**** BEGIN LOGGING AT Thu Apr 03 02:59:59 2014
Apr 03 06:56:13 <mem47> Morning, I've have a question. I'm trying to create a recipe that fetchs from a git server some kernel module sources, that should be place in a fixed folder in the kernel's workdir, and I would like to add it to the kernel recipe that I have. Is this possible?
Apr 03 07:14:27 <jmleo> Hi there
Apr 03 07:14:47 <jmleo> I have a fetch issue, and I can't find out what I did wrong...
Apr 03 07:15:05 <jmleo> Fetcher failure: Unable to find revision 4d42368214f2fd102b2d163c086e0f2d8c166dc6 in branch master even from upstream
Apr 03 07:15:07 <jmleo> Function failed: Fetcher failure for URL: 'git://github.com/xianyi/OpenBLAS.git;protocol=https'. Unable to fetch URL from any source.
Apr 03 07:15:31 <jmleo> but when I go to github and check the log, this is the correct commit
Apr 03 07:32:34 <mem47> I found the solution, in the SRC_URI variable adding the "destsuffix" with the desire folder
Apr 03 09:09:18 <ducuy> hi all, when i used bitbake linux-yocto, the populate_sysroot seem not do anything, so tmp/sysroot/<platform>/usr/src/kernel is empty
Apr 03 09:09:48 <ducuy> and i have to use bitbake linux-yocto -c populate_sysroot -f, then it work
Apr 03 09:10:17 <ducuy> does anyone know i missing something in configuration?
Apr 03 09:33:11 <bluelightning> morning all
Apr 03 09:45:21 <linu1> hi all i got the following error while compiling kernel 3.6.9, and i have used poky with meta-atmel. please see http://pastebin.com/qDu0Wpxm
Apr 03 09:50:07 <jvuo> hi all! question about kernel module development: I'm trying to add  my own module to kernel with .bbappend file and patch to Kconfig, result is after compile my_modules-dirty and my_modules in deploy folder. And also in rootfs there is /lib/modules/kernel_version-dirty/ and /lib/modules/kernel-version/. This dirty version is 'right one' and includes modules I need. How can I resolve this dirty issue? I'm using Arago project.
Apr 03 09:54:41 <linu1> hi, anybody,have you seen my question can you please help me.
Apr 03 10:08:05 <jmleo> Hi all
Apr 03 10:08:40 <jmleo> I finally thought I had the solution I needed, because in yocto (will, in meta-openembedded) there is a libeigen recipe
Apr 03 10:09:05 <jmleo> This is EigenBLAS library, and it should be a very good way to have optimal matrix multiplication with python-numpy
Apr 03 10:09:22 <jmleo> sadly, even if I can compile libeigen, I cannot have it in my image
Apr 03 10:09:35 <jmleo> and I am not sure that python-numpy will see it either
Apr 03 10:10:01 <jmleo> ERROR: libeigen not found in the base feeds
Apr 03 10:11:20 <bluelightning> jmleo: maybe that is the wrong package name - see the actual packages produced by the libeigen recipe (in its workdir)
Apr 03 10:12:04 <jmleo> tmp/work/cortexa9hf-vfp-neon-poky-linux-gnueabi/libeigen/
Apr 03 10:13:10 <ducuy> jmleo: you may not include layer in bblayer.conf
Apr 03 10:14:06 <jmleo> sources/meta-openembedded/meta-oe/recipes-support/libeigen/ ==> this means layer to include is meta-oe ?
Apr 03 10:14:25 <jmleo> BBLAYERS = " \
Apr 03 10:14:26 <jmleo> ${BSPDIR}/sources/meta-openembedded/meta-oe
Apr 03 10:19:39 <ducuy> jmleo: hope your BBLAYERS is correct. i can build with my BBLAYERS += '/local/users/uvan/workspace/yocto/meta-openembedded/meta-oe'
Apr 03 10:22:54 <bluelightning> ducuy: I don't think that's the problem, you wouldn't get to the error jmleo is getting without at least the recipe being available
Apr 03 10:29:55 <Xz> I have a 'make xyz' command that returns non-zero status
Apr 03 10:30:03 <Xz> How do I fail Yocto build if that happens?
