**** BEGIN LOGGING AT Tue Feb 15 02:59:57 2022 Feb 15 03:31:11 emulator build finally finished. it built a vmdk but not an .ovf. did I do something wrong? Feb 15 04:50:23 Did you build the appliance? Feb 15 04:50:38 There should be vmdk and ovf inside archive Feb 15 06:00:11 It should be called luneos-dev-emulator-qemux86-64-0-0.tar.gz Feb 15 07:48:41 Morning@! Feb 15 07:49:40 Tofe: Seems we might need to fix the :webos in webruntime.inc, wam,bb as well. I hacked around it in wam,bb by the looks of it by forcing the correct inherit. Feb 15 07:52:53 Morning! Feb 15 07:54:30 Herrie: yes, we'd need that... I don't know how widely this is used though Feb 15 07:57:06 Tofe: A quick grep in meta-webosose only shows matches in webruntime.inc and webos_qmake6.bbclass though Feb 15 07:57:14 A grep for ":webos" that is Feb 15 07:57:55 Not sure how it ended up in the LSM recipe then Feb 15 07:58:56 Ah it seems there might be more in JaMa's Honister one Feb 15 07:59:14 Ah it's because the _ to : conversion ;) Feb 15 07:59:17 https://github.com/webosose/meta-webosose/blob/master/meta-webos/recipes-webos/luna-surfacemanager/luna-surfacemanager.bb#L68 Feb 15 07:59:22 ah yes, you found out too Feb 15 08:00:41 if you can redo the grep I'm interested in the stat Feb 15 08:02:52 Seems we're good in general... It's in some other bits such as systemd, pulseaudio, com,palm.service.devmode and the packagegroup-webos-extended Feb 15 08:02:59 And wpa-supplicant Feb 15 08:03:16 ok, so not very widely used then Feb 15 08:03:40 For what we use now: webruntime.inc, webos_qmake6.bbclass, luna-surfacemanager.bb and wam.bb Feb 15 08:03:46 So yes not widely used Feb 15 08:04:19 We don't have a ":luneos", do we ? Feb 15 08:05:15 Or we could make so that ":webos" is also active on luneos, but I don't know how Feb 15 08:10:46 Tofe: No we don't I'm also not really sure how ot enable :webos Feb 15 08:10:58 It could be DISTRO = "webos" in webos.conf Feb 15 08:11:00 But not sure Feb 15 08:11:53 I'm sure JaMa knows about this Feb 15 08:12:03 JaMa: ^ Feb 15 12:00:00 no, don't change the DISTRO, webos can be added to DISTROOVERRIDES, but I would rather change the imported recipes to set PACKAGECONFIG without the override Feb 15 12:01:03 in OSE it's used just to change something only for OSE and not for other DISTROs which are based on OSE (and even there it doesn't make much sense as it could be overriden from "higher" DISTROs) Feb 15 12:11:16 ok, let's juste remove the override in our recipes then Feb 15 12:12:31 right Feb 15 13:38:22 Herrie: the emulator worked great. installed the museum from sdk, picked an app from the museum, installed from Preware. Feb 15 13:38:41 so the new appinstalld problems are specific to Tenderloin Feb 15 13:47:25 I'm a bit lost as to why it's specific to a device Feb 15 13:48:51 maybe building wrong version .. :) Feb 15 14:08:02 build per the wiki for all three devices! Feb 15 14:10:30 that doesn't prove anything when make update doesn't work correctly for you Feb 15 14:12:21 why would that matter if i pulled fresh bits? Feb 15 14:12:39 because you don't know if they are fresh Feb 15 14:13:01 but what do I know taking care of OE builds for last 15+ years Feb 15 14:13:47 how could they not be fresh bits? all the folders and the cache were wiped out. there were no bits left. i started from scratch. where would old bits have come from? Feb 15 14:15:11 i started from: wget https://raw.github.com/webOS-ports/webos-ports-setup/testing/Makefile Feb 15 14:15:20 what should I be doing different? Feb 15 14:16:40 codepoet: there are some pitfalls with webos-ports-setup and the "make update": for example, if some changes were done in meta-webos-ports, "make update" will not necessarily get you to the lastest