**** BEGIN LOGGING AT Tue Oct 27 02:59:59 2015
Oct 27 03:44:56 <turbo-joe> sent output in pm qknight
Oct 27 03:44:58 <turbo-joe> ok sry about that
Oct 27 03:45:06 <turbo-joe> Linux arm 4.1.10-ti-r21 #1 SMP PREEMPT Fri Oct 9 06:05:02 UTC 2015 armv7l armv7l armv7l GNU/Linux
Oct 27 03:45:07 <turbo-joe> http://paste.ubuntu.com/12976941/
Oct 27 03:45:07 <turbo-joe> thats the previous command you asked
Oct 27 03:45:07 <turbo-joe> zcat
Oct 27 03:45:17 <turbo-joe> I'm so sorry
Oct 27 03:45:17 <turbo-joe> I never used paste.ubuntu.com before
Oct 27 03:45:17 <turbo-joe> qknight: can you see that something is missing, or maybe something went wrong when I flashed the BBB?
Oct 27 03:45:17 <turbo-joe> and I truly apologize for the spam, I honestly didn't know better as ignorant as that sounds.
Oct 27 04:48:56 <qknight> hm
Oct 27 04:49:04 <qknight> how to setup my own meta-foo
Oct 27 04:49:14 <qknight> it seems not to find my own bb files tough
Oct 27 04:49:19 <turbo-joe> I thought I lost ya for a sec
Oct 27 04:51:47 <turbo-joe> is there an image of ubuntu with a gui? even an older one.
Oct 27 06:20:38 <LetoThe2nd> Ox4: good morning. if i'm not answering here, chances are that i'm not around. generally, keeping support in this channel is always good because it doesn't limit the answers to what one supporter know, and others can profit too
Oct 27 06:20:50 <LetoThe2nd> Ox4: so if nobody answers here, maybe try the ML or such.
Oct 27 06:23:20 <LetoThe2nd> qknight: for creating your own meta layer, the easiest way is to just use yocto-create-layer and then add it int your bblayers.conf
Oct 27 07:54:12 <Ox4> LetoThe2nd: ML?
Oct 27 07:55:34 <LetoThe2nd> Ox4: mailing list
Oct 27 08:06:21 <mckoan> good morning
Oct 27 08:13:54 <Ox4> ok, one more try :) On my board I have serveral partitions: 1) for kernel with partition size 8Mb and 2) for rootfs - 116 Mb. I have to create one img file which includes kernel and rootfs. How can I do that?
Oct 27 08:14:16 <bluelightning> morning all
Oct 27 08:14:55 <bluelightning> Ox4: assuming you're using a recent version, you could use wic
Oct 27 08:15:22 <bluelightning> http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#creating-partitioned-images
Oct 27 08:26:11 <Ox4> bluelightning: yes, I am using git version
Oct 27 08:28:53 <Ox4> bluelightning: currently I am using dd utility
Oct 27 08:29:04 <Ox4> bluelightning: but thank you for the link
Oct 27 08:29:56 <bluelightning> with wic you can still produce a file that you dd to the device,
Oct 27 08:31:31 <Ox4> bluelightning: yes, I am investigating the source of wic
Oct 27 08:46:38 <lazao> hi
Oct 27 08:53:00 <parrot2> bluelightning: ping
Oct 27 08:54:23 <bluelightning> parrot2: pong
Oct 27 08:55:32 <parrot2> bluelightning: not sure if that's your area of responsibility, what does EXTRA_OECONF_BASE, EXTRA_OECONF_INITIAL and INTERMEDIATE mean in http://git.yoctoproject.org/cgit.cgi/poky/plain/meta/recipes-devtools/gcc/gcc-4.9.inc
Oct 27 08:56:39 <bluelightning> parrot2: it has to do with the separate times that we build glibc, though the "intermediate" is obsolete I think
Oct 27 08:57:00 <lazao> is there a way to have something like the bb.utils.contains by giving a wildcard?
Oct 27 08:58:57 <parrot2> bluelightning: so if I plan to enable libgomp, is it not the best way if I simply replace --disable-libgomp in that inc file with --enable-libgomp ?
