**** BEGIN LOGGING AT Tue May 30 03:00:05 2017 May 30 06:03:51 morning all May 30 08:47:34 Hey, i get this error: "Files/directories were installed but not shipped in any package:" Followed by 447 Folders/Files that are apparently not packaged. Does anyone know how to fix this? Because i dont think just adding all these 447 files to FILES_PN is the way to go May 30 08:52:24 rj_: depends, should the files be shipped in the package? you might want to split them up into different packages. Also FILES_... can have patterns in it, eg. "/a/b/*" so you shouldn't need to specify all the files individually. May 30 08:54:06 nrossi: it seems i had some argument wrong which gave the wrong path. It's fixed now. Thanks anyway ^^ May 30 09:02:24 Hi guys :) May 30 09:03:10 I'm having trouble trying to save a busybox configuration, I'm trying to have a release version of busybox and a debug version. But IDK how to save them... Any leads? May 30 09:07:29 [13:28:07] hello bluelightning did you send the patch about upper case PN files? (the variables not being evaluated when PN is upper case) May 30 09:07:33 :) May 30 09:15:46 YCN-: i dont know if i understand you correctly, but it is possible to append an own defconfig file in your own layer, which shadows the default defconfig May 30 09:18:34 YCN-: ah and any file ending in '.cfg' is appended to the config file May 30 09:19:12 okay I'll try that way I think that I find some interesting leads in the manual. But I still have to try May 30 09:19:26 I have to admit that it's not as easy as it seems ahah May 30 09:22:17 I don't really get how busybox configuration file are saved after a changement. I can't find them. May 30 09:22:32 i have a problem genarating a SDK. im getting: binutils-crosssdk-x86_64-linaro-2.25-r2016.10 do_populate_sysroot_setscene: The recipe binutils-crosssdk-x86_64 is trying to install files into a shared area when those files already exist. May 30 09:35:15 MarcWe: is mingw involved? If not, which files is it saying are duplicated? May 30 09:36:00 RP: no Mingw involved yes it tells me what files if i remove them the build continues May 30 09:36:58 But man removing the files is not the correct thing i think May 30 09:37:10 MarcWe: correct. Which files are these? May 30 09:37:57 can it be that binutils is biting arago-tmp-glibc/sysroots/x86_64-linux/usr/share/gdb/syscalls/gdb-syscalls.dtd arago-tmp-glibc/sysroots/x86_64-linux/usr/include/gdb/jit-reader.h arago-tmp-glibc/sysroots/x86_64-linux/usr/share/gdb/syscalls/***-linux.xml May 30 09:38:03 ** are some architectures May 30 09:39:57 MarcWe: well, binutils-2.28.inc has a do_configure_prepend which does rm ${S}/gdb May 30 09:40:11 MarcWe: question is therefore why linaro doesn't do that May 30 09:40:20 gdb should be built by the gdb recipe May 30 10:12:09 LocutusOfBorg: I don't think so but I am aware it will not work, and there isn't a workaround May 30 10:12:40 uppercase overrides are not supported and that breaks uppercase letters in recipes and distro / machine names May 30 10:18:40 bluelightning, if you remember (I think I discussed that with you IIRC), we agreed about sending a patch like May 30 10:19:08 bb.warn ehi, you are using upper case PN recipes, something bad might happen, e.g. your postinst variables not being evaluated May 30 10:19:37 this is not about supporting it, but giving a notice to the developer about stop using it May 30 10:19:53 does this sounds like a plan? May 30 10:21:00 yes we should definitely do that May 30 10:21:12 in fact, I shall open a bug May 30 10:24:17 https://bugzilla.yoctoproject.org/show_bug.cgi?id=11592 May 30 10:24:18 Bug 11592: enhancement, Undecided, ---, ross.burton, NEW , Add QA check for uppercase recipe names May 30 10:27:39 thanks, will subscribe to it May 30 10:52:25 hi May 30 10:52:52 I cannot compile with