**** BEGIN LOGGING AT Fri May 29 02:59:57 2020 May 29 05:26:31 Morning! May 29 05:26:49 Tofe: Did some minor fixes again and kicked off a new build just now, let's see how it goes now May 29 06:35:23 Morning! May 29 07:01:41 Tofe: https://github.com/webOS-ports/meta-webos-ports/commits/herrie/zeus-acg-qt5.15-large May 29 07:02:18 Emails still not showing up, but seems it was filecache permission missing for mojomail, just added that and will retest now May 29 07:03:31 Also C+DAV not working yet somehow, account gets added it seems but doesn't show up in the app and also nothing synced May 29 07:03:36 Might be similar issue somewhere May 29 07:03:57 Your new implementation is nice, but will right away show all missing permissions as well, where previously these got ignored ;) May 29 07:08:25 hehe, let's hide back all this work :) May 29 07:09:24 Well it's good it all shows up now ;) May 29 07:09:32 Should be sorted at some point anyway May 29 07:10:39 Would it be possible to compact a little bit the commit history, for example merge all the "fixup" commits without description ? You know how to do that with "git rebase -i" ? May 29 07:10:49 Yes May 29 07:10:57 I do git rebase -i quite often ;) May 29 07:11:04 I'll do that once done with all May 29 07:11:14 Ok, great May 29 07:12:03 Hmmz it now tells me it's syncing (email) account and sync completed, but no mails still :S May 29 07:12:06 Let's check logs agian May 29 07:12:36 email app not retrieving account info or something, maybe May 29 07:16:52 Logs don't show anything... I see the acount, just no email list... May 29 07:22:24 Maybe I should check ls-monitor instead May 29 07:23:24 Tofe: I'm seeing this in journal by the way: May 29 09:15:11 qemux86-64 LunaWebAppManager[785]: WARNING: 09:15:11.171: qrc:/qml/ApplicationContainer.qml:99: ReferenceError: Screen is not defined May 29 07:25:57 oh ? surprising it only shows up now May 29 07:26:08 it's probably a missing qml import May 29 07:26:42 And also this one: May 29 09:09:13 qemux86-64 luna-next[591]: WARNING: 09:09:13.234: file:///usr/palm/luna-next/shells/card/qml/LaunchBar/LaunchBar.qml:120: TypeError: Cannot read property 'verticalCenter' of null May 29 07:30:49 And all those warnings about implicitly defined functions, but we already have a fix for that May 29 07:33:35 "parent" is null ? weird... May 29 07:35:30 Tofe: Yeah :S May 29 07:35:37 I saw that one before, not sure where it comes from May 29 08:02:17 Ah, I think I understand; I shouldn't use "parent" in the top-level element of a delegate item May 29 08:14:09 Tofe: I pulled the fix into my acg webappmanager branch as well, so I can build it there too. May 29 08:21:32 ok ! May 29 08:46:02 Tofe: Seems your fix doesn't work for launchbar... Everything looks funny, big etc and I see this in log: May 29 10:42:51 qemux86-64 luna-next[413]: WARNING: 08:42:51.872: file:///usr/palm/luna-next/shells/card/qml/LaunchBar/LaunchBar.qml:117:13: QML Item: Cannot anchor to an item that isn't a parent or sibling. May 29 08:55:59 damn May 29 09:01:52 Why my email is not working is a bit of a mystery to me at the moment May 29 09:01:56 I'll dig a bit more in logs etc May 29 09:55:10 Ok getting somewhere with email now it seems May 29 09:55:16 I get a list of items after reboot May 29 09:55:23 Just no email body yet, but hints in logs May 29 10:06:15 Yup that works now too :0 May 29 10:06:16 :) May 29 10:26:38 good work :) May 29 10:35:40 Now if I could get C+DAV to work too that would be good ;) May 29 11:50:32 Herrie: is herrie/zeus-acg-qt5.15-large the most complete version of all ACG changes or do I need to merge something else into my builds? May 29 11:59:20 JaMa: for now yes May 29 11:59:29 It's updated on the fly though May 29 11:59:43 Going through logs and pushing changes all the time May 29 11:59:51 But it should be pretty good May 29 12:00:15 I. E. I can add my IMAP/SMTP account and receive emails May 29 12:02:43 any plans to rebase it on latest zeus? there are some conflicts with the bbclasses already merged in zeus May 29 12:03:27 Herrie: ^I think we should rebase, squash some commit together and PR the whole thing May 29 12:03:27 JaMa: yes it should be rebased and squashed a bit May 29 12:03:59 JaMa: if you could rebase that would be great, I'm always fighting with that May 29 12:04:06 Squashing I can do May 29 12:04:17 Or I can give it a try May 29 12:04:34 Herrie: do you have pending local modifications ? May 