**** BEGIN LOGGING AT Tue May 21 09:33:43 2019 May 21 09:35:35 Seems the logger is back online ;) May 21 09:35:54 Jenkins too May 21 09:36:22 webos-ports.org and webos-internals.org too May 21 09:37:32 cool, will trigger the builds soon May 21 09:38:36 Not sure what's all running & whats not May 21 09:38:40 But seems some things are back May 21 09:38:46 novaldex: ^ May 21 10:06:56 morning all, yes we've been working crazily May 21 10:09:18 suffice to say, all our shuffling didn't quite work as we planned & we had to do some more May 21 10:09:24 but yes, things should be up again May 21 10:09:37 and i've even been able to patch Jenkins (well, it was effectively offline for a bit..) May 21 10:10:11 novaldex: Any idea what was the issue in the end? May 21 10:10:18 Or the root cause is still not solved? May 21 10:11:14 we know we have a storage problem to one of our hosts, so we're running slightly reduced by having VMs on a different host May 21 10:11:29 that will need resolving, but it'll probably be next week at the earliest May 21 10:11:44 OK, well hopefully by that time the new disks arrive as well :P May 21 10:11:57 during the migration one of our DNS servers had a hissy fit & fell over, hence nothing was resolving for a bit May 21 10:12:04 that's what we fixed today May 21 10:14:01 new disks, hope so! May 21 10:45:59 prserv wasn't running again since kaylee reboot, I've started it now May 21 10:53:26 Herrie: I've merged other 3 PRs, can you please rebase https://github.com/webOS-ports/meta-webos-ports/pull/353 or should I? May 21 11:33:03 I've rebased it now May 21 11:36:51 JaMa, was gone for a bit. Now having quick lunch, should be back behind a pc in 10 or so mins May 21 11:38:03 np, all merged now May 21 11:46:10 Great :) May 21 11:47:51 JaMa: Everything pending incl. qt5 updates and meta-oe has been bumped/merged now? May 21 11:49:27 yes, I'll do the builds now May 21 11:49:36 after fixinig one more issue in master signatures May 21 11:49:55 http://jenkins.nas-admin.org/job/LuneOS/view/unstable/job/luneos-unstable_workspace-compare-signatures/48/console May 21 11:54:53 JaMa: Happy to test the fix locally as well ;) May 21 11:54:57 Get my builder working a bit :P May 21 11:55:40 it's in jansa/master, simple enough for my builder to test :) May 21 11:56:02 even with load 30+ it should finish soon May 21 11:58:00 Not sure why you did https://github.com/webOS-ports/meta-webos-ports/commit/3e8c286ce9f4ed914b64ebd00eb33b9c2ba6e138 and https://github.com/webOS-ports/meta-webos-ports/commit/d0fcf000f3cf038e8d0d6cee9d27665ed64e30a8 ? May 21 11:58:05 No description there ;) May 21 12:01:38 that's only in jansa/master, right? May 21 12:02:53 it's to include all debug symbols and dev packages in the image (useful when debugging something, but extremly big resulting image which doesn't fit on some of our targets, that's why I've reverted it later after the qemux86 issue was fixed) May 21 12:04:32 20:17 <+JaMa> it does, quite a bit 2.9G Mar 30 14:00 luneos-dev-emulator-qemux86--1.0-r0-20190330062814.zip May 21 12:04:35 20:17 <+JaMa> is from 7.7G vmdk May 21 12:04:38 20:18 <+JaMa> without -dbg and -dev package the size difference is still significant May 21 12:04:41 20:18 <+JaMa> 3.7G Mar 30 05:57 luneos-dev-image-qemux86.rootfs.ext4 May 21 12:04:43 20:19 <+JaMa> 1.4G Mar 30 05:59 luneos-dev-image-qemux86.rootfs.wic.vmdk May 21 12:04:46 20:19 <+JaMa> 454M Mar 30 06:03 luneos-dev-emulator-qemux86--1.0-r0-20190330042119.zip May 21 12:05:06 16:28 <+JaMa> I haven't tried master + 5.12 + new libhybris yet, because dbg-pkg included in my local build