**** BEGIN LOGGING AT Fri Jan 28 02:59:57 2022 Jan 28 07:04:12 Herrie: no, for the moment I only implemented "severe" and "limit" Jan 28 07:04:19 a bit at random Jan 28 07:05:16 "severe" if the free space is under the minimum, "limit" if the percentage of used space is nearly at maximum Jan 28 07:05:36 could it should be some other criteria Jan 28 09:14:49 Tofe: OK Jan 28 09:14:53 Morning BTW Jan 28 09:15:13 I'm fighting a bit with all the changes I made... I have 2 issues at the moment, just not really sure where they come from.... Jan 28 09:15:24 1. Seems that the systemd service doesn't really work or dies... Jan 28 09:15:36 I need to restart it with systemctl restart appinstalld Jan 28 09:15:43 Which isn't the end of the world of course for now Jan 28 09:15:58 Afterwards ls-monitor -i will give me the available API calls and I can make them Jan 28 09:16:06 Just the actuall install fails at the moment Jan 28 09:16:43 And I wasn't really able to track it down yet Jan 28 09:26:56 For systemd, is the service installed (systemctl enable appinstalld) ? Maybe appinstalld crashes, and there ni a "no restart" policy in the systemd service file ? Jan 28 09:29:48 Yeah I guess so Jan 28 09:29:52 I can see it's started Jan 28 09:30:28 https://github.com/webOS-ports/appinstalld2/blob/webOS-ports/webOS-OSE/files/systemd/appinstalld.service#L28 Jan 28 09:30:32 Should restart on failure Jan 28 09:30:44 I'm getting some JSON errors in log for appinstalld though Jan 28 09:30:50 Just not being able to track it down yet Jan 28 09:31:58 Jan 28 01:21:47 qemux86-64 systemd[1]: Started webos - "appinstalld.service". Jan 28 09:32:01 So that seems to work Jan 28 09:32:19 Ah I guess this is not helping: {"SERVICE_NAME":"com.palm.appInstallService","EXE":"/usr/sbin/LunaAppManager","APP_ID":"(null)","PID":370} Service not listed in service files (cmdline: /usr/sbin/LunaAppManager) Jan 28 09:35:19 Let me grep around my image where that comes from Jan 28 09:35:33 that's a LS2 error Jan 28 09:36:03 so it's a missing declaration in luna-service2 subfolder Jan 28 09:36:09 I don't remember which one though :/ Jan 28 09:36:15 Yeah Jan 28 09:36:31 This is the full log: https://paste.ubuntu.com/p/wCYM7gMRFF/ Jan 28 09:36:39 It's right after it starts the service Jan 28 09:36:42 All files seem to be there Jan 28 09:42:26 OK got 3 leads I'm following up Jan 28 09:43:20 Could be something is making a call to com.palm.appInstallService still while we only register the com.webos.appInstallService Jan 28 09:44:28 seems systemui, luna-appmanager and org.webosports.service.ipkg make a call to that still Jan 28 09:45:45 interesting com.webos.service.attachedstoragemanager has the same kind of issue Jan 28 09:50:34 Tofe: Well seems that they hardcode some names in OSE... Jan 28 09:50:42 Most we can get away by registering under both names Jan 28 09:50:53 yup Jan 28 09:55:33 speaking of which, I wonder if attachedstoragemanager can provide the same functionality as my fsCheck service Jan 28 10:02:34 https://github.com/torvalds/linux/tree/8bcab0346d4fcf21b97046eb44db8cf37ddd6da0 lol and a bit worrying :/ Jan 28 10:03:41 JaMa2: Yeah LOL I saw that one too Jan 28 10:04:07 One wonders if Linus was drunk or someone got access to some magic keys Jan 28 10:04:16 MAybe he was on Czech absynth ;) Jan 28 10:11:14 Yes, and to know that GitHub knows about it and classifies it as "won't fix", it's a bit deceiving Jan 28 10:54:51 yes, maybe we should start signing our commits, but as very few people actually check the signature it would be just for us to feel better about this github fsck-up Jan 28 11:28:46 well there isn't any real harm done, it's just