**** BEGIN LOGGING AT Fri Oct 02 02:59:59 2015
Oct 02 06:55:17 <mckoan> good morning
Oct 02 06:57:15 <RagBal> What is the best task to prepend if I want to copy a device tree file to a kernel source dir? I use linux-dtb.inc
Oct 02 07:23:18 <bluelightning> morning mckoan, all
Oct 02 08:37:53 <RagBal> What is the best task to prepend if I want to copy a device tree file to a kernel source dir? I use linux-dtb.inc <<< will do_install_prepend do?
Oct 02 08:44:25 <bluelightning> RagBal: by definition do_install would happen after compilation - is that what you want?
Oct 02 08:46:40 <RagBal> Device trees won't be compiled during do_compile right?
Oct 02 08:46:56 <bluelightning> I honestly don't know, I don't deal with device trees often myself
Oct 02 08:47:11 <bluelightning> I would advise looking at what linux-dtb.inc actually does to know for sure
Oct 02 08:54:19 <RagBal> It appends both do_install and do_deploy, during do_install it actually compiles the device tree. So prepending to do_install seems right
Oct 02 09:20:46 <bboozzoo> need advice, I'm using meta-gateworks, and they explicitly add thigs like fsl-rc-local, hostapd, wpa_supplicant, dropbear, iproute2 etc. in MACHINE_EXTRA_RDEPENDS, IMO that's wrong, and I'd like to override it at least by adding MACHINE_EXTRA_RDEPENDS_remove = '...', what would be a good place to do it?
Oct 02 09:20:58 <bboozzoo> I'd like to avoid adding it in local.conf if possible
Oct 02 09:21:38 <bluelightning> bboozzoo: your custom distro config would be the best place...
Oct 02 09:22:29 <rob_w> anybody know me a good backlight governor .. able to interface xorg and store settings etc ?
Oct 02 09:22:42 <bboozzoo> hm ok, can I inherit a distro like classes? I already have a custom distro using systemd that's defined elsewhere and it's shared between a couple of boards
Oct 02 09:23:16 <bluelightning> bboozzoo: no but you would do something like MACHINE_EXTRA_RDEPENDS_remove_machinename = "..."
Oct 02 09:23:27 <bluelightning> so it would only apply to that specific machine
Oct 02 09:24:28 <bboozzoo> bluelightning: sounds good, thanks
Oct 02 09:26:33 <raykinsella78> bluelightning: Got time for a git question?
Oct 02 09:26:34 <bboozzoo> btw. does it actually make any sense to add pkgs like dropbear, wpa_supplicant to MACHINE_EXTRA_RDEPENDS? I'd normally expect only firmware packages listed there
Oct 02 09:26:52 <bluelightning> bboozzoo: no, that isn't good practice IMO
Oct 02 09:27:03 <bluelightning> bboozzoo: that's distro policy being forced into the machine definition
Oct 02 09:27:10 <bluelightning> raykinsella78: sure
Oct 02 09:27:25 <bboozzoo> bluelightning: that's what I thought, thanks a lot :)
Oct 02 09:28:01 <raykinsella78> bluelightning: how do I list local changes on my branch.
Oct 02 09:28:28 <raykinsella78> bluelightning: when I merge upstream to my local branch my local changes get pushed way down in the git log.
Oct 02 09:28:52 <raykinsella78> bluelightning: these are my changes pending upstream, how do I pull them out - so I know what is local and what is upstream
Oct 02 09:28:57 <raykinsella78> bluelightning: make sense?
Oct 02 09:30:11 <bluelightning> raykinsella78: the answer I think if I understand what you're saying is don't merge, rebase instead
Oct 02 09:30:31 <bluelightning> that way your local changes will always be on top
Oct 02 09:31:52 <raykinsella78> bluelightning: I usually use rebase to merge local changes into a single commit, reword commits etc.
Oct 02 09:32:44 <bluelightning> right, but when you do git-pull the default behaviour is merge rather than rebase
Oct 02 09:32:58 <raykinsella78> bluelightning: thats silly - https://robots.thoughtbot.com/keeping-a-github-fork-updated
Oct 02 09:33:01 <bluelightning> there's a --rebase option though I believe
Oct 02 09:33:26 <raykinsella78> bluelightning: fetch and rebase .... great.