Oct 27 08:59:35 <parrot2> I do plan a bbappend on this but I'm not sure if the autotools will be happy with both --enable and --disable there
Oct 27 09:00:12 <bluelightning> parrot2: libgomp is already enabled when we build glibc properly, AFAIK
Oct 27 09:00:30 <bluelightning> parrot2: I don't think you want to enable it for glibc-initial
Oct 27 09:02:32 <parrot2> bluelightning: I don't think libgomp is there since I couldn't find it in the deploy/rpm directory. Also, my stock core-image-sato-sdk don't have it either
Oct 27 09:04:02 <bluelightning> parrot2: actually, it seems like libgomp comes from gcc-runtime not glibc
Oct 27 09:04:33 <bluelightning> parrot2: if you build gcc-runtime, or add libgomp to some packagegroup or your image, you will have it
Oct 27 09:05:44 <LetoThe2nd> concerning this gcc fun - is there a known goot way to have gcc-runtime and libgfortran for native to depend on?
Oct 27 09:05:52 <LetoThe2nd> s/goot/good/
Oct 27 09:06:12 <bluelightning> wouldn't you be relying on the host for that?
Oct 27 09:06:24 <LetoThe2nd> hm, would i?
Oct 27 09:07:10 <bluelightning> if it's native, and it comes from gcc, I would have thought so
Oct 27 09:07:40 <bluelightning> I'm not an expert in that area by any means though
Oct 27 09:08:13 <LetoThe2nd> hmhm
Oct 27 09:08:58 <parrot2> bluelightning: I guess I will add libgomp to CORE_IMAGE_EXTRA_INSTALL then ?
Oct 27 09:09:08 <bluelightning> parrot2: that should do it yes
Oct 27 09:09:14 <parrot2> just want to make sure I dont hit my head with hammer again
Oct 27 09:10:11 <LetoThe2nd> bluelightning: then why are we building all those native tools anyways and not relying on the host?
Oct 27 09:10:14 <bluelightning> interestingly though, libgomp and libgomp-dev are in the RRECOMMENDS for packagegroup-core-sdk which should be in core-image-sato-sdk (unless you've removed tools-sdk from IMAGE_FEATURES, that is)
Oct 27 09:10:36 <bluelightning> LetoThe2nd: because a lot of the time we've found unfortunately we can't rely on the host
Oct 27 09:11:02 <bluelightning> I think it's possible we might re-evaluate some of the things we're building
Oct 27 09:11:07 <LetoThe2nd> bluelightning: thats what i mean.
Oct 27 09:17:24 <parrot2> hmm...I tried out this recipe https://github.com/bmwcarit/meta-ros/blob/master/recipes-extended/freeglut/freeglut_2.8.1.bb and everything went well except I couldn't find the resultant rpm package within the tmp/deploy/rpm . Any idea? I first thought it was missing pkgconfig and I inherited pkgconfig as well. But no luck :-(
Oct 27 09:17:28 <lazao> help anyone?
Oct 27 09:20:51 <LocutusOfBorg1> bluelightning, do you have yocto write permissions on git?
Oct 27 09:20:56 <LocutusOfBorg1> hi BTW
Oct 27 09:21:00 <LocutusOfBorg1> "# opkg-utils is for update-altnernatives :("
Oct 27 09:21:04 <LocutusOfBorg1> there is a typo
Oct 27 09:21:08 <LocutusOfBorg1> http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/conf/layer.conf
Oct 27 09:26:36 <bluelightning> LocutusOfBorg1: I don't, I have the ability to send patches, just like all of us ;)
Oct 27 09:29:06 <bluelightning> lazao: there is no function to do that, you can use whatever python functions you like in in-line python, so you could define a function with def ... and then call it
Oct 27 09:32:59 <bluelightning> parrot2: you should look into the logs and packages-split under the workdir for the recipe
Oct 27 09:36:34 <lazao> bluelightning: thanks, so you mean there is no way to have my function properly defined when executing bitbake -e myrecipe? i need to define a python function to override the variable then call it in a do_task?
