**** BEGIN LOGGING AT Wed Mar 11 02:59:58 2015
Mar 11 07:38:32 <ronob> hi guys
Mar 11 07:39:16 <ronob> can you check this error while giving  bitbake -v tegra-genivi-image in 64 bit ubuntu 14.04 machine
Mar 11 07:40:29 <ronob> ERROR: Function failed: do_configure (log file is located at <>/..tmp/work/x86_64-linux/cmake-native/2.8.11.2-r0/temp/log.do_configure.12076)
Mar 11 08:00:31 <ronob> hello anybody here
Mar 11 08:01:11 <ronob> ERROR: Function failed: do_configure (log file is located at <>/..tmp/work/x86_64-linux/cmake-native/2.8.11.2-r0/temp/log.do_configure.12076)
Mar 11 08:01:16 <ronob> please have a look
Mar 11 08:55:01 <Euan> Hi, I am trying to setup package management for my image. I have set PACKAGE_CLASSES ?= "package_ipk", EXTRA_IMAGE_FEATURES = "package-management" and FEED_DEPLOYDIR_BASE_URI = "http://192.168.7.1/repo" in local.conf but /etc/opkg/base-feeds.conf is 0 bytes. Have I missed something?
Mar 11 09:48:54 <funman> should core-image-minimal have a dhcp client?
Mar 11 09:49:47 <funman> my odroid-c1 board is booting but without anything being displayed over hdmi or network (or serial) I can't check if it's working properly
Mar 11 09:53:10 <bluelightning> morning all
Mar 11 10:05:22 <funman> well, not sure i make it past u-boot
Mar 11 10:05:41 <funman> better wait for the serial cable to arrive in the mail
Mar 11 10:06:25 <ndec> Euan: not quite sure about FEED_DEPLOYDIR_BASE_URI, i generally use PACKAGE_FEED_URIS for that.
Mar 11 10:09:18 <Euan> ndec: Thanks will give that a try
Mar 11 10:37:42 <Euan> ndec: PACKAGE_FEED_URIS worked perfectly, thanks again.
Mar 11 11:11:39 <_4urele_> hi
Mar 11 11:41:30 <simonl> Is there a way to tell the systemd bbclass to only enable a service, not (re)start it?
Mar 11 11:48:36 <rburton> simonl: see the class source - SYSTEMD_AUTO_ENABLE *actually* controls whether its restarted on install
Mar 11 11:48:41 <rburton> oh no it doesn't
Mar 11 11:48:43 <rburton> hm
Mar 11 11:48:55 <rburton> no, it appears that enabled services are started on install
Mar 11 11:49:09 <rburton> which does seem sensible, why would you not want it started not but started on reboot?
Mar 11 12:19:31 <soderstrom> I've set up a u-boot recipe in a clone of the meta-xilinx layer. The recipe requires the 'meta-xilinx/conf/machine/include/machine-xilinx-default.inc' which in turn depends on 'meta-xilinx/recipes-bsp/device-tree/device-tree.bb'. This depends on 'dtc-native' in yocto. My problem is.. the dtc is never built when running "bitbake u-boot-xlnx" it is unable to find the device tree compiler "dtc: command not found". Any idees on what is going on here? The de
Mar 11 12:20:54 <simonl> rburton: because of reboot installs/uppgrades.
Mar 11 12:21:41 <simonl> We have services that handle devices that need to be flashed on upgrades, which can't be done while normal services are running
Mar 11 12:22:08 <rburton> simonl: doesn't systemd have an exclusive thing you can use for that?
Mar 11 12:22:30 <simonl> rburton: yes, http://www.freedesktop.org/wiki/Software/systemd/SystemUpdates/
Mar 11 12:23:07 <simonl> which we use, though quite trivially by installing the packages themselves in that mode
Mar 11 12:23:45 <simonl> which AIUI is the intended usage. So that critical binaries etc are not replaced while the whole system is running
Mar 11 12:25:26 <simonl> anyway, what happens is, some service is updated and the postinst script starts the service which pulls in other services that it needs and so on
Mar 11 12:25:35 <simonl> rburton: It wreaks havoc :P
Mar 11 12:26:16 <rburton> hm, yeah
Mar 11 12:27:54 <rburton> patch systemd.bbclass? :)
Mar 11 12:28:16 <rburton> assuming systemctl isoloate doesn't stop other services starting afterwards
Mar 11 12:28:45 <simonl> I'm considering it, but I'd rather keep the poky tree itself unpatched... can a bbappend replace a function?
