**** BEGIN LOGGING AT Wed Feb 18 03:00:00 2015
Feb 18 07:53:57 <RagBal> What recipes do I need to include in my image for maliit in QT5?
Feb 18 09:10:30 <ionte> hi. anyone involved with meta-ti? i need kernel 3.16 or later because of some bugfixes, but meta-ti only supports up to 3.14. is there any work in progress on this?
Feb 18 09:12:12 <abelloni> probably not, it is not an LTS kernel
Feb 18 09:28:15 <bluelightning> morning all	
Feb 18 10:18:36 <Nilesh_> Does any one know how to enable Qsqlite ?? i created a .bbappend file in meta-qt with the content "PACKAGECONFIG_append = " sql-sqlite" . But it seems to not build the Qsqlite? I'm i missing something
Feb 18 10:20:29 <AndersD> Nilesh_, if you add a qtbase_XXX.bbappend with that line, it should build the Qt sql-plugin.
Feb 18 10:20:40 <AndersD> I know that we're using it in our own bbappend
Feb 18 10:21:45 <AndersD> Well, we're using PACKAGECONFNIG += "sql-sqlite" but it shouldn't matter
Feb 18 10:21:50 <Nilesh_> AndersD:ok , I'll give a fresh bitbake and check out.
Feb 18 10:55:40 <pohly> Does OE package feed support include any kind of package signing or at least securing the https download? Does smart.py require valid  SSL certificates? Python for a long time by default allowed https without valid certificates.
Feb 18 11:34:39 <Lern-X> Hello, I have something to ask about bitbake errors. Is it the good channel ?
Feb 18 12:13:06 <_dalton> hey
Feb 18 12:13:19 <_dalton> I am trying to compile hello world using bitbake
Feb 18 12:13:37 <_dalton> I ran into an error
Feb 18 12:14:00 <_dalton> would anyone help me out if i paste the error here?
Feb 18 12:14:07 <_dalton> thanks...
Feb 18 12:38:08 <AndersD> _dalton, post the error and the recipe on pastebin (a lot easier to read them there), and a link here
Feb 18 13:20:31 <pev> 'noon
Feb 18 13:20:55 <pev> Daft question I know, but is there a way to *remove* something from an RDEPENDS in a bbappend?
Feb 18 13:21:28 <abelloni> RDEPENDS_remove = "stuff" ?
Feb 18 13:21:32 <pev> Point in case, I've added i2c-tools from meta-openembedded to my target and it's pulled in perl and 40 odd extra dependancies
Feb 18 13:21:45 <pev> (for an irrelevent feature)
Feb 18 13:21:52 <pev> Cool, I'll try that thanks
Feb 18 13:26:22 <pev> Also, I've never looked into stripping out udev within yocto before - is it a big deal or straightforward?
Feb 18 13:43:21 <bluelightning> I suspect you will find just removing an item from RDEPENDS won't work for that, you will have to attack how the unwanted dependencies get there in the first place
Feb 18 13:50:14 <abelloni> bluelightning: if it depends on perl only to run a script, I believe it doesn't do any harm
Feb 18 13:50:32 <abelloni> the main utilities from i2c-tools are in c
Feb 18 13:52:15 <bluelightning> abelloni: right, that's not really what I meant - whether or not touching RDEPENDS in that way depends on how that RDEPENDS got there in the first place
Feb 18 13:52:43 <bluelightning> I wonder if we adjusted the packaging in master to address this
Feb 18 13:53:30 <abelloni> bluelightning: maybe with a PACKAGECONFIG ?