Apr 03 10:30:23 <Xz> I execute that 'make xyz' in added, non-default step
Apr 03 10:31:18 <linu1> http://pastebin.com/qDu0Wpxm can you please tell me this error
Apr 03 10:32:02 <ducuy> Xz: thought you have to use bitbake, not make
Apr 03 10:35:26 <rburton> Xz: hm, it should be failing.  pastebin the task?
Apr 03 10:35:40 <linu1> bluelightning, any idea ?
Apr 03 10:37:44 <bluelightning> Xz: try set -e at the top of the shell function
Apr 03 10:38:27 <bluelightning> linu1: have you modified that file?
Apr 03 10:43:35 <linu1> bluelightning, linux-yocto-custom_3.6.9.bb ?  , yes i just modified for patch. please see http://pastebin.com/qJukLzw0
Apr 03 10:47:45 <bluelightning> linu1: you aren't by chance setting DTS_FILE to something in your config are you?
Apr 03 10:48:17 <bluelightning> (or KERNEL_DEVICETREE for that matter)
Apr 03 10:51:33 <linu1> bluelightning, yes i forgot to add dts file,it is now compiling.
Apr 03 10:52:09 <bluelightning> linu1: I believe that part of the code has been changed since the version you're running, hopefully it can't blow up like that anymore
Apr 03 10:53:14 <bluelightning> linu1: btw in case it wasn't clear, DTS_FILE isn't meant to be set, KERNEL_DEVICETREE is though
Apr 03 11:14:01 <kiran__> Can you help abt ERROR: ExpansionError during parsing /home/kirankumar/work/genivi/dora.10.0.1/poky/meta/recipes-core/meta/external-sourcery-toolchain.bb: Failure expanding varia      ble sysroot_stage_all: ShellSyntaxError: LexToken(LPARENS,'(',0,0)
Apr 03 12:30:21 <Denwid> Hi. What's up with +=, =+  and  _append? Im certain it is documented somewhere but I can't find it, does anybody have a pointer?
Apr 03 12:33:36 <bluelightning> Denwid: here's the latest in-development bitbake manual, it covers those things: http://www.yoctoproject.org/docs/1.6/user-manual/user-manual.html
Apr 03 12:40:33 <deuter> hi, I have very slow exa on last vivante drivers
Apr 03 12:40:54 <deuter> how could I test vivante rendering, glxinfo and glxgears work ok
Apr 03 12:46:50 <Denwid> Thanks, bluelightning!
Apr 03 12:47:39 <linu1> hi all i got the following error while compiling the kernel please see http://pastebin.com/9PB8pdgZ can help me
Apr 03 13:27:13 <jmleo> bluelightning, it does not work... and I can't find out why...
Apr 03 13:28:11 <jmleo> it seems to be the correct name
Apr 03 13:28:14 <jmleo> libeigen
Apr 03 13:28:22 <bluelightning> jmleo: what other packages does the libeigen recipe produce?
Apr 03 13:28:45 <jmleo> bluelightning, how can I know that easily ?
Apr 03 13:29:27 <bluelightning> jmleo: have a look under the packages-split directory under the workdir for libeigen
Apr 03 13:29:47 <jmleo> libeigen  libeigen-dbg  libeigen-dev  libeigen-doc  libeigen-locale  libeigen-staticdev
Apr 03 13:38:14 <Xz> rburton: bluelightning: you're right. My make xyz was returning 0 - so build didn't fail. Now I made sure it returns non-zero and build fails, thanks.
Apr 03 13:51:08 <Xz> can I use '_append_class-target' for every variable in Yocto?
Apr 03 13:51:41 <Xz> I have an additional step 'execute_tests' in binutils bbappend
Apr 03 13:51:50 <Xz> however original binutils is extended to native as well
Apr 03 13:52:06 <Xz> I want my tests to be executed only for non-native version :)
Apr 03 13:53:46 <rburton> Xz: yes
Apr 03 13:54:09 <rburton> (see ptest, which does exactly this)
Apr 03 13:56:53 <Xz> so 'do_execute_tests_append_class-target() { ...}' looks valid?
Apr 03 13:57:08 <bluelightning> rburton: not quite, this is running tests on the host, not the target
Apr 03 13:57:21 <Xz> bluelightning: rburton: yes, tests on the host. It's sth like unit tests
Apr 03 13:57:24 <rburton> bluelightning: similar problem, i meant
Apr 03 13:57:36 <rburton> oh, hm.
