**** BEGIN LOGGING AT Mon Feb 08 02:59:58 2016 Feb 08 08:40:05 Hi. I facing absolute path issue in target rootfs as shown in below link Feb 08 08:40:06 http://pastebin.com/15B1D4ba Feb 08 08:40:15 How do I fix it? Feb 08 08:53:07 rtr_: Its the ln call, you are setting the symbolic path to include ${D}, 'ln -s ${D}${libdir}/libmosquitto.so.1'. It should just be ${libdir}... Feb 08 11:32:27 anyone using meta-erlang? Feb 08 12:50:16 is there a way of getting all DEPENDS done before running a specific task? Feb 08 13:27:04 mborzecki: If you mean finding all the dependencies, maybe this is of some use to you: https://www.yoctoproject.org/docs/2.0/bitbake-user-manual/bitbake-user-manual.html#generating-dependency-graphs Feb 08 13:33:10 jynik: i was looking for something like this: do_rebar_deps[deptask] = "do_populate_sysroot" which I found in bitbake's manual :) but thanks anyway Feb 08 13:39:54 mborzecki: Ahh cool. Sorry I misunderstood. Feb 08 14:17:24 hi Feb 08 14:17:35 I have this in my .bbapend file: http://pastebin.com/xBf2pGbS Feb 08 14:17:49 i wonder why isn't the symbolic link created? Feb 08 14:19:41 because do_install has nothing to with images Feb 08 14:19:49 IMAGE_ROOTFS is likely unset when do_install runs Feb 08 14:19:55 so youre attempting to create symlinks on your host Feb 08 14:20:07 recipes generate packages, packages make images Feb 08 14:20:19 do_install is used to generate a package, so is isolated from image Feb 08 14:22:12 ok, but i'm struggling to get my service to start automatically Feb 08 14:22:26 any ideas, i can past the whole "experiment" Feb 08 14:22:28 just a sec Feb 08 14:22:51 rburton: http://pastebin.com/XtnTBD8n Feb 08 14:22:53 the systemd class has instructions Feb 08 14:23:06 so i tried overwriting postinst() Feb 08 14:23:07 etc Feb 08 14:23:09 but nothing Feb 08 14:23:23 it's really frustrating Feb 08 14:23:53 1 SYSTEMD_PACKAGES += "${PN}-telnetd" Feb 08 14:23:53 2 SYSTEMD_SERVICE_${PN} += "busybox-telnetd.service" Feb 08 14:23:58 what package is busybox-telnetd in? Feb 08 14:24:02 PN or PN-telnetd? Feb 08 14:24:30 to be honest I'm not sure what I'm doing there Feb 08 14:24:47 SYSTEMD_PACKAGES lists the packages where systemd units are to be found Feb 08 14:24:48 i triead it in every way Feb 08 14:25:04 ok, but what do the stuff after do Feb 08 14:25:06 SYSTEMD_SERVICE_[package] is where you say what services it can find in what pacakge Feb 08 14:25:12 i see Feb 08 14:25:20 so you probably meant SYSTEMD_SERVICE_${PN}-telnetd Feb 08 14:25:25 (this is documented) Feb 08 14:25:54 hi zeddii Feb 08 14:25:54 probably it is, but for someone who just sees yocto the first time Feb 08 14:26:00 it's confusing Feb 08 14:26:29 so how can I make it start then? Feb 08 14:27:24 start automatically i mean, from the systemd_postinst() in systemd.bbclass it should automatically enable those that have the "enable" set Feb 08 14:27:28 if i understand right Feb 08 14:27:30 * zeddii is off planes .. finally Feb 08 14:27:47 SYSTEMD_AUTO_ENABLE_${PN}-telnetd = "enable" should do it in theory Feb 08 14:29:30 yet it doesn't Feb 08 14:30:57 hmm Feb 08 14:31:18 but now that I fixed the SERVICE it's there Feb 08 14:31:21 thanks rburton Feb 08 14:32:14 now at least it's in multi-user.target.wants Feb 08 14:39:11 does anyone know when is the right time to start telnet? I've tried After=network.target but it gets killed somehow so probably it should be after something else Feb 08 15:00:46 bleh... buildhistory-diff doesn't work when the buildhistory is stored in a submodule, because gitdb returns a Repo instance corresponding to the main repo if you give it a submodule path Feb 08 15:01:36 7 Feb 08 15:01:54 8 Feb 08 15:02:20 (yeah, this is what happens when you think you have shift pressed down, sorry) Feb 08 15:09:11 Hi, I have added a non-root user to my system and try to run a Qt/QML-app with eglfs with this user, but getting segfaults. If I run the same app as root, it works. Feb 08 15:09:15 Any ideas? Feb 08 15:59:08 frsc: Feb 08 15:59:14 whoops Feb 08 16:06:34 khem`: those lovely people at uclibc-ng deleted their 1.0 branch that we were using Feb 08 16:12:50 (fix sent, easy fix) Feb 08 16:18:29 rburton: since intuition is clearly not enough, is this enough http://www.yoctoproject.org/docs/1.6/dev-manual/dev-manual.html for getting the hang of yocto? Feb 08 16:18:54 blotunga: yes, but make sure you use the right version. 