Feb 18 13:54:08 <abelloni> I would say that the main complain I hear about oe/yocto is about the number of dependencies
Feb 18 13:54:14 <bluelightning> well, in OE-Core master (i2c-tools got moved to OE-Core at some point recently, in dizzy perhaps, the perl-requiring scripts were moved to a separate package
Feb 18 13:54:34 <bluelightning> it is a common complaint I agree
Feb 18 13:55:18 <bluelightning> it's not always easy to remove a dependency, since they come in through different means and not all of them are added conditionally
Feb 18 13:55:39 <bluelightning> I think we could develop some better tools to handle some situations in future
Feb 18 13:56:07 <abelloni> the one that surprises me the most is qemu-native :)
Feb 18 13:59:55 <bluelightning> yeah, there are two reasons for that - for qemu* machines it's a bit more understandable (and comes in via POKYQEMUDEPS); but otherwise we have it around for running certain postinstall scripts that just have to be run with the target architecture
Feb 18 14:07:39 <pev> That's OK, I'm back-burnering it for a bit and will re-visit :-)
Feb 18 14:12:51 <pev> Another interesting question - in my layer i have several image types. I'd like to munge /etc/issue in a different way for each image to make it obvious what you're booting into. However, doing a bbappend for base-files isn't image specific. Is it "nice" to directly append into issue from a do_install task in the image recipes or is there a cleaner way to do this?
Feb 18 14:15:28 <bluelightning> well, it wouldn't be do_install, it would need to be a function called from ROOTFS_POSTPROCESS_COMMAND
Feb 18 14:15:58 <bluelightning> if you are using runtime package management you'd have to mitigate any future updates to base-files wiping out the customisations applied there
Feb 18 14:17:20 <pev> No runtime package management
Feb 18 14:17:27 <pev> I'm not that much of a masochist :-)
Feb 18 14:17:40 <bluelightning> ok, then at least that's one issue you don't have to worry about :)
Feb 18 14:18:40 <bluelightning> so yeah, your function within the image would just modify ${IMAGE_ROOTFS}${sysconfdir}/issue as desired, and that should do the trick
Feb 18 14:19:45 <pev> And I guess issue.net too? (not that ive seen it before but it gets changed in parallel)
Feb 18 14:21:56 <bluelightning> pev: right, yes
Feb 18 14:22:10 <bluelightning> issue.net is for remote logins
Feb 18 14:46:20 <abelloni> bluelightning: I have a hard question :)
Feb 18 14:46:30 <bluelightning> sure, I love hard questions :)
Feb 18 14:47:27 <abelloni> is it possible to make a PACKACONFIG depending on the presence of another package ?
Feb 18 14:47:31 <abelloni> to be clear
Feb 18 14:47:43 <abelloni> qtwebkit needs icu to be set in qtbase
Feb 18 14:50:32 <abelloni> that one is a bit tricky because qtwebkit itself will not fail
Feb 18 14:50:48 <abelloni> but anything trying to use qtwebkit will fail
Feb 18 14:51:05 <bluelightning> yeah, JaMa and I have talked about this kind of problem in qt5 before
Feb 18 14:51:32 <abelloni> I'm wondering whether we should add icu to PACKAGECONFIG_DEFAULT
Feb 18 14:52:14 <bluelightning> there is no definitive mechanism for handling it, but you could in theory conditionally add something to PROVIDES in qtbase such that the qtwebkit build would immediately fail if it hasn't been configured in the right way
Feb 18 14:52:45 <abelloni> because now, it means that stock poky is failing to build any image having an application using qtwebkit
Feb 18 14:53:50 <bluelightning> that's certainly not ideal...
Feb 18 14:53:57 <abelloni> buildroot is handling that a bit more gracefully ;)
Feb 18 14:54:20 <abelloni> maybe we should build all qt components from the same recipe
Feb 18 14:54:37 <abelloni> that would allow to add -icu when qtwebkit is selected
Feb 18 14:54:40 <bluelightning> how is it handled there? by a config option that they share (which we could do), or by enabling icu by default?
Feb 18 14:55:14 <abelloni> from any package, they can check whether another package is selected
Feb 18 14:55:47 <abelloni> that is kind of how they handle the DISTRO_FEATURES
Feb 18 14:55:57 <bluelightning> ok, that wouldn't really work with the way our system is designed
Feb 18 14:56:04 <abelloni> like "is openssl selected, is udev selected, ..."
