**** BEGIN LOGGING AT Tue Nov 29 02:59:56 2022 Nov 29 06:34:21 Morning! Nov 29 06:35:10 Tofe: Well that's quite some improvement already! I have doubts SSD will arrive today looking at current status, but we'll see. Nov 29 07:07:45 Herrie: well, it needs to cross border overseas it seems ;) Nov 29 07:09:17 however, there's often a 24h delay between reality and the webpage status Nov 29 07:20:17 Tofe: We'll see Nov 29 07:40:52 Seems meta-webosose was pushed as well so building webOS OSE 2.19 here now with Chromium 94 Nov 29 09:18:44 ok, old motherboard is on its way back Nov 29 09:26:49 Tofe: Well good the new one works at least :D Nov 29 09:28:43 NodeJS takes forever to build here, not sure what's up ;) Nov 29 09:30:41 Compile for NodeJS takes about 1 hour here it seems for OSE-devel image :S Nov 29 09:31:18 I've pushed my OSE fork as well now, but the diff in kirkstone is quite small now as expected Nov 29 09:31:51 JaMa: Nice! Nov 29 09:32:02 Any idea why my NodeJS might be so extremely slow? Nov 29 09:32:10 It took over an hour to compile.... Nov 29 09:32:17 With -b 16 and -p 16 Nov 29 09:32:37 As per OSE building guide Nov 29 09:33:05 it took 9 minutes for me yesterday, not sure why it would be slow for you Nov 29 09:33:56 don't you have more than 16 threads? Nov 29 09:34:02 Well I do ;) Nov 29 09:34:15 But OSE guide tells me to set it to 16 ;) Nov 29 09:34:41 well that's just wrong Nov 29 09:35:20 OK ;) Nov 29 09:35:52 Any recommendations for what values to use then ? Nov 29 09:36:25 -l for PARALLEL_MAKE and 8 for BB_NUMBER_THREADS Nov 29 09:37:03 ICECC_DISABLED = "1" Nov 29 09:37:03 BB_PRESSURE_MAX_CPU = "1000" Nov 29 09:37:03 BB_NUMBER_THREADS = "8" Nov 29 09:37:03 BB_NUMBER_PARSE_THREADS = "64" Nov 29 09:37:03 PARALLEL_MAKE = "-j 70 -l 140" Nov 29 09:37:06 #PARALLEL_MAKE:pn-qtwebengine = "-j 40" Nov 29 09:37:08 PARALLEL_MAKE:pn-webruntime = "-j 40" Nov 29 09:37:15 is what I use for 64 thread 3970x Nov 29 09:39:08 JaMa: OK I should be able to use similar on 2990WX Nov 29 09:39:12 In local.conf right? Nov 29 09:41:59 yes Nov 29 09:42:26 or ~/webos-global.conf or TOPDIR/webos-local.conf Nov 29 09:42:36 as local.conf is overwritten by mcf Nov 29 09:42:50 JaMa: OK Nov 29 09:42:57 Let me try that Nov 29 09:43:01 For LuneOS same values I guess? Nov 29 09:43:20 yes Nov 29 09:44:21 OK restarted webruntime build Nov 29 09:44:26 Should help a bit I guess with time Nov 29 09:44:43 New laptop should be here anytime soon too :D Nov 29 09:47:36 Herrie: oooh first time I have a lower build time than you ! \o/ Nov 29 09:49:10 hehe, you have faster cores than both of us now :) Nov 29 09:52:40 Tofe: Well not for long :P Nov 29 09:52:56 webruntime seems will take about 30 mins here Nov 29 09:53:15 JaMa: about OSE recipes you were talking with Herrie on Sunday: do you think overriding directly their recipe, plus some bbappend and bbclass overrides in meta-webos-ports ? Nov 29 09:53:25 +would work Nov 29 09:53:30 Not sure that's "normal" Nov 29 09:53:49 JaMa: Is there something like "uninherit" in bb? Nov 29 09:54:30 Herrie: I was thinking about an empty bbclass of the same name in our meta layer Nov 29 09:54:47 It should take priority, if I understand correctly Nov 29 09:56:04 Tofe: Ah OK, yes also true... Nov 29 09:56:09 Didn't think about that Nov 29 09:56:44 I'd prefere a less hacky