**** BEGIN LOGGING AT Tue Apr 25 03:00:03 2017 Apr 25 04:56:14 Quite some more failing here... LSB Apr 25 04:56:15 qtbase Apr 25 04:56:20 But ran out of disk space Apr 25 04:56:25 LSB seems to be an issue for sure Apr 25 04:56:30 qtbase trying again now Apr 25 04:56:57 For LSB you need to remove the patch https://github.com/openembedded/openembedded-core/blob/master/meta/recipes-extended/lsb/lsb_4.1.bb#L18 Apr 25 06:54:35 JaMa: "work/x86_64-linux/ncurses-native/6.0+20161126-r0/temp/run.do_siteconfig_gencache.26806: gen-site-config: not found" <-- do you know what this error means? I do have gen-site-config in meta-oe/scripts... Apr 25 06:57:44 Maybe it was some remaining conflicting cache; I'll retry after a good cleanup Apr 25 07:03:23 nope Apr 25 07:09:34 ok this can be solved by defining WEBOS_EXTRA_PATH="" in local.conf, but it's still a bit weird Apr 25 07:24:09 Tofe: My qt had some issues due to running out of space Apr 25 07:24:19 I put it to run again this morning Apr 25 07:24:35 ubuntu camera bits were complaining about qt multimedia Apr 25 07:24:46 So I manually forced build of qtmultimedia Apr 25 07:25:02 Could be due to my build running out of space somehow. Apr 25 07:28:42 I'm building from scratch, from pyro updated by JaMa Apr 25 07:29:26 Tofe: Ah Apr 25 07:29:34 I had the one from before JaMa's update Apr 25 07:29:43 You're doing Mako or Hammerhead? Apr 25 07:29:54 hammerhead Apr 25 07:30:04 I think you're doing mako right? Apr 25 07:31:30 Tofe: Yup Apr 25 08:27:42 I wonder, what is the mozjs package? Have we always had this one? Apr 25 08:28:34 what's issue with mozjs? Apr 25 08:29:01 yesterday I got some fix for that "[oe] [meta-oe][PATCH] mozjs: prevent target CFLAGS etc being passed to the host" Apr 25 08:29:17 if it's something you were seeing with pyro builds Apr 25 08:29:44 my pyro build over night wasn't very successfull, might be the WEBOS_EXTRA_PATH breaking the HOSTTOOLS filtering Apr 25 08:30:09 Tofe: mozjs might be the mozilla js for pdf.js used by pdf app? Apr 25 08:31:08 no issue, just pure curiousity Apr 25 08:31:49 Herrie|Pre3: ah yes, makes sense Apr 25 08:33:03 I feel like OE guys have put progress bars and percentages everywhere they could :) That's quite nice Apr 25 08:35:19 Tofe: I'm not 100% sure about that mozjs but that's the only one I can think of right now Apr 25 08:37:14 I've pushed 2 more fixes to meta-webos-ports/pyro and restarted my build Apr 25 08:44:35 JaMa: I had issues with alsa-lib before your fixes yesterday. I added 2 patches to herrie/pyro. Might be good to double check if they're included in your latest patches. Apr 25 08:46:01 JaMa: Ah seems they're there now with 8f7d832 Apr 25 08:46:36 JaMa: Nevermind that comment Apr 25 08:46:41 I need to recheck tonight Apr 25 08:47:13 I also had lsb from oe-core misbehaving (see log) Apr 25 08:50:34 Herrie|Pre3: I had seen patch issues with alsa and lsb as well Apr 25 08:50:50 so maybe they still aren't included in pyro branch Apr 25 08:51:33 JaMa: Is there an easy way to drop a single patch from oe-core without having to fork it completely? Apr 25 08:52:03 SRC_URI patch? Apr 25 08:52:59 many things can be undone with .bbappends, but depends on what kine of patch you meant Apr 25 08:53:46 still missing in pyro: Patch 0001-ucm-fix-handling-of-config-files-when-file-type-is-n.patch does not apply (enforce with -f) Apr 25 08:55:12 ok, I just stumbled on the lsb issue; I'll then remove https://github.com/openembedded/openembedded-core/blob/master/meta/recipes-extended/lsb/lsb_4.1.bb#L18 Apr 25 08:55:36 I basically want to remove the patch that Tofe linked just not Apr 25 08:55:41 s/not/now Apr 25 08:58:25 why not remove that patch from our bbappend? ./meta-luneos/recipes-extended/lsb/lsb_4.1.bbappend Apr 25 08:59:01 Oh, I didn't even see we had a bbappend there :p Apr 25 09:00:16 both alsa-lib and lsb fixed just now