29 12:04:59 I can, but I wasn't following the changes, so it will take me probably much longer than one of you May 29 12:05:44 I'll do it then May 29 12:05:45 e.g. "if "com.palm." in app_id:" added and then removed in bbclass May 29 12:06:01 Tofe: thanks! May 29 12:06:23 I'll just wait to make sure Herrie doesn't have pending modifications to push May 29 12:08:15 ok, my quick rebase is in jansa/zeus-acg-qt5.15-large but not sure if this is what you wanted in the end May 29 12:09:18 JaMa: well , I can also do it now and push the result under a tofe/... branch May 29 12:11:35 I have a few pending let me push them now May 29 12:13:31 OK pushed May 29 12:15:18 JaMa: rebased: https://github.com/webOS-ports/meta-webos-ports/commits/tofe/zeus-acg-qt5.15-large May 29 12:15:27 (with the very last commit of Herrie :) ) May 29 12:16:44 Maybe it's a good time to start a fresh build here May 29 12:19:57 let's go May 29 12:22:05 Herrie: core-apps failing ? May 29 12:22:15 ^fetch task May 29 12:25:06 pushed a fix on my branch May 29 12:26:59 Damn, cardshell too May 29 12:28:28 pushed too May 29 12:31:25 Ah, I understand, you've switched the srcrevs of those two May 29 12:32:00 An easy mistake with such a big commit tree May 29 12:32:17 JaMa: ^ should be fine now May 29 12:45:45 Hi. Attempted qtweb-engine compilation third time. After 2.5 hrs, progresss still at 0%. So i closed that by keyboard interrupt. It seems bottlenect is Ram (4gb only). Swap usage climbs to 5.5-6GB. Idle system ram usage is 550Mb. So its is clear that compilation was going buy surprisingly slow. May 29 12:45:48 Any other solution? May 29 12:45:49 Can I use compiled version of some other system from your guys? I mean webengine compiled for rosy or some other snapdragon? Since architecture and instruction set is same along with Operating system, compiled file could also be same most probably (I hope web engine is independent of bluetooth and other sensors) GPU and cores since share same instruction set, isn't compilate file same???. May 29 12:47:33 Sstate server sync could be boon but since they are not configured, so that option is gone May 29 12:49:54 Atleast i should be then able to attempt Even if webengine file is not 100% correct, I just hope it doesn't brick my device. What will happen at most? Web browser (plus other apps dependent on chromium) doesn't work? Still i'd love to run webOS May 29 12:49:55 Tofe: Ah could be I copied & pasted in wrong recipes :P May 29 12:51:16 BT40: the fileserver now has sstate for latest zeus builds, but your build probably doesn't have latest changes like last 2 qt upgrades, so it won't be able to reuse that May 29 12:51:52 Tofe: Let me take your branch and clean it up ;) May 29 12:51:58 sstate is much more picky for reuse than "architecture and instruction set is same along with Operating system", it needs every metadata input (and all it's dependencies) to be identical May 29 12:52:01 Herrie: sure May 29 12:52:02 ok. any other hack?????/ May 29 12:52:02 :) May 29 12:53:09 JaMa: yes, but even so, xiaomi-land is nearly identical to xiaomi-rosy, so I think it should work May 29 12:53:47 Tofe: in runtime yes, but without locked sstate I'm pretty sure something in sstate signature is different between these 2 May 29 12:54:46 damn May 29 12:55:45 but then there's no point in exposing the sstate on a server... May 29 12:56:58 can we trick bb in some manner that qtengine has been compiled successfully? Even in reality it doesn't. I mean using WebOS as feature phone without browser? This is possible only if rest of webOS core can work without webEngine May 29 12:57:07 Tofe: if he has latest zeus, then there might be some sstate reuse May 29 12:57:26 JaMa: ah but yes, I meant after rebasing on top of zeus May 29 12:57:34 Tofe: or if we keep multiple versions of sstate on fileserver (instead of prunning all older and keeping only the latest to save disk space) May 29 12:58:19 BT40: ok, here's what you can try: make sure you have pushed your current changes somewhere -- if would be really sad to lose that May 29 12:58:48 then update to latest testing and rebase your meta-webos-ports and meta-smartphone repos May 29 12:59:27 How to upfate to latest zeus? any guide for layman like me? May 29 12:59:49 ehm I don't use layman, I do that by hand May 29 13:00:10 newbie like me :) May 29 13:00:47 I guess you initialized the source code using our webosports-setup script ? May 29 13:01:14 I followed