way, but not sure there's one Nov 29 09:58:11 you can override whole bbclass in some other layer, but you cannot uninherit it and you cannot e.g. undo :remove as well Nov 29 10:00:05 e.g. recently discussed (elsewhere) meta-webos/recipes-qt/qt6/qtbase_git.bbappend:PACKAGECONFIG:remove = "widgets" Nov 29 10:00:17 you cannot undo without BBMASK of this .bbappend and introducing your own Nov 29 10:00:33 ah, I didn't think of that Nov 29 10:01:00 though maybe we're fine with the current :remove in ose Nov 29 10:02:04 there is still more :remove in OSE than I would like to see: meta-webosose $ git grep :remove | grep -v WEBOS_PREFERRED_PROVIDER_FOR_BASH | grep libc-glibc -v | grep -v .patch: | wc -l Nov 29 10:02:08 67 Nov 29 10:02:45 mostly because it's used only as a base for LGE internal builds and which doesn't cause issues in internal builds is not really considered when adding :remove Nov 29 10:03:27 they even used BBMASK in some places just to undo some :remove Nov 29 10:03:51 and then re-introduce it in some other build which depends on both layers Nov 29 10:04:18 as we have 4 DISTROs each depending on each other with OSE being the "base" Nov 29 10:04:29 I see, a bit of a mess Nov 29 10:06:43 yes, everything is possible but with some mess and at some point the mess is worse than forking it Nov 29 10:08:15 with fork you can at least clealy see the diff with git and even rebase if some of the changes are upstreamed, but with a lot of bbappends, overlayed recipes and bbclasses and BBMASKs, you can pretty much trust only bitbake -e to see what's going on Nov 29 10:09:37 but we'd have tons of merges at each rebase, I guess Nov 29 10:10:06 ah, maybe not if we just empty the annoying bbclass instead of removing the inherit clause Nov 29 10:11:06 "annoying bbclass" like what? Nov 29 10:11:21 I think we don't want the submission one, do we ? Nov 29 10:12:18 Tofe: Well the enhanced submission one is actually easier when we want to update OSE components Nov 29 10:12:19 Imho Nov 29 10:12:21 well, that's up to you, but without this WEBOS_VERSION variables won't work and all recipes will need to be modified anyway to set PV and SRCREV Nov 29 10:12:52 JaMa: That's what I did previously and it was quite tedious to set PV and SRCREV everywhere :P Nov 29 10:12:52 and that will be PIA to update in .bbappends Nov 29 10:14:15 I rarely imported recipes from webosose actually, it's mainly Herrie who did most of the work on that aspect :p So I don't know that much about what we need to change when importing something currently Nov 29 10:15:18 Tofe: Mainly add the PV, SRCREV and remove some of the inherits is what I did Nov 29 10:15:40 Herrie: "Your parcel has crossed the border ; Time of delivery: today" they're quite optimistic Nov 29 10:16:08 Tofe: Yeah I know.... Nov 29 10:16:10 I doubt it... Nov 29 10:16:27 wait&see Nov 29 10:17:23 UPS had 10:45 to 14:45 as delivery window, now says "The delivery date will be provided as soon as possible.", but shows "out for delivery". Last time UPS guy didn't come to the house and just dropped it @ depot Nov 29 10:18:44 carriers are a lost cause Nov 29 10:20:18 DHL & PostNL are pretty OK here usually Nov 29 10:20:40 UPS used to be good, but I noticed they have a new delivery guy here which isn't that good as previous one Nov 29 10:24:36 The status