a URL pointing to a pending PR Jan 28 12:03:55 yes, but the author isn't correct and it's shown in the official repo (instead of some random fork Jan 28 12:06:29 https://github.com/webosose/meta-webosose/commit/799b24a4d66a11796920767a74ca8e339ec192c7 this commit never existed in official webosose/meta-webosose (not even as part of pending PR), but it's possible to reference it from there even when it belongs to my shr-project/meta-webosose fork, at least in this case there is a warning "This commit does not belong to any branch on this repository, and may Jan 28 12:06:35 belong to a fork outside of the repository." Jan 28 12:06:43 not sure how they got rid of this warning in kernel case Jan 28 12:07:32 aha, because it's a blob not commit being referenced Jan 28 12:07:52 https://github.com/torvalds/linux/commit/8bcab0346d4fcf21b97046eb44db8cf37ddd6da0 shows the same warning Jan 28 13:46:42 Tofe: Had some debugging time in the plane today, will soon compile so new appinstalld that might fix things a bit Jan 28 13:48:57 flying anywhere fun? Jan 28 14:08:49 Just back home for a change Jan 28 14:09:22 Should put the house on AirBnB not using it a whole lot these days with all traveling Jan 28 14:09:38 where's home? Jan 28 14:18:17 Amersfoort, Netherlands Jan 28 14:33:02 very cool Jan 28 15:01:42 OK seems my builder decided to go down Jan 28 15:01:54 I'll be home in a few hours so can check toen Jan 28 15:01:56 then Jan 28 15:02:20 I guess I need to put Wake On LAN Jan 28 15:02:38 travel safe Jan 28 15:40:25 codepoet80: Thanks. I think I might be a bit closer to solving the appinstalld mystery but we'll see later today Jan 28 15:41:32 Turned out it was making calls that didn't exist in luna-sysmgr/appmgr but exist in SAM in webOS OSE, so needed to hack around them a bit for now until we find some migration path to SAM Jan 28 15:53:54 SAM is luna-sysmgr replacement in OSE basically and LG split out a lot of the functionalities into separate components Jan 28 15:54:42 But a lot of the luna-sysmgr things that were "mobile" specific seem to have been removed as well, so migration is not straightforward Jan 28 15:55:35 I guess we at some point might strip luna-sysmgr from the stuff that's also in SAM and let them exist side by side. Or add the functionality to SAM Jan 28 19:07:52 Tofe: Did you check the luna-downloadmgr? Do the API calls work for you? I couldn't get them to work at my end earlier this week Jan 28 20:24:37 Hmmz still no dice with appinstalld2: https://bpa.st/FWCA Jan 28 20:59:57 I can see I was on little coffee... If I place ApplicationInstallerUtility in /usr/sbin I should update that in appinstalld2 of course too.. Thought I did it, but seems I have 2 forks locally, probably did it in the wrong one Jan 28 21:10:02 Ah wait it's still in /usr/bin.... Jan 28 21:10:13 Must be missing a recipe update Jan 28 21:15:39 Nope I did that, weird Jan 28 21:15:53 https://github.com/webOS-ports/applicationinstallerutility/commit/7d97bf4fb3dfe3054a0c34975f16bc380a3f815b Jan 28 21:22:55 OK I'm getting somewhere now.... Really need to get back on coffee, been almost without for the past 2 weeks Jan 28 21:43:14 Herrie: well it looks like LunaDownloadMgr isn't started at boot Jan 28 21:43:20 so that doesn't help Jan 28 21:44:22 Herrie: but I'll travel a bit by train this week-end, I'll check on the pinephone I bring with me Jan 28 21:57:33 Tofe: Ah OK Jan 28 21:57:37 That could be it just... Jan 28 21:57:45 Not really sure what's up with this appinstalld2 issue Jan 28 21:59:32 I did quite a bit of that today too ;) Car, plane, bus, train, bus, train, train, train, bus ;) Jan 28 21:59:34 Long trip Jan 28 