EXTRA_IMAGE_FEATURES set with "tools-profile" May 30 10:53:05 the compiler says: ../../../../valgrind-3.11.0/none/tests/arm/vcvt_fixed_float_VFP.c:13:1: internal compiler error: Max. number of generated reload insns per insn is achieved (90) May 30 10:53:55 i've never had this kind of error before May 30 10:54:58 I'm on krogoth May 30 12:53:01 Anyone a clue how to fix this? https://pastebin.com/mBTA90ZX May 30 13:36:50 is it possible to build the esdk with meta-qt5 ?! May 30 14:10:30 Trying to build the ivi-image in meta-ivi using yocto pyro branch, one of the recipes(CAPI-Core) which needs JRE fails because it cant find the JRE in the 'recipe-sysroot-native', but till the morty branch it properly getting the JRE on the host system with the 'SANITY_REQUIRED_UTILITIES += "java"' tweak. May 30 14:12:08 Any idea about the changes in pyro related to sanity class ? May 30 14:28:31 jo90: not sure if you can find something here downloads.yoctoproject.org/releases/yocto/yocto-2.3/RELEASENOTES May 30 14:34:35 lsandov: Thank you. Already checked the release notes, but didn't find anything related. I was trying to find someone who uses JRE from host with SANITY_REQUIRED_UTILITIES May 30 15:55:41 jo90: has the layer been updated to work with pyro? There were extensive changes to switch to recipe specific sysroots in pyro and I suspect JRE would need tweaking to work with that May 30 15:57:11 hi, i created a layer with a u-boot-mydistro.bb inside, where i placed a new SRC_URI to a local u-boot.tgz, then i added PREFERRED_PROVIDER_virtual/bootloader="u-boot-mydistro" and UBOOT_CONFIG="mydistro_defconfig" to local.conf . when i run bitbake u-boot-mydistro it unpacks the u-boot.tgz and seemed to use the config file, but there is no u-boot binary ? when i run bitbake u-boot-mydistro -c devshell and run there make May 30 15:57:11 mydistro_defconfig and make it returns, bad value -march= May 30 15:58:49 vm_: I think you have to look at the recipe you are adding this might be having issues May 30 15:59:42 khem, but what issues ? in the recipe is no var for -march, also this should be set by yocto May 30 16:01:46 vm_: there was en error with armv5 May 30 16:02:20 march in binutils few releases ago, I think it was due to some memory corruption when running assembler May 30 16:02:34 not sure if you are running into that May 30 16:02:46 usually a reboot of build machine helped May 30 16:02:54 but your case may be different May 30 16:03:38 i think it is using the wrong compiler, i build imx6qsabresd and it is armv7a May 30 16:05:14 just running 'make' in a devshell is unlikely to succeed in any recipe, since it doesn't pass along EXTRA_OEMAKE May 30 16:05:24 run ../temp/run.do_compile instead May 30 16:13:14 kergoth, with bitbake ? May 30 16:13:29 I don't understand the question May 30 16:13:55 there's a shell script in ../temp/ relative to the dir you're in inside the devshell, the script is named 'run.do_compile' May 30 16:14:04 bitbake wrote that script out and then ran it to run the do_compile task May 30 16:14:07 running "bitbake tmp/.../temp/run.do_compile " or just ".../tmp/ru.." May 30 16:14:07 you can run that same script yourself May 30 16:14:10 no May 30 16:14:15 neither May 30 16:14:26 *inside the devshell*, run ../temp/run.do_compile May 30 16:14:33 ok May 30 16:14:39 or outside, just run bitbake May 30 16:14:44 but just running 'make' won't do May 30 16:15:27 when i run bitbake u-boot-mydistro it doesnt show any errors, but no binary May 30 16:23:28 why are in tmp/work 5 different directories for one machine? May 30 16:28:08 in different tmp/work/* dirs are different programs, is yocto building one image with different compilers May 30 16:30:37 can i change the SRC_URI with a .bbappend file ? May 30 17:06:25 you should also ensure that it respect the CC/CXX etc from env and not