**** BEGIN LOGGING AT Thu Sep 25 03:00:01 2014
Sep 25 08:38:30 <bluelightning> morning all
Sep 25 08:38:54 <florian> morning bluelightning
Sep 25 08:40:06 <Net147> if there are multiple entries in the LICENSES field separated by space, does it mean OR or AND?
Sep 25 08:42:38 <bluelightning> morning florian, Net147
Sep 25 08:42:47 <Net147> bluelightning: hi bluelightning
Sep 25 08:43:09 <bluelightning> Net147: neither, you should use & or | as separators
Sep 25 08:43:39 <Net147> bluelightning: I am seeing this for recipes in oe-core...
Sep 25 08:44:47 <Net147> bluelightning: also, in license.manifest "AFL-2 | GPLv2+" in recipe becomes "AFL-2 GPLv2+" in the manifest which is a bit confusing
Sep 25 08:48:04 <bluelightning> Net147: on the face of it that would seem wrong...
Sep 25 08:48:13 <bluelightning> Net147: which recipes have that in it?
Sep 25 08:48:46 <Net147> bluelightning: nss for example has "MPL-1.1 GPL-2.0 LGPL-2.1"
Sep 25 08:49:04 <bluelightning> ok, then that should be fixed
Sep 25 08:49:21 <bluelightning> I wonder if that generates a warning
Sep 25 08:53:46 <Net147> bluelightning: that's the only one that I could see just quickly scrolling through the list of grepped LICENSE entries
Sep 25 08:54:05 <Net147> bluelightning: would be nice to have a warning or error
Sep 25 08:54:29 <Net147> bluelightning: should I create a bug report for license.manifest?
Sep 25 08:54:50 <bluelightning> Net147: yes, please do
Sep 25 08:55:08 <Net147> bluelightning: nss as well?
Sep 25 08:55:11 <bluelightning> and for whatever reason there doesn't seem to be a warning... I could have sworn it used to throw up a warning
Sep 25 08:55:36 <bluelightning> well I'd like to think patching the nss recipe could be done quicker than filing a report ;)
Sep 25 08:55:52 <bluelightning> I guess you'd have to examine the actual licensing conditions though
Sep 25 09:07:01 <Net147> bluelightning: bug reports created
Sep 25 09:07:06 <bluelightning> Net147: thanks
Sep 25 09:17:11 <frsc> Hi! I've got a problem during build of a recipe. Somehow oe_runmake creates -I/usr/include as a parameter for make instead of -I/${sysroot}/usr/include. I can see this when I look into log.do_compile. Can anyone help?
Sep 25 09:38:07 <bluelightning> frsc: it seems unlikely that the system would be injecting that... is there perhaps an EXTRA_OEMAKE value being set in the recipe?
Sep 25 10:00:51 <b00^wk> does the Build Applicance image can only do limited things ?
Sep 25 10:01:27 <b00^wk> i meant , it one can only build limited type of images with it
Sep 25 10:06:04 <frsc> bluelightning: i tried to override the include path by using EXTRA_OEMAKE, but it does not work. Without using EXTRA_OEMAKE there's still that wrong include path generated!
Sep 25 10:09:35 <bluelightning> frsc: is it possible that the makefile itself contains that?
Sep 25 10:09:51 <bluelightning> b00^wk: there is no in-built limitation, no
Sep 25 10:11:56 <b00^wk> bluelightning, seems fails to make some images, as i tried few times now, the qemu-arm  , minimal image, fails to build.    fails with error no syslinux.  i tried to add that package , still fails.   qemu x86 minimal build was fine
Sep 25 10:14:56 <bluelightning> b00^wk: that's odd... it really should not be building syslinux at all for an arm machine
Sep 25 10:15:50 <bluelightning> what do you have selected for image output types? (I think it's under options or advanced options in hob)
Sep 25 10:16:15 <b00^wk> bluelightning,  i have hddimg, bz2,  ext3
Sep 25 10:16:39 <bluelightning> right, hddimg is intended for x86 machines, that's probably the cause of the error
Sep 25 10:17:05 <b00^wk> bluelightning, but if i have no hddimg, how am i suppose to transfer it to the target flash say ?