Oct 27 09:38:45 <bluelightning> lazao: no, there definitely is
Oct 27 09:38:58 <bluelightning> you don't need a task to do this
Oct 27 09:41:01 <lazao> bluelightning: so, if i define a python var_replacement() {} function, how do i call it in my recipe?
Oct 27 09:41:35 <bluelightning> lazao: see meta/recipes-extended/pam/libpam_1.2.1.bb for an example of doing this (search for "get_multilib_bit")
Oct 27 09:42:07 <LocutusOfBorg1> well, sending a patch for a typo fix seems an overkill
Oct 27 09:42:08 <LocutusOfBorg1> :)
Oct 27 09:42:52 <lazao> thanks!
Oct 27 09:46:49 <bluelightning> LocutusOfBorg1: I've got a patch here locally, I can send it out with my next set after the release if nobody gets to it first
Oct 27 09:48:22 <LocutusOfBorg1> wonderful
Oct 27 09:48:44 <LocutusOfBorg1> do you use git send mail to send patches? I tried to use it for a patch on meta-raspberrypi
Oct 27 09:48:57 <LocutusOfBorg1> I admit I like more github pull requests
Oct 27 09:49:27 <bluelightning> I use either git-send-email or our create-pull-request/send-pull-request scripts
Oct 27 09:49:31 <bluelightning> O
Oct 27 09:49:49 <bluelightning> (the latter wrap git-format-patch and git-send-email anyway)
Oct 27 09:50:24 <bluelightning> if you've not yet seen it : http://www.openembedded.org/wiki/How_to_submit_a_patch_to_OpenEmbedded
Oct 27 09:51:24 <jmleo> Is it possible to replace a class like we can replace a bb in custom recipes ? say I want to use a custom classes/devshell.bbclass and create a meta-mymeta/classes/devshell.bbclass will it be used ?
Oct 27 09:52:08 <bluelightning> jmleo: if you ensure your layer's directory appears first in BBPATH
Oct 27 09:52:33 <bluelightning> jmleo: because of order of operations, that can be dependent on the order the layer appears in BBLAYERS in bblayers.conf
Oct 27 09:52:53 <jmleo> bluelightning: ok, thanks
Oct 27 09:53:33 <bluelightning> jmleo: because of that I usually recommend that you avoid doing that; use an additional class, or send a patch for the one you want to modify
Oct 27 09:54:09 <bluelightning> in the end we all benefit if the classes are as flexible as is practical, and you avoid maintenance burden in the future
Oct 27 10:01:03 <xerent> hey guys, I need to put together a kernel patch, and I was planning to do that by diffing kernel sources from work folder. problem is I need to clean that up first, e.g. i'd like for yocto to perform the do_unpack and do_patch steps only
Oct 27 10:01:07 <xerent> any tips?
Oct 27 10:01:30 <fredcadete> bitbake linux-yocto -c cleansstate
Oct 27 10:01:36 <fredcadete> bitbake linux-yocto -c patch
Oct 27 10:01:44 <fredcadete> I guess
Oct 27 10:06:49 <xerent> seems to work :)
Oct 27 10:06:56 <xerent> thanks!
Oct 27 10:27:15 <fredcadete> no problem
Oct 27 10:44:03 <[Sno]> during simplifying meta-java bootstrap I hit (hopefully) one remaining problem
Oct 27 10:44:25 <[Sno]> there're several recipes providing virtual/javac-native
Oct 27 10:44:47 <[Sno]> and there can be only one ...
Oct 27 10:46:45 <[Sno]> since jikes ecj-bootstrap-native provides virtual/javac-native (Java 1.5), Icedtead7-native provides virtual/javac-native (Java 1.7) and openjdk-8-native provides it (Java 8) - and openjdk-8 requires icedtea7 which requires ecj-bootstrap-native which requires jikes-initial-native - how give a user the chance to pick one in the middle and say: that's enough for me
Oct 27 11:43:23 <mtownsend> Hi, I'm using PACKAGE_CLASSES ?= "package_deb" which works great and gives me dpkg, is there a way to stop it from installing apt to?  I'm using Daisy.