online guide which is written for hammerhead built May 29 13:01:28 Ok, this hasn't changed since then May 29 13:02:05 so, go to the parent directory, where you have the "common" directory and a Makefile May 29 13:02:21 (and webos-ports too) May 29 13:02:43 opened May 29 13:03:22 now do a "make update" May 29 13:03:26 One more question, isn't my build based on latest zeus? May 29 13:03:44 I fear not, JaMa updated zeus a few hours ago May 29 13:03:51 it's a moving target May 29 13:04:10 make update..... I already ran that command during my first time environment setup May 29 13:04:27 Ok will run again May 29 13:04:39 yes, it'll fetch the latest layer description, and use that to update all the repos in webos-ports/ May 29 13:04:53 Should i need to enable sstate which i disabled earlier in config before running update? May 29 13:05:10 no, that's independant of bitbake, it's just git stuff May 29 13:06:08 WARNING: You have different layers.txt. May 29 13:06:08 Please sync it from common directory or call update-conffiles to replace all config files with new versions. May 29 13:06:17 Rest process is complete May 29 13:06:23 ok, do make update-conffile May 29 13:06:27 +s May 29 13:06:47 and then make update again (I always do that to be sure it's happy) May 29 13:07:08 +s is after running above command? May 29 13:07:18 or append in above May 29 13:07:21 it's update-conffileS May 29 13:07:28 okay May 29 13:07:46 done May 29 13:07:53 it says files upto date May 29 13:08:06 good, now let's check your repos in webos-ports May 29 13:08:49 the status of git in openembedded-core should be branch zeus, nothing to commit May 29 13:08:55 same for meta-openembedded May 29 13:09:10 meta-qt5 should be on master, also clean May 29 13:09:32 How to check above......the status of git in openembedded-core May 29 13:09:39 any command? May 29 13:09:46 "git status" from withing the directory May 29 13:09:46 or by opening some ile? May 29 13:10:16 "On branch zeus -- nothing to commit, working tree clean" May 29 13:10:32 Command 'qit' not found, did you mean: May 29 13:10:42 git, not qit :) May 29 13:11:30 fatal: not a git repository (or any parent up to mount point /) May 29 13:11:30 Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set). May 29 13:11:56 I mean, go into openembedded-core/ first May 29 13:11:59 Do i need to run it from webos-ports directory? May 29 13:12:04 ok May 29 13:12:09 from webos-ports/openembedded-core/ May 29 13:12:46 On branch zeus May 29 13:12:46 Your branch is up to date with 'origin/zeus'. May 29 13:12:46 nothing to commit, working tree clean May 29 13:12:52 great May 29 13:13:20 now, the tricky part May 29 13:13:31 go into meta-webos-ports May 29 13:14:01 what does a "git status" give you here May 29 13:14:14 shatll i do same for these also before proceeding to tricky part? same for meta-openembedded May 29 13:14:15 meta-qt5 should be on master, also clean May 29 13:14:45 it's optional, but yes you can do that first May 29 13:15:17 Your branch is up to date with 'origin/zeus'. May 29 13:15:17 Untracked files: May 29 13:15:30 nothing added to commit but untracked files present (use "git add" to track) May 29 13:15:46 6 files are red highlighted May 29 13:16:16 untracked files are always red; but they could just be temporary files you left when editing something May 29 13:16:40 Tofe: OK cleaned it up, should I PR it? May 29 13:17:04 luna-next-conf/land/ May 29 13:17:05 , nyx-modules/land.cmake May 29 13:17:05 nyx-conf/land/ May 29 13:17:05 luna-sysmgr-conf/land/ May 29 13:17:05 pulseaudio-distro-conf/land/ May 29 13:17:05 no, there's nothing to PR, now you are equivalent to origin/zeus May 29 13:17:05 ofono-conf/land/ May 29 13:17:17 These are untracked ones May 29 13:17:22 ah ok, sorry, you're in meta-webos-ports May 29 13:18:32 at some point it would be good to PR all this work yes May 29 13:18:43 if only just to have a copy on a server May 29 13:19:21 anyway, it seems the script managed updating the repo alright May 29 13:19:28 so now, about meta-smartphone May 29 13:20:35 it could be that the status is a bit more messy May 29 13:22:57 and the next step/? May 29 13:23:13 go into meta-smartphone/ and go "git status" May 29 13:23:18 s/go/do/ May 29 13:23:36 meta-webos-ports/ looks alright, from what you said May 29 13:25:10 Tofe/JaMa: There we go: https://github.com/webOS-ports/meta-webos-ports/pull/434 May 29 13:25:55 Herrie: woohoo May 29 13:26:44 Herrie: wait, it's