Sep 25 10:17:11 <bluelightning> although ideally Hob would check for and disallow that currently invalid combination
Sep 25 10:17:24 <b00^wk> i read syslinux works for ARM' too
Sep 25 10:17:31 <b00^wk> bluelightning,  what booloader is used for ARM ?
Sep 25 10:17:36 <bluelightning> it depends entirely on the board you are booting on...
Sep 25 10:17:49 <bluelightning> which board would that be?
Sep 25 10:17:58 <b00^wk> at the moment none :P
Sep 25 10:18:05 <b00^wk> bluelightning,  so yocto has no default arm bootloader ?
Sep 25 10:18:10 <bluelightning> absolutely we do
Sep 25 10:18:14 <b00^wk> :D
Sep 25 10:18:14 <bluelightning> u-boot for example
Sep 25 10:18:24 <b00^wk> yea sounds familiar
Sep 25 10:19:09 <b00^wk> bluelightning, if i leave out hddimg,  then what am i suppose to do with the bz2 say it makes ?
Sep 25 10:19:19 <bluelightning> for qemuarm (or indeed any qemu machine), you just need a kernel and an ext3 image, that's all qemu needs to boot
Sep 25 10:19:33 <bluelightning> the kernel will be produced automatically so you just need to select ext3
Sep 25 10:20:15 <b00^wk> ok, but , suppose, i do have an arm based system, and i do have a microsd flash i want to boot from
Sep 25 10:20:25 <b00^wk> how am i suppose to then put that system there, if i have to hddimg
Sep 25 10:20:49 <b00^wk> besides manually copying that fs and installing bootloader
Sep 25 10:21:41 <b00^wk> bluelightning,  when i made qemu x86, if i'm not mistaken, qemu lanuched it off the hddimg
Sep 25 10:22:54 <b00^wk> i'm probably wrong, but i looked inside hddimg at least :)
Sep 25 10:23:11 <frsc> bluelightning: I'm new to autotools. Where do I have to look for the wrong include path? I've got GNUmakefile.am, GNUmakefile.in and configure.ac
Sep 25 10:23:57 <b00^wk> i mostly used vbox to do that.  so i take a raw image, like hddimg yocto makes, convert it to vbox vmdk (or whatever) image, and then boot it in vbox
Sep 25 10:24:15 <b00^wk> but its exactly same image, not different
Sep 25 10:24:29 <bluelightning> b00^wk: I think for most non-x86 machines we don't currently have a built-in way of producing an SD card image, in the past that has been a separate script shipped with the BSP for the machine (because it actually has to be prepared specifically for the SD card size)
Sep 25 10:25:39 <bluelightning> b00^wk: we do now have an image creation tool (wic) which can do those tasks, but there is no integration of wic into Hob and SD card support really has only just been added to wic in any case in master
Sep 25 10:25:59 <b00^wk> bluelightning,  ok.
Sep 25 10:26:05 <bluelightning> b00^wk: yes but in that case the virtualbox machine is an x86 machine right?
Sep 25 10:26:17 <b00^wk> bluelightning, yea,  yes
Sep 25 10:26:33 <b00^wk> bluelightning,  but if you take raw hddimg, basically dd dump,  and say syslinux arm version
Sep 25 10:26:37 <b00^wk> would that be any different ?
Sep 25 10:27:12 <bluelightning> I don't have any experience with syslinux on arm, and limited experience with SD card formatting across different arm boards to be honest
Sep 25 10:27:23 <b00^wk> bluelightning, on the other note :
Sep 25 10:27:34 <b00^wk> bluelightning,  if i wanted to modify how that x86 image is made,
Sep 25 10:27:39 <b00^wk> where would i look and tchange what ?
Sep 25 10:27:50 <bluelightning> frsc: I'd start by just doing a grep of all of those for the offending -I/usr/include
Sep 25 10:27:52 <b00^wk> bluelightning,  i looked at what Hob made me by default, and i saw this :
Sep 25 10:28:24 <b00^wk> bluelightning,   a fat16, syslinux,  a rootfs imsage on fat16, which then mounted in initrd time
Sep 25 10:28:39 <b00^wk> bluelightning,  what do i do if i want different way to do this  here ?
Sep 25 10:28:58 <b00^wk> bluelightning,  dont' want loopback, want straight fs on target
Sep 25 10:29:15 <bluelightning> in that case, wouldn't you just use the ext3 image?