22:02:55 Hmmz OK install now seems to work but crashes the kernel afterwards Jan 28 22:03:02 I guess that's the SMACK bits maybe they added Jan 28 22:04:05 https://paste.ubuntu.com/p/7V4BfSVVBY/ Jan 28 22:04:06 Herrie: I don't know how you manage that kind of trip without coffee :) Jan 28 22:04:25 Well I had 1 espresso somewhere in the afternoon around 16:30 I think Jan 28 22:04:34 And that on 4,5 hrs of sleep, quite proud of myself LOL Jan 28 22:04:38 appinstalld2 crashes the kernel ? wow Jan 28 22:05:23 I had about 1L expresso just sitting at home :) Jan 28 22:06:47 JaMa2: isn't expresso supposed to be a small quantity?... Jan 28 22:07:01 Tofe: https://github.com/webOS-ports/appinstalld2/blob/6b2da0ad7f1c70e522638180db8467e1c180db6b/files/conf/appinstalld-conf.json.in#L20-L23 Jan 28 22:07:11 I guess when I get rid of SMACK it might work better Jan 28 22:07:12 that's why I ask wife every 5 minutes to bring another one :) Jan 28 22:07:14 I already removed other bits Jan 28 22:07:32 Herrie: for the crash I'm a bit clueless; the crash occurs in ext4 kernel code, not sure what kind of app bug could trigger that Jan 28 22:07:36 My laptop is 1m from the coffee machine Jan 28 22:07:58 smack maybe? Since it's some security module Jan 28 22:08:08 I can remove it easily from the steps to test anyway Jan 28 22:08:19 wife got new coffee machine last year and I still don't like to operate it, it's all GUI without CLI.. Jan 28 22:08:47 Herrie: maybe you're right yes, it can't be just a coincidence Jan 28 22:10:04 JaMa2: ah, those new coffee machines with no GPIO and no support of RFC 2324 Jan 28 22:23:09 HTCPCP/1.0 :) Jan 29 01:05:38 hey herrie, saw the update, thanks! Jan 29 01:05:40 codepoet-x1: I just updated the issue in the tracker Jan 29 01:05:47 :) Jan 29 01:05:51 Was a tough cookie, but seems we're getting close Jan 29 01:05:59 Still some fine tuning, but well Jan 29 01:06:10 anyway for me to patch it in? or do i have to wait and do a full build? Jan 29 01:06:28 I uploaded a full emulator image with all included Jan 29 01:07:43 ok, sweet. i'll try it out tomorrow Jan 29 01:08:38 It basically includes these changes: https://github.com/webOS-ports/appinstalld2/pull/4/files, https://github.com/webOS-ports/luna-appmanager/pull/31/files and quite some other changes in meta-webos-ports too Jan 29 01:09:54 wild how intense that was! Jan 29 01:09:57 thanks for working on it Jan 29 01:12:31 Anything that touches such core functionality entails quite a bit usually Jan 29 01:12:52 Here's my current Enyo project: http://wosa.link/image.php?am9ud2lzZXx6N2U5ODlkNw Jan 29 01:12:59 Hardly ever a quick 1:1 replace and it's working usually lots of tinkering Jan 29 01:13:15 its just web services, so it shouldn't have any trouble on LuneOS Jan 29 01:13:26 Nice Jan 29 01:13:35 Well nothing should have trouble on LuneOS Jan 29 01:13:43 But we need to test and debug Jan 29 01:13:49 Cookie issue next I guess Jan 29 01:14:19 ya, that one is gonna be a problem :-/ Jan 29 01:14:39 BUT if permissions are being parsed right now, maybe they'll just work...? Jan 29 01:15:56 Seems not, but it could be something else with sandboxing or such Jan 29 01:16:00 I guess we can just debug Jan 29 01:16:09 But let's take one at the time Jan 29 01:16:16 Also downloadmanager is almost back too Jan 29 01:16:28 agreed! Jan 29 01:16:32 and good news :) Jan 29 01:17:41 well it's beer o'click here. have a good night, and thanks again Jan 29 01:17:41 Most things aren't hard to fix, there's good logging in the OS Jan 29 01:18:00 Yeah had short nights past days, so going to catch up on some missed hours ;) **** ENDING LOGGING AT Sat Jan 29 02:59:57 2022