Oct 27 11:47:33 <bluelightning> mtownsend: you could set ROOTFS_PKGMANAGE in the image to just "dpkg"
Oct 27 11:53:06 <mtownsend> bluelightning: Thank you, sounds promising, I'll give it go.
Oct 27 12:00:40 <codewarrior_> can someone expain me a difference between do_rootfs[postfuncs] and ROOTFS_POSTPROCESS_COMMAND
Oct 27 12:01:11 <codewarrior_> are the functions there called in the same point?
Oct 27 12:01:31 <codewarrior_> if so which one I should use?
Oct 27 12:04:44 <fredcadete> if I remember correctly, do_rootfs[postfuncs] calls every function in ROOTFS_POSTPROCESS_COMMAND
Oct 27 12:04:50 <bluelightning> er, no
Oct 27 12:04:59 <bluelightning> codewarrior_: if you want to do anything with the actual image contents you will want ROOTFS_POSTPROCESS_COMMAND
Oct 27 12:04:59 <fredcadete> gaaah ok
Oct 27 12:05:30 <bluelightning> do_rootfs[postfuncs] will only get called after all of the image construction and image file generation has been done
Oct 27 12:05:39 <bluelightning> by which time it'll be too late to actually influence it
Oct 27 12:07:18 <soderstrom> I'm working on adding support for NextThingCo C.H.I.P in the meta-sunxi layer. But I'm having issues with getting the files generated under tmp/work/chip.../linux-chip/4.2.0-r0/image into the final root filesystem. I'm currently generating one ext4 system for local mounting and viewing, and also an rootfs.ubi that should be placed on the device. However in both of these the contents from ..4.2.0-r0/image
Oct 27 12:07:24 <soderstrom>  is missing. I feel like I am missing something really obvious, am I?  Is my qeustion related to the current discussion?
Oct 27 12:07:48 <codewarrior_> bluelightning: so ROOTFS_POSTPROCESS_COMMAND is called before do_rootfs[postfuncs]
Oct 27 12:07:57 <codewarrior_> those sound bit similar
Oct 27 12:08:21 <codewarrior_> and it's hard to find good explanation when each "hook" is run
Oct 27 12:08:25 <LetoThe2nd> soderstrom: well does your image recipe contain kernel, devicetree, modules and all that? otherwise they're not pulled in
Oct 27 12:08:41 <bluelightning> codewarrior_: postfuncs are a generic mechanism, available for any function
Oct 27 12:08:55 <bluelightning> codewarrior_: but just like everywhere else, the function itself runs before all of its postfuncs
Oct 27 12:09:11 <codewarrior_> and what if I will use do_rootfs[prefuncs] and corresponding ROOTFS_PREPROCESS_COMMAND
Oct 27 12:09:53 <bluelightning> prefuncs happen before anything is done for the image generation
Oct 27 12:10:08 <bluelightning> see http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#var-ROOTFS_POSTPROCESS_COMMAND and the vars that surround it if you haven't already
Oct 27 12:10:18 <codewarrior_> but are the two I've mentioned executed in different point
Oct 27 12:10:29 <bluelightning> if you could explain what you're trying to do perhaps I could offer advice
Oct 27 12:10:32 <codewarrior_> or it doesn't really matter
Oct 27 12:10:33 <soderstrom> My linux recipe looks like this: https://github.com/soderstrom-rikard/meta-sunxi/blob/master/recipes-kernel/linux/linux-chip_4.1.0.bb
Oct 27 12:10:52 <bluelightning> codewarrior_: absolutely they are, that's what I said
Oct 27 12:11:03 <codewarrior_> bluelightning: I want to store image manifest
Oct 27 12:11:08 <LetoThe2nd> soderstrom: not the kernel recipe defines what goes into the image, the image recipe does.
Oct 27 12:11:12 <bluelightning> codewarrior_: what kind of manifest?
Oct 27 12:11:35 <LetoThe2nd> soderstrom: by default, an image does *not* contain kernel and such, it has to be specifically included.
Oct 27 12:11:35 <soderstrom> but in the machine config I do a lot of wierdness: https://github.com/soderstrom-rikard/meta-sunxi/blob/master-soderstrom/conf/machine/chip.conf  (wrong branch on previous link)
Oct 27 12:11:37 <codewarrior_> like file with device name, build number and time
Oct 27 12:11:52 <bluelightning> codewarrior_: into the image?