not rebased ? why is there some qt stuff ? May 29 13:26:45 I'm sure there's stuff to fix down the line, but this should really cover most already May 29 13:26:53 Tofe: I took your branch.... May 29 13:26:57 And then cleaned it up May 29 13:27:18 I can drop the Qt commits if needed May 29 13:27:34 I'd prefer understanding why they're there in the first place May 29 13:28:00 https://pastebin.com/qh8n88Gj This is output in meta smartphone May 29 13:31:01 Herrie: I must have rebased on an older zeus somehow... oh, well. May 29 13:31:32 BT40: looks good, I think you're good to go May 29 13:31:43 re-enable sstate, and give it a try :) May 29 13:32:08 how much time can i expect?? just rough idea would be appreciated May 29 13:32:33 Herrie: rebase again :) May 29 13:32:47 JaMa, Herrie: GH can do it for us too May 29 13:33:15 I'll rebase&merge it, if there is no objection May 29 13:33:27 BT40: let's see... 8 hours ? May 29 13:33:32 i saw log of do compile for webengine, 41 tasks were completed in 2.5 hrs today. Can u guys do me a favour by telling total quantity iof tasks from your log file? just last task numder would be sufficient. May 29 13:33:38 or less, if sstate works May 29 13:34:05 BT40: Currently 4 running tasks (6574 of 6892) 95% May 29 13:34:13 8 hts for qt engine only???? May 29 13:34:17 hrs May 29 13:34:22 hahaha May 29 13:34:33 without sstate it would be 200hrs then May 29 13:34:33 probably more.. May 29 13:34:41 BT40: no no, I hadn't taken sstate into account May 29 13:35:09 Tofe/JaMa: Would be good to review the recipe changes... I did some things that JaMa might not agree with ;) May 29 13:35:11 if sstate works, a lot of the aarch64-halium stuff should be shared, qt included May 29 13:35:19 Tofe, without sstate, 2.5 hrs lead to 0% progress, 4 hrs 1% progress May 29 13:35:45 These numvers by meare for qt engine compilation, not overall May 29 13:35:54 BT40: that's with swap too May 29 13:36:00 these numbers by me-----typo May 29 13:36:06 yes May 29 13:36:08 hahaha May 29 13:36:23 cpu usage was mere 1.5% May 29 13:36:27 hahaha May 29 13:36:28 with sstate, it's just downloading the thing, mainly May 29 13:36:43 ok. will give it try right now May 29 13:36:59 Don't forget to activate sstate May 29 13:37:17 Herrie: I'll have a look. A loooong look. May 29 13:38:00 will come back later on. Compilation load freezes my PC, sometimes disconnect automatically from irc. So in order to reduce bloat in this channel of connecting and disconnecting, i think i should go offline May 29 13:39:08 Not these.Tofe. BT40: Currently 4 running tasks (6574 of 6892) 95% May 29 13:39:17 These are mine 98% complete May 29 13:39:20 Herrie: I've rebased your branch/PR May 29 13:39:39 i am talking about within log file of qtwebengine folder with do compile name May 29 13:40:19 Herrie: any plan to merge herrir/acg branches into the default webosports/master branches? May 29 13:40:25 i want to know how many sub processes/sub commands do webengine compilation have? May 29 13:40:54 BT40: you should see it in the log.do_compile (it depends on how many were already built before) May 29 13:41:42 JaMa: At some point yet, but seeing we know this works now, we prefer to merge as is and later do PR's in each repo and update the recipes back May 29 13:41:54 Just so we don't forget things in the process May 29 13:41:59 oh yes. forgot that. Thanks..(it depends on how many were already built before May 29 13:42:30 ok. see u later on May 29 13:44:11 Herrie: ok, as long as we'll eventually merge it back to better see what changes it introduces May 29 13:44:48 Herrie: and there is already too many different branches everywhere (even when we relly support only one build for all components) May 29 13:45:49 meta-webos-ports$ git grep WEBOS_GIT_PARAM_BRANCH | wc -l May 29 13:45:49 79 May 29 13:45:52 used to be < 5 May 29 13:46:37 JaMa: I'll get to that ;) May 29 13:46:40 martin@jama:/OE/meta-webos-ports$ git grep WEBOS_GIT_PARAM_BRANCH | grep -c herrie/acg May 29 13:46:43 43 May 29 13:46:46 martin@jama:/OE/meta-webos-ports$ git grep WEBOS_GIT_PARAM_BRANCH | grep -c OSE May 29 13:46:49 23 May 29 13:46:59 You want everything in master? May 29 13:47:00 Or.... May 29 13:47:01 ? May 29 13:47:14 ehm webos-ports/master that is May 29 13:47:48 meta-luneos/recipes-luneos/apps/org.webosports.app.phone.bb:WEBOS_GIT_PARAM_BRANCH = "tofe/acg" May 29 13:47:51 meta-luneos/recipes-luneos/apps/org.webosports.app.settings-qml.bb:WEBOS_GIT_PARAM_BRANCH = "qml-based" May 29 13:47:54 meta-luneos/recipes-luneos/libwebos-application/libwebos-application.bb:WEBOS_GIT_PARAM_BRANCH = "tofe/acg" May 29 13:47:57 meta-luneos/recipes-luneos/luneos-components/luneos-components.bb:WEBOS_GIT_PARAM_BRANCH = "tofe/acg" May 29 13:48:00 meta-luneos/recipes-luneos/webos-keyboard/webos-keyboard.bb:WEBOS_GIT_PARAM_BRANCH = "herrie/qt-5.15" May 29 13:48:03 meta-luneos/recipes-webos/luna-qml-launcher/luna-qml-launcher.bb:WEBOS_GIT_PARAM_BRANCH = "herrie/work" May 29 13:48:06 meta-luneos/recipes-webos/luna-sysmgr/luna-sysmgr.bb:WEBOS_GIT_PARAM_BRANCH = "herrie/ls2-acg-fixes" May 29 13:48:09 meta-luneos/recipes-webos/luna-systemui/luna-systemui.bb:WEBOS_GIT_PARAM_BRANCH = "herrie/ls2-acg-wam" May 29 13:48:12 meta-luneos/recipes-webos/nodejs-module-webos/nodejs-module-webos.inc:WEBOS_GIT_PARAM_BRANCH = "tofe/nodejs12" May 29 13:48:15 meta-luneui/classes/webos_ports_fork_repo.bbclass:WEBOS_GIT_PARAM_BRANCH ?= "webOS-ports/master" May 29 13:48:18 meta-luneui/classes/webos_public_repo.bbclass:WEBOS_GIT_PARAM_BRANCH ?= "master" May 29 13:48:21 meta-luneui/classes/webos_public_repo.bbclass:WEBOS_GIT_BRANCH ?= ";branch=${WEBOS_GIT_PARAM_BRANCH}" May 29 13:48:24 meta-luneui/recipes-qt/qt5/qtwebengine_git.bbappend:WEBOS_GIT_PARAM_BRANCH = "webOS-ports/master" May 29 13:48:27 yes, we should eventually merge everything to webOS-ports/master May 29 13:49:33 or just master if it's identical with OSE and just mirrored to webOS-ports fork May 29 13:52:51 JaMa: OK I can do that going forward after all is merged & switched back. Do it in 1 go May 29 13:57:10 Sstate summary: Wanted 238 Found 216 Missed 22 Current 2315 (90% match, 99% complete) May 29 13:57:26 * BT40 would love to understand above numbers May 29 13:57:57 Means that from 238 packages, 216 were found and downloaded and 22 needed to be build/rebuild May 29 13:58:05 So that's pretty good May 29 13:58:12 99% complete... Does this refer to compilation tasks around 6500? May 29 13:58:16 BT40: https://www.openembedded.org/pipermail/openembedded-core/2018-August/273336.html May 29 13:59:01 actually the longer answer https://www.openembedded.org/pipermail/openembedded-core/2018-August/273335.html May 29 14:00:06 JaMa, Thank you. Now feeling pumped May 29 14:00:42 But still woried, one of you said that 25% build time is for web engine only some days ago. May 29 14:01:02 so 99% is actually 74% :) May 29 14:01:32 this is number of sstate archives, not related to build time percentage finished or not May 29 14:11:01 I got this error. Anything serious? ERROR: qtdeclarative-5.15.0+gitAUTOINC+0fafffedf4-r0 do_package_write_ipk_setscene: Fetcher failure: Unable to find file file://6f/sstate:qtdeclarative:aarch64-halium-webos-lin........ May 29 14:11:18 ERROR: qtdeclarative-5.15.0+gitAUTOINC+0fafffedf4-r0 do_package_write_ipk_setscene: No suitable staging package found May 29 14:11:38 nope May 29 14:11:57 It will rebuild normally May 29 14:12:50 though it's a bit surprising it didn't find that one May 29 14:13:41 4 threads too slow for sstate, underutilizing my bandwidtha and processor.. will bosst it to 12. May 29 14:19:58 my guess is the connection was refused when trying to fetch it from file server May 29 14:20:38 it happens often when when file server isn't under chinese ddos May 29 14:20:51 even when May 29 14:27:35 is this irc chat logged on your/freenode servers? In future if we need to access due to lost local log file? May 29 14:29:19 yes, it is logged May 29 14:29:50 BT40: see http://logs.nslu2-linux.org/livelogs/webos-ports/ May 29 14:30:16 great May 29 14:35:58 can i get some article explaining commands like write_ipk_cache, do package setscene etc. I only can understand do fetch and do compile or somewhat do package. May 29 14:38:45 SetScene basically takes what is alraedy in cache and unpacks it in the build directories May 29 14:40:07 write_ipk_cache I don't know exactly May 29 14:44:05 Bad news for me: ERROR: qtwebengine-5.15.0+gitAUTOINC+7b53f3018b_25e1fa7d83-r0 do_package_setscene: No suitable staging package found May 29 14:44:28 this error came after somewhat downloading of file....around 25% May 29 14:45:48 write_ipk_cache doesn't exist :) May 29 14:47:56 sorry couldn't get May 29 14:48:00 NOTE: Sstate package /home/hextreme/luneos/mi-env/webos-ports/sstate-cache/20/sstate:qtwebengine:aarch64-halium-webos-linux:5.15.0+gitAUTOINC+7b53f3018b_25e1fa7d83:r0:aarch64-halium:3:2095bad80b259eaf68eeca94057ae864a66e8885e3c2aba5dbe3b7b117a4f7cf_package.tgz