Sep 25 10:29:30 <b00^wk> bluelightning,  yes but i do want the hddimg too
Sep 25 10:29:37 <b00^wk> because it has all of it
Sep 25 10:29:43 <b00^wk> bluelightning,  no ?
Sep 25 10:30:00 <bluelightning> I think you can use wic to create flexible partition layouts with the bootloader included
Sep 25 10:30:16 <b00^wk> bluelightning,  it has bootloader too.  and, i may want to partition the raw image, and different partiitions
Sep 25 10:30:27 <b00^wk> wic, ok
Sep 25 10:50:51 <gatis1> hello, I was wondering if there is a way to disable md5sum check for SRC_URI=file ?
Sep 25 10:57:41 <rburton> gatis1: there isn't a checksum on file: uris
Sep 25 10:57:59 <rburton> what's the error you're seeing
Sep 25 10:58:51 <gatis1> rburton, there is an error when it compares md5sum of the downloaded file
Sep 25 10:58:58 <gatis1> checksum mismatch
Sep 25 10:59:15 <rburton> so not a file: uri
Sep 25 10:59:18 <gatis1> because fonts from google https://www.google.com/fonts/download?kit=3hvsV99qyKCBS55e5pvb3ltkqrIMaAZWyLYEoB48lSQ always generate new package
Sep 25 10:59:18 <rburton> no, you can't turn that off
Sep 25 10:59:22 <gatis1> so md5sum changes
Sep 25 10:59:27 <rburton> hah, thanks google
Sep 25 11:00:35 <gatis1> rburton, hm, this kind if sucks, is there a workarounds how i could get those fonts on my image? since downloading from recipe does not seem possible
Sep 25 11:01:08 <rburton> that's a good argument for adding an option though
Sep 25 11:01:31 <rburton> write your own do_fetch
Sep 25 11:02:23 <gatis1> i started to use yocto just couple days back, so might too challenging :)
Sep 25 11:03:19 <rburton> do_fetch() { cd ${WORKDIR} ; wget ${SRC_URI} }
Sep 25 11:03:20 <rburton> might work
Sep 25 11:03:47 <gatis1> will try that
Sep 25 11:04:02 <rburton> bonus points for using wget -c and checking the file exists so it doesn't download if its already there
Sep 25 11:04:29 <rburton> super bonus points for pretending to be the fetcher, downloading into ${DL_DIR} and copying it to ${WORKDIR}
Sep 25 11:04:58 <rburton> gatis1: please file a bug against bitbake for a "skip checksum" option with google webfonts as a reason
Sep 25 11:19:07 <gatis> rburton, thanks for the tips, i will create a bug report
Sep 25 11:19:39 <rburton> thanks
Sep 25 12:25:55 <frsc> bluelightning: concerning my question 2 hours ago: I tried to find the erroneous include path, but couldn't find anything. I see "includedir" beeing set to /usr/include globally in bitbake.conf, but I guess thats correct and the sysroot path should get prepended to it.
Sep 25 12:30:50 <frsc> bluelightning: ok. Not completely true: includedir is set in configure script
Sep 25 12:59:53 <Piziwate> Hello ! I've a stange issue with a custom machine .conf... This is my configuration file (based on x86)
Sep 25 13:00:08 <Piziwate> PREFERRED_PROVIDER_virtual/kernel ?= "linux-yocto"
Sep 25 13:00:12 <Piziwate> PREFERRED_VERSION_linux-yocto ?= "3.14%"
Sep 25 13:00:18 <Piziwate> COMPATIBLE_MACHINE_sdn2020 = "sdn2020"
Sep 25 13:00:25 <Piziwate> KMACHINE_sdn2020 = "sdn2020"
Sep 25 13:00:31 <Piziwate> TARGET_ARCH = "i586"
Sep 25 13:00:37 <Piziwate> DEFAULTTUNE ?= "core2-32"
Sep 25 13:00:42 <Piziwate> require conf/machine/include/tune-core2.inc
Sep 25 13:00:49 <Piziwate> require conf/machine/include/genericx86-common.inc
Sep 25 13:01:23 <Piziwate> After "bitbake core-image-minimal" i got QA Issue: Architecture did not match (3 to 62) on /work/sdn2020-poky-linux/linux-yocto/3.14+gitAUTOINC+09424cee64_144595ef62-r0/packages-split/kernel-vmlinux/boot/vmlinux-3.14.0-yocto-standard
Sep 25 13:02:00 <Piziwate> I just want to create a custom machine conf, but this is an x86 32bit arch
Sep 25 13:02:04 <bluelightning> I don't think you should be setting COMPATIBLE_MACHINE and KMACHINE in your machine.conf
Sep 25 13:02:29 <bluelightning> I suspect that TARGET_ARCH shouldn't be set there either, it'll be set automatically via the tune files
Sep 25 13:03:02 <Piziwate> If I remove TARGET_ARCH and COMPATIBLE_MACHINE... the main recipe doesn't buid the kerneé ! :(
Sep 25 13:03:06 <Piziwate> kernel... sorry
Sep 25 13:03:56 <bluelightning> I believe KMACHINE and COMPATIBLE_MACHINE should be set in your kernel recipe
Sep 25 13:05:01 <Piziwate> But I use the standard yocto kernel recipe...
