**** BEGIN LOGGING AT Thu Oct 08 02:59:58 2015
Oct 08 03:08:36 <seebs> So, in cross-canadian.bbclass:
Oct 08 03:08:41 <seebs> cronss_canadian_bindirlinks () {
Oct 08 03:08:46 <seebs>   for i in linux ${CANADIANEXTRAOS}
Oct 08 03:08:48 <seebs> should this perhaps be
Oct 08 03:08:54 <seebs> for i in ${TARGET_OS} ${CANADIANEXTRAOS}
Oct 08 03:08:55 <seebs> ?
Oct 08 03:09:06 <seebs> (Mostly for RP, but maybe someone else understands this code better than I do.)
Oct 08 03:31:04 <namra> hm how to properly set enviroment variables for the cross-compiler? i.e. want to build a package with autotools, it depends on libxml2, thus i add it to DEPENDS in the recipe but it fails finding libxml/parser.h and other during the compilation.
Oct 08 03:31:53 <namra> adding "export CFLAGS_append = " -I/usr/include/libxml2" results in an error during do_configure() saying that this location is unsafe for cross-compilation
Oct 08 03:32:21 <namra> does it refer to the local systems /usr/include/libxml2 and not in tmp/sysroot/machine/?
Oct 08 03:33:21 <namra> what is then the purpose of the sysroot option?
Oct 08 03:39:28 <namra> nvm, STAGING_INCDIR
Oct 08 03:52:21 <ryanstur1er> Is there an easy way to view the .bb recipe for a specific package with all of the .bbappends applied?
Oct 08 03:53:09 <ryanstur1er> I am (attempting to) override a variable (enable a systemd service) in a .bbappend, and my service is not getting enabled, so I am suspcieious of the overlay
Oct 08 03:53:21 <ryanstur1er> (perhaps a layer priority issue, although I believe I have that set correctly)
Oct 08 03:54:03 <ryanstur1er> Do the ultimate values of the variables in the recipe appear in the logs somewhere?  (I didn't see this when I looked)
Oct 08 03:55:12 <ryanstur1er> @namra does your recipe inherit from any classes?
Oct 08 04:01:54 <namra> ryanstur1er: yes autotools
Oct 08 04:03:12 <ryanstur1er> can you post the recipe using pastebin so I can take a look?
Oct 08 04:03:40 <ryanstur1er> oh crying baby, afk
Oct 08 04:03:48 <namra> yes, but as mentioned above, solved the issue
Oct 08 04:03:49 <namra> hehe
Oct 08 04:05:08 <namra> ryanstur1er: http://pastebin.com/rL7NdDhE
Oct 08 05:51:41 <drou> hi guys
Oct 08 05:52:15 <drou> has anyone experienced this kind of error: | cp: cannot stat '/home/cedric/fido/build/tmp/work/i686-linux/libgpg-error-native/1.18-r0/libgpg-error-1.18/src/syscfg/lock-obj-pub..h': No such file or directory ?
Oct 08 05:53:07 <drou> i'm trying to build libgpg-error-native (on fido) but there is probably something wrong with my host configuration. lock-obj-pub..h should be something like lock-obj-pub.<arch>.h
Oct 08 07:08:06 <frsc> otavio: Regarding this: http://comments.gmane.org/gmane.linux.embedded.yocto.meta-freescale/16166. How to debug things like that? MACHINE_SOCARCH should be set in fsl-dynamic-packagearch, but somehow it's empty.
Oct 08 07:09:18 <frsc> Doing "bitbake alsa-lib -e | grep MACHINE_SOCARCH" shows me MACHINE_SOCARCH_mx6="armv7a-vfp-neon-mx6", but no MACHINE_SOCARCH_mx6ul
Oct 08 07:28:58 <pegu`> Is there support for MACHINE=socfpga_arria10 in meta-altera?
Oct 08 07:41:13 * LetoThe2nd gets ready to troll YPDD
Oct 08 07:43:30 <Crofton> w00t
Oct 08 07:43:43 * Crofton is trying to wake up
Oct 08 07:44:30 <LetoThe2nd> \m/
Oct 08 07:45:03 <namra> did someone experince the same error, when building openjdk-7-jre from the meta-java layer?
Oct 08 07:45:05 <namra> tmp/sysroots/x86_64-linux/usr/bin/ecj-initial is a version of gcj... no
Oct 08 07:45:17 <namra> tmp/sysroots/x86_64-linux/usr/bin/ecj-initial works... configure: error: The Java compiler /home/otti/Work/MedicusTek/iot/yocto/builds/pigeon-atom_baytrail/tmp/sysroots/x86_64-linux/usr/bin/ecj-initial failed (see config.log, check the CLASSPATH?)
Oct 08 07:47:11 <namra> tmp/work/x86_64-linux/classpath-native/0.99-r6.0/temp/run.do_configure.24029:1 exit 1 from
Oct 08 07:50:13 <namra> hm: tmp/sysroots/x86_64-linux/usr/bin/ecj-initial: line 4:  6107 Segmentation fault      (core dumped) ${RUNTIME} -Xmx512m -cp ${ECJ_JAR} org.eclipse.jdt.internal.compiler.batch
Oct 08 07:50:14 <namra> .Main ${1+"$@"}
Oct 08 07:54:44 <neg> hi, is there a way I can create a pseudo image (.tar.gz) containing the content of two packages? I tried doing a normal image (inherit image) but in addition to the package files I get etc/ld.so.cache and lib/modules/ stuff
Oct 08 07:56:28 <neg> What I'm trying to do is creata a .tar.gz file with uboot, uImage and a config file that can be handed to a developer extracted and burned to the device
Oct 08 08:52:22 <namra> damn howto build meta-java classpath package, it's always segfaulting :/
Oct 08 09:12:56 <maxin> namra: please file a bug.. I got that failure now...