Apr 03 13:57:58 <rburton> that syntax probably won't work
Apr 03 13:58:07 <Xz> bluelightning: rburton: so these are tests run on host machine, but using target binaries :)
Apr 03 13:58:10 <rburton> actually, might.
Apr 03 13:58:13 <rburton> Xz: try it ;)
Apr 03 13:58:24 <Xz> rburton: I hope not to blow up the server I'm on
Apr 03 13:58:38 <rburton> Xz: remember that's two operators, _append and _class-target
Apr 03 14:09:22 <ronanguirec__> Hi all,
Apr 03 14:10:43 <ronanguirec__> does someone already use getfattr/setfattr, ... inside a devshell ?
Apr 03 14:11:34 <ronanguirec__> ie, is pseudo work with extend attribute?
Apr 03 14:37:37 <JaMa> RP: btw I've just aborted few jenkins jobs and some bitbake threads were left behind (with latest bitbake)
Apr 03 14:37:55 <JaMa> RP: I'll try to send more details later, but now jenkins will be busy for many days
Apr 03 14:57:35 <Xz> should I put my SELECTED_OPTIMIZATION into machine.conf file?
Apr 03 14:58:16 <Xz> as it is CPU specific
Apr 03 15:02:44 <vicky__> hi. I am facing a problem while building my custom image. It returns  error: file /usr/bin/iconv conflicts between attempted installs of libiconv-1.14-r1.cortexa9_vfp_neon and eglibc-utils-2.18-r1.cortexa9_vfp_neon
Apr 03 15:11:53 <bluelightning> vicky__: basically the libiconv recipe and eglibc should not be built in the same build (since eglibc provides iconv functionality internally)
Apr 03 15:12:15 <bluelightning> the libiconv recipe is really intended for use with other libc implementations e.g. uclibc
Apr 03 15:13:27 <vicky__> k. So eglibc-utils is enough?
Apr 03 15:13:34 <bluelightning> vicky__: it should be
Apr 03 15:14:00 <bluelightning> Xz: I think that's the kind of thing that should be in the distro config really; it can still be overridden such that it's machine-specific
Apr 03 16:09:40 <internet_explore> Hey.
Apr 03 16:09:51 <internet_explore> Does anyone know how I can get commit access to oe-alliance-core?
Apr 03 16:10:09 <internet_explore> Or at least be able to send in a commit or a 'pull request' (I don't remember the terminology here) so they can review it.
Apr 03 16:11:11 * kergoth has no idea what oe-alliance-core even is, but from some quick googling, ran across https://github.com/oe-alliance/oe-alliance-core, which as with any other github repo, supports forking anjd pull request generation.. whether thats the correct method or not he has no idea, contact the layer maintainers
Apr 03 16:11:42 <internet_explore> kergoth, Fair enough, cheers. Just wondered, and I remember OE contributers hanging around here.
Apr 03 16:12:03 <kergoth> you'd have better luck with non-yocto oe stuff in the oe channel, i expect
Apr 03 16:13:11 <internet_explore> kergoth, Ah, of course... Sorry, I just joined #yocto out of habit.
Apr 03 16:15:06 <kergoth> no worries
Apr 03 16:18:03 <bluelightning> internet_explore: oe_alliance isn't associated with OE btw, that seems to be another group
Apr 03 16:18:18 <bluelightning> (AFAIK)
Apr 03 16:18:19 <internet_explore> bluelightning, Yea, it's what I turned up.
Apr 03 16:18:36 <internet_explore> bluelightning, I won't hold you to it.
Apr 03 16:19:07 <internet_explore> bluelightning, I've forked their repo, I'll fix all the issues I can find that are preventing compatability with "dora" (If that's not an actual release of OE, it's from the Freescale BSP).
Apr 03 16:19:22 <rburton> what the hell is oe alliance?
Apr 03 16:19:38 <internet_explore> From what I can tell, a group of developers on github.
Apr 03 16:20:53 <rburton> ah, it's not a fork of oe-core, that's good
Apr 03 16:21:00 <rburton> from the name it seemed like a fork
Apr 03 16:21:11 <internet_explore> Ah, yea... that'd suck.