Mar 11 12:44:28 <joshuagl> you could create your own systemd-foo.bbclass that chains up to the stock systemd.bbclass, but that could soon get messy
Mar 11 13:32:18 <simonl> joshuagl: right. Perhaps patching systemd.bbclass is better
Mar 11 13:36:49 <rburton> simonl: your layer can replace systemd.bbclass
Mar 11 13:37:10 <rburton> and send the patch so you don't have to maintain it for long :)
Mar 11 13:38:09 * simonl wonders why he didn't think of that
Mar 11 13:38:21 <simonl> rburton: sounds like a way forward :)
Mar 11 13:44:58 <bluelightning> note that whether or not a class can be replaced will depend on BBPATH, which in turn depends on how your layer.conf extends it and the order in which your layer appears within bblayers.conf
Mar 11 13:45:22 <rburton> details
Mar 11 13:45:28 <rburton> bluelightning: shouldn't you be on holiday?
Mar 11 13:45:35 <bluelightning> rburton: I'm back today :p
Mar 11 13:45:38 <rburton> ah
Mar 11 13:45:41 <rburton> welcome
Mar 11 13:46:19 <rburton> look, a joshuagl!
Mar 11 13:46:45 <bluelightning> details yes, but you have to be aware of them otherwise you can be confused as to why your modified class isn't being picked up (and you may not get any notification if it isn't)
Mar 11 13:46:54 <bluelightning> to be honest, it's not something I really recommend
Mar 11 13:53:49 <simonl> bluelightning: I would have assumed it was subject to the layer priority... but since it is a path variable I guess it makes sense
Mar 11 13:54:17 <bluelightning> simonl: right, the layer priority doesn't influence it
Mar 11 13:54:48 <bluelightning> something we probably ought to fix if/when we rework how layers work in a future release
Mar 11 13:56:16 <simonl> right.
Mar 11 13:59:10 <paulg> anyone reported/fixed the vim wreckage before I look into it?
Mar 11 13:59:21 <paulg> ERROR: No checksum specified for /home/paul/poky/build/downloads/vim_vim.googlecode.com_.hg._v7-4-481.tar.gz, please add at least one to the recipe:
Mar 11 14:05:34 <T0mW> I have code in a git repo that has been updated outside of my yocto project.  If I run a 'bitbake -c clean FOO', will the image build for FOO cause a pull from the repo to get any updates?  IOW, if I have a cloned git repository are part of a recipe, how can it sync with the remote?
Mar 11 14:06:31 <T0mW> Up to this point, I've been doing a 'bitbake -c cleansstate FOO' which I belive will delete the cloned copy.
Mar 11 14:10:27 <T0mW> nm, found the answer.
Mar 11 14:11:02 <rburton> ${AUTOREV} if you want it to always build the latest revision for development purposes
Mar 11 14:12:07 <T0mW> rburton: ah, yeah, I see I put that in my recipe. thanks
Mar 11 15:21:26 <soderstrom> regarding the issue I mentioned earlier, if I first do "bitbake dtc-native" followed by "bitbake u-boot-xlnx" everything build fine. I still think the chain of dependencies for 'u-boot-xlnx' looks ok. But when issuing the 'bitbake u-boot-xlnx' it never builds the device tree compiler. At first i thought I was having some cache issues, but even on a clean build this issue remains. Is this a known problem for 'meta-xilinx' on branch 'dizzy'?
Mar 11 16:18:51 <TuTizz> Hi all, what is the better way to chown a file on a recipe.bbappend when the do_install_append is already implement in the recipe.bb (I would like to chown foo /etc/default/ntpdate in ntp_4.2.6p5.bbappend)
Mar 11 16:20:53 <rburton> write another do_install_append()
Mar 11 16:21:15 <rburton> do_install_append says "append this to do_install", so you can have many of them
Mar 11 16:21:43 <TuTizz> it will not remove the recipe.bb one?
Mar 11 16:23:51 <TuTizz> I thought that will rewrite it
Mar 11 16:29:50 <rburton> otavio: can you test the uboot rename patch?
