**** BEGIN LOGGING AT Thu Feb 16 03:00:02 2017 Feb 16 03:24:44 reddy_: what kind of data are you looking for. There are MACHINE_OVERRIDES for designating machine specific things Feb 16 03:30:43 i guess original patch is obsoleted; when/which branch does gcc-6 patch hit? Feb 16 03:32:27 otavio: same breakage in boundary u-boot-imx6 and mainline Feb 16 03:33:23 and the recipe for u-boot-boundary never gets built (only the scripts part) so i'm not sure anyone noticed Feb 16 03:34:37 nerdboy: oe/gcc-6-branch Feb 16 03:34:41 u-boot-imx6 on github has not exposed the issue tracker Feb 16 03:34:44 is where the patch should be proposed to Feb 16 03:35:11 khem`: meaning what exactly? Feb 16 03:35:21 in my github gcc repo Feb 16 03:35:28 if thats what you were asking for Feb 16 03:35:49 you already merged it Feb 16 03:36:02 i meant where does it go from there Feb 16 03:36:19 it can go upstream into gcc Feb 16 03:36:24 if its deemed needed Feb 16 03:36:38 otherwise I will keep rebasing it for OE as long as we support gcc6 Feb 16 03:37:51 i didn't look that far, since toolchain devs are still maintaining it Feb 16 03:38:11 it helps me when I maintain gcc in OE Feb 16 03:38:19 ^^ assumed they would've already done that Feb 16 03:42:34 usually it depends on type of change many times they are distro specific Feb 16 03:44:41 not distro-specific in this case, rather system-built-with-specific Feb 16 03:44:54 *with-pie even Feb 16 08:53:39 Hello, I have a problem with fetching via gitsm ( git with submodules ), the unpack process is copying my full git repo inside my workdir, which takes really long. It is possible to genereate a symlink ? Feb 16 08:58:26 I have a recipe for a nodejs application. The recipe was made following this wiki article: https://wiki.yoctoproject.org/wiki/TipsAndTricks/NPM#Non-Registry_Workaround Feb 16 08:59:47 I can bitbake the recipe without error but the main ipk package is not created... any suggestions? Feb 16 09:17:34 mrpelotazo: generally: look in the work directory, e.g. check if package/ contains the files you'd expect. Check logs in $WORKDIR/temp/ Feb 16 09:47:15 jku: package/ is empty and deploy-ipks has *-dbg and *-dev ipks but they are empty. No main ipk package there. Feb 16 09:55:09 mrpelotazo: probably there's just nothing mathing the FILES to be installed Feb 16 09:57:13 mrpelotazo: depending on the module, and especially if it is something containing native code, i'd suggest manually setting up a do_install task and corresponding FILES variable Feb 16 10:26:15 Hi! I'm in the process of documenting myself on best practices for speeding up Yocto/OE workflow when projects, users, and build hosts gets all >1 in a company environment. I have a basic understanding of DL_DIR, PREMIRRORS and sstate-cache concepts and know that they can be shared at some point, but getting confused how to move on. First topic first: how to reuse my currently already quite big DL_DIR thing Feb 16 10:26:21 that I have been sharing myself for years between projects (with user and build host all =1). I'm reading more in depth about things like 'own-mirror' and PREMIRRORS, but still not seeing the best approach. Anyone willing to help by sharing their experience? Feb 16 10:54:56 Hi everybody! I have the following issue: I want to debug a multithreaded application on an embedded system running Yocto. I have installed 'gdb_linaro-7.8-r2014.09.0_armv7at2hf-vfp-neon.ipk'. GDB seems to work, but is not able to debug other threads, because multi thread support needs a library. I would like to know what I need to do to get '/lib/libthread_db.so.1' or so. Feb 16 10:55:49 I am runnning Yocto 1.7 Feb 16 11:55:10 nerdboy: send an email to our mailing list, I can copy them Feb 16 14:49:25 q:q Feb 16 15:02:42 marquiz: rpm signing should be fixed now Feb 16 15:02:57 marquiz: I force pushed the branch as usual Feb 16 16:09:06 Hello - is there a way to put dependencies in pkg_postinst_${PN} scripts? Feb 16 16:25:58 Circuitsoft: yes - add them to RDEPENDS Feb 16 17:24:20 joshuagl: thanks for the sanity-check on the pseudo thing, I apparently didn't check the right thing. Feb 16 17:26:01 seebs: no problem :-) Feb 16 