constructs its own May 30 17:18:48 RP: I am planning on adding support for runit ( init system ), since systemd is getting too complex and demanding on flash size for small systems, runit is oriented towards smaller footprint yet provide application life cycle management May 30 17:19:07 http://smarden.org/runit/ May 30 17:19:49 I think it could be a good alternative to folks to move away from sysvinit without converting their religion May 30 17:21:24 khem: :p right May 30 17:25:42 I see beaglebone boot quicker with runit then systemd May 30 17:26:00 its because it just isnt clogging the memory bus not that its faster May 30 18:10:00 khem: it does sound interesting and I have my worries about systemd... May 30 18:10:19 khem: adding in another init system has its own concerns though, e.g. testing May 30 18:28:37 Hi all, i have a little question ! May 30 18:28:55 In a motherboard card, what does it means 1333(O.C)/1066/800 ? May 30 18:29:11 Does it means that the motherboard should be overclocked to use a 1333Mhz RAM? May 30 18:29:34 off-topic May 30 18:38:28 it means you have ancient arse DDR2 stuff and need to upgrade. :) May 30 18:39:12 <--- keepin it classy. ;-) May 30 18:44:44 RP: I think yes it does. Although it will be time well spent May 30 19:23:52 Hi folks, I'm trying to use runqemu-export-rootfs to export a filesystem and have a real system boot it. Right now, I'm seeing the kernel get to the point where it is looking for the rootfs, and then timeout. Trying to mount locally with the options that runqemu-export-rootfs prints out fails with "Protocol family not supported". Examining the kernel source code for nfs shows no handling for the nfsoptions May 30 19:23:54 "nfsprog" and "mountprog" (which runqemu-export-rootfs includes in it's recommended option line), so I'm not sure how it would ever manage to handle those options. Am I missing something about how this is supposed to work? I'm using a custom kernel on morty. May 30 19:49:59 paulg : But it's DDR3 stuff May 30 19:51:19 paulg : I'm a bit confusing because CPU is 1333Mhz in bus, and i have two memory bar of 1066Mhz and it says by the motherboard that 1333Mhz(O.C.)... In internet i checked that it means that you need to config manually to the BIOS for telling that you are using 1333 or in the bios i have only 1067 and 800 available... But in the systeminfo it tells me that i am in 1333 Mhz.... Confusing May 30 19:55:53 paulg : I will check documentation about fsb , northbridge etc t May 30 20:26:17 Ok, so removing the -x and -y (NFS and MOUNTD RPC port) from the unfsd call gets a local mount working. Any idea why we're passing a custom rpc ports here? May 30 20:38:06 Still can't get my remote device to mount it though: still times out trying to mount root and panics :( May 30 20:49:04 interesting... appears to be something related to having more than 1 route (I'm connecting the device via a second network interface with a static ip) May 30 20:49:12 boot works now, though :) May 30 23:56:48 how does one build boost-python with python2 instead of python3 in recent yocto? May 30 23:57:04 khem: is it broken / disabled? May 31 02:04:03 is it a bad idea to use the github copy of yocto to build from? May 31 02:05:46 not githib, this one git://git.yoctoproject.org/poky May 31 02:09:38 meta-openembeded i go from github, same as meta-qt5 May 31 02:17:48 berndhs: the github copy of meta-openembedded is synced automatically, so no issue there May 31 02:18:14 AFAIK meta-qt5's canonical location is on github May 31 02:18:22 so no issue there either May 31 02:19:42 it almost completes building rpi-test-image now May 31 02:20:17 last few tasks if blows up, license_deployed_manifest May 31 02:21:11 no such file or directory on recipeinfo, those look like licensing files **** ENDING LOGGING AT Wed May 31 03:00:02 2017