Mar 11 16:38:42 <rburton> Crofton: yocto powers facebook!
Mar 11 16:46:13 <Crofton|work> We need Yocto inside stickers
Mar 11 16:46:23 * Crofton|work trolls Intel
Mar 11 16:46:45 <Crofton|work> zecke found that
Mar 11 17:06:53 <rburton> i'd love to have that discussion with legal about producing yocto inside stickers
Mar 11 17:07:26 <rburton> facebook's ops team do seem pretty awesome
Mar 11 17:07:52 * rewitt goes to work at facebook
Mar 11 17:08:15 <otavio> rburton: yes; I can. I will do it later today
Mar 11 17:09:41 <tomaz> people - I'm trying to create a new meta layer to compile samba 4.2, based on the meta-oe one, that has samba 3.6
Mar 11 17:09:43 <Crofton|work> I've already trooled the advocacy people by suggesting those
Mar 11 17:09:48 <Crofton|work> I've already trolled the advocacy people by suggesting those
Mar 11 17:10:06 <otavio> rburton: facebook is using yp?
Mar 11 17:10:15 <rburton> https://code.facebook.com/posts/1601610310055392/introducing-openbmc-an-open-software-framework-for-next-generation-system-management
Mar 11 17:10:15 <Crofton|work> The Dec fanboys I used to work with managed to find some cat litter inside stickes and put them on PC's
Mar 11 17:10:17 <tomaz> I'v tried to set the bb file to inherti waf and removed the do_configure so it would call the waf correctly
Mar 11 17:10:26 <tomaz> but now I'm getting
Mar 11 17:10:34 <Crofton|work> Yes world, meta-facebook is a thing
Mar 11 17:10:41 <rburton> our work here is done
Mar 11 17:10:42 <tomaz> line 104: /data/yocto/poky/build-qemu/tmp/work/i586-poky-linux/samba/4.2.0-r0/samba-4.2.0/source/waf: No such file or directory
Mar 11 17:11:16 <tomaz> anyone can give me a clue? the documentation says to only use 'inherti waf' and things should work magically
Mar 11 17:11:56 <benjamirc> rburton: perhaps if Microsoft delivers images through Yocto I would consider our work done
Mar 11 17:14:21 * Crofton|work wonders why we haven't updated samba?
Mar 11 17:17:03 <tomaz> Crofton|work, that's what I'm trying to do... I'd appreciate a bit of help - very new to yocto / poky and the docs are a bit dry
Mar 11 17:17:45 <Crofton|work> are there a lot of differences?
Mar 11 17:18:38 <tomaz> Crofton|work, the buildsystem changed from autotools to waf
Mar 11 17:18:53 <Crofton|work> oh lovely
Mar 11 17:19:00 <Crofton|work> yet another build system
Mar 11 17:19:04 <tomaz> and I can't seem to make waf works ( by using inherti waf ) because it says
Mar 11 17:19:22 <tomaz> build-qemu/tmp/work/i586-poky-linux/samba/4.2.0-r0/samba-4.2.0/source/waf: No such file or directory
Mar 11 17:19:26 <Crofton|work> don't you mean "inherit waf"
Mar 11 17:19:40 <tomaz> that - sorry, my english is not perfect :)
Mar 11 17:19:44 <Crofton|work> ok
Mar 11 17:20:13 <tomaz> I can compile another app that's waf based ( glcompbench, from the meta-oe ) without issues
Mar 11 17:20:39 <Crofton|work> where is the waf file in the samba source
Mar 11 17:20:50 <tomaz> samba/buildtools/bin/waf
Mar 11 17:21:07 <tomaz> but I think that if I used 'inherit waf' it should use the yocto waf one, right?
Mar 11 17:21:26 <Crofton|work> so it looks like the waf file isn't where oe expects it
Mar 11 17:21:42 <Crofton|work> read waf.bbclass and see if there is a variabel to set
Mar 11 17:22:50 <tomaz> ugh
Mar 11 17:22:53 <tomaz> just doing that
Mar 11 17:23:14 <tomaz> it doesn't tries to find the waf, nor it uses the system one, it expects the project to have waf on the source directory
Mar 11 17:23:49 <tomaz> if I try to search for the waf file and use that one if it doesn't find in the current folder, would that be upstreamed?