does not exist May 29 14:48:07 this is in log file May 29 14:48:11 Herrie: nodejs-module-webos-sysbus.bb change is maybe suspicious, couldn't we keep the common .inc JaMa put in place a few days ago ? May 29 14:48:26 Tofe: Ehm yes I guess so... May 29 14:48:31 I probably forgot that one May 29 14:48:33 Let me check May 29 14:49:33 BT40: it could be a server issue too, interrupting the download May 29 14:49:44 for qtwebengine it's quite worth a retry May 29 14:50:11 JaMa: for sstate, does it retry when download failed ? can it be tweaked ? May 29 14:50:16 ok. will try again after some time May 29 14:51:40 Herrie: "ln -svnf libluna-prefs.so.3 ${D}${libdir}/libluna-prefs.so.2" ... ugh... is that the right fix ? what apps are still looking for so.2 ? May 29 14:52:14 Herrie: otherwise, we'll have to merge the branches, but I don't have any other remark May 29 14:52:42 Tofe: Not sure, but I couldn't get it to work without somehow.... May 29 15:01:15 Herrie: I'll first finish my build here, then I can look into it May 29 15:17:33 Tofe: OK fixed the nodejs-modules-sysbus one May 29 15:36:13 2020-05-29 15:31:24 (129 KB/s) - Connection closed at byte 43130851. Giving up. May 29 15:36:25 seems server side issues May 29 15:36:52 webengine downloading failed at 12% May 29 15:37:03 will try early morning May 29 15:37:19 hope traffic well be less on server May 29 15:52:36 Tofe; I've force pushed this bump to meta-webos-ports: https://github.com/webOS-ports/imlibpurpleservice/commit/184aba264d9c4e3734f4f7619614f6bc282ace6c. Seems I forgot "com.palm.imlibpurple/onEnabled" May 29 15:54:03 Nothing too shocking though, just had some funny errors in logs after adding Skype account May 29 15:54:34 ok May 29 15:56:57 Tofe: no it doesn't retry imho (because it often might be just removed during the build) May 29 15:57:40 Herrie: do you remember why pmloglib-provate now need pkgconfig inherit May 29 15:59:58 JaMa: I'm not 100% sure, I got a few funny errors. There were a few more recipes where I removed it May 29 16:00:07 Otherwise my builds would fail May 29 16:09:01 They were related to each other that's what I remember May 29 16:09:23 I thought it was weird because it built fine previously in Zeus but well May 29 16:12:23 I can build from scratch without it, so I think it's still not needed May 29 16:13:19 it uses pkgconfig only when BUILD_PRIVATE isn't set May 29 16:21:31 JaMa: OK not sure why I had this issue. Feel free to remove it then May 29 16:53:29 ok, removed together with that libluna-prefs.so.3 symlink May 29 16:56:58 Qt upgrade builds should finish on jenkins tomorrow, I think ACG is good enough for merge after that? It can be incrementally improved later May 29 17:05:19 JaMa: Yes it should be fine May 29 17:05:25 I needed that symlink, not sure why May 29 17:05:31 I think on runtime I got some errors May 29 17:09:03 If memory serves me right May 29 17:14:58 if something fails in runtime, we should find out why it wasn't rebuilt against correct version May 29 17:16:15 Well we didn't migrate everything from OSE yet May 29 17:17:41 but we don't use any prebuilt binaries, so everything should link against the right version, if it doesn't then something is really bad (as it shouldn't be possible to link with wrong version in new yocto with RSS) May 29 17:18:01 either you have to have the dependency or it won't be in RSS and linking will fail May 29 17:18:50 my guess is some local tampering you did in the WORKDIR :) May 29 17:21:06 JaMa: Could be not sure May 29 17:23:05 Also building here locally now. Testing + pending PR May 29 17:47:40 atleast firefox can resume download from where it failed .... hope this trick can work May 29 17:47:52 will try downloading sstate files with firefox. In commandline, it failed 5 times after 15-20%. Always started from scratch instead of resuming download. Hope firefox can fix this :) May 29 17:50:54 Herrie: quick grep in my local luneos-dev-image doesn't show anything depending on libluna-prefs.so.2 May 29 17:51:53 JaMa: OK will check runtime here May 29 17:59:09 Damn, out of space May 29 17:59:20 Damn(bis), just before the end of qtwebengine May 29 17:59:33 it will recover from that May 29 17:59:52 well it depends how I make space May 29 17:59:58 right :) May 29 18:00:39 didn't expect you to do something as reckless as wipping its workdir May 29 18:00:59 emulator builds