Oct 27 12:12:01 <codewarrior_> place it somewhere in /etc/manifest or something
Oct 27 12:12:05 <codewarrior_> yup
Oct 27 12:12:09 <codewarrior_> into the image
Oct 27 12:12:20 <codewarrior_> but as the very last step
Oct 27 12:12:21 <bluelightning> codewarrior_: so it turns out we already have functionality to do this that is fairly easy to enable and customise
Oct 27 12:12:27 <codewarrior_> so that I know all the details
Oct 27 12:12:45 <bluelightning> codewarrior_: meta/classes/image-buildinfo.bbclass
Oct 27 12:13:51 <codewarrior_> I will take a look
Oct 27 12:14:16 <codewarrior_> but is it possible to store build time, device name and some custom fields
Oct 27 12:15:09 <bluelightning> yes, it's easy to do that assuming those are all represented by variable values
Oct 27 12:15:34 <bluelightning> you just set your own value for IMAGE_BUILDINFO_VARS listing the names of the variables you want put in the file
Oct 27 12:16:03 <bluelightning> DATETIME would give you the build date/time
Oct 27 12:16:35 <codewarrior_> is DATETIME the same as used to create the name of the image?
Oct 27 12:17:21 <bluelightning> yes
Oct 27 12:17:28 <soderstrom> LetoThe2nd: Ah, well that do make since :) Thanks
Oct 27 12:17:40 <LetoThe2nd> soderstrom: have fun
Oct 27 12:18:00 <soderstrom> LetoThe2nd: Always :P
Oct 27 12:21:45 <codewarrior_> bluelightning: thanks for your help
Oct 27 12:38:29 <mtownsend> bluelightning: Your suggestion work a treat, many thanks!!
Oct 27 12:41:54 <bluelightning> mtownsend: no problem
Oct 27 12:51:33 <mtownsend> dpkg is complaining that ldconfig is not on the path, is this part of busybox? do I need to alter it's configuration?
Oct 27 12:53:08 <bluelightning> mtownsend: at least when using the default libc of glibc, /sbin/ldconfig is installed from the glibc package
Oct 27 12:54:19 <fredcadete> mtownsend: another possibility is that /sbin is not on the path when you launch dpkg
Oct 27 12:54:47 <mtownsend> bluelightning: I'm using eglibc from the linaro external toolchain I think.  I'm a bit stuffed then.
Oct 27 12:55:50 <bluelightning> mtownsend: hmm ok, could be that they package that differently? I'm not very familiar with that
Oct 27 12:56:14 <bluelightning> mtownsend: can you verify /sbin/ldconfig is not there?
Oct 27 12:56:28 <mtownsend> fredcadete: find / -name "ldconfig" returned nothing
Oct 27 12:57:00 <mtownsend> I'll have a dig around in the linaro stuff to see if it's there somewhere.
Oct 27 13:00:56 <mtownsend> The linaro toolchain has sbin/ldconfig in it's installed directory but it's not been packaged up by Yocto, weirdly the man pages were :)
Oct 27 13:02:50 <bluelightning> er, under normal conditions that would generate a warning/error at packaging time
Oct 27 13:02:52 <mtownsend> Would something like PREFERRED_PROVIDER_ldconfig = "external-linaro-toolchain" work?
Oct 27 13:02:58 <bluelightning> I doubt it
Oct 27 13:03:55 <bluelightning> PREFERRED_PROVIDER only works when there are actual providers of that item along with things depending on it, I don't think that's true for ldconfig
Oct 27 13:04:35 <bluelightning> (when I say actual providers, I mean where more than one recipe states that it PROVIDES that thing)
Oct 27 13:05:55 <mtownsend> ok found the external-linaro-toolchain recipe. I'll see if I can append it and get it to package up ldconfig.