Mar 11 17:24:49 <Crofton|work> a good solution that builds samba-4.x would likely be accepted
Mar 11 17:25:02 <tomaz> Crofton|work, :)
Mar 11 17:25:05 <tomaz> you just made me smile.
Mar 11 17:25:51 <Crofton|work> rburton, seems to be the waf expert
Mar 11 17:46:11 <Crofton|work> shoud I file a bug about zip not installing unzip over the busybox unzip?
Mar 11 17:51:49 <bluelightning> Crofton|work: yes
Mar 11 17:51:57 <Crofton|work> ok
Mar 11 17:52:04 <bluelightning> (I guess we just need an alternative set up)
Mar 11 17:52:11 <bluelightning> (for unzip)
Mar 11 17:52:14 <Crofton|work> I'd try and fix it, but I'm afraid I will forget
Mar 11 17:52:22 <Crofton|work> before I get around to fixing
Mar 11 17:52:26 <bluelightning> thx
Mar 11 17:59:39 <Crofton|work> low hanging fruit: https://bugzilla.yoctoproject.org/show_bug.cgi?id=7446
Mar 11 17:59:40 <yocti> Bug 7446: normal, Undecided, ---, saul.wold, NEW , Installing the zip package in a file system leaves unzip linked to busybox
Mar 11 17:59:51 * Crofton|work hopes
Mar 11 18:21:28 <rburton> Crofton|work: *so* not the waf expert
Mar 11 18:21:39 <Crofton|work> rofl
Mar 11 18:21:44 <Crofton|work> I just looked at the commit
Mar 11 18:21:53 <rburton> i hate waf as much as i hate any other build system that isn't automake
Mar 11 18:22:23 <Crofton|work> so say we all
Mar 11 18:23:07 <Crofton|work> tomaz is tryin gto buils samba-4.x which uses it and the waf file isn't where the class expects it
Mar 11 18:27:59 <tomaz> rburton, Crofton|work : I'v successfully made it find the correct waf in a way that it will not break other projects
Mar 11 18:28:10 <tomaz> rburton, I hate automake, I love CMake. ;p
Mar 11 18:28:25 <tomaz> now, I'm facing *another* issue,
Mar 11 18:28:28 <tomaz> | Checking for library python2.7                                                    : yes
Mar 11 18:28:28 <tomaz> | Checking for program python2.7-config                                             : /usr/bin/python2.7-config
Mar 11 18:28:28 <tomaz> | Checking for custom code                                                          : Could not find the python development headers
Mar 11 18:28:29 * Crofton|work htes cmake
Mar 11 18:29:03 <tomaz> I'v tried to find " *python*.bb*" looking for python-dev, python_dev, pydev or anything like that, but failed.
Mar 11 18:29:12 * rink_ hates cmakes lack of documentation
Mar 11 18:29:17 <rink_> the system itself is actually pretty okay-ish
Mar 11 18:29:51 <kergoth> tomaz: you arent' inheriting the right classes to find or depend on python-native
Mar 11 18:30:38 <tomaz> aha, I was looking for python-dev, but it should be python-native
Mar 11 18:30:39 <tomaz> thakns.
Mar 11 18:31:21 <Crofton|work> inherit python-native?
Mar 11 18:31:23 <kergoth> depending on it alone won't do it, it's binaries are installed to a separate prefix. you need to inherit pythonnative
Mar 11 18:32:06 <Crofton|work> right no - :)
Mar 11 18:32:12 <kergoth> :)
Mar 11 18:33:20 <tomaz> I'm loving yocto, it's good, well structured. but it's *SO* huge that somtimes find the correct answer for a simple question is hard =s
Mar 11 18:33:49 <kergoth> heh, yeah, that's pretty common. high flexibility, high power, but high complexity
Mar 11 18:36:07 <tomaz> whoa, if I use python-native, I get *lots* of errors.
Mar 11 18:36:19 <kergoth> what do you mean by 'use'?