makes the image too big for sideload May 21 12:05:09 16:29 <+JaMa> sideload-host file size -1583068846 block size 65536 May 21 12:05:11 16:29 <+JaMa> file has too many blocks (4294943141) May 21 12:05:14 small refresh from Mar30 Apr02 May 21 12:05:46 Ah ok:) May 21 12:05:51 Clear May 21 12:05:56 I remember that bit May 21 12:06:05 Seems I broke something with the rdxd somehow May 21 12:07:33 OK I think I found it, let me quickly check & PR May 21 12:08:21 agreed, looks like missing PROVIDES May 21 12:09:55 Yeah cleaned the recipe a bit too much :P May 21 12:09:58 Herrie: rwhitby says he'll forward on any donations May 21 12:10:14 Herrie: why yave you added webos_machine_impl_dep inherit in rdxd? May 21 12:10:40 There we go: https://github.com/webOS-ports/meta-webos-ports/pull/358 May 21 12:11:23 JaMa: Yeah that shouldn't be there I guess May 21 12:11:28 Copied from another recipe probably :S May 21 12:11:52 yes we don't want to make pmlogdaemon MACHINE_ARCH May 21 12:12:31 JaMa: Fixed May 21 12:12:39 Force-pushed the change, so it's in same PR now May 21 12:13:26 novaldex: Great... any word if something arrived from rwhitby? May 21 13:01:03 JaMa: https://github.com/webOS-ports/meta-webos-ports/pull/359 May 21 13:01:45 I saw the license one before, must have forgotten to change that one when migrating changes from builder to machine where I did the PR. May 21 13:02:05 The qtlocation-luneos-plugin I didn't see before but popped up now. I guess something changed at qt side here May 21 13:02:28 your local.conf changed, but the fix isn't correct May 21 13:03:20 Ah wait May 21 13:03:28 Yeah I need to revert that one in local.conf :S May 21 13:10:11 Not sure the position suffix is required, we didn't have that before. My local.conf is back to original, which means qtbase is rebuilding May 21 13:10:16 Should know quick enough ;) May 21 13:10:18 Few mins May 21 13:11:21 it would be nice to fix with the variable, in case we will use e.g. qt6 eventually May 21 13:11:45 but I really hope that we'll be able to upgrade all qt* at the same time, so we could keep them in normal native bindir May 21 13:12:07 JaMa: Yeah will replace it anyway with the right variable from the bbclass May 21 13:12:15 Just I didn't write the original recipe either ;) May 21 13:12:45 Just fixing what I come across along the way. Happy for suggestions on how things can be better of course :) May 21 13:12:57 Don't mind to be the janitor ;) May 21 13:13:18 the bbclass is already inherited through qmake5 May 21 13:14:06 Yes May 21 13:18:43 do you have some vmdk image already built on your builder? Can you try to compress it with zip, tar czvf, tar cjvf and share the time needed? I'm comparing the same on my builder to see if we should replace zip-native with something faster / more efficient May 21 13:19:43 and so far no big gain, 2:14 with zip, 1:40 with gzip, 3:44 with bzip2 May 21 13:20:44 but I still have OSE build running on background so it's even slower than it should be May 21 13:27:39 JaMa: Let me try to see the difference at my end May 21 13:30:40 I have some qtwebengine building in background now. I'll let it finish first ;) May 21 13:35:14 ok May 21 13:36:49 it gets interesting here only when using pigz or pbzip2 https://paste.ubuntu.com/p/gfz4qV35zN/ May 21 14:07:53 I get: https://bpaste.net/show/7bb8d6a8d03b May 21 14:08:30 So pigz seems to help here as well May 21 14:21:08 Herrie: ok, thanks May 21 14:22:50 Just I needed to install both pigz and pbzip2 on my machine, so I guess if we go down that route, same on our Jenkins builders ;) May 21 14:28:04 no, I'm