Apr 03 16:21:22 <internet_explore> I wouldn't use it then, but their 'oe-alliance-core' repo contains some recipe that we bbappend to.
Apr 03 16:21:46 <internet_explore> So, now that I'm upgrading poky itself, I need to fix all the deprecated stuff in the recipies we append to that came from oe-alliance-core...
Apr 03 16:23:07 <rburton> fwiw dora is the latest stable release of oe-core
Apr 03 16:23:14 <internet_explore> Ah, good.
Apr 03 16:23:36 <internet_explore> It's simple stuff, really, just f.x replacing "task.bbclass" with "packagegroup.bbclass"
Apr 03 16:23:54 <internet_explore> But I still need some way to carry these changes, and would rather not have yet another company-internal fork of everything... But I'll figure something out, cheers guys.
Apr 03 16:30:26 <mikejanzen> Slightly off-topic question: I have an image running on IMX6 hardware that inherits "core-image" and doesn't include X11/WL/DFB.  At runtime, the attached HDMI monitor turns off every 5 minutes or so (unless a key is pressed or I "echo 0 > /sys/class/graphics/fb0/blank").  How do I stop the image from turning off the monitor?  Or does anyone know where the code would be that turns off the monitor
Apr 03 16:30:26 <mikejanzen> ?  Thanks.
Apr 03 16:31:31 <rburton> mikejanzen: as you can stop it by using a sysfs tweak, its the kernel turning it off
Apr 03 16:32:07 <mikejanzen> Thanks for the pointer.
Apr 03 19:08:20 <florin_> hi everybody
Apr 03 19:08:39 <florin_> I have an issue setting the build dir for a package using an .bbappend
Apr 03 19:08:55 <florin_> I am redefininf the B variable but it does not work
Apr 03 19:09:00 <florin_> I am redefining the B variable but it does not work
Apr 03 19:09:20 <florin_> so: B = "${WORKDIR}/whatever"
Apr 03 19:09:26 <florin_> does not change anything
Apr 03 19:09:56 <florin_> I do a bitbake -e on the .bbappend-ed package and still see B unchanged
Apr 03 19:10:18 <kergoth> look at the comments above the var in bitgbake -e, where it tells you where it was defined
Apr 03 19:10:33 <kergoth> most likely B_pn-recipename was set in the confi gmetadata, overriding B
Apr 03 19:13:45 <florin_> it's weston BUILD dir that I am trying to change
Apr 03 19:14:04 <florin_> I don't see B_pn-weston set anywhere
Apr 03 19:14:22 <kergoth> again, use bitbake -e to see exactly what line in what file defined it the way it is
Apr 03 19:16:26 <rburton> florin_: poky and other distros will set B=WORKDIR/build
Apr 03 19:16:36 <rburton> florin_: i do have to ask why you want to change B for weston
Apr 03 19:17:24 <florin_> exactly, I want it built in the ${S} directory
Apr 03 19:17:55 <florin_> rburton, and can't make B change
Apr 03 19:20:23 <florin_> rburton, I am using a different SRC_URI from a git, set S to point to workdir/git but then bitbake creates a build dir into which I am cd'ed at configure
Apr 03 19:20:33 <florin_> and the build dir is empty
Apr 03 19:20:38 <florin_> hence it fails at configure
Apr 03 19:21:43 <kergoth> We woudn't set B if the project didn't support separate build dir vs source dir.
Apr 03 19:22:20 <rburton> florin_: it should still build fine from git
Apr 03 19:22:26 <rburton> what's the error log?
Apr 03 19:22:36 <rburton> there might be a bug in the recipe that needs a tweak, instead of changing B
Apr 03 19:23:06 <rburton> (but do use -e to verify what is overriding the setting of B, you should be able to change that from a bbappend)
Apr 03 19:23:37 <florin_> it does build fine from git; it works if I just copy the entire recipe and modify SRC_URI and S accordingly; but I want to use a bbappend to be cleaner
Apr 03 19:23:54 <florin_> problem is that it cd's me in a empty directory called build
Apr 03 19:24:00 <kergoth> that's not a problem.