17:26:44 denix: do you happen to know the status of beaglebone black atm? is meta-ti still the correct place for support, and do you happen to know the rough delta between the amsdk and the public layers? :) Feb 16 17:32:26 kergoth: yes, BBB is still supported in meta-ti. is there something in particular you are looking for? Feb 16 17:32:59 nope, just remembered there being things like meta-beagleboard and whatnot and wanted to check the current status offhand, coming back to it after a while Feb 16 17:33:00 thanks Feb 16 17:34:00 kergoth: amsdk was one of the "products" built on meta-ti and arago. the naming changed few years ago to "TI Processor SDK", but it's basically the same Feb 16 17:35:01 kergoth: meta-beagleboard was the community layer to add cape support on top of meta-ti. afaik, it's been abandoned once beagle switched to debian few years ago Feb 16 17:35:07 ahh Feb 16 17:35:11 okay, that clarifies things, thanks Feb 16 18:40:26 hi all Feb 16 18:42:02 Can some one point me how to access architercture specific variable in the recipe file . Eg . how can i know its ARm or X86 in my recipe file ?? Please point me to a document/recipe is appreciated Feb 16 18:43:26 see the docs about OVERRIDES Feb 16 18:44:03 ok thanks Feb 16 18:44:06 do overrides work on ARCH? I thought from experience they only worked on MACHINE. Feb 16 18:44:26 maybe they are _meant_ to work on ARCH but don't. :) Feb 16 19:54:47 otavio: i found their support@ email form Feb 16 19:57:08 got a couple of weird replies already, not really liking their level of comprehension so far... Feb 17 00:24:21 paulg: bitbake -e | grep OVERRIDES= -C5 shows "${TARGET_OS}:${TRANSLATED_TARGET_ARCH}:build-${BUILD_OS}:pn-${PN}:${MACHINEOVERRIDES}:${DISTROOVERRIDES}:${CLASSOVERRIDE}:forcevariable${LIBCOVERRIDE}" Feb 17 00:24:42 i.e. TARGET_ARCH is an override Feb 17 00:26:22 RP1, yeah, I'll have to dig up and re-test my use case ; from what I recall -- everything I saw made it _look_ like it should work; but it simply would not work. Feb 17 00:26:49 I was trying to something as simple as limiting grub to x86_64 and i386 builds. Feb 17 00:27:56 s/to/to do/ Feb 17 00:29:13 paulg: that is usually with COMPATIBLE_HOST Feb 17 00:35:47 RP1, yeah I originally did see COMPATIBLE_HOST examples, and may have fallen into that trap originally but zeddii and jwessel were looking at the same changes and didn't see why it was working for MACHINE and not arch. Which made me feel slightly less like a knuckle dragging mouth breather in a puddle of self drool. :) Feb 17 00:36:00 I'll revisit it tomorrow. Feb 17 00:39:31 It is probably a common enough example that we should have a documented example for it once everything is sorted. Feb 17 00:39:54 ...as the earlier query here on #yocto reinforces. Feb 17 00:58:30 paulg: not sure you'd do knuckle dragging ;-) Feb 17 01:01:48 RP1, perhaps, but I'd rather be on the modest side of the Dunning-Kruger effect... Feb 17 01:59:11 what is the issue with my bitbake recipe using cmake build, can some one let me know? Feb 17 01:59:16 QA Issue: ABC: Files/directories were installed but not shipped in any package: Feb 17 02:00:08 Please set FILES such that these items are packaged. Alternatively if they are unneeded, avoid installing them or delete them within do_install. Feb 17 02:28:05 trollkarlen1: it just told you what to do.. Feb 17 02:34:12 I donot want to delete them. Also I am setting FILES_${PN} += "${sysconfdir}/myfile" but still getting the error. Feb 17 02:36:49 then your file isn't actually in sysconfdir Feb 17 02:37:18 also, all of sysconfdir is packaged by default, there's no need to add that at all Feb 17 02:39:44 The files it is telling are Feb 17 02:39:46 /usr/etc Feb 17 02:39:46 /usr/etc/file1 Feb 17 02:39:46 /usr/etc/file2 Feb 17 02:39:46 /usr/etc/file3 Feb 17 02:39:46 /usr/etc/file4 Feb 17 02:40:20 Does that means a relative path from my build directory right? Feb 17 02:44:42 and the files are present here tmp/work/-poky-//-/package/usr/etc/file1 Feb 17 02:56:11 /usr/etc is not /etc Feb 17 02:56:17 sysconfdir is /etc Feb 17 02:59:37 kergoth: oh, ok. let me try. **** ENDING LOGGING AT Fri Feb 17 03:00:00 2017