Oct 02 09:37:25 <raykinsella78> bluelightning: thanks for the help
Oct 02 10:09:44 <ohmy> Hello everybody
Oct 02 10:09:59 <ohmy> i'm stuck since three days i hope i can get some help here
Oct 02 10:10:16 <ohmy> i'm using yocto for IMX6 board and i'd like to use Qt over Wayland
Oct 02 10:10:36 <ohmy> While Wayland Weston build and run correctly i'm stuck with QtWayland
Oct 02 10:10:59 <ohmy> When Qt is build the wayland plugin isnt and i dont know how to make it
Oct 02 10:11:02 <ohmy> any idea please ?
Oct 02 12:18:17 <ohmy> :(
Oct 02 12:21:10 <bluelightning> ohmy: AIUI from reading, qtwayland *is* the Qt wayland plugin
Oct 02 12:22:15 <ohmy> bluelightning: yes, bitbake qtwayland does not output any error, but Qt5 build does not include the plugin
Oct 02 12:22:41 <bluelightning> what does that mean? where are you looking?
Oct 02 12:25:30 <ohmy> bluelightning: i'l looking in Qt5 (/usr/lib/qt5) i have wayland-decoration-client wayland-graphics-integration-client wayland-graphics-integration server folders
Oct 02 12:25:32 <ohmy> bluelightning: but
Oct 02 12:27:09 <ohmy> bluelightning: when i run examples from HW (-platform wayland or by exporting the required QT_QPA_ variable) i have ther following error "This application failed to start because it could not find or load the Qt platform plugin "wayland"."
Oct 02 12:29:18 <bluelightning> ok, so have you ensured that qtwayland is actually included in your image?
Oct 02 12:29:27 <bluelightning> just building it doesn't do that
Oct 02 12:30:28 <bluelightning> in case you haven't seen it: http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#usingpoky-extend-customimage
Oct 02 12:43:05 <ohmy> bluelightning: thank you so much indeed i was flashing the wrong image, now i can run Qt5_cinematic experience and it segfaults :( i want to kill myself :(
Oct 02 12:43:43 <bluelightning> ohmy: doh... please don't kill yourself though, it's not that bad ;)
Oct 02 12:46:48 <ohmy> bluelightning: i've been using framebuffers, X11 for years, everyone says that wayland rox and i'm really frustrated now, weston compositor runs perfectly, i can run the weston-* samples, when i run Qt based applications they simply segfault
Oct 02 12:50:29 <bluelightning> unfortunately this is well outside my realm of expertise... a segfault at that point could be caused by all sorts of things
Oct 02 12:51:25 <ohmy> bluelightning: yes i know, of course the frustrating thing is all the videos running the same applications using the same yocto meta fsl arm branch and the same HW targets
Oct 02 12:51:50 <ohmy> bluelightning: i mean there's too many videos showing that everything works well
Oct 02 12:57:29 <Crofton|work> ohmy, try asking on th elist, maybe someone there has done it
Oct 02 12:58:03 * Crofton|work is packed and hopes to escape before the flood
Oct 02 12:58:21 <bluelightning> Crofton|work: flood?
Oct 02 12:58:34 <Crofton|work> Heavy rain, flooding around here this weekend
Oct 02 12:58:47 <bluelightning> Crofton|work: ah ok... good luck and see you in Dublin
Oct 02 12:59:14 <Crofton|work> yes, I have a spare day in the schedule, hopefully it is not swallowed up by travel issues
Oct 02 12:59:16 <ryansturmer> if I have a recipe that is added to an image, how is the version of the recipe to use specified?
Oct 02 12:59:23 <Crofton|work> demo is even done
Oct 02 12:59:33 <Crofton|work> ryansturmer, the latest
Oct 02 12:59:42 <ryansturmer> The latest is always used?
Oct 02 12:59:50 <Crofton|work> barring preferred version and defualt preference settings
Oct 02 12:59:57 <ryansturmer> Is it possible to override that?
Oct 02 12:59:58 <bluelightning> and layer priorities
Oct 02 13:00:27 <ryansturmer> This intel edison package I've downloaded is poky-1.7.something and in the IMAGE_INSTALL they add e2fsprogs
Oct 02 13:00:33 <ryansturmer> which comes in at version 1.26
Oct 02 13:00:42 <ryansturmer> even though the newest version is much newer than that
Oct 02 13:00:54 <Crofton|work> is there a ecipe for a later one?
Oct 02 13:01:09 <Crofton|work> urg, what release name does 1.7 mean?