updates from PostNL are weird. With other carriers often can see "where" it's located in terms of city etc Nov 29 10:25:17 yes, and LaPoste can't tell where the package is currently :p Nov 29 10:28:03 with modified recipes you would still need to set own WEBOS_VERSION and SRC_URI in .bbappend push rebased annotated tag for new submission from upstream Nov 29 10:30:00 "quite tedious to set PV and SRCREV everywhere" with WEBOS_VERSION it's pretty much just (PV, SRCREV) = WEBOS_VERSION.split('_'), the migration from P_V to WEBOS_VERSION was more annoying :) Nov 29 10:31:36 rebasing (re-creating this with some scripts) for over a year was pain https://github.com/webOS-ports/meta-webos-ports/commit/b4c3946d8bcf79a83c06c051e163539088c9f5b8 Nov 29 10:31:39 sorry I forgot, but why don't we use WEBOS_VERSION on our copy of webosose recipes ? Nov 29 10:32:35 because it forces you to create annotated tags every time you want to change SRCREV Nov 29 10:32:52 not a bad thing, just more tedious than simple SRCREV bumps we were using Nov 29 10:33:42 do we often use more recent upstream versions than the last tagged one? Nov 29 10:34:31 I'm looking at recipes-webos/cjson/cjson.bb in your commit, we don't have any bbappend there I think, so I'm not sure what I'm missing Nov 29 10:45:16 which commit? the one from 2013? Nov 29 10:45:48 yes :) Nov 29 10:46:32 but I feel I should first have a more serious look at what we modify in our copies, before bothering you with obvious question Nov 29 10:48:43 there was a plan (which I've already forgot) to use WEBOS_VERSION in WP as well (based on https://github.com/webOS-ports/meta-webos-ports/commit/70e6faa968a8be2d92a8d2d3f14467e3717b3a5a), but then we decided to use PV + SRCREV like every other OE layer Nov 29 10:49:44 and did https://github.com/webOS-ports/meta-webos-ports/commit/1a748c2d957b6b82002b18b8105befec09cb3f32 Nov 29 10:52:55 that's a huge commit :) Nov 29 11:12:15 nodejs build with webruntime building in parallel (OSE image build with mickledore): Nov 29 11:12:18 webosose/mickledore $ grep Elapsed BUILD/buildstats/20221129093009/nodejs-16.18.1-r0/do_compile Nov 29 11:12:21 Elapsed time: 884.81 seconds Nov 29 11:17:09 We used to use tags in the past for our components, but dropped that due to manpower mainly I think Nov 29 11:17:27 There should be no cases where we use newer tags vs OSE for their components Nov 29 11:18:08 but we apply new commits in some of the components, right? Nov 29 11:18:26 so we would need to tag them with new submission Nov 29 11:20:32 JaMa: The stuff we add is usually LuneOS specific, so not upstreamable, so we could suffice with patches Nov 29 11:21:29 It shouldn't be numerous in most cases either, just a few patches Nov 29 11:21:39 I don't expect we get series of 40 patches ;) Nov 29 11:21:53 2-3 for most components, some others maybe 10 max Nov 29 11:31:00 heh build almost like mine https://youtu.be/xsA-yv0lC6w, the same case, cpu, mb, similar AIO, just a bit weaker PSU and GPU **** ENDING LOGGING AT Tue Nov 29 11:32:10 2022 **** BEGIN LOGGING AT Tue Nov 29 11:34:13 2022 Nov 29 11:48:23 $21000 that's a lot ;) Nov 29 11:49:20 I recall mine was something like EUR 5500 or so Nov 29 11:50:38 and here I am, with my 600€ rig Nov 29 11:52:02 Tofe: Well CPU & Memory were 4K already here.... Nov 29 11:53:45 Memory got a