Mar 11 18:36:29 <tomaz> inherit pythonnative
Mar 11 18:36:45 <tomaz> let me pastebin-it
Mar 11 18:37:35 <tomaz> if I *don't* inherit pythonnative, it finds python and only fails when it tries to find the dev-headers
Mar 11 18:37:38 <tomaz> if I use, this happens:
Mar 11 18:37:49 <tomaz> https://paste.kde.org/pzdvbjq44
Mar 11 18:38:25 <kergoth> add two lines: export BUILD_SYS and export HOST_SYS
Mar 11 18:38:37 <realBigfoot> guys does bbappend has to be in the same folder of the original recipe ?
Mar 11 18:38:48 <Crofton|work> realBigfoot, no
Mar 11 18:39:03 <kergoth> the distutils/setuptools classes handle it automatically, but non-distutils-based-buildsystems need to export them themselves if they use the site package to get at paths
Mar 11 18:39:29 <realBigfoot> Crofton|work, how would yocto know that we are appending things for the recipe INSERT_NAME_HERE ?
Mar 11 18:39:50 <Crofton|work> this is ain your own layer?
Mar 11 18:40:00 <kergoth> bbappend filesystem path doesnt' matter, its filename does, along with its inclusion in BBFILES in the layer.conf
Mar 11 18:40:13 <tomaz> kerogoth: thanks.
Mar 11 18:40:18 <kergoth> purely filename based with regard to matching up with the recipe
Mar 11 18:40:20 <kergoth> tomaz: np
Mar 11 18:40:39 <tomaz> I'm upgrading the waf.bbclass with all the info that you guys are providing me, and I'll upstream it as soon as I have something that compiles.
Mar 11 18:40:49 <kergoth> ah, nice
Mar 11 18:40:49 <realBigfoot> Crofton|work, yes
Mar 11 18:40:54 <realBigfoot> kergoth, I see... thanks
Mar 11 18:42:28 <tomaz> kergoth, I'v added those two lines in waf_fo_configure in the waf.bbclass, but same error raised
Mar 11 18:42:35 <tomaz> waf_do_configure*
Mar 11 18:42:54 <kergoth> no, either the export lines are done in the recipe, not the task, or you'd need to specify their value
Mar 11 18:43:07 <kergoth> 'export' at the bitbake level controls whether they end up defined in the shell at all
Mar 11 18:43:26 <kergoth> so either export there, or in the shell function do e.g. BUILD_SYS="${BUILD_SYS}" etc
Mar 11 18:43:35 <kergoth> that's twhat the distutils classes do, feel free to read them
Mar 11 18:43:57 <tomaz> reading them.
Mar 11 18:45:02 <tomaz> but this means that *every* app that uses waf would need to export those two variables
Mar 11 18:45:04 <tomaz> right?
Mar 11 18:45:19 <tomaz> aaaand, exporting those two in the recipe gives me the same error.
Mar 11 18:46:34 <kergoth> I don't understand the question. classes can export just fine
Mar 11 18:46:42 <kergoth> anything a recipe can do, a class can do
Mar 11 18:46:44 <kergoth> same syntax
Mar 11 18:47:34 <tomaz> well, I added in the waf_do_compile, on the top of the class, and on the recipe. nothing changed, same error.
Mar 11 18:47:46 <tomaz> on the recipe - samba_4.2.0.bb
Mar 11 18:51:55 <kergoth> given the error is about those variables being undefined, that seems quite unlikely
Mar 11 18:51:58 <tomaz> AH, worked by inheriting distutils
Mar 11 18:52:10 <tomaz> good. <3
Mar 11 18:52:16 <tomaz> now to the next error. :3
Mar 11 18:52:29 <kergoth> if you don't use distutils, you shouldn't be inheriting the distutils classes
Mar 11 18:52:58 <kergoth> at least, not distutils.. i could see an argument for distutils-base
Mar 11 18:53:06 * kergoth shrugs
Mar 11 18:59:35 <tomaz> well, will try the later as soon as the compilation fails here
Mar 11 19:02:44 <tomaz> did.
Mar 11 19:02:49 <tomaz> now to the next few errors
Mar 11 19:08:13 <Xz> hi guys, what's the best place for 'PREFERRED_VERSION_uclibc' variable?
Mar 11 19:08:48 <Xz> can I do it e.g. in layer.conf?