Oct 02 13:01:28 <bluelightning> specifically, a recipe available in your configuration
Oct 02 13:01:41 <ryansturmer> so I'm here:
Oct 02 13:01:52 <ryansturmer> http://recipes.yoctoproject.org/rrs/recipes/1.7/All/
Oct 02 13:01:59 <ryansturmer> and I see a newer version
Oct 02 13:02:06 <ryansturmer> not sure if that means what I think it means
Oct 02 13:02:25 <Crofton|work> https://www.yoctoproject.org/docs/1.6/bitbake-user-manual/bitbake-user-manual.html
Oct 02 13:02:32 <Crofton|work> search on PREFERRED_VERSIO
Oct 02 13:02:35 <bluelightning> ryansturmer: try bitbake-recipes show-recipes e2fsprogs
Oct 02 13:02:46 <bluelightning> er
Oct 02 13:02:50 <bluelightning> that would be
Oct 02 13:02:54 <bluelightning> bitbake-layers show-recipes e2fsprogs
Oct 02 13:06:40 <ryansturmer> hmm, that shows the correct version
Oct 02 13:06:48 <ryansturmer> so perhaps I was looking at the wrong version number
Oct 02 13:07:01 <ryansturmer> I am trying to rebuild the package again now, and will reexamine the error
Oct 02 13:07:02 <ryansturmer> thanks!
Oct 02 13:11:27 <ryansturmer> Yep.
Oct 02 13:11:30 <ryansturmer> Don't know what I was thinking
Oct 02 13:11:46 <ryansturmer> That is absolutely the correct version number :P
Oct 02 13:12:00 <ryansturmer> I'm still getting all these multiple definition errors, but that maybe I can solve.
Oct 02 13:24:40 <ryansturmer> Has anyone else run into this?  e2fsprogs is failing with about a zillion multiple definition errors.
Oct 02 13:24:41 <ohmy> Crofton|work: i've posted 3 messages on meta-fsl-arm since yesterday, without answer for the moment
Oct 02 13:25:26 <ryansturmer> I've read that gcc5 changed the way inlines work when it moved to the new C standard, and to get the old behavior, you have to pass -fgnu89-inline to the compiler.
Oct 02 13:25:34 <ryansturmer> but to no avail.
Oct 02 13:27:33 <ohmy> btw i have tried to rebuild the filesystem many times, even if i have added qtwayland to IMAGE_INSTALL_append section, the plugins are put into my sysroot but never copied to the generated .sd image, anyone knows how to fix this ?
Oct 02 13:28:58 <ohmy> AFAIK  IMAGE_INSTALL_append is enough to force a package to be depolyed on the rootfs i'm i wrong ?
Oct 02 13:31:13 <bluelightning> ryansturmer: is this e2fsprogs-native?
Oct 02 13:31:17 <ryansturmer> yes
Oct 02 13:31:46 <bluelightning> honestly I'd save your sanity and just build on a host distro that doesn't impose gcc5 for now
Oct 02 13:32:01 <bluelightning> since you're stuck with 1.7 for the moment
Oct 02 13:32:38 <bluelightning> ohmy: yes... you have a leading space in the value you're appending right?
Oct 02 13:33:08 <ryansturmer> Ah I had a feeling you'd say that.
Oct 02 13:33:53 <ryansturmer> it's been so educational, though! :P
Oct 02 13:34:37 <ryansturmer> Nothing gets you familiar with a build system like having *all the problems*
Oct 02 13:37:16 <ohmy> bluelightning: yes of course
Oct 02 13:37:32 <bluelightning> ryansturmer: gcc5 has been a royal pain for everyone it seems :/
Oct 02 13:38:46 <bluelightning> ohmy: ok, well that will definitely include it in the image... whether or not it appears in the sd card image depends on how that's produced; it's been a while since I have used meta-fsl-arm so I'm not sure specifically in that case
Oct 02 13:39:55 <bluelightning> ryansturmer: which distro are you using out of interest?
Oct 02 13:44:43 <ryansturmer> Archlinux
Oct 02 13:45:05 <ryansturmer> I'm super happy to continue slogging through it - it's fun finding these issues, although I do have to get real work done, so I'm creating a VM
Oct 02 13:45:15 <ryansturmer> Archlinux is sort of your worst case test platform
Oct 02 13:45:19 <ryansturmer> because it's a rolling release distro
Oct 02 13:45:49 <ryansturmer> What ubuntu release is recommended?