lot cheaper ;) Nov 29 11:53:53 yup, to my delight Nov 29 12:01:53 where did you see the price? mine wasn't 21K as well Nov 29 12:02:12 ah title Nov 29 12:02:31 they had to use very expensive gloves to build it :) Nov 29 12:07:41 JaMa: Any thoughts on "could not read from the boot medium! System halted" in VirtualBox? Nov 29 12:07:59 Not sure why they cannot just provide an appliance like we do Nov 29 12:08:59 Nevermind, user errror :P Nov 29 12:09:12 wrong root= parameter in vmdk image most likely, are you using sata/scsi for virtual drive? Nov 29 12:09:24 Seems "floppy" is my first option in the list in VBox :P Nov 29 12:09:53 https://github.com/webosose/meta-webosose/commit/698315915de146c957d4ec11bc3a9b4dcc57cde5 Nov 29 12:09:53 And I blindly deselected the hard drive :P Nov 29 12:09:58 it now boots Nov 29 12:10:01 ok Nov 29 12:13:00 New launcher looks too bright for my liking Nov 29 12:13:07 Those Asians LOL Nov 29 12:14:39 they made a kawai theme? Nov 29 12:16:26 Tofe: https://imgur.com/a/ER1P16R Nov 29 12:17:33 wasn't new launcher in previous release already? Nov 29 12:18:12 yes, since builds/webosose/master/492 Nov 29 12:19:09 I'm not very thrilled by this new visual identity; it works, but doesn't look very special Nov 29 12:19:29 JaMa: Probably, but I didn't build OSE for a while Nov 29 12:19:33 Not very thrilled either Nov 29 12:19:34 But well Nov 29 12:19:56 Luckily we don't need to use LG's UI :P Nov 29 12:19:57 I don't like visuals, but at least it shows launcher icons right away instead of people searching for manual to at least see the launcher :) Nov 29 12:20:31 JaMa: Yeah, but closing apps is still a pain in VBox Nov 29 12:21:30 JaMa: completely agree with this, my first reaction to the previous UI was "did it crash? nothing shows up" Nov 29 12:22:26 LOL even in OSE I see ACG permission errors Nov 29 12:22:28 and to make things worse even the way how to show the launcher changed few times, I remember Home key, Win key and swipe from bottom Nov 29 12:22:48 Ah it's Win key now Nov 29 12:25:31 Nov 29 04:14:33 qemux86-64 bootd[702]: [] [pmlog] LS_REQUIRES_SECURITY {"SERVICE":"com.palm.webappmanager","CATEGORY":"/","METHOD":"getBootStatus"} Service security groups don't allow method call. Nov 29 12:25:33 And a few more Nov 29 12:26:12 It does look quite serious Nov 29 12:27:04 And there are a few more Nov 29 12:28:36 Asked them on Skype how they accept PR's. Nov 29 12:29:54 Tofe: Full log FYI: https://pastebin.ubuntu.com/p/4GF3qfVmQd/ Nov 29 14:00:56 And no sign of UPS yet Nov 29 14:51:32 Tofe: Package arrived in France... I doubt for a delivery today, but who knows :P Nov 29 17:37:52 Herrie: I didn't see anything yet; LaPoste just says it's now managed by them Nov 29 17:38:10 So likely to arrive tomorrow, I'd say Nov 29 17:38:35 And then i'll finally be able to benchtest a full rebuild :) Nov 29 17:42:33 Tofe: Yeah same here I guess Nov 29 17:52:08 "Out for delivery" since 4:07 and "The delivery date will be provided as soon as possible." Nov 29 19:23:44 Tofe: "package was approved by customs" around 17:00 Nov 29 19:24:49 yes, I also saw this; a bit surprised that there had been a custom at all, but well Nov 29 19:41:18 Tofe: So I guess tomorrow it will arrive then **** BEGIN LOGGING AT Wed Nov 30 01:05:40 2022 **** ENDING LOGGING AT Wed Nov 30 02:59:56 2022