in pyro branch Apr 25 09:00:42 ok great Apr 25 09:06:01 next step now: glib-gettextize, needed by our pidgin-sipe Apr 25 09:08:32 might be we need gettext-native, or glib-native, not sure yet Apr 25 09:10:22 or inherit gettext Apr 25 09:10:52 ah, didn't know that, seems cleaner Apr 25 09:11:58 nope, doesn't work... but it worked with glib-2.0-native Apr 25 09:12:03 (in depends) Apr 25 09:14:57 now I have the issue with qtubuntu-camera, "Unknown module(s) in QT: multimedia", a bit weird Apr 25 09:15:23 Tofe: That one I had too Apr 25 09:15:37 Seems it might be the order in which things get build Apr 25 09:15:48 I.e. it only shows now because of clean build Apr 25 09:15:56 I saw qtmultimedia earlier though Apr 25 09:16:03 Seems it might be a missing depends on qtmultimedia or something Apr 25 09:16:19 that would be most likely, if it's missing Apr 25 09:16:41 as pyro now uses separate sysroots, it's much more strict on that aspect Apr 25 09:17:57 Tofe: Could also be that it was present before in our builds and we didn't rebuild from scratch. I've seen that happen before :P Apr 25 09:18:15 yes, qtubuntu-camera is pretty recent Apr 25 09:18:40 and even when rebuilding from scratch, if qtmultimedia was build before by chance, then we won't see the mistake Apr 25 09:18:56 because of the common sysroot Apr 25 09:37:04 Tofe: Yeah that's what I mean ;m) Apr 25 09:37:06 ;) Apr 25 09:37:36 That's why sometimes it's good to build from scratch. I found other similar issues before this way :P Apr 25 11:00:33 last issue I see for now is qemux86 mesa build meta-luneos/recipes-graphics/mesa/mesa_10.3.7.bb:do_configure looking into that one now Apr 25 11:01:13 Herrie|Pre3: with RSS (recipe specific sysroots) and HOSTTOOLS filtering, there should be a lot fewer issues discovered in clean builds Apr 25 11:02:32 as each component build basically starts from clean sysroot now anyway Apr 25 11:03:59 JaMa: Well I fixed quite some of those in the past when I did my initial builds. Apr 25 11:05:17 mesa should be fixed in pyro branch Apr 25 11:22:06 ah, nodejs-module-webos-pmlog: xxd: not found Apr 25 11:26:11 JaMa: ^ looks like you already tackled some issues regarding this : https://patchwork.openembedded.org/patch/137943/ Apr 25 11:27:16 Not sure what I should do to solve the xxd dependency... add vim-native in DEPENDS ? something else ? Apr 25 11:58:35 yes, vim-native, I'll do it (I already did in our internal repos) Apr 25 11:59:00 ok thanks Apr 25 11:59:39 added in pyro Apr 25 12:05:20 ah, looks like webos-keyboard doesn't build anymore due to a hunspell upgrade Apr 25 12:06:25 Tofe: Hmmmz that's a tricky one Apr 25 12:06:30 Any log? Apr 25 12:07:53 Herrie|Pre3: I can fix it easily actually Apr 25 12:14:04 I made a quick&dirty fix to continue the build, but I'll have to do a cleaner one once home Apr 25 12:19:04 Tofe: OK :P Apr 25 12:20:36 Herrie|Pre3: http://paste2.org/aGKJE9hc Apr 25 12:29:52 mmh if the recipe doesn't find pkg-config, what should I do ? inherit something special, or put a DEPENDS on pkg-config ? Apr 25 12:31:00 let's try inherit pkgconfig. Apr 25 12:38:16 That was a pretty good guess. Apr 25 12:44:50 Tofe: inherit pkgconfig Apr 25 12:45:02 ah, nearly finished, but rootfs failed Apr 25 12:45:04 "package packagegroup-core-boot-1.0-r17.2.hammerhead requires systemd, but none of the providers can be installed" Apr 25 12:45:08 I've already pushed fix for sdl2-opengles test if that's where you were seeing it Apr 25 12:45:18 JaMa: yes, that was it Apr 25 12:45:32 I'm a bit behind NOTE: Running task 6379 of 6716 Apr 25 12:45:54 I'm a bit surprised we don't include systemd :) Apr 25 12:49:48 ah, ok, maybe we didn't migrate this: http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#selecting-an-initialization-manager Apr 25 12:51:18 which part? we have: Apr 25 12:51:19 WEBOS_DISTRO_FEATURES_INIT = "systemd" Apr 