finished on jenkins, now pine and rpi* May 29 18:01:06 I'm just wiping the pinephone's aarch64 May 29 18:01:14 which is on an old zeus anyway May 29 18:01:23 and not the same arch as tissot May 29 18:01:24 My build should be done soon here too May 29 18:01:29 End of qtwebengine May 29 18:01:59 Then another 75 taksa May 29 18:02:47 * BT40 wonders how many builds each of you pro guys might be making monthly May 29 18:03:13 I'll have 1000+ :) May 29 18:03:40 BT40: me, not that many, I often hack on the same build for a couple of weeks May 29 18:03:57 1000 monthly? JaMa or cumulative until now May 29 18:04:12 monthly, but across many servers May 29 18:04:22 wao May 29 18:04:25 yeah, JaMa is off-chart May 29 18:05:13 * BT40 is getting desperate now......to know which devices are targeted. Are they IOT..... May 29 18:05:28 ok I recovered 83GB, should be enough! May 29 18:05:34 * BT40 is sure 1000s aren't the phones May 29 18:05:46 I've jenkins with 356 different jobs which I trigger as often as they are being finished May 29 18:06:22 9 more and I can have different job for each day in year :) May 29 18:06:24 BT40: it's not only LuneOS, it's many different Yocto-related targets May 29 18:06:36 okay May 29 18:06:53 right and also different versions of each so it multiplies quickly May 29 18:07:24 looks like I'm generating around 10TB of sstate montly May 29 18:07:51 You're an SSD killer May 29 18:08:03 great.....jama the legend. Even thee pro word is not sufficient May 29 18:09:06 Tofe: sstate is on fileserver with some old disk array (not SSDs for sure) where it seems to have surprising lifetime May 29 18:09:14 JaMa: qtwebengine_git.bb:do_package_qa failed on my side, host contamination :( May 29 18:09:32 but be because the populate_sysroot was interrupted brutally May 29 18:09:37 even the server itself is ancient from 2008 with 2.6 kernel and still runs May 29 18:09:41 Tofe: what file? May 29 18:09:49 looks like all of them May 29 18:09:56 qtwebengine: /usr/lib/qml/QtQuick/Pdf/libpdfplugin.so is owned by uid 1000, which is the same as the user running bitbake. This may be due to host contamination May 29 18:10:03 ^hundreds of those May 29 18:10:09 Tofe: yeah might be, pseudo probably didn't have enough space to dump database when it got kille :/ May 29 18:10:12 d May 29 18:10:24 I hope sstate will be able to recover May 29 18:10:27 you can add INSANE_SKIP to ignore it in this run May 29 18:10:52 JaMa: no worth trying to clean and reuse sstate ? May 29 18:10:54 probably not, if the pseudo db got broken, then the broken one will be in sstate archive May 29 18:11:01 ah, damn May 29 18:11:25 worth adding INSANE_SKIP in local conf and if it reruns just package_qa then you're fine May 29 18:12:02 if it starts re-running something else, then clean and sstate is the other option May 29 18:12:08 ok May 29 18:14:22 it's not "host-user-contaminated" I have to add ? May 29 18:15:09 if it didn't resolve the error, try also ERROR_QA_remove_pn-qtwebengine = "host-user-contaminated" May 29 18:15:53 better, thanks May 29 18:16:39 that was the last step before do_rootfs, would have been frustrating to rerun it May 29 18:17:40 well your files in rootfs might be owned by your host user running bitbake May 29 18:17:54 if it causes some issues in runtime, you'll have to chown May 29 18:18:41 are you running with rm_work? May 29 18:19:56 rerunning do_install and follow-up tasks might also help to recreate the pseudo database (if the build directory wasn't removed already) May 29 18:21:40 I don't do rm_work May 29 18:21:51 I'll try to force do_install May 29 18:22:05 but first I'll check the actual owner of the files May 29 18:22:10 then when you have an image ready try bitbake -f -c install qtwebengine May 29 18:22:39 image just finished, I restarted install yes May 29 18:23:05 let's remove the QA overrides too... May 29 18:23:13 and then drop the ERROR_QA and run the rest May 29 18:25:27 if it works then remove the taint from -f (tmp-glibc/stamps/*/qtwebengine/*.do_install.taint) May 29 18:25:28 nope :( May 29 18:25:39 again host-user-contaminated? May 29 18:25:43 yes May 29 18:25:58 did it run do_package again? May 29 18:26:08 yes, and also populate_sysroot May 29 18:26:37 qtwebengine: /usr/src/debug/qtwebengine/5.15.0+gitAUTOINC+7b53f3018b_25e1fa7d83-r0/build/src/core/release/gen/media/mojo/mojom/content_decryption_module.mojom-blink.h is owned by uid 1000, which is the same as the user running