Sep 25 13:07:59 <bluelightning> Piziwate: yes, but I think you should have a bbappend for that which configures the kernel appropriately for your board
Sep 25 13:08:32 <Piziwate> Ok ! I understand... I'll try it !
Sep 25 13:09:00 <bluelightning> but still, TARGET_ARCH shouldn't be set in there either
Sep 25 13:10:32 <Piziwate> Ok ! Thank you !
Sep 25 13:49:19 <b00^wk> tired to giit clone pokey release , i get connection refused. is that myside problem ?
Sep 25 13:49:23 <b00^wk> tried
Sep 25 13:55:54 <b00^wk> git clone -b daisy git://git.yoctoproject.org/poky.git .....   fatal: unable to connect... errno=Connection refused .
Sep 25 13:57:16 <b00^wk> but i can access yoctoproject website fine .
Sep 25 13:59:21 <bluelightning> b00^wk: well, it does seem to work here...
Sep 25 13:59:28 <b00^wk> ok
Sep 25 14:19:39 <challinan> can anyone explain this advisory message:
Sep 25 14:19:41 <challinan> NOTE: Inode check for /scratch/working/build-minnowmx/tmp is unavaliable, will remove it from disk monitor
Sep 25 14:21:38 <rburton> challinan: the filesystem that directory is on doesn't report inode count, so the checker disables itself
Sep 25 14:21:52 <rburton> challinan: i guess your /scratch is a btrfs partition?
Sep 25 14:23:29 <rburton> you could argue that the monitor should silently disable itself
Sep 25 14:23:30 <challinan> well I'll be darned, yes it is
Sep 25 14:23:45 <rburton> because it just leads to people turning up here looking confused :)
Sep 25 14:23:57 * challinan is working on a VM that m halstead built
Sep 25 14:24:04 <challinan> heh
Sep 25 14:25:25 <rburton> just fixing the message and making it disapepar
Sep 25 14:25:35 <challinan> nice!
Sep 25 14:25:42 <rburton> btrfs doesn't run out of inodes so there's no point monitoring them
Sep 25 14:26:24 * challinan has never used btrfs - until now, unknowingly ;)
Sep 25 15:36:18 * kergoth wonders if meta-minnow is still broken
Sep 25 15:36:59 <bluelightning> koen sent a patch to just disable the emgd stuff today
Sep 25 15:38:07 <rburton> it still mentions emgd in the image afaict, so don't expect images that work to pop out
Sep 25 15:38:10 <rburton> but at least it parses
Sep 25 15:40:21 <bluelightning> the question I have is, isn't emgd still needed on that board?
Sep 25 15:53:58 <bluelightning> Darren just commented on this on the meta-intel list
Sep 25 16:37:59 <sgw_> otavio: you around?  Can you have a look at bug #6760
Sep 25 16:38:00 <yocti> Bug https://bugzilla.yoctoproject.org/show_bug.cgi?id=6760 normal, Undecided, ---, otavio, NEW , i.MX6: not possible to run X11 without tearing / with proper vsync
Sep 25 17:09:37 <xerent> Does anyone have a working recipe for kernel module pcspkr? I'm trying to make the PC speaker work.