Oct 27 13:48:40 <hanDerPeder> hi, is it possible to have several git repos in SRC_URI? trying to compile a linux kernel with some backports
Oct 27 13:54:34 <bluelightning> hanDerPeder: it is, yes
Oct 27 13:54:57 <jku> hanDerPeder: yes. see example with named URIs in http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#new-recipe-fetching-code
Oct 27 13:55:56 <bluelightning> jku: actually I don't think that section quite covers multiple git URIs, though perhaps it should
Oct 27 13:56:37 <bluelightning> hanDerPeder: you just need to add ;destsuffix onto one (or both) of the SRC_URI entries, to tell it to check out the files to a subdirectory other than "git"
Oct 27 13:56:48 <bluelightning> ;destsuffix=subdirname
Oct 27 13:57:12 <jku> oh that's true
Oct 27 14:02:14 <soderstrom> LetoThe2nd: Thanks again for the help, my rootfs now looks the way I expected it too ^^.
Oct 27 14:02:59 <LetoThe2nd> soderstrom: :-)
Oct 27 14:17:34 <Arjenlodder> I hope I'm at the right place. I ordered a mediabox with a Yocto linux on it.
Oct 27 14:20:27 <LetoThe2nd> Arjenlodder: besides some minor nomenclature, go ahead and state your case of emergency, then wait for rescue :-)
Oct 27 14:23:36 <bluelightning> FYI Arjenlodder and I have been discussing this on #poky
Oct 27 14:23:41 <bluelightning> though this is a better channel
Oct 27 14:23:55 <Arjenlodder> Yeah, that's why I stopped, sorry
Oct 27 14:24:44 <Arjenlodder> I need to install GCC somehow, but it has no package in the smart repository
Oct 27 14:25:55 <hanDerPeder> bluelightning: thank you
Oct 27 14:26:35 <LetoThe2nd> Arjenlodder: so what you're looking for now is a package repository for your box, right?
Oct 27 14:26:47 <Arjenlodder> I guess so, yes
Oct 27 14:28:09 <LetoThe2nd> well then - not knowing any details about your thing, the canonical way would be replicating the distribution that came with it first, and and then build whatever you want for it.
Oct 27 14:30:15 <bluelightning> or poke the provider of the distro to provide package feeds and/or proper instructions on how to use them
Oct 27 14:30:27 <LetoThe2nd> or that :-)
Oct 27 14:35:13 <qknight> LetoThe2nd: yeah, thanks for that layer hints. i had that and it didn't find a package there because foo-1.2.bb vs foo_1.2.bb issue. stupid me
Oct 27 14:37:34 <LetoThe2nd> qknight: ah, nice to hear you found it.
Oct 27 14:38:11 <qknight> LetoThe2nd: yes! ;-)
Oct 27 14:38:12 <qknight> thanks
Oct 27 14:39:05 <qknight> $perl $perlFlags -e 'use DBI; use DBD::SQLite;' fails on the ubuntu host when doing cross compliation of nix to PPC64
Oct 27 14:39:21 <qknight> i'm not sure, do i have to install DBI on the host to get that working?
Oct 27 14:39:39 <xerent> I've created and applied a patch to kernel sources which is applied OK, but building the kernel fails. the error message is kind of vague. it ends after  'LD  drivers/gpu/built-in.o' line and make returns Error 2. Patch only applied to hid driver. How can I get more information about what's failing really
Oct 27 14:43:35 <bluelightning> xerent: we build the kernel in parallel, is it possible the error got printed some way up in the logs?
Oct 27 14:44:41 <xerent> dunno. I just did bitbake linux-yocto-rt so that's the only thing building
Oct 27 14:48:36 <bluelightning> xerent: yes I know, I meant make parallelism rather than multiple tasks
Oct 27 14:50:00 <xerent> yes, I got it
Oct 27 14:50:09 <xerent> I found the real error, thanks :)
Oct 27 14:51:34 <bluelightning> no problem :)
Oct 27 15:31:56 <mtownsend> Is it ok to have ${@bb.utils.contains("TCMODE", "external-linaro-toolchain", "external-linaro-toolchain-utils", "" ,d)} in IMAGE_INSTALL variable of my distro conf?
Oct 27 15:32:11 <qknight> http://paste.ubuntu.com/12980313/ <- why does this only create nix-dev and nix-dbg but not a package called nix?