Feb 18 14:56:17 <abelloni> yeah
Feb 18 14:56:39 <abelloni> they just check whether qtwebkit is selected in the qtbase recipe and add -icu in that case
Feb 18 14:56:47 <JaMa> gentoo has nice syntax for this
Feb 18 14:57:08 <JaMa> but until we have that, reading README is the next best thing :)
Feb 18 14:57:44 <bluelightning> we should be able to do better right now
Feb 18 14:57:56 <JaMa> but we don't have even versioned dependencies
Feb 18 14:57:57 <abelloni> yeha, I'm actually concerned about having poky failing to build
Feb 18 14:58:30 <abelloni> and I'd like to not mess with the qt PACKAGECONFIG in a BSP layer
Feb 18 14:58:46 <bluelightning> failing to build is OK if it fails early enough with a message that gives the user enough of an idea of how to deal with the problem
Feb 18 15:02:13 <bluelightning> and, setting this should be the domain of your distro layer rather than the BSP ;)
Feb 18 15:03:31 <abelloni> bluelightning: obviously, that is why I want to avoid it
Feb 18 15:04:09 <abelloni> but honestly,  it seems that people just like to use poky instead of creating their own distro
Feb 18 15:04:42 <bluelightning> that's fine for getting started... but I worry if people think that's OK for anything more than that
Feb 18 15:05:27 <abelloni> I saw a few customers using it in production
Feb 18 15:05:43 <bluelightning> oh, I'm positive there are
Feb 18 15:05:59 <abelloni> Maybe it should be documented a bit more that you should actually define your own distro
Feb 18 15:06:34 <bluelightning> well, we did add a section to the manual stating exactly that a release or two ago
Feb 18 15:07:04 <bluelightning> we could always do more (make it easier, more obvious, etc.)
Feb 18 15:07:49 <abelloni> hum, maybe a warning when building pky ;)
Feb 18 15:07:52 <abelloni> poky
Feb 18 15:08:01 <abelloni> with a link to the doc
Feb 18 15:08:18 <abelloni> I didn't see that section can you give me a pointer ?
Feb 18 15:08:23 <bluelightning> http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#creating-your-own-distribution
Feb 18 15:08:31 <abelloni> thanks
Feb 18 15:08:35 <bluelightning> the wording wasn't quite as I remembered it to be honest
Feb 18 15:11:01 <bluelightning> it may be part of another area about development vs. production; e.g. in development it's perfectly fine to use debug-tweaks, but in production you may not want to allow logging in as root with no password ;)
Feb 18 15:11:21 <bluelightning> at the moment it's easy to forget to remove that
Feb 18 15:25:10 <kergoth> The other day I started poking at minisat and picosat solvers, and started experimenting with the possibility of using one of them to resolve the deps in a Packages file for construction of an eclipse p2 repository based on it.. but it made me wonder about using such a solver for bitbake rather than home-rolling it
Feb 18 16:23:54 <pev> Does bitbake cleansstate imply clean?
Feb 18 16:25:20 <paulg> I think so ; pretty sure you actually see it run clean on the console if you are watching it.
Feb 18 16:25:35 <pev> Cool, cheers...
Feb 18 16:44:50 <funman> Is there a method for submitting a BSP to be published on git.yoctoproject.org ? specifically for hardkernel boards (I see a couple of machine specific layers on github)
Feb 18 16:45:29 <abelloni> I guess you'd have to become a yocto project member
Feb 18 16:45:57 <abelloni> what's wrong with being on github ?
Feb 18 16:46:19 <abelloni> if it is registered at layers.openembedded.org/layerindex/, people will find it
Feb 18 16:46:29 <funman> ah that would work too I guess
Feb 18 16:46:49 <funman> i'm mostly concerned with people starting their own board specific layer and not sharing the work
Feb 18 16:47:21 <funman> 'Submit layer' is what I was looking for, thank you
Feb 18 16:47:44 <Crofton|road> funman, that is good news
Feb 18 16:47:58 <Crofton|road> where is the bsp now?