Mar 11 19:22:43 <tomaz> Building the AD DC requires GnuTLS (eg libgnutls-dev, gnutls-devel) for ldaps:// support and for the BackupKey protocol
Mar 11 19:22:50 <tomaz> but I added the PACKAGE += " gnutls"
Mar 11 19:22:56 <tomaz> PACKAGES*
Mar 11 19:27:15 <kergoth> i think your'e misunderstanding the variables
Mar 11 19:27:22 <kergoth> PACKAGES is what binary packages will be emitted by your recipe
Mar 11 19:27:27 <kergoth> if your recipe isn't gnutls, that's wrong
Mar 11 19:27:29 <kergoth> you want DEPENDS
Mar 11 19:31:38 <tomaz> oh
Mar 11 19:31:44 <tomaz> I was misunderstanding.
Mar 11 20:08:42 <_obad_> hi.  does anyone know of an easy way of listing the contents of an ipk?  (i.e. without using ar & tar to extract & list)
Mar 11 20:09:31 <mario-goulart> _obad_: dpkg -c <ipk file>
Mar 11 20:09:51 <_obad_> mario-goulart: argh I have to use fedora at work :(
Mar 11 20:10:07 <dgm816> anyone know if there is a package that holds /etc/services? i've got ntpd installed, but i have to manually add ntp to it to make it work..
Mar 11 20:10:20 <toxickore> _obad_: could be worst, what about using winME
Mar 11 20:10:40 <_obad_> toxickore: tall building. window.
Mar 11 20:10:47 <mario-goulart> _obad_: ar p <ipk file> data.tar.gz | tar tvzf -
Mar 11 20:11:28 <_obad_> mario-goulart: thanks!! that worked.  I didn't know about the 'p' command
Mar 11 20:11:41 <mario-goulart> yw!
Mar 11 20:12:08 <dgm816> ah.. its netbase..
Mar 11 20:20:56 <kergoth> it's handy setting up shell scripts/aliases for examining ipks like that
Mar 11 20:21:48 <kergoth> https://github.com/kergoth/dotfiles/blob/master/scripts/ipkinfo, https://github.com/kergoth/dotfiles/blob/master/scripts/ipkcontents, https://github.com/kergoth/dotfiles/blob/master/scripts/ipkunpack
Mar 11 20:27:03 <tomaz> 'configure finished sucessfully: Error do_configure' o_O
Mar 11 20:27:09 <tomaz> https://paste.kde.org/ptufeud9b
Mar 11 20:27:16 <tomaz> fighting this for about 20 minutes now =/
Mar 11 20:39:04 <tomaz> I'v used the command "          waf configure --prefix=${prefix} ${EXTRA_OECONF} -b ${S} " , but the ${S} is the HOST, not the CHROOT one,
Mar 11 20:39:17 <tomaz> and it's giving me an error while trying to compile.
Mar 11 20:39:27 <tomaz> maybe I should stop trying to use a different folder for compilation.
Mar 11 20:40:12 <kergoth> we don't use chroot..
Mar 11 20:41:28 <tomaz> kergoth, then I have no idea why it failed. :D
Mar 11 20:41:42 <tomaz> if I removed that -b ${S} and changed another thing, it works
Mar 11 20:42:10 <tomaz> uh, no, it doesn't.
Mar 11 20:42:11 <kergoth> we don't really support docbook documentation generation at this time
Mar 11 20:42:15 <kergoth> you should figure out how to disable it
Mar 11 20:42:32 <tomaz> kergoth, okay, I'll disable it, but it's not failing because of that
Mar 11 20:42:35 <tomaz> that's just a warning
Mar 11 20:43:12 <kergoth> that path doesn't look like a hsot include or library path to me
Mar 11 20:43:25 <kergoth> methinks the check is confused, or the error is misleading, one of the two
Mar 11 20:43:49 <kergoth> generally that error is for when it starts poking at /usr/include or so
Mar 11 20:43:54 <tomaz> hhmmm..
Mar 11 21:08:43 <tomaz> hm......
Mar 11 21:08:54 <tomaz> maybe I should use --cross-compile option on Samba 4.2
Mar 11 21:10:23 <tomaz> what should be the compiler to pass to --cross-execute= ? ${CC} ²
Mar 11 21:10:24 <tomaz> ?
**** ENDING LOGGING AT Thu Mar 12 02:59:58 2015