Sep 25 17:09:55 <xerent> I already added 'beep' app but that requires a driver (of course)
Sep 25 17:12:10 <k-s> guys, why does kmod_git.bb contains INHIBIT_PACKAGE_STRIP based on ptest_enabled? ptest is enabled by default and that blocks kmod being stripped, which adds too much to its size
Sep 25 17:23:20 <dimonic> In order to build with directfb, I have DISTRO_FEATURE_remove x11; DISTRO_FEATURE_append " directfb". How should this really be done?
Sep 25 17:24:22 <dimonic> I mean it seems a little presumptuous in an embedded world to always assume X11?
Sep 25 17:28:18 <bluelightning> dimonic: basically we configure something that provides a reasonable baseline to test the functionality that the system provides; a large portion of the recipes we ship are X11-based and thus that needs to be tested
Sep 25 17:28:55 <bluelightning> dimonic: if you're doing much more than just trying the system out you should really have your own distro config in which you can customise DISTRO_FEATURES to better suit your requirements
Sep 25 17:29:05 <bluelightning> there's a section on that in the manual: http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#creating-your-own-distribution
Sep 25 17:43:52 <dimonic> bluelightning: Fair enough. I was just wandering. I have been diving into Yocto for about 2 weeks now and I still feel like a n00b. I just received the book from Amazon, but it doesn't seem much more than the online stuff.
Sep 25 17:45:05 <dimonic> Although it means that the provided "core-image-directfb" will not build as it stands.
Sep 25 17:45:26 <bluelightning> right... to be honest I'm not sure how useful core-image-directfb really is these days
Sep 25 17:45:48 <bluelightning> it was created back when gtk+ had support for directfb output, but AIUI that was deprecated
Sep 25 17:45:59 <dimonic> bluelightning: I am about to find out!
Sep 25 17:52:46 <dimonic> bluelightning: It seems we are missing a dependency on libvivante-dfb-mx6.
Sep 25 17:53:35 <bluelightning> I'm guessing that is with meta-fsl-arm?
Sep 25 17:54:01 <dimonic> yes, on imx6qsabresd
Sep 25 17:54:14 <bluelightning> otavio: ^
Sep 25 17:55:48 <dimonic> Our use case is to get Kivy working as a development platform. Pygame works. Kivy dies when calling opengl with a weird op-code message.
Sep 25 17:56:51 <dimonic> I was trying the iOS route - which has Kivy use SDL with directfb for graphics output, Opengl from mesa for other stuff. Except that we can't build the frame-buffer stuff either.
Sep 25 18:09:17 <zecke> hi!
Sep 25 18:09:45 <zecke> how comes the srcdir!=builddir changes in dizzy cause so much fallout? I thought if "make distcheck" works.. src!=build would just work?
Sep 25 18:10:23 <kergoth> i expect its silly developers using realtive paths rather than $(top_srcdir) and whatnot, but *shrug*
Sep 25 18:10:56 <zecke> kergoth: yes, what is puzzling me is that make distcheck creates  build/ directory within the sourcedirectory as well
Sep 25 18:11:46 <zecke> the stuff I fixed in Osmocom was either #include "../../config.h" or not searching generated files in builddir.. I am just very puzzled why make distcheck doesn't find such issues. :)
Sep 25 18:47:22 <dimonic> otavio: Any chance of getting core-image-directfb working for imx6sabresd? It is missing "libvivante-dfb-mx6".
Sep 25 19:36:42 <davis_> hi
Sep 25 20:59:08 <challinan> halstead: ready for snapshot, but I need to get access again as soon as possible
Sep 25 20:59:45 <halstead> Okay.
Sep 25 20:59:52 <halstead> About an hour.
Sep 25 21:17:53 <challinan> thanks
Sep 25 21:20:42 <Crofton> arg, why aren't my k config frags making it in to .config
Sep 25 21:22:00 <kergoth> non-linux-yocto kernel?