commit of the testing branch Feb 15 14:17:59 but if you used the same recipes to build all 3 builds, then for me your builds should be ok; I'll try to test on my TP at home too Feb 15 14:19:18 yes, followed the steps from the wiki to build hammerhead, tenderloin and emulator. i don't know enough to try anything different :-P Feb 15 14:19:39 codepoet: maybe you can share your build with me somehow ? Building a TP image will take me a whole day :p Feb 15 14:20:32 yes! i can Feb 15 14:24:50 uploading now Feb 15 14:31:06 https://1drv.ms/u/s!Av5IQUxnr8DUj5k-REyx0mdeh98x0Q?e=mIN5Px Feb 15 14:32:10 codepoet: thanks, will download at home in a few hours Feb 15 17:06:15 codepoet: Really weird issue, good emulator is working for you as well. It's my main go-to development target usually Feb 15 17:06:24 Easy and quick to debug things on a laptop Feb 15 17:06:35 No cables needed etc, just lacking the touch Feb 15 17:07:44 def good for debugging! daily use on tenderloin is my goal, though! Feb 15 17:20:57 It's probably the android side that's somehow bugged Feb 15 17:21:03 But hard to say for now Feb 15 17:21:22 I noticed it's less stable compared to other targets in general Feb 15 17:31:16 codepoet: ok, it's downloaded, now flashing Feb 15 17:32:32 fingers crossed its just something i did wrong! Feb 15 17:38:00 yes I hope it'll be easy to explain (and fix) Feb 15 17:39:21 codepoet: could you remind me where I can find the latest ipk for museum ? Feb 15 17:40:17 last time I fixed the remaining issues manually, but I guess there's a new version now Feb 15 17:42:08 yup, hold on Feb 15 17:42:32 http://stacks.webosarchive.com/luneos/com.palm.app-museum2_2.9.0_beta_all.ipk Feb 15 17:42:39 perfect thanks Feb 15 17:43:21 or if you want to build from source: https://github.com/codepoet80/webos-catalog-frontend Feb 15 17:43:38 Not sure I even know how to do that, to be honest :D Feb 15 17:43:59 (which is a shame, of course) Feb 15 17:44:29 https://github.com/webOS-ports/webos-ports-sdk Feb 15 17:44:35 run install-linux.sh Feb 15 17:44:50 cd to the folder the source is in, then: lune-run . Feb 15 17:44:59 or: lune-run /path/to/folder Feb 15 17:45:08 ok, I thought it would be harder Feb 15 17:45:41 anyway, I'm now booting your image, let's see how it goes Feb 15 17:46:40 ah, I forgot I never disabled systemd-udev-settle on luneos Feb 15 17:46:51 so the boot time is still horrible Feb 15 17:48:44 right, you still have to manually turn on wifi, and ya boots take foreverrrr Feb 15 17:54:40 AppInstallD TASK_ERROR {"app_id":"com.palm.app-museum2","error_code":-5,"error_text":"Failed to extract package"} mmh Feb 15 17:54:55 ya, that's what I'm seeing. Feb 15 17:54:58 also: tar: ./control: Cannot change ownership to uid 0, gid 992: Operation not permitted Feb 15 17:55:02 glad its not just me Feb 15 17:57:15 so here's the explaination (just my own guess): /dev/mapper/store-media on /media/cryptofs type vfat <-- vfat doesn't have uids/gids Feb 15 17:58:06 let's see what precise "tar" command is issued by appinstalld2 Feb 15 18:01:38 is this fatal for tar? I think it should be only warning Feb 15 18:02:46 https://paste2.org/xDcK0JAH Feb 15 18:03:01 JaMa: seems to be fatal, unfortunately, I'm surprised too Feb 15 18:06:27 JaMa: --no-same-owner is the default when tar isn't started as root, otherwise it's --same-owner Feb 15 18:07:35 I guess we should always pass --no-same-owner in appinstalld2, I don't see when that would be a good idea Feb 15 18:18:17 Herrie: https://github.com/webOS-ports/appinstalld2/pull/6 this should help Feb 15 18:21:18 yay, thanks! Feb 15 18:22:30 I think it's safe to merge, let's do that Feb 15 18:25:33 Herrie: https://github.com/webOS-ports/meta-webos-ports/pull/490 Feb 15 18:25:48 ^ codepoet: if you want you can do this change locally and rebuild the TP image Feb 15 18:25:59 the rebuild should be quick, it's just appinstalld2 Feb 15 18:26:12 so now I look to JaMa, should I just "make update" then start the build? Feb 15 18:30:43 ok, i tried that. but it didn't seem to do anything Feb 15 18:30:58 "Attempted 7913 tasks of which 7913 didn't need to be rerun" Feb 15 18:31:06 so I think my box didn't get anything new Feb 15 18:35:24 oh i see, sorry. i have to make the change locally, missed that. will try! Feb 15 18:40:50 Tofe: merged Feb 15 18:40:56 JaMa: thanks Feb 15 18:41:07 codepoet: please retry make update Feb 15 18:41:39 yay! will do as soon as this meeting is over Feb 15 18:42:02 :) Feb 15 18:58:44 ok, make update pulled a change for appinstalld2 and build is started! Feb 15 19:00:52 got a bunch of warnings about appinstalld2 going backwards, but it looks like its building it anyway Feb 15 19:09:48 built! deploying Feb 15 19:13:50 that's probably because you don't have PRservice enabled in your builds Feb 15 19:14:07 You can also simply pull a single IPK and install it with "opkg install --force-reinstall --force-downgrade myipkname.ipk" so you don't need to flash the whole image Feb 15 19:17:15 Thanks Herrie. JaMa how do I enable PRservice? Feb 15 19:27:17 ok, it worked! Herrie: I had to do a manual rescan before it would launch the newly installed app, though. Feb 15 19:36:45 Preware doesn't have the problem, so I'll just add the scan back to the SDK for now. Feb 15 19:39:16 Open Source question: I was going to work on the Photos & Video app next -- its a little rough. Is there any reason not to borrow heavily from the webOS Photos and Videos app? Is it not fair game for OSS? Feb 15 19:44:36 codepoet: PRSERV_HOST = "localhost:0" in conf/local.conf Feb 15 19:45:21 thanks, i'll try that! Feb 15 20:19:16 JaMa: I'm trying to understand an issue with luna-surfacemanager, which comes down to an error in qt-features-webos-native: "Unknown module(s) in QT: input_support" Feb 15 20:19:34 but I tried to track down this module in qtbase, and all seem to indicate that it's built Feb 15 20:20:21 Tofe: It might be OSE's maliit version somehow? Feb 15 20:20:52 HerrieTP_: no no, it's just that module here: https://github.com/qt/qtbase/blob/5.15/src/platformsupport/input/input-support.pro Feb 15 20:21:26 it's needed to build generate_qmap binary, which is later used during the do_configure of LSM for compiling a keymap file Feb 15 20:22:20 at first I thought that there was a missing flag in our qtbase PACKAGECONFIG, but no, it's fine Feb 15 20:24:21 Do we need a keymap file in general? Feb 15 20:24:52 not sure :) I just want to give LSM all it needs, before tinkering with our cardshell Feb 15 20:25:12 mmmh I don't find any libQtInputSupport.so in our rootfs... Feb 15 20:31:21 I have tmp-glibc/work/aarch64-webos-linux/qt-features-webos/1.0-r8/recipe-sysroot/usr/include/QtInputSupport but not tmp-glibc/work/x86_64-linux/qt-features-webos-native/1.0-r8/recipe-sysroot-native/usr/include/QtInputSupport so I guess the issue is in the qtbase-native configure somehow Feb 15 20:31:34 ok, I'll just try without the keymap Feb 15 20:32:42 Tofe: is it failing to build qt-features-webos-native or this call "${STAGING_DIR_NATIVE}${OE_QMAKE_PATH_QT_BINS}/generate_qmap ${D}${datadir}/webos-keymap/webos-keymap.qmap" failing to execute with this error during luna-surfacemanager do_install? Feb 15 20:33:31 it "succeeds" in building qt-features-webos-native, but just because the qmake error I pasted isn't fatal Feb 15 20:33:59 so I have no generate_qmap native binary Feb 15 20:34:09 and do_install of LSM fails too, yes, as a consequence Feb 15 20:34:48 can you try to enable evdev PACKAGECONFIG