Apr 03 19:24:08 <kergoth> that's exactly what it should do
Apr 03 19:24:08 <florin_> and then it tries to do a autoreconf there
Apr 03 19:24:19 <kergoth> autotools.bbclass runs autoreconf in ${S}, not ${B}
Apr 03 19:24:19 <florin_> I only have a configure.ac
Apr 03 19:24:30 <kergoth> afaik, anyway
Apr 03 19:25:41 <florin_> from log.do_configure: autoreconf: 'configure.ac' or 'configure.in' is required
Apr 03 19:25:55 <florin_> and from run.do_configure
Apr 03 19:26:33 <florin_> cd '/media/hdd/build/tmp/work/i586-poky-linux/weston/1.4.0-r1/build'
Apr 03 19:26:36 <florin_> do_configure
Apr 03 19:26:56 <florin_> but the build directory only has the aclocal-copy directory
Apr 03 19:27:09 <florin_> and errors because of that
Apr 03 19:30:24 <florin_> I just can't see how bitbake -e tells me what actually sets B
Apr 03 19:30:40 <florin_> I am looking at that output and see B is set wrong
Apr 03 19:32:12 <kergoth> maybe you're on really old poky. nowadays bitbake -e tells you exactly how your variables got set to what they are, in the comments above the variable
Apr 03 19:35:38 <rburton> kergoth: configure runs cd B; S/configure
Apr 03 19:37:26 <florin_> I think I am old :)
Apr 03 19:37:37 <florin_> haven't looked at bitbake -e for a while
Apr 03 19:37:53 <florin_> did not know it was also saying how vars are set
Apr 03 19:38:06 <rburton> that's new in the last release
Apr 03 19:38:07 <florin_> and I was just grepping in the bitbake -e output
Apr 03 19:38:10 <florin_> henc seeing nothing
Apr 03 19:38:14 <rburton> ah, yeah, don't do that :)
Apr 03 19:38:23 <rburton> the new -e output is awesome
Apr 03 19:38:24 <florin_> so I found it
Apr 03 19:38:26 <florin_> it was SEPB
Apr 03 19:38:46 <rburton> i'd have said that bbappend should override that
Apr 03 19:38:50 <rburton> so maybe we need more ?= somewhere
Apr 03 19:39:27 <rburton> my real concern is why configure fails with a git clone in separate-build
Apr 03 19:39:29 <kergoth> rburton: i know. he's saying it's trying to run autoreconf in ${B} rather than ${S}, so ${S}/configure wasn't being created. autoreconf can't be run in ${B}, as configure.ac isn't there
Apr 03 19:39:35 <kergoth> yeah, that shouldn't be failing
Apr 03 19:39:43 * kergoth shrugs
Apr 03 19:40:06 <rburton> maybe there's a bootstrap issue, would need to see the log really
Apr 03 19:40:26 <florin_> the run.do_configure?
Apr 03 19:40:46 <rburton> so autoreconf is ran in $S
Apr 03 19:40:50 <rburton> florin_: log.
Apr 03 19:42:13 <kergoth> damn, implicit exception chaining in python 3.3 looks nice
Apr 03 19:42:23 * kergoth wanders off for a bit
Apr 03 19:43:25 <florin_> log says:
Apr 03 19:43:28 <florin_> DEBUG: Executing python function sysroot_cleansstate
Apr 03 19:43:29 <florin_> DEBUG: Python function sysroot_cleansstate finished
Apr 03 19:43:29 <florin_> DEBUG: SITE files ['endian-little', 'bit-32', 'ix86-common', 'common-linux', 'common-glibc', 'i586-linux', 'common']
Apr 03 19:43:29 <florin_> DEBUG: Executing shell function autotools_preconfigure
Apr 03 19:43:29 <florin_> DEBUG: Shell function autotools_preconfigure finished
Apr 03 19:43:31 <florin_> DEBUG: Executing python function autotools_copy_aclocals
Apr 03 19:43:32 <florin_> DEBUG: Python function autotools_copy_aclocals finished
Apr 03 19:43:34 <florin_> DEBUG: Executing shell function do_configure
Apr 03 19:43:36 <florin_> autoreconf: 'configure.ac' or 'configure.in' is required
Apr 03 19:43:39 <florin_> exit 1 from
Apr 03 19:43:46 <florin_> autoreconf --force -v --install
Apr 03 19:43:48 <florin_> and that is it
Apr 03 19:45:35 <rburton> is that the entire log?