Oct 27 15:35:03 <LetoThe2nd> qknight: maybe it doesn't install anything into the pathes matched by FILES (purely guesswork!)
Oct 27 15:35:22 <qknight> LetoThe2nd: yes, i might have screwed the 'make install'
Oct 27 15:35:25 <mtownsend> qknight: what's in the tmp/work/../nix/image directory?
Oct 27 15:36:08 <qknight> rpm -qlp for /home/linux/QorIQ-SDK-V1.8-20150619-yocto/build_t4240rdb-64b_release/tmp/work/ppc64e6500-fsl-linux/nix/1.10-r0/deploy-rpms/ppc64e6500/nix-* shows it is basically empty rpm files, the nix-dbg contains some headers
Oct 27 15:36:18 <qknight> which is odd
Oct 27 15:36:44 <qknight> mtownsend: which directory exactly?
Oct 27 15:38:41 <mtownsend> qknight: Where nix is built, it's exact path will depend on your conf.  but it should be in something like tmp/work/xxx/nix/yyy/image
Oct 27 15:39:26 <mtownsend> where xxx is target architecture, mines cortex9t2hf-vfp-neon-oe-linux-gnueabli
Oct 27 15:39:30 <mtownsend> yyy will be version
Oct 27 15:40:51 <mtownsend> My understanding which may be wrong :) is that this is where the files get installed to.
Oct 27 15:41:16 <mtownsend> I'm thinking your do_install is overridding things and not installing to the correct directory.
Oct 27 15:56:54 <fredcadete> Hi. I'm trying to build gstreamer1.0-plugins-ugly with "x264", and getting a QA "installed but not shipped" error
Oct 27 15:57:31 <fredcadete> does someone know if it's working for them?
Oct 27 15:57:37 <fredcadete> or it was just me that screwed up
Oct 27 16:01:08 <mtownsend> Does anyone know the best way of including a package if external-linaro-toolchain is being used.  I tried using TCMODE and bb.utils.contain but this doesn't work.
Oct 27 16:04:26 <fredcadete> in case anyone's interested, I had to add FILES_gstreamer1.0-plugins-ugly-x264 += "${datadir}/gstreamer-${LIBV}/presets/GstX264Enc.prs" to fix the QA error with x264
Oct 27 16:06:17 <mtownsend> +-
Oct 27 16:13:51 <mtownsend> ignore that +-, i must have accidentally leant on the keyboard.
Oct 27 16:17:01 <mtownsend> I was checking the wrong value in TCMODE, with the right value "external-linaro" it works
Oct 27 16:19:33 <qknight> mtownsend: thanks. i think https://nixos.org/wiki/How_to_install_nix_in_home_(on_another_distribution) says i have to set the prefix also, so maybe /usr/local/bin in yocto just won't get me any package called nix-1.10
Oct 27 16:19:48 <qknight> mtownsend: so your assumption with wrong directory holds true
Oct 27 16:20:48 <mtownsend> qknight: Is there not a nix recipe already that you can use?
Oct 27 16:21:26 <qknight> mtownsend: for yocto?
Oct 27 16:21:28 <qknight> no
Oct 27 16:26:24 <mtownsend> qknight: I've just had a look at nix and didn't realise it was a Linux distro. Can Yocto build another Distro?
Oct 27 16:28:01 <rburton> mtownsend: yocto builds whatever distro you want, it isn't a distro itself
Oct 27 16:28:32 <xerent> if you want ubuntu, yocto won't build that for you though
Oct 27 16:28:41 <xerent> it's a tool for creating your own linux distributions
Oct 27 16:33:53 <qknight> mtownsend: i have yocto on ppc64 and i want to deploy apps there using nix
Oct 27 16:34:04 <qknight> mtownsend: nix is a package manager and nixos is using nix as a distribution
Oct 27 16:34:18 <qknight> mtownsend: i want to bootstrap nix with a ppc64 toolchain on my target
Oct 27 16:34:53 <mtownsend> qknight: ok got you.  how does it build nix package manager, cmake, autotools etc?