Oct 02 13:46:28 <ryansturmer> Is ubuntu "destkop" recommended over "server"
Oct 02 13:46:50 <bluelightning> I don't think there's much difference as far as we're concerned
Oct 02 13:47:15 <ryansturmer> I'm grabbing the 14.04.3 LTS of ubuntu desktop, unless there's a compelling reason to go with the server version, etc.
Oct 02 13:47:18 <ryansturmer> ok.
Oct 02 13:47:21 <bluelightning> that should be fine
Oct 02 13:47:40 <bluelightning> officially we don't support rolling release distros for exactly this reason
Oct 02 13:48:03 <bluelightning> (though some do build on them)
Oct 02 13:48:38 <ryansturmer> In general I can roll with the issues associated with rolling release
Oct 02 13:48:47 <ryansturmer> but gcc5 is a mess for this kind of thing
Oct 02 14:13:40 <_taw_> ohmy: same problem here, bitbake -e says _append packages are in the IMAGE_INSTALL, but nothing gets installed
Oct 02 15:09:03 <sa2ajj> pidge: hi. i've been reading the http://git.yoctoproject.org/cgit/cgit.cgi/yocto-autobuilder/tree/docs/YoctoAutobuilderDevelopersDocument.html and i wonder if it was converted from something else?
Oct 02 15:09:27 <sa2ajj> on a related, would it be useful if it's converted to sphinx (my tool of choice for documentation)
Oct 02 15:12:11 <sa2ajj> s/related/related note/
Oct 02 15:15:54 <pidge> sa2ajj, yes, there is a google doc. I want to eventually convert it to pydoc since that's what the project uses.
Oct 02 15:16:39 <sa2ajj> any reason not to use sphinx? :)
Oct 02 15:18:07 <sa2ajj> i wonder since i think sphinx became more or less standard way to document python projects
Oct 02 15:19:44 <sa2ajj> on a slightly related note, is that google doc linked anywhere?
Oct 02 15:30:46 <sa2ajj> pidge: would you still consider a sphinx version?  i'm looking at yocto autobuilder and to improve a bit the formatting of the manual (something are just very difficult to read)
Oct 02 15:40:14 <pidge> sa2ajj: sure, I'd consider. send a proposal to the list. cc me elizabeth.flanagan@intel.com
Oct 02 15:40:25 <sa2ajj> ok, thanks
Oct 02 15:40:37 <sa2ajj> a link to the original google doc would help :)
Oct 02 15:42:02 <ryansturmer> Friggin ubuntu build is working fine :)
Oct 02 15:42:08 <ryansturmer> (So far)
Oct 02 16:48:08 <kergoth> oops, forgot to support nobranch=1 in the shallow git support
Oct 02 16:48:17 * kergoth mutters
Oct 02 16:49:35 <kergoth> since i heavily prune the shallow repo to make sure it doens't keep around stuff it shouldn't, i have to make sure we have a ref pointing at the srcrev, otherwise it'll be removed.. so in the case where the commit is a tag i need to make sure i create the tag ref, and also handle both annotated and non-annotated, and abort if nobranch=1 and the commit isn't a tag, since we won't be able to make it shallow without a ref..
Oct 02 16:49:40 <kergoth> although.. hmm
Oct 02 20:27:36 <ebolton> hey all, found a strange bug...I've added cmake and bison to an SDK. If I install the SDK to a non-default directory and try to build a cmake project that uses bison....bison looks for the M4 scripts in the DEFAULT SDK dir and fails
Oct 02 20:27:59 <ebolton> I can't see anything obvious in the bison or autoconf recipes that would cause this, so I'm not sure where it's coming from
Oct 02 20:29:34 <kergoth> installing the sdk should run a relocation script to replace hardcoded path references in any text files / scripts. binaries usually need wrappers that it can adjust the paths in
Oct 02 21:02:59 <bbhoss> Hey all. I'm using an intel edison with the latest image intel distributes, and it seems to be missing the install command. Is this normal, and how do I fix it?
Oct 02 21:04:54 <kergoth> Doesn't surprise me at all, unless you're doing development on target it's unlikely you'd need it, and we have images specifically for that
Oct 02 21:13:43 <bbhoss> kergoth: well it has make, cc, and friends along with a full node.js environment, lol
Oct 02 21:14:17 <bbhoss> I'm actually trying to run erlang on the poor thing
Oct 02 21:14:18 <kergoth> are you talking about the host or the target?
Oct 02 21:14:27 <kergoth> and which image is it for the target?