Apr 03 19:46:04 <florin_> so I guess I would just stick to using the bbappend to override SEPB which works
Apr 03 19:46:06 <florin_> yes
Apr 03 19:46:18 <rburton> can you pastebin your bbappend please
Apr 03 19:50:04 <florin_> I do have included poky/meta/conf/distro/include/seperatebuilddir.inc
Apr 03 19:50:21 <florin_> which may be the cause
Apr 03 19:51:13 <florin_> seperatebuilddir.inc does at line 689
Apr 03 19:51:14 <florin_> B_pn-weston = "${SEPB}"
Apr 03 19:51:15 <rburton> that's the cause of B!=S, but that should work fine
Apr 03 19:51:25 <rburton> that's my point, a git clone of weston should still build
Apr 03 19:51:31 <rburton> we *want* B!=S
Apr 03 19:51:42 <rburton> so please pastebin your bbappend so i can replicate
Apr 03 19:51:52 <florin_> http://pastebin.com/9j4NRDgA
Apr 03 19:52:13 <kergoth> why arey ou manually calling autoreconf?
Apr 03 19:52:15 <kergoth> no wonder it doesn't work
Apr 03 19:53:01 <rburton> yeah don't do that
Apr 03 19:53:10 <rburton> just set SRC_URI and be done
Apr 03 19:53:31 <rburton> all you need in that is SRREV and SRC_URI
Apr 03 20:03:30 <florin_> the package has a autogen.sh
Apr 03 20:03:35 <florin_> which calls autoreconf
Apr 03 20:03:45 <florin_> I thought I should also call them first
Apr 03 20:03:54 <florin_> oevrlooked the main recipe inherits autotools
Apr 03 20:03:59 <kergoth> all our recipes get autoreconf run by default by autotools.bbclass
Apr 03 20:04:13 <florin_> so yeah, work with only setting SRCREV, SRC_URI and S
Apr 03 20:04:18 <kergoth> doing it yourself just duplicates taht, and runs it in the wrong path, which was breaking you
Apr 03 20:04:32 <florin_> makes sense now
Apr 03 20:04:37 <florin_> thanks guys
Apr 03 22:14:41 <Lucretia> hi just trying to get a beaglebone test build going using instructions from http://blog.sola-dolphin-1.net/archives/4632604.html
Apr 03 22:14:43 <Lucretia> using dora
Apr 03 22:16:11 <Lucretia> getting an error http://pastebin.com/sqRzU9mk as meta-oe/conf/layer.conf is not there
Apr 03 22:16:15 <Lucretia> any ideas?
Apr 03 22:26:12 <kroon> Lucretia, so it really is there ?
Apr 03 22:26:25 <Lucretia> eh?
Apr 03 22:26:52 <Lucretia> $ ls ../meta-oe/conf/layer.conf
Apr 03 22:26:52 <Lucretia> ls: cannot access ../meta-oe/conf/layer.conf: No such file or directory
Apr 03 22:27:21 <Lucretia> this is from the beagleboneblack dir
Apr 03 22:28:20 <kroon> my meta-oe has a conf/layer.conf in it at least
Apr 03 22:31:45 <Lucretia> got it from https://github.com/openembedded/meta-oe.git
Apr 03 22:32:59 <JaMa> my too, but nto in ../
Apr 03 22:33:00 <bluelightning> Lucretia: surely you need one directory below that i.e. meta-oe/meta-oe ... ?
Apr 03 22:33:51 <Lucretia> ahhh, might've typed it in wrong, been ages since i touched this
Apr 03 22:33:55 <jesselang> Lucretia, did you add the meta-oe/meta-oe directory to your bblayers.conf?
Apr 03 22:34:31 <Lucretia> yup that works
Apr 03 22:34:40 <Lucretia> missed the extra /meta-oe
Apr 03 22:34:43 <Lucretia> duh
Apr 03 22:57:55 * kergoth wonders when daisy will be branched
Apr 04 01:28:02 <jesselang> I'm trying to build an image for the beagleboneblack following these instructions: http://blog.sola-dolphin-1.net/archives/4632604.html  One task fails compiling mesa with wayland: http://pastebin.com/JdqwxuVJ Any ideas?
**** ENDING LOGGING AT Fri Apr 04 02:59:59 2014