**** BEGIN LOGGING AT Wed Oct 30 02:59:57 2019 Oct 30 08:10:47 Tofe: When you got a chance, can you test http://build.webos-ports.org/luneos-stable-staging/wip/images/tissot/ ? Oct 30 08:14:25 yes, that'll be tonight Oct 30 08:24:47 Tofe: Thnx Oct 30 08:24:50 I'll test others in meanwhile Oct 30 08:25:29 Hi Oct 30 08:26:13 Herrie: didn't you run feeds_new_staging job for 056 before doing the 056 builds? Oct 30 08:27:30 http://build.webos-ports.org/luneos-stable-staging/056/ is completely empty Oct 30 08:28:05 should I fix it by adding 057 and updating changelog for 056 as accidentally empty? Oct 30 08:30:13 ah, looks like there's a missing step in our how-to :) Oct 30 08:32:35 JaMa: I guess not ;) Oct 30 08:32:46 Yes we need to add a step it seems Oct 30 08:35:31 https://www.webos-ports.org/wiki/Pre_and_Post_Release_To_Do looks correct Oct 30 08:35:57 first build the images in wip, then run the new-staging jobs to "close" the wip Oct 30 08:37:09 I can try to revert the 056 changes, but I didn't do it for long time and I might break it even more (there are some steps to undo on milla as well as bonaire) Oct 30 08:37:29 creating new 057 might be easier fix Oct 30 08:41:12 I don't see any need to stay on 056 Oct 30 08:41:38 it's just an internal counter, right? Oct 30 08:48:34 057 is fine with me Oct 30 08:48:40 Yeah just internal counter Oct 30 09:01:04 should I merge all 4 to public feed and release 057? Oct 30 09:03:46 I'm also updating the tags in webos-ports-setup, I've already tagged the release before last update from Herrie Oct 30 09:05:37 + 2830ba1...e03feec 2019-10-24-055-eggnoglatte -> 2019-10-24-055-eggnoglatte (forced update) Oct 30 09:05:40 + 1b6721b...688e8e6 eggnoglatte -> eggnoglatte (forced update) Oct 30 09:05:48 done and pushing to public in progress Oct 30 09:06:07 http://jenkins.nas-admin.org/job/LuneOS/view/stable/job/luneos-stable_feeds-sync-to-public/50/console Oct 30 10:39:54 JaMa: Sorry was in meeting. What you mean exactly by "should I merge all 4 to public feed and release 057?" Oct 30 11:17:19 this http://jenkins.nas-admin.org/job/LuneOS/view/stable/job/luneos-stable_feeds-sync-to-public/50/console Oct 30 11:17:22 JaMa: What was the surprise? I didn't see anything too shocking in the new release? Except for that it's 2.0 and some new roadmap? Oct 30 11:17:46 it rsyncs these 4 feeds from luneos-stable-feed to luneos-stable Oct 30 11:19:04 surprise in OSE 2.0 release? not much yeah, just rpi4 and few features for auto which might explain the launcher/lsm changes you were asking about Oct 30 11:20:41 JaMa: Thnx for cleanup Oct 30 11:23:25 Well seeing that they're @ Chromium 72 and ahead of us is a surprise :D Oct 30 11:24:48 After the release will give a go to see if I can get the OSE WAM & Chromium in our build. I did some quick sideloading of some of our apps yesterday and it looked pretty OK actually on OSE emulator. Some things wouldn't start due to missing bits, but Enyo 2.0 apps were all fine in general. Just Enyo 1.0 core-apps had some issues. Oct 30 11:26:43 But this was mainly due to missing services, permissions etc Oct 30 11:35:34 should we release the current stable (057) now? Oct 30 11:35:48 and can I merge zeus upgrade to testing? :) Oct 30 11:36:27 without Qt upgrade for now (it's more useful to upgrade Yocto), because Qt 5.12 is LTS and 5.13.2 still wasn't release yet Oct 30 11:36:59 I've already tested emulator and it seems to work as well as last testing build Oct 30 11:38:14 JaMa: Let me check Tenderloin & Hammerhead (flashing Tenderloin now). Should know the answer in about 30 minutes after some testing Oct 30 11:38:23 Then we should be good to close this chapter and move forward Oct 30 11:38:38 cool, thanks Oct 30 11:39:17 I hope there's not funny systemd and glibc issues causing havoc with 3.4 kernels, otherwise we should be fine hopefully Oct 30 11:39:30 I guess we should also look into GCC8 again? Seems we're on 7.3 now? Oct 30 11:46:06 we are? I thought you've already merged the upgrade to gcc8 already, let me check Oct 30 11:46:36 my last warrior build was with gcc8 Oct 30 11:47:00 JaMa: Ah OK Oct 30 11:47:04 http://jenkins.nas-admin.org/job/LuneOS/view/stable/job/luneos-stable_hammerhead/lastBuild/consoleFull also shows gcc 8.3 so we're good there Oct 30 11:47:15 Ah OK Oct 30 11:47:18 Good :) Oct 30 11:47:21 1 worry less Oct 30 11:47:47 Tenderloin seems OK, flashing Hammerhead now. Oct 30 11:48:00 unstable is on 9.2, but I haven't seen any runtime issues caused by gcc/glibc, this time Oct 30 11:48:12 except in OSE webruntime :) Oct 30 11:48:41 Tofe: You OK with finalizing the release images, seeing that the issues on Qemux86-64, Hammerhead and Tenderloin would be solved? I wouldn't expect anything new and odd on Tissot then? Oct 30 11:49:09 Herrie: go ahead! Oct 30 11:59:21 Hammerhead seems OK too Oct 30 11:59:39 JaMa: Let's go ahead :) Oct 30 12:03:27 ok, added in jenkins queue after unstable mido build Oct 30 12:03:51 what about switching testing to zeus? Oct 30 12:04:23 JaMa: I have no issues with this in general Oct 30 12:04:29 Any important updates? Oct 30 12:05:06 https://lists.yoctoproject.org/pipermail/yocto-announce/2019-October/000169.html Oct 30 12:06:13 nothing super important for us I would say, but the hash equivalence would help a lot with build times later when enabled (after fixing some remaining issues) Oct 30 12:06:40 and I've enabled uninative with zeus, which allows us to use different host distros while reusing sstate from jenkins Oct 30 12:07:51 I just hope it won't break yet our old kernels Oct 30 12:08:30 Herrie: can you test unstable build on tenderloin when you have some spare time? Oct 30 12:09:47 JaMa: can the "Build change equivalence" detect unneeded rebuild when modifying mesa/libhybris ? Oct 30 12:11:07 not right now Oct 30 12:11:27 but it should detect when something was rebuilt without any modification in the output Oct 30 12:12:15 do you know how that works ? it hashes the "images" output folder ? Oct 30 12:14:12 I think it hashes the content of sstate archives (so packages-split for package, sysroot content for populate_sysroot etc) and then it allows to re-use binary equivalent sstate archive instead of triggering casade rebuilds of whatever depends on this Oct 30 12:14:43 ah, ok, nice Oct 30 12:15:09 e.g. if you rebuild zlib, because of small change which doesn't affect its binary output, you won't need to rebuild everything which depends on zlib, just because sstate signature was changed Oct 30 12:16:11 but there are still some smaller issues (like not rebuilding something which should be rebuilt in some corner cases or still rebuilding too much), so it's not enabled by default, but it's definitely step in right direction Oct 30 12:17:48 so typically, it will rebuild a direct impact, but if that rebuild ends with the same result, it won't propagate it Oct 30 12:19:18 JaMa: Tenderloin one seems to be from 10-Oct, is that correct? Oct 30 12:22:57 should be good enough for test Oct 30 12:23:05 Tofe: yes Oct 30 12:23:37 another build is in queue now, but it will be a while until it gets to tenderloin Oct 30 12:24:26 or there should be a bit newer hammerhead build for unstable Oct 30 12:24:53 Oct 28th Oct 30 12:25:14 Yeah will test Hammerhead first Oct 30 12:25:25 Also on 3.4 kernel I guess in this build Oct 30 12:26:00 yes linux-lg-hammerhead-3.4.0+gitrAUTOINC+2cb0f6675c Oct 30 12:32:17 Well if one 3.4 works other one is likely to work as well or with a very minor patch Oct 30 12:50:18 JaMa: Hammerhead boots to GUI Oct 30 12:51:07 Audio works, wifi seems to work Oct 30 12:51:14 Rotation works Oct 30 12:51:29 This is with luneos-dev-package-hammerhead-unstable-0-87.zip Oct 30 12:53:45 Vibration works as well Oct 30 12:55:20 sounds good, should I stop unstable builds and do it in testing instead? Oct 30 12:55:29 Yeah guess so. Oct 30 12:56:25 Looks in same or at least very similar state v.s. what's in Warrior and release now Oct 30 12:57:30 ok, from now please use meta-webos-ports zeus instead of warrior Oct 30 12:58:00 OK :) Oct 30 12:58:11 JaMa: Can you cherry pick the PR's that are pending for Warrior? Oct 30 12:58:16 will merge the 2 PRs yes Oct 30 12:58:27 NExt release will be "Eiskaffee' Oct 30 12:58:32 "Eiskaffee" even Oct 30 13:05:10 with release out, I'll delete the testing/unstable builds for pyro, rocko, sumo, thud, OK? Oct 30 13:06:00 or even warrior as well Oct 30 13:07:07 Yeah why not Oct 30 13:07:17 If need be we could recreate them when needed Oct 30 13:07:46 1.4G luneos-testing-pyro/ Oct 30 13:07:46 9.2G luneos-testing-sumo/ Oct 30 13:07:46 1.5G luneos-unstable-rocko/ Oct 30 13:07:46 1.1G luneos-unstable-sumo/ Oct 30 13:07:49 9.4G luneos-unstable-thud/ Oct 30 13:07:51 1.6G luneos-unstable-warrior/ Oct 30 13:07:54 104G luneos-testing-warrior/ Oct 30 13:08:18 Tofe: should I keep your manually uploaded pinephone builds in luneos-testing-warrior? Oct 30 13:11:19 46bis ? yes Oct 30 13:11:37 the rest can go Oct 30 13:15:47 JaMa: 46bis should stay until we have a build with the newly merged PR for sensorfw Oct 30 13:16:09 novaldex|away: http://issues.webos-ports.org/ seems down again Oct 30 13:16:11 otherwise there will be no working build for pinephone Oct 30 13:16:31 ok Oct 30 13:17:26 Tofe: How can we build working builds on Jenkins? Oct 30 13:18:19 Herrie: for pinephone ? seeing that sensorfw PR has been merged, it's just starting a testing build Oct 30 13:19:04 shouldn't even take long Oct 30 13:19:29 (but that PR might break other targets, that's why I didn't want it just before the release) Oct 30 13:20:59 Tofe: OK good :) Oct 30 13:25:31 saved quite a bit of space on file server: Oct 30 13:25:34 54G luneos-stable Oct 30 13:25:34 113G luneos-stable-staging Oct 30 13:25:34 0 luneos-testing Oct 30 13:25:34 420M luneos-testing-warrior Oct 30 13:25:37 8.0K luneos-testing-zeus Oct 30 13:25:39 0 luneos-unstable Oct 30 13:25:41 6.6G luneos-unstable-zeus Oct 30 13:25:44 0 luneos-zeus Oct 30 13:25:46 as well as on bonaire: Oct 30 13:41:44 12K luneos-shared Oct 30 13:41:44 1.7G luneos-stable Oct 30 13:41:44 5.4G luneos-testing Oct 30 13:41:44 3.1G luneos-unstable Oct 30 13:41:54 release job finished Oct 30 13:42:14 biggest release yet :) Oct 30 13:42:15 7.2G releases/eggnoglatte Oct 30 13:43:19 /dev/xvdc1 985G 580G 405G 59% /home/jenkins Oct 30 14:12:18 /dev/xvdb1 985G 155G 788G 17% /home Oct 30 14:12:59 looks like new git can now push from shallow clones, that allows to shallow clone buildhistory repos and makes them significantly smaller and faster to handle, will recreate all 3 workspaces with them Oct 30 14:40:55 JaMa: Ooh nice Oct 30 14:41:14 We do shallow clones for the Halium bits, saves quite some downloading ;) Oct 30 14:41:20 Especially when there are tons of branches etc Oct 30 14:44:15 https://github.com/webOS-ports/jenkins-jobs/commit/78a747fd890fbaa724906d2e7499b5d6f2a6f7f3 Oct 30 15:13:12 Ooh nice :) Oct 30 15:25:34 ooh nice Oct 30 16:13:37 JaMa: Should I make zeus the default branch in meta-webos-ports? Oct 30 16:15:50 yes, good idea Oct 30 16:16:24 new testing builds are running now, building from scratch so it will take a while Oct 30 16:18:25 JaMa: Haven't build tested these yet, but should be pretty straight forward: https://github.com/webOS-ports/meta-webos-ports/pull/401 Oct 30 16:19:05 Herrie: pmlog* is going away soon (OSE using journald), not sure if it's worth updating to their version now Oct 30 16:20:46 JaMa: Yeah true, but well, better to get rid of the OWO version ;) Oct 30 16:20:52 Code changes were very minimal anyway Oct 30 16:21:28 Just noticed I forgot them previously somehow Oct 30 16:21:42 ok, just saying that it might be even better to get rid of it completely :) Oct 30 16:23:06 completely? no logs - no problem to choose provider :D Oct 30 16:24:34 just stupid joke Oct 30 16:24:44 but i hate binary logs fwiw Oct 30 16:33:03 it cannot be much worse than current state where something logs to messages, something to legacy-log and everything complains about journald failing to write logs Oct 30 16:44:12 JaMa: Building qemux86-64 at home too Oct 30 16:52:44 JaMa: I agree on log status, it's weird needing to look in 3 different places... Oct 30 16:52:57 At least on our LuneOS devices currently Oct 30 16:53:04 Where we do have journald **** ENDING LOGGING AT Thu Oct 31 02:59:58 2019