Sep 25 21:22:27 <Crofton> yes
Sep 25 21:22:41 <Crofton> from daisy
Sep 25 21:22:46 <Crofton> the other files get in ok
Sep 25 21:22:54 <Crofton> just the one I just made
Sep 25 21:24:08 <Crofton> arg
Sep 25 21:24:17 <Crofton> need CONFIG_ in front of option
Sep 25 21:24:40 <Crofton> is it Friday yet
Sep 25 21:24:40 <kergoth> heh, oops
Sep 25 21:24:50 <kergoth> if only
Sep 25 21:25:00 <zenlinux> it might be in Japan?
Sep 25 21:25:39 <Crofton> I do not know how I would cope without #oe and its troublesome stepchild #yocto
Sep 25 21:28:05 <Crofton> now if I could just get git into the sysroot
Sep 25 21:31:15 <rburton> oh hi zenlinux
Sep 25 21:31:23 <zenlinux> heya ross!
Sep 25 21:32:02 <rburton> Crofton: isnt that just a matter of removing the git-native in assume-provided
Sep 25 21:38:13 <Crofton> I suppose I should try this
Sep 25 21:38:42 <Crofton> I understand why it is there, but it seems like we should be able to use build machien git for fethcer
Sep 25 21:38:55 <Crofton> and populate the sysroot when needed
Sep 25 21:41:38 <Crofton> is there a way to remove something from ASSUMEPROVIDED, without deiting bitbake.conf?
Sep 25 21:51:48 <kergoth> i don't see why you don't jjust let cmake use the git on the host. we require it anyway.. :)
Sep 25 21:51:54 <kergoth> either that, or remove it from ASSUME_PROVIDED
Sep 25 21:58:56 <Crofton> we try pretty hard to tel cmake not to search the build system in cmake.bbclass
Sep 25 21:59:48 <Crofton> basically, I am having to change stuff in oe-core because some guys want to bury the git hash in a binary so they know where the binery came from
Sep 25 22:00:03 <Crofton> instead of using the normal means of reporting versions
Sep 25 22:00:11 <Crofton> lazy/careless developers
Sep 25 22:58:56 <return_to_sender> Hi all, quick question.  I have a yocto setup bulding a 3.14 kernel for the Wandboard (fsl-community-bsp) that I want to switch to the 3.0.35 boundary devices kernel.  Is it possible to change the kernel without cloning a new repo, perhaps through local.conf?
Sep 25 23:02:49 <dvhart> return_to_sender, the MACHINE typically defines PREFERRED_PROVIDER_virtual/kernel
Sep 25 23:04:35 * ikey|afk hypnotizes dvhart into giving him a minnowboard
Sep 25 23:06:13 <return_to_sender> OK, I'll start there.  I should be able to override PREFERRED_PROVIDER_virtual/kernel in local.conf though, right?
Sep 25 23:10:20 <rburton> ikey|afk: you might want to specify minnowmax, there's plenty of minnowboards spare now :)
Sep 25 23:10:33 <ikey|afk> oh yeah. that one
Sep 25 23:10:34 <ikey|afk> :D
Sep 25 23:11:15 * ikey|afk needs shiny things
Sep 25 23:14:13 <rburton> i should get mine out of its box
Sep 25 23:14:33 <rburton> speaking of shiny the xeons should be arriving soon!
Sep 25 23:14:42 * rburton hugs at a safe distance the shiny xeons
Sep 25 23:15:21 <ikey|afk> I got some Xeons
Sep 25 23:15:24 <ikey|afk> wanna trade?
Sep 25 23:15:58 <rburton> almost certainly not
Sep 25 23:16:15 <rburton> Xeon DP Ivy Bridge-EP E5-2690 v2 LGA2011 3.0GHz 25MB 130W 10 cores x 2
Sep 25 23:16:29 <ikey|afk> but i haz 20.
Sep 25 23:16:30 <ikey|afk> :(
Sep 25 23:17:07 <ikey|afk> need to start using that server more tbh
Sep 25 23:17:16 <ikey|afk> was using it as an irc client one week.
Sep 25 23:17:20 <rburton> !
Sep 25 23:17:25 * ikey|afk shrugs
Sep 25 23:17:28 <ikey|afk> vnc :P
Sep 25 23:18:09 <rburton> bakeotron is off, time to for bed
Sep 25 23:18:19 <ikey|afk> yeah i need to do that soon.
Sep 25 23:18:25 <ikey|afk> o/
**** ENDING LOGGING AT Fri Sep 26 03:00:00 2014