 18 17:06:46 yes May 18 17:07:33 But I see in journal that luna-appmanager opens com.palm.systemui: May 18 15:04:33 qemux86-64 LunaAppManager[412]: SYSMGR PERF: APP LAUNCHED app id: com.palm.systemui, start time: 207117 May 18 17:09:51 too bad I can't reproduce the issue with luna-send May 18 17:09:54 So I guess it's the call there (and maybe in other .js files somehow that don't get an id May 18 17:11:04 I mean, if it's an app that calls the API, not having a service name isn't that surprising May 18 17:13:46 Tofe: True, maybe LS2 should have some more output in logs May 18 17:14:00 let me hack that a bit May 18 17:14:14 You can anyway do a ls-monitor -f com.palm.systemservice and in another terminal do a systemctl restart luna-appmanager May 18 17:14:19 That should trigger the calls May 18 17:15:02 And luna-webappmanager as well May 18 17:15:06 Those trigger the calls at my end May 18 17:15:33 ok May 18 17:15:38 https://paste.ubuntu.com/p/P6z4chx4Ck/ May 18 17:15:44 last 4 are from luna-webappmanager May 18 17:16:33 So seems luna-webappmanager and luna-appmanager are not sending any ID's. Might be a similar solution required as for the QML? May 18 17:16:38 We had similar there if I recall May 18 17:16:51 could be May 18 17:20:24 oh well ... LunaSysService [] LS_REQUIRES_SECURITY {"SENDER":"(null)","RECEIVER":"com.webos.service.systemservice","SERVICE":"(null)","CATEGORY":"/time","METHOD":"getSystemTime"} Service security groups don't allow method call. May 18 17:23:19 Yes that's what I'm getting May 18 17:23:33 This one is a bit more specific v.s. getPreferences, so therefore easier to find it's source May 18 17:24:42 I've added the source :p May 18 17:52:45 Curious what you did May 18 18:14:02 Just this: https://paste.ubuntu.com/p/FTZTVkMGCt/ May 18 19:24:47 Ah I Somehow thought you did more already May 18 19:24:59 It's just the logging now I understand May 18 20:19:28 Tofe: Actually it seems OSE guys added back these bits after they first removed it (see history): https://github.com/webosose/luna-service2-security-conf/commit/e807cd9f527c17f801ae7498dce72dda3250cd53 May 18 20:19:38 And it seems our calls are in there as well, so might be for a reason ;) **** ENDING LOGGING AT Tue May 19 02:59:57 2020