25 12:51:19 WEBOS_DISTRO_FEATURES_INIT_CONSIDERED = "sysvinit" Apr 25 12:51:24 for last couple years Apr 25 12:52:15 commit dc04615fc27d23a97c17a2bafd2d92a9e55983c5 Apr 25 12:52:15 Author: Martin Jansa Apr 25 12:52:15 Date: Tue Jan 15 23:17:39 2013 +0100 Apr 25 12:52:15 webos-ports: modify distro configuration Apr 25 12:52:28 * switch init manager from upstart to systemd Apr 25 12:52:41 Do we need ' VIRTUAL-RUNTIME_init_manager = "systemd" ' ? Apr 25 12:53:19 meta-luneos/conf/distro/include/luneos-preferred-providers.inc:VIRTUAL-RUNTIME_init_manager = "systemd systemd-compat-units systemd-machine-units" Apr 25 12:53:47 Oh. mmh. Maybe I'm not on the right track here, then Apr 25 12:54:26 I'll check if my rootfs fails the same Apr 25 12:54:32 ok Apr 25 12:54:40 now building qtwebengine.. :) Apr 25 14:32:10 Tofe: the rootfs issue looks like some kind of opkg (or it's new libsolv implementation), because both systemd and systemd-analyze packages were correctly created and are available in the feed Apr 25 14:32:16 will debug a bit more Apr 25 14:33:00 ok thanks Apr 25 14:37:44 it might be also related to 3 changes I had to do in order to fix opkg in morty with special OPKGLIBDIR (which we still use in luneos): Apr 25 14:37:47 Backport following changes until they are merged in morty branch and Apr 25 14:37:49 oe-core revision is upgraded to include them: Apr 25 14:37:52 http://lists.openembedded.org/pipermail/openembedded-core/2017-March/134520.html Apr 25 14:37:55 http://lists.openembedded.org/pipermail/openembedded-core/2017-March/134519.html Apr 25 14:37:58 http://lists.openembedded.org/pipermail/openembedded-core/2017-March/134521.html Apr 25 14:38:01 re-trying with default OPKGLIBDIR now Apr 25 14:42:50 didn't help Apr 25 14:48:05 mmh Apr 25 14:51:56 I've disabled libsolv for now, which allows the image to be built Apr 25 14:52:00 all fixes in pyro branch now Apr 25 14:52:20 at least for just finished the build for qemux86 Apr 25 14:52:35 ok, I'll try for hammerhead Apr 25 15:00:31 rpmdeps seems to have big issues with my rootfs Apr 25 15:01:14 I'll look at that when at home, it's really not handy from here Apr 25 16:29:09 JaMa, Herrie|2: https://github.com/webOS-ports/meta-webos-ports/pull/220 Apr 25 16:32:43 ... and N5 rootfs done! I had introduced a mistake when doing my own little fixes :p Apr 25 16:38:36 ... bootloop :( Apr 25 16:40:51 ahah, yey systemd... https://github.com/systemd/systemd/issues/5236 Apr 25 16:48:08 https://paste2.org/VcY68P8p Apr 25 16:50:22 Looks like using a recent systemd is even more problematic than a recent glibc... systemd 232 wants at least a 3.12, which de facto excludes it for all the current mobile platforms... Apr 25 16:59:36 Tofe: This invocation ID should be merely an error Apr 25 17:01:12 Ehm warning I mean Apr 25 17:01:30 yup Apr 25 17:01:36 still, the bootloop is real Apr 25 17:01:37 But yeah their 3.12 requirement is a pain Apr 25 17:01:53 Let me update my mako Apr 25 17:01:57 Ehm I mena build Apr 25 17:03:08 I had Mako booting before Apr 25 17:08:43 Seems it'll be a while Apr 25 17:08:48 4000 tasks to finish first Apr 25 18:24:13 OK 500 to go :P Apr 25 18:24:46 Looks like we have some cleaning up to do in some recipes as well :P https://bpaste.net/show/dc41404b799e Apr 25 19:03:42 Another 60 incl qtwebengine :P Apr 25 19:03:57 Tofe: I THINK I tried the systemd 231 downgrade but not 100% sure Apr 25 19:04:05 I have some vage memory of that Apr 25 19:04:19 s/vage/vague Apr 25 19:06:42 Tofe: Yup... "Jan 08 15:21:16 Mako stopped bootlooping with systemd downgrade" Apr 25 19:42:54 I recall I tried with 231 Apr 25 21:24:52 The only problem with using 231 is the high severity fixes; otherwise I wouldn't really bother Apr 25 21:25:16 And it looks like (but not sure) that the problems is with udev - which is part of systemd now **** ENDING LOGGING AT Wed Apr 26 03:00:00 2017