adding dependency on pigz-native May 21 14:28:25 can you check the size of your vmdk (or all files)? just to compare with my local build? May 21 14:28:39 in warrior it should be almost the same, but my build was 32bit May 21 14:30:24 https://github.com/webOS-ports/meta-webos-ports/commit/98aff30c0b817300e2fc5f9f76fc8cb2587413da May 21 14:32:29 https://bpaste.net/show/b806bd4ec76d May 21 14:32:41 This is Warrior, but 64 bits May 21 14:32:48 pbzip2 gives best size May 21 14:33:03 together with bz2 May 21 14:39:41 thanks commit message updated May 21 14:42:08 :) May 21 14:48:05 Herrie: no, not yet, I asked ka6sox to do some checking to make sure it was still an acceptable method May 21 14:49:19 thinking of Necunos May 21 14:52:27 novaldex: Yes May 21 14:52:28 Good May 21 14:53:03 * halfhalo yawns in his annual irc visit May 21 14:53:20 hey halfhalo May 21 14:53:35 Hola May 21 14:54:13 How is Life at LG nowadays? May 21 14:58:23 halfhalo: Long time no see ;) May 21 15:00:36 Si senor! May 21 15:00:56 JaMa: On Uriparser: Agree. Same would apply to pidgin 2.13 May 21 15:01:06 If you could send it upstream to mailing list would be great May 21 15:03:07 Same for libpng May 21 15:03:18 I just did these because they were due for an update ;) May 21 15:03:28 I didn't check upstream to be honest for them May 21 15:13:26 halfhalo: almost the same, it still not responding etc :) May 21 15:13:54 Herrie: the easiest way is to check upstream in https://layers.openembedded.org/layerindex/branch/master/recipes/ May 21 15:14:13 Herrie: see jansa/master I'll add meta-luneos-backports-2.8 for recipes backported from newer Yocto May 21 15:14:26 to make it easier to cleanup meta-lune* when we upgrade to newer Yocto May 21 15:14:42 it's the same as what I've added to meta-webosose basically May 21 15:17:09 JaMa: Makes sense :) May 21 15:25:32 I'll push my changes before proper test to get them in the current build, I promise to fix possible breakage soon :) May 21 15:37:26 JaMa: Added pidgin 2.13 in there as well now May 21 15:37:35 With updated patch which seems I forgot :P May 21 15:37:37 Failed on fuzz May 21 15:39:06 Local build running with these changes May 21 15:46:25 JaMa: What was blocking GCC 8? Our kernels? May 21 15:47:44 kernels on android devices yes May 21 15:49:41 JaMa: OK May 21 15:52:18 Herrie: "Pending sending pidgin 2.13 to upstream for 2.8 release." May 21 15:52:35 this isn't how that layer works, it should be moved there only after it's accepted upstream May 21 15:52:47 see description in https://layers.openembedded.org/layerindex/branch/master/layer/meta-luneos-backports-2-8/ May 21 15:53:37 let me get it upstreamed first, then I'll move it to backports layer May 21 15:56:10 JaMa: OK ;) May 21 16:20:17 Herrie: the kernel + gcc-8 issue described a bit here https://wiki.webos-ports.org/wiki/Repository_Layout May 21 16:27:21 JaMa: Yes I remember, Will have a look and see if I can get something to build. The fixes aren't indeed always very straightforward May 21 16:27:39 And that would only allow building. Could still give runtime issues (quite common with GCC upgrade...) May 21 16:28:21 Seems your changes to Warrior triggered a pretty much complete rebuild again somehow (or my local.conf ;)). Anyway it's running. May 21 16:30:06 the libpng causes quite a lot May 21 16:30:25 that's why I wanted to move it to the backports layer before it's triggered on slow jenkins May 21 16:31:01 JaMa: Ah OK May 21 16:47:40 Herrie: and what wasn't invalidated by your libpng (e.g. qtbase and everything depending on it) was invalidated by my cleanup of EXTENDPRAUTO