in qtbase-native bbappend? Feb 15 20:35:14 and are you using webos-oe-g++ PLATFORM in qtbase now? Feb 15 20:35:19 can do yes, and it'll probably will help Feb 15 20:35:42 hum, not sure, I guess not, as I don't know what it is about Feb 15 20:36:20 in meta-qt5 I've added "oe" platform, LGE did few more modifications and call it webos-oe-{g++,clang} now Feb 15 20:37:57 we're not tweaking the PLATFORM in our bbappend, so we're still on "oe" I think Feb 15 20:38:34 https://github.com/webosose/meta-webosose/commit/7ac55bdb522893c9b6b984863a4e8a86687d1858#diff-7c1468fc17535047c29cb333d556c8b953f8a107261fa65d98ecfab3e44e6afa Feb 15 20:39:29 but the evdev is more likely the issue Feb 15 20:39:59 the platform mostly adds -DQ_OS_WEBOS and the input_support most likely doesn't use that Feb 15 20:40:20 yes I agree Feb 15 20:40:52 I see we use PACKAGECONFIG_DISTRO in our qtbase bbappend, not sure what difference that makes w.r.t. just PACKAGECONFIG Feb 15 20:41:03 it used to be enabled here https://github.com/webosose/meta-webosose/commit/7ac55bdb522893c9b6b984863a4e8a86687d1858#diff-a6175dcdd9e77b38cfdabda0e1a5560f5675fc056ddd887468f91e343eb32f69L8 Feb 15 20:41:15 see meta-qt5 recipe Feb 15 20:41:46 https://github.com/meta-qt5/meta-qt5/blob/master/recipes-qt/qt5/qtbase_git.bb#L70 Feb 15 20:42:03 the final PACKAGECONFIG is just defined by bunch of PACKAGECONFIG_* variables Feb 15 20:42:16 ah ok, thanks, I see Feb 15 20:42:33 I'll set PACKAGECONFIG_DISTRO in the -native too then, for consistency Feb 15 20:43:04 no, see https://github.com/meta-qt5/meta-qt5/blob/master/recipes-qt/qt5/qtbase-native_git.bb Feb 15 20:43:23 it actually isn't even PACKAGECONFIG option Feb 15 20:43:23 oh, damn Feb 15 20:43:24 ok Feb 15 20:43:39 QT_CONFIG_FLAGS:append = " -evdev" Feb 15 20:44:09 or PACKAGECONFIG_CONFARGS ? Feb 15 20:44:29 that would work, but QT_CONFIG_FLAGS is clearer in this case Feb 15 20:44:35 right Feb 15 20:45:19 or you can add PACKAGECONFIG[evdev] in the bbappend and enable it Feb 15 20:45:40 I'll keep it simple :) Feb 15 20:45:43 ok let's go for a full rebuild Feb 15 20:47:19 did you update meta-qt5 and meta-webos-ports already? Feb 15 20:47:33 there might be few more qtbase changes from me (from yesterday) Feb 15 20:47:43 no, I'm still on herrie's LSM branch, which is a few days behind Feb 15 20:53:59 -evdev wasn't enough to get QtInputSupport Feb 15 20:54:27 I tried -xkbcommon, but then I have some libs missing... maybe I could try -libinput Feb 15 20:56:35 evdev should have been enough, though Feb 15 21:01:19 ah https://github.com/qt/qtbase/blob/5.15/src/src.pro#L228 it require also the GUI :( Feb 15 21:08:14 but wait, this is absurd, the dependency here is only to avoid duplicating this https://github.com/qt/qtbase/blob/5.15/src/platformsupport/input/evdevkeyboard/qevdevkeyboard_defaultmap_p.h Feb 15 21:08:59 let's just patch the qt-features-webos-native tool and voilĂ ... Feb 15 21:44:32 ok, I've got generate_qmap Feb 15 22:03:20 Herrie: https://github.com/webOS-ports/meta-webos-ports/pull/491 Feb 15 22:04:07 and that's all for tonight :p Feb 15 22:29:47 the PR LGTM but will leave it for Herrie as it's his branch Feb 15 22:43:23 FYI: Next week is feature freeze for 3.5, our next LTS release Feb 16 02:39:14 I opened two new issues. Neither are blockers, but both have significant impact on app compatibility: Feb 16 02:40:14 ApplicationEvents aren't firing: https://github.com/webOS-ports/luneos-testing/issues/13 Feb 16 02:40:26 Apps can't save preferences: https://github.com/webOS-ports/luneos-testing/issues/14 Feb 16 02:40:57 Test and repro apps are linked, source code is available! Feb 16 02:43:46 repro-able on tenderloin, hammerhead and emulator **** ENDING LOGGING AT Wed Feb 16 02:59:56 2022