Oct 02 21:14:37 <kergoth> not familiar with what intel ships, offhand
Oct 02 21:14:44 <bbhoss> target
Oct 02 21:14:50 <bbhoss> just fooling around right now
Oct 02 21:15:03 <bbhoss> but yeah, everything compiles but install is missing so I can't install it
Oct 02 21:15:48 <bbhoss> probably will end up making my own bsp but I don't have a linux machine right now
Oct 02 21:16:35 <kergoth> looks like install is usually from coreutils
Oct 02 21:16:38 <kergoth> hmm
Oct 02 21:17:42 <bbhoss> kergoth: looks like installing that package did the trick
Oct 02 21:17:50 <kergoth> cool
Oct 02 21:18:26 <bbhoss> thanks
Oct 02 21:21:20 <kergoth> np
Oct 02 21:39:03 <dasabhi> hey guys i am trying to cross compile again, using kernel version 3.8.7
Oct 02 21:39:23 <dasabhi> last time some one suggested something that fixed my compiling issues
Oct 02 21:39:46 <dasabhi> i am having some issues again when i run the following command make ARCH=i386 CROSS_COMPILE=i586-poky-linux- -j8
Oct 02 21:40:02 <dasabhi> this is the error message i see http://www.lick.moe/paste/ee9841594f9eb0c2b9b8f4003f3483d6
Oct 02 21:50:36 <kergoth> sounds like you should google. this channel isn't for general embedded linux help..
Oct 02 22:02:46 <kergoth> hm, wonder if the issue with piping bitbake output messing up the terminal ever got fixed
Oct 02 22:02:59 <seebs> Hmm. Doesn't sound familiar.
Oct 02 22:03:26 <seebs> Although... I think bitbake does stuff based on is-a-tty, and bitbake writes to both stderr and stdout, but pipes by default do not include stderr, so I would sort of expect it to cause weirdness.
Oct 02 22:06:41 <kergoth> it should really be checking the tty it's writing to. if it's writing color to stderr, then stderr needs to be a tty
Oct 02 22:06:46 <kergoth> and the same for stdout
Oct 02 22:07:54 <seebs> yeah
Oct 02 22:09:31 <kergoth> oh, right, i have a branch that fixes that while adding the stdout/stderr split to scriptutils
Oct 02 22:09:35 <kergoth> but i never moved it to bitbake
Oct 02 22:09:41 <kergoth> well, fixes it for the color bits at least
Oct 02 22:09:47 <kergoth> there's other bits done by the UI
Oct 02 22:10:15 <kergoth> https://github.com/kergoth/openembedded-core/commit/6a7666e83c44ac56a09aec5619e44379d167972f constructs two formatters instead of using the same formatter for both stdout and stderr, and https://github.com/kergoth/openembedded-core/commit/232185c2493208aa41983f92c73024cb1ccdebb6 enables color based on isatty for each handler's individual stream
Oct 02 22:52:06 <neverpanic> Does pseudo actually work on OS X at the moment?
Oct 02 23:30:39 <seebs> neverpanic: sorta?
Oct 02 23:30:51 <seebs> At least, last I checked, although the xattrdb stuff probably won't.
Oct 02 23:31:01 <seebs> It has not been heavily-tested, but I do periodically run some basic tests on it.
Oct 02 23:32:42 <neverpanic> seebs: well, bad news is, El Capitan breaks it for any binary that comes with the system.
Oct 02 23:32:50 <seebs> I am not surprised.
Oct 02 23:32:58 <seebs> That seemed like a really likely security thing for them to add sooner or later.
Oct 02 23:33:06 <neverpanic> so anything in /usr/bin and /bin
Oct 02 23:34:07 <neverpanic> There is an easy workaround that I may up implementing for my use case, though, and that's having a "shadow copy" of /, wrapping execve and posix_spawn and whenever a protected binary is called, copy it to the shadow copy (which will strip the protection) and run it form there
Oct 02 23:38:22 <kergoth> glad they added the protection, should have done it sooner. though it does interfere with certain apps capabilities, for the best in the long run
Oct 02 23:52:51 <kergoth> damnit, using --no-sysroot-suffix for meta-sourcery is breaking my build, which doesn't make any sense
Oct 02 23:52:54 <kergoth> hmmm
Oct 03 00:17:42 <kergoth> whew, think i finally have the kinks worked out - https://github.com/MentorEmbedded/poky/compare/master...kergoth:git-shallow
Oct 03 00:17:46 * kergoth does more testing
**** ENDING LOGGING AT Sat Oct 03 02:59:58 2015