you imported from webosose :) May 21 16:48:05 but luckily it causes mostly just do_package to be re-executed, so without rm_work it shouldn't take too long May 21 16:49:06 jenkins is still working on unstable sstate check http://jenkins.nas-admin.org/job/LuneOS/view/unstable/job/luneos-unstable_workspace-compare-signatures/51/console May 21 17:10:19 If we were to cleanup blacklist, how to go about this? May 21 17:10:55 See what's blacklisted but building in meta-oe/oe-core and unblacklist it or... ? May 21 17:55:52 yes, but why do you want to clean it now? May 21 17:56:21 as mentioned in commit message I think it's good to have a bit smaller world (until we need something which is blacklisted) May 21 17:57:09 I was planing to eventually drop all entries and start again, but that's really boring way to spend time May 21 17:58:17 and also quite slow process, because you fix or blacklist the first couple failing, then it will detect few more failing + many more blocked now because their dependencies were blacklisted May 21 17:59:02 so you rinse and repeat until bitbake world builds for all supported MACHINEs without any errors from build failure nor dependency tree parsin May 21 17:59:29 when I was doing this for recipes broken with multilib few weeks ago it took maybe 30+ iterations to catch all of them May 21 17:59:50 so I don't plan to do that again any time soon :) May 21 19:23:04 2 small issues in emulator appliance, I'm testing the fix will push it soon May 21 19:23:48 JaMa: what kind of issues? May 21 19:28:00 JaMa: I noticed that for the 64 bits version the ovf states it's 32 bits Linux. Which I don't think really makes much difference, but thought it might be good to mention just in case May 21 19:28:28 Technically it's not correct at least May 21 19:28:34 tar doesn't follow symlink (so resulting tar.gz is _really_ small now) and forgot to redirect pigz output May 21 19:28:38 so do_deploy fails May 21 19:29:41 JaMa: Ah ;) May 21 19:29:49 Yeah that doesn't help ;) May 21 19:29:53 fixed in meta-webos-ports May 21 19:31:24 http://paste.ubuntu.com/p/pfNvSRHpPR/ is the current difference in .ovf, but I was planing to refresh it from new VirtualBox once we all have Vbox-6 May 21 19:36:26 this part https://github.com/webOS-ports/meta-webos-ports/commit/a9df8945f3e8c2934c65700c52cd9a0520769853 might fix it May 21 19:39:21 Already running mine in VBox 6 May 21 19:54:15 and this is the refreshed config: https://github.com/webOS-ports/meta-webos-ports/commit/f33a5c9dda859290962b7f64a5cf300589f87929 May 21 19:54:32 will split it into refresh and rename as GH doesn't detect it as a rename May 21 19:55:49 https://github.com/webOS-ports/meta-webos-ports/commit/70e0056750d97be112f26f88c394259805de7bfe May 21 20:03:06 but even without any of these ovf modifications, it doesn't start in my VBox (my local build) May 21 20:04:06 sorry, my bad, I was testing luneui-example image, which I know is broken May 21 20:04:10 the full image wasn't built yet May 21 20:21:57 JaMa: It could very well be I broke something somewhere as well, but should be something minor to fix May 21 20:27:17 there might still be some issue, but my local test doesn't prove anything May 21 20:31:19 Bedtime here, feel some bug coming up, so trying to get an extra hour of sleep :p May 21 20:31:31 Some testing tomorrow :) May 21 21:19:59 :) May 21 21:20:01 gnight May 21 22:41:29 signature issues all fixed in unstable: http://jenkins.nas-admin.org/job/LuneOS/view/unstable/job/luneos-unstable_workspace-compare-signatures/53/console **** ENDING LOGGING AT Wed May 22 03:02:00 2019