bitbake. This may be due to host contamination May 29 18:26:47 hmm bummer May 29 18:26:58 yup May 29 18:27:22 you can still try clean and sstate if you're ready to let it build over night if it wents wrong May 29 18:28:13 I'd prefer having a clean image, if I have to debug a strange think like the .so.2 -> .so.3 Herrie was describing May 29 18:28:13 thing* May 29 18:28:51 yeah May 29 18:29:10 Tofe: I can put qemu on ftp May 29 18:29:16 It's ready May 29 18:29:16 btw why aren't you using the DISKMON? it should be enabled by default and stop the build before this happens May 29 18:29:32 JaMa: because... I don't know it exists ? May 29 18:29:53 nope, clean doesn't help. go, cleansstate. May 29 18:30:07 Tofe: https://github.com/webOS-ports/webos-ports-setup/commit/9d7630eef62d5df28f8390dab79fd4fdda068ce6 May 29 18:31:01 Tofe: On FTP now May 29 18:31:07 Herrie: thanks May 29 18:31:33 smugling luneos images :) May 29 18:31:35 This is latest Zeus testing so incl 5.15 final + acg May 29 18:32:19 JaMa: Yes via my router's FTP server May 29 18:32:27 Herrie: without the .so.3 hack ? May 29 18:32:33 Can build a lot quicker vs Tofe LOL May 29 18:32:48 Tofe: I think JaMa removed it already? May 29 18:33:09 from the PR, but don't know if you updated it in your build May 29 18:33:10 yes, I didn't know what version of the build you were using May 29 18:33:24 at least it's easy to find and undo in rootfs May 29 18:33:37 right. Let's smuggle. May 29 18:35:13 JaMa: I added DISKMON; was using a 6 year old local.conf, so... May 29 18:36:02 JaMa: it should be without the symlink May 29 18:38:13 Tofe: :) better check what else you were missing :) May 29 18:39:23 actually, nothing, it was your first commit on that file since 2014 May 29 18:47:08 JaMa: I took it after your force push May 29 18:47:13 So should be good May 29 18:53:39 It boots here, FirstUse behaves, now bathing children so no chance to test further for now May 29 18:53:50 I'm at the same step May 29 18:56:02 What else I did for testing was adding email (IMAP/SMTP), emails were received after a reboot May 29 18:56:11 user.warning luna-next [] LS_REQUIRES_SECURITY {"SERVICE":"com.palm.applicationManager","CATEGORY":"/","METHOD":"focusApplication"} Service security groups don't allow method call. May 29 18:56:18 Sending email causes crash May 29 18:56:43 Tofe: Saw that one and bit puzzled May 29 18:56:59 Things LOOK OK at my end May 29 18:57:09 at my end too May 29 18:59:50 user.warn ls-hubd: [] [pmlog] ls-hubd LSHUB_NO_ROLE_FILE {"APP_ID":"org.webosports.app.testr"} No role file for application id: "org.webosports.app.testr" May 29 19:00:56 well well, it's pretty good overall May 29 19:01:57 and the libluna-prefs symlink isn't there, so it's clean May 29 19:10:02 Tofe: I swear I fixed Testr maybe missed something in bb May 29 19:11:39 ahah, I do believe you :) May 29 19:23:52 Tofe: seems I might have forgotten an inherit webos_app there May 29 19:27:17 Since I see SRCREV bumped and branch switched in bb May 29 19:33:19 that would explain it yes May 29 19:41:47 My Google GMail C+DAV gets added successfully but no contacts & calendars synced when I tried with previous image May 29 19:42:50 Probably something small since logs don't show much. I guess ls-monitor might help, though on qemu it's flooded with failed com.palm.telephony calls May 29 19:43:22 It's good to have it in Qemu because without it we had other issues just I guess we need to tweak it a little so it doesn't spam May 29 19:43:36 we could address that yes May 29 19:56:54 It would help debugging using ls-monitor without filtering for a specific service which is currently a bit hard because of the flooding May 29 20:04:40 I think we could improve ls-monitor a bit, concerning the output May 29 20:05:11 fixed column width, exclusion filter... May 29 20:05:59 maybe even colored output May 29 20:07:33 Well they have a minimal option which is fixed columen I think May 29 20:10:07 the compact output ? May 29 20:10:22 I don't want things to be compact, I want them to be aligned May 29 20:12:46 So... we have a go for the PR merge, or ?... May 29 20:13:16 That can way until tomorrow, we're not in a hurry May 29 20:19:35 Tofe: I'll force push the bb for Testr I guess otherwise Ok May 29 20:19:51 JaMa wants to merge sometime tomorrow I think May 29 20:19:58 Fine by me May 29 20:20:16 Yeah also not a fan of compact, but it is aligned at least **** ENDING LOGGING AT Sat May 30 02:59:59 2020