Feb 18 16:48:57 <funman> Crofton|road: I have https://github.com/MarkusEich/meta-odroid which should work for 2 boards I don't have, I will give some news when I get it working for odroid-c1 and/or odroid-xu3
Feb 18 16:50:32 <funman> I never worked on a BSP but since afaiu the u-boot and kernel sources come from the same tree it should just be a matter of giving it the right config
Feb 18 16:51:02 <funman> I hope to use https://fosdem.org/2015/schedule/event/starting_yocto/ as well
Feb 18 16:51:44 <abelloni> hehe
Feb 18 16:57:10 <bluelightning> FWIW, anyone can have a repo on git.yoctoproject.org for YP/OE related stuff
Feb 18 16:57:36 <bluelightning> but to be fair setting one up on github is slightly easier in that you don't have to ask anyone :)
Feb 18 16:58:10 <bluelightning> and yes, if you want people to find your layer, please submit it to the layer index
Feb 18 16:58:58 <bluelightning> whoever submits it there though should be the maintainer or acting on the maintainer's behalf
Feb 18 17:02:11 <abelloni> funman: for a bsp layer, maybe it can be better to have a look a my elc talk, rather than the one at fosdem
Feb 18 17:02:49 <abelloni> http://free-electrons.com/pub/conferences/2014/elc/belloni-yocto-for-manufacturers/belloni-yocto-for-manufacturers.pdf
Feb 18 17:11:57 <Lern-X> Hello everyone, I have some questions to ask about a problem concerning bitbake. Is it the good channel ?
Feb 18 17:13:56 <bluelightning> Lern-X: certainly, please ask your question
Feb 18 17:15:56 <Lern-X> Okay, thanks. So, I want to build a tizen-common-img_minimal-dev image thanks to the command bitbake -k [recipe]. (here the recipe is tizen-common-imge_min-dev).
Feb 18 17:16:16 <Lern-X> Some errors occured during the building
Feb 18 17:16:18 <Lern-X> build*
Feb 18 17:17:13 <Lern-X> First, do_patch seems to have bugued :
Feb 18 17:17:21 <Lern-X> ERROR: Task 95 (virtual:native:/home/dvegara/Documents/tizen-distro/meta-tizen/meta-tizen-common-base/recipes-devtools/rpm/rpm_git.bb, do_patch) failed with exit code '1'
Feb 18 17:17:41 <Lern-X> And then, others functions, like do_prep or do_patch, bugued.
Feb 18 17:18:16 <bluelightning> pohly: this may be a question for you ^
Feb 18 17:18:51 <Lern-X> So I think I should re-build with -v option, to know more about the problem ? Or is there just a package to install int my Fedora 21 ?
Feb 18 17:19:03 <Lern-X> in*
Feb 18 17:23:19 <funman> abelloni: was the fosdem talk more general audience?
Feb 18 17:24:04 <abelloni> funman: they are overlapping but they don't take the same approach
Feb 18 17:27:44 <Lern-X> bluelightning: my problem isn't very common ?
Feb 18 17:28:11 <bluelightning> Lern-X: it's almost certainly specific to building tizen with meta-tizen, which is not something I have tried
Feb 18 17:28:31 <bluelightning> Lern-X: if nobody here can help you may try asking on the tizen mailing lists
Feb 18 17:31:06 <Lern-X> bluelightning: However, I don't think I used meta-tizen ?
Feb 18 17:34:17 <Lern-X> bluelightning: Hmm, should I ask ontizen mailing list, or on yocto ones ?
Feb 18 17:37:51 <bluelightning> Lern-X: it's mentioned in the error you pasted, so I can only assume you did
Feb 18 17:40:40 <Lern-X> bluelightning: ah okay, so I'll use mailing lists^^
Feb 18 17:41:50 <Lern-X> thank you even though :)
Feb 18 18:22:30 <_dalton_> hey,
Feb 18 18:22:48 <_dalton_> If anyone is free, please help me compile hello world
Feb 18 18:25:19 <neverpanic> _dalton_: as you've been told before: "_dalton, post the error and the recipe on pastebin (a lot easier to read them there), and a link here"
Feb 18 18:41:33 <_dalton_> i am sorry, will do that in a sec
Feb 18 18:42:21 <_dalton_> http://pastebin.com/Eyz1A4sd
Feb 18 18:43:04 <_dalton_> its just that i saw the logs, but now that i am in my hostel, away from the workstation, i had to go through the recepie making again...