Oct 27 16:35:25 <mtownsend> qknight: What happens if you comment out your do_install?
Oct 27 16:38:39 <hanDerPeder> im trying to add some backports for my kernel, seems that kconf/conf gets compiled for target, not host. how can I change this?
Oct 27 16:44:28 <hanDerPeder> do kernel recipes switch between native and cross toolchains?
Oct 27 17:08:32 <JaMa> so good feeling to delete 4K+ oe-core e-mails.. pity it won't help getting my pending patches merged or at least getting some feedback on them
Oct 27 17:09:01 <bluelightning> JaMa: what did you just delete?
Oct 27 17:09:23 <JaMa> my local oe-core e-mail folder
Oct 27 17:09:33 <bluelightning> I hope you're not giving up...
Oct 27 17:09:49 <JaMa> bluelightning: I did gave up couple months ago
Oct 27 17:10:06 <rburton> JaMa: the thumb stuff is in a local branch but just been too busy to test/review it properly
Oct 27 17:10:25 <JaMa> but this cleanup has nothing to do with it.. it was just selective cleanup, because mutt was starting to open that folder a bit slower
Oct 27 17:10:49 <rburton> JaMa: was there anything out important and outstanding?  didn't spot anything apart from the thumb things in your world build summary last time i read it.
Oct 27 17:11:15 <bluelightning> FWIW, I am really pushing hard to get this patch process improved (by pushing I mean attempting to muster resources to solve it); I can't promise anything but I am working on it
Oct 27 17:11:26 <JaMa> http://lists.openembedded.org/pipermail/openembedded-core/2015-September/111123.html feedback would be nice
Oct 27 17:12:48 <JaMa> hmm this one was merged, yay!
Oct 27 17:20:27 <nine_pt_> I have a recipe when execute bitbake -c devshell package it have my patches applied, when generate the ipk don't have it ...
Oct 27 17:21:43 <nine_pt_> I don't have any error on log,
Oct 27 17:39:05 * armpit hopes he's not on the deleted email list
Oct 27 17:39:36 <Crofton> zeddii, you awake?
Oct 27 17:57:11 <jatac> what's the best way to copy an image onto a target machine and boot from a hard disk? I used the .hddimg but ran into a "Waiting for removable media..."
Oct 27 21:32:19 <qknight> hey. i compile a c++ project natively on the yocto target platform and it tells me:
Oct 27 21:32:24 <qknight> src/libutil/types.hh:5:18: fatal error: string: No such file or directory
Oct 27 21:32:31 <qknight> so what am i missing?
Oct 27 22:33:23 <qknight> hey, i need some pointers:\
Oct 27 22:33:48 <qknight> https://community.freescale.com/thread/309578 tells me how to use meta-toolchain-qt5
Oct 27 22:34:03 <qknight> and i'm using poky but with daisy repos
Oct 27 22:35:06 <qknight> anyway, the documentation mentiones the poky-eglibc-x86_64-arm-toolchain-qt-1.4.1.sh    SDK, here it is actually called: fsl-networking-eglibc-x86_64-ppc64e6500-toolchain-QorIQ-SDK-V1.8.sh
Oct 27 22:35:32 <qknight> and another thing which bugs me, they both seem to have a very differnet FS structure
Oct 27 22:36:05 <qknight> for instance, in QtCreate i'm supposed to set the compiler to something like: /opt/poky/1.4.1/sysroots/x86_64-pokysdk-linux/usr/bin/armv7a-vfp-neon-poky-linux-gnueabi/arm-poky-linux-gnueabi-
Oct 27 22:37:29 <qknight> but in my installation from fsl-networking-eglibc-x86_64-ppc64e6500-toolchain-QorIQ-SDK-V1.8.sh i have /opt/fsl-networking/QorIQ-SDK-V1.8/
Oct 27 22:39:44 <qknight> and i can't even find an executable file ending with 'gnueabi-'
Oct 27 22:40:05 <qknight> and the ones i find have arm in their name but i'm actually deploying to PPC64
Oct 27 22:48:50 <qknight> ... confusion ...
**** ENDING LOGGING AT Wed Oct 28 02:59:58 2015