1.6 is quite old! Feb 08 16:18:59 ah Feb 08 16:19:06 google threw that up Feb 08 16:19:49 yeah google does that Feb 08 16:19:51 sometimes i wonder if google's search is getting worse every year Feb 08 16:20:17 i guess we need a sitemap file to point google at the important bits Feb 08 16:20:25 :D Feb 08 17:43:22 RP: i think we really need to give some thought to how to let someone undo a _remove. the permanence is a problem. if a distro wants to remove a packageconfig by default, and doesn't want to override packageconfig entirely for a recipe to avoid future compatibility issues, there's no recourse but to use _remove, and then local.conf can't undo it, nor can an append Feb 08 17:43:46 though if we allow backfilled packageconfigs, it could be more kosher to override a recipe packageconfig from a distro .conf Feb 08 17:44:04 kergoth, ya.. that is hte one place I've had problems with _remove as well.. Feb 08 17:44:07 * kergoth ponders Feb 08 17:45:07 kergoth: I have also wondered about that. I concluded we could at least expose python api which killed the overrides Feb 08 17:45:23 * RP had something like that at some point, not sure what happened to it Feb 08 17:45:27 i was thinking that too, then at least we could use anonymous python in a bbappend to counteract it Feb 08 17:46:10 I'm thinking it's rare enough (at this point) that direct python intervention might be adequate.. Feb 08 17:47:33 my concern is it could become more common, i know my distro is gradually accruing more packageconfigs to tweak recipe behavior, particularly as more and more configs are added to recipes. but you could make an argument that a user wanting to undo that should create their own distro instead of trying to do it in local.conf, so.. maybe that's okay Feb 08 17:48:47 only slightly related, but i think we should seriously consider making it best practice to use neither ?= nor ??= for PACKAGECONFIG in a recipe Feb 08 17:48:57 https://bugzilla.yoctoproject.org/show_bug.cgi?id=9057 Feb 08 17:48:59 Bug 9057: critical, Medium+, 2.1 M3, saul.wold, NEW , u-boot on Edison BSP fails due to "binutils: Upgrade to 2.26" Feb 08 17:49:00 the fact that a user setting 'PACKAGECONFIG' in local.conf will break the universe is not good Feb 08 17:49:00 Fix it fix it fix it! Feb 08 17:49:13 distros will use pn overrides anyway.. Feb 08 17:52:40 kergoth: yes, true Feb 08 17:52:58 kergoth: If you want to write that up with some expanded reasoning, I think we could do that Feb 08 17:53:07 * kergoth nods, will do Feb 08 18:08:25 Hi Guys, does the bug in https://bugs.tizen.org/jira/browse/BTY-23 look familiar to anyone? Feb 08 18:09:15 I'm still unable to get past "Computing transaction...error: Can't install XXXX@cortexa9hf_vfp_neon: no package provides YYYY". Where XXXX and YYYY seem to always be python3- Feb 08 18:09:30 and that tizen bug sounds similar Feb 08 19:25:27 khem`: any idea what 'checking dynamic linker characteristics... configure: error: Link tests are not allowed after GCC_NO_EXECUTABLES.' means? Feb 08 19:25:42 (gcc-runtime:do_configure, specifically) Feb 08 19:26:05 hmm Feb 08 19:26:12 are you on latest master ? Feb 08 19:26:20 if not then please update Feb 08 19:26:46 secondly is your build with security flags Feb 08 19:26:52 enabled or disabled ? Feb 08 19:27:17 no security flags, fairly recent master, but I'm experimenting with static, so it's certainly broken by what I'm doing, but was hoping for some background on the error Feb 08 19:38:03 Is there some way to specify a fixed size for a wic partition? Seems that if you --source + --size, it adds some extra, unexplained, padding of some sort. My ext4 image (--source) is 290MB, without a --size, it creates a 317MB partition, with a --size of 1900, it creates a 2590MB partition. weird Feb 08 19:39:24 been struggling with looking at the python code for wic (and friends) for a couple of days over this. Is there some other doc on wic other than the yocto mega-manual? Feb 08 22:42:48 Hi, I presume Upstream-Status: Pending in a patch means "it was submitted to the appropriate project, but not applied yet", right ? Feb 08 22:43:11 what is the status for "it was never submitted to upsteam, but it should be" ? Feb 08 22:43:25 I found patches for dpkg , which are Pending , but were never submitted to DPKG Feb 08 22:46:12 no Feb 08 22:46:20 pending means "we should do something with this" Feb 08 22:46:45 (http://www.openembedded.org/wiki/Commit_Patch_Message_Guidelines#Patch_Header_Recommendations Feb 08 22:46:48 *sigh* ... Feb 08 22:46:57 rburton: thanks, looks like I should do something about it Feb 08 22:47:15 some stuff that's pending should be marked as Inappropriate, some should be submitted upstream Feb 08 22:47:24 some are dirty hacks for problems that need to be fixed upstream in a better way Feb 08 22:47:39 rburton: I'll try to remember and poke the dpkg maintainer tomorrow about the pending ones, thanks Feb 08 22:47:57 cool Feb 08 22:57:57 rburton: even better, looks like the dpkg maintainer is willing to check the Pending ones Feb 08 23:00:16 hooray Feb 08 23:00:27 who is dpkg maintainer these days Feb 08 23:00:42 (you can tell i'm a lapsed DD in that I don't know anymore) Feb 08 23:00:49 in my mind, it's still keybuk Feb 08 23:01:12 rburton: I think Guillem Jover Feb 08 23:01:40 rburton: I'm still a newcomer to debian development ; I've been using it for almost 10 years tho Feb 08 23:03:59 you can have some of the packages which i'm still the "maintainer" for if you want! :) Feb 08 23:04:01 * rburton -> bed Feb 08 23:04:13 sjolley: promise i'll be canceling our 1:1 tomorrow! Feb 08 23:05:11 rburton: gnight ;-) **** ENDING LOGGING AT Tue Feb 09 02:59:59 2016