Feb 18 18:43:24 <_dalton_> will now pase the recipe
Feb 18 18:44:22 <_dalton_> ok, here is the recipie
Feb 18 18:44:26 <_dalton_> http://pastebin.com/puV5y1QC
Feb 18 18:45:41 <_dalton_> i tried the steps here:
Feb 18 18:45:45 <_dalton_> http://www.yoctoproject.org/docs/1.6/bitbake-user-manual/bitbake-user-manual.html#hello-world-example
Feb 18 18:51:05 <neverpanic> looks like your $PATH is wrong
Feb 18 18:51:23 <_dalton_> waita sec, leeme check
Feb 18 18:51:59 <_dalton_> this is my PATH
Feb 18 18:52:01 <_dalton_> BBPATH="/home/dalton/poky/helloworld/"
Feb 18 18:52:15 <neverpanic> that's BBPATH, not PATH
Feb 18 18:52:20 <_dalton_> ok...
Feb 18 18:52:28 <neverpanic> follow A.3. again.
Feb 18 18:52:32 <_dalton_> u see, im a bit of a noob
Feb 18 18:52:36 <_dalton_> ok
Feb 18 18:53:30 <_dalton_> actually, i was lazy
Feb 18 19:02:26 <_dalton_> hey
Feb 18 19:02:38 <_dalton_> now i get a new error
Feb 18 19:02:50 <_dalton_> uploading to pastebin so that you could take a look
Feb 18 19:03:33 <_dalton_> http://pastebin.com/zvk9V7xv
Feb 18 19:06:39 <_dalton_> in the mean time, con u briefly tell be how to compile a .c file?
Feb 18 19:06:46 <_dalton_> *can
Feb 18 19:11:28 <neverpanic> that looks like an inconsistent working copy, like mixing different versions of bitbake; I'm not sure though
Feb 18 19:11:47 <neverpanic> And I don't have the time to explain you how to compile a .c file, but maybe somebody else has
Feb 18 19:13:22 <_dalton_> ok
Feb 18 19:13:35 <_dalton_> i just colned poky
Feb 18 19:14:05 <_dalton_> so still there is aposssiblinily of existance of different versions of poky?
Feb 18 19:27:32 <abelloni> JaMa: I just sent my patches for meta-qt5, I hope I took the correct approach
Feb 18 19:30:53 <JaMa> abelloni: I don't see them in ML yet, did you send them to oe-devel ML?
Feb 18 19:31:32 <abelloni> I used what is in the readme :)
Feb 18 19:31:52 <abelloni> that's oe-devel
Feb 18 19:33:48 <JaMa> ok maybe just mailman has longer delay today
Feb 18 19:38:22 <abelloni> or is it subsribers only ?
Feb 18 19:38:30 <abelloni> subscribers
Feb 18 19:38:50 <JaMa> abelloni: 1/2 just arrived
Feb 18 19:39:41 <JaMa> and I don't like it :) please move -r to separate variable (like we have OE_QMAKE_DEBUG_OUTPUT) and just set it to empty in recipes which shouldn't receive -r
Feb 18 19:44:26 <JaMa> abelloni: and please squash 2/2 to it when sending v2
Feb 18 19:50:33 <abelloni> just to be clear no qt recipe should use -r
Feb 18 19:50:45 <abelloni> So I'll set it from qt5.inc
Feb 18 20:39:27 <macbug> Hello, i'm working with a Intel® Atom™ E3845 atm. I can't give out any more hardware spec's so i appologise. My problem occurs during the boot, the screen turns black and then the unit seems to die. I'm using the intel vanilla bsp and a bsp from the manufacturer for the module the cpu is mounted on. I'm courius to know any hints of howto debug this issue.
Feb 18 21:02:01 <abelloni> JaMa: setn v2 which I did not call v2 :)
Feb 18 21:02:41 <JaMa> abelloni: thx, will look tomorrow
Feb 18 21:26:23 <Matulis> hey all, I swapped my kernel and I am getting the following error when trying to build my image: kernel-module-ipv6 not found in the base feeds
Feb 18 21:28:21 <rewitt> Matulis: What do you mean exactly when you say you swapped your kernel? That makes it sound like something has an RDEPENDS on the ipv6 module but the ipv6 module isn't getting built.
Feb 18 21:29:34 <Matulis> so I changed the machine conf to point to a different kernel recipe, the image had an IMAGE_INSTALL += kernel-modules
Feb 18 21:30:40 <rewitt> Try deleting tmp and rebuilding
Feb 18 21:31:16 <Matulis> ok
Feb 18 21:32:44 <rewitt> The documentation doesn't say that is necessary when changing kernels like that, but I THINK I remember someone suggesting I do that if I changed the kernel.
Feb 18 21:33:12 <Matulis> okay sounds good, i will update here if it works or not
Feb 18 21:46:24 <Matulis> is there a way to specify which version of gcc I would liek yocto to use?
Feb 18 22:02:32 <Matulis> rewitt, it looks like it didnt work
Feb 18 22:05:45 <rewitt> if you run bitbake -g and look at the package-depends.dot then you should be able to tell who has the dependency on the ipv6 module
Feb 18 22:10:24 <Matulis> hmm, theres no ipv6 in the package-depends.dot file
Feb 18 22:10:43 <Matulis> err, whoops opened task-depends
Feb 18 22:11:28 <Matulis> i mean my image depends on it, because I include all kernel modules
Feb 18 22:11:45 <rewitt> That's a meta package, it only contains the modules that the kernel actually built
Feb 18 22:13:07 <rewitt> is ipv6 in your DISTRO_FEATURES?
Feb 18 22:13:43 <rewitt> "bitbake -e image | grep "^DISTRO_FEATURES="
Feb 18 22:14:30 <Matulis> yes its in there
Feb 18 22:14:33 <rewitt> because what it sounds like is happening is that ipv6 is on in DISTRO_FEATURES and hence a dependency on kernel-module-ipv6 is coming from packagegroup-base.bb but your kernel isn't configured to build the ipv6 module
Feb 18 22:14:46 <Matulis> ah okay
Feb 18 22:15:03 <Matulis> so I can just defconfig the kernel to include ipv6 right?
Feb 18 22:15:25 <Matulis> ls
Feb 18 22:15:31 <rewitt> However that is done, I'm not extremely familiar with the mechanism :)
Feb 18 22:16:56 <rewitt> But, ipv6 is tristate which means it's possible that it's getting built into the kernel instead of as a module
Feb 18 22:17:47 <rewitt> And if that's the case you just need to get rid of the kernel-module-ipv6 dependency if you don't want it built as a module.
Feb 18 22:18:10 <rewitt> So either way, it's just whichever fits your goals best.
Feb 18 22:18:10 <Matulis> looks like i didnt have it selected at all in the defconfig
Feb 18 22:18:59 <rewitt> Yeah that would do it
Feb 18 22:20:46 <Matulis> yeah sorry I keep bugging you guys, still new to the embedded world, so its hard for me to tell whether its a linux or a yocto issue
Feb 18 22:22:17 <rewitt> Well each time you should learn more of how to debug, so eventually you'll be able to figure things out yourself. :)
Feb 18 22:30:54 <Matulis> yeah, every time someone responds I add a new tool to my collections ;)
Feb 18 22:36:23 <Matulis> okay rewitt, seems like I have more modules missing but now I know what to do, thanks!
Feb 18 23:42:19 <eicordob> Hi, anybody knows what is the meaning of the variable "$$@" inside a Makefile?
Feb 18 23:57:00 <rewitt> https://www.gnu.org/software/make/manual/html_node/Automatic-Variables.html
Feb 18 23:57:12 <rewitt> and the extra $ is https://www.gnu.org/software/make/manual/html_node/Secondary-Expansion.html#Secondary-Expansion
Feb 19 02:10:22 <nrossi> :q
Feb 19 02:10:30 <nrossi> whoopies :|
**** ENDING LOGGING AT Thu Feb 19 02:59:58 2015