Oct 08 09:14:13 <namra> maxin: want to try it with classpath 0.98 but bitbake doesn't want to use it. specified PREFERRED_VERSION_classpath = "0.98" in build/conf/local.conf
Oct 08 09:14:22 <namra> and running bitbake classpath-native
Oct 08 09:15:22 <maxin> namra: available version is 0.99 now (if you are in master branch)
Oct 08 09:16:24 <namra> you're a right and there's only 0.93 besides 0.99
Oct 08 09:17:58 <namra> maxin: where can one file an issue for that?
Oct 08 09:19:09 <maxin> namra: bugzilla.yoctoproject.org
Oct 08 09:19:22 <namra> thank you
Oct 08 09:33:34 <namra> maxin: https://bugzilla.yoctoproject.org/show_bug.cgi?id=8460
Oct 08 09:33:37 <yocti> Bug 8460: major, Undecided, ---, saul.wold, NEW , meta-java classpath-native segmentation fault
Oct 08 10:06:16 <namra> i'm confused Oo. cacao-initial-native depens on classpath but builds without it Oo
Oct 08 11:15:53 <lzm> how can i compare version numbers in a recipe? for example, V="2.24", i want to set BLA = ${V} > "2.23" ? "A" : "B";
Oct 08 11:43:16 <niteshnarayanlal> Hi
Oct 08 11:43:47 <niteshnarayanlal> I am writing a recipe in which I want to package some libraries to /usr/lib
Oct 08 11:44:15 <niteshnarayanlal> I tried doing  install -d ${D}${libdir}
Oct 08 11:44:15 <niteshnarayanlal>   install -m 755 myLib1.so ${D}${libdir}
Oct 08 11:44:15 <niteshnarayanlal>   install -m 755 myLib2.so ${D}${libdir}
Oct 08 11:44:20 <otavio> frsc: hello
Oct 08 11:44:28 <niteshnarayanlal> but didn't get any success
Oct 08 11:44:30 <otavio> frsc: which machine did you use/
Oct 08 11:45:09 <frsc> otavio: imx6ulevk
Oct 08 11:45:20 <otavio> frsc: let me check here
Oct 08 11:45:43 <otavio> frsc: which package?
Oct 08 11:46:01 <frsc> otavio: alsa-lib
Oct 08 12:04:45 <bluelightning> lzm: something like BLA = "${@'A' if d.getVar('V', True) > '2.23' else 'B'}"
Oct 08 12:05:16 <bluelightning> lzm: (that's just python within the ${@...})
Oct 08 12:39:28 <otavio> frsc: it works for me
Oct 08 12:39:39 <otavio> frsc: please paste your local.conf file please
Oct 08 12:40:07 <otavio> frsc: [4] ~/src/yocto/build/tmp/deploy/ipk/cortexa7hf-vfp-neon-mx6ul/alsa-lib_1.0.29-r0.1_cortexa7hf-vfp-neon-mx6ul.ipk [4]
Oct 08 12:44:21 <frsc> otavio: Ok. Thanks for investigating. This is my local.conf: http://pastebin.com/Un5qAJKt
Oct 08 12:45:09 <otavio> frsc: line 8 is wrong
Oct 08 12:45:52 <otavio> frsc: use ':production'
Oct 08 12:46:38 <otavio> frsc: but ideally you would do: MACHINEOVERRIDES_append = ":production" or use the distro one
Oct 08 12:47:45 <frsc> otavio: ah. I see, one colon too much. Thanks for the hint...
Oct 08 12:47:49 <neg> hi is there a way I can create a deploy archive (.tar.gz) containg only the content of two packages? I'm trying to collect firmware that is built as part of my linux system
Oct 08 12:48:47 <LocutusOfBorg1> neg, maybe by overriding do_rootfs?
Oct 08 12:49:33 <LocutusOfBorg1> or by using an imagefile.bb without any external dependency
Oct 08 12:49:46 <neg> LocutusOfBorg1: 'inherit image' and then override do_rootfs?
Oct 08 12:50:15 <frsc> otavio: but that doesn't solve the MACHINE_SOCARCH issue. Do you have an idea where to look for this?
Oct 08 12:50:34 <otavio> frsc: have you change the machine file somehow?
Oct 08 12:50:45 <otavio> frsc: and which branches are you using?
Oct 08 12:52:18 <LocutusOfBorg1> neg, I guess so
Oct 08 12:53:10 <ryanstur1er> I am attempting to enable a systemd service using a .bbappend file that is disabled by default in the base recipe
Oct 08 12:53:17 <ryanstur1er> the base recipe inherits from systemd
Oct 08 12:53:27 <ryanstur1er> and the SYSTEMD_AUTO_ENABLE="disable"
Oct 08 12:53:43 <ryanstur1er> so in my .bbappend, I am setting SYSTEMD_AUTO_ENABLE to "enable"
Oct 08 12:53:56 <ryanstur1er> but it's not doing the trick, and I'm wondering if this is how the universe works at all
Oct 08 12:54:22 <ryanstur1er> I can bitbake-layers flatten and see that my bbappend file is being applied to the base recipe
Oct 08 12:54:33 <frsc> otavio: No I didn't edit the machine file. I'm using current master of meta-oe, meta-fsl-arm, poky, meta-qt5 plus one custom layer...
Oct 08 12:55:03 <LocutusOfBorg1> neg, be sure you don't have any *IMAGE_FEATURES in the conf file
Oct 08 12:55:14 <ryanstur1er> but I don't see the symbolic link I expect in the final image under /etc/systemd/system/multi-user.target.wants
Oct 08 12:57:00 <otavio> frsc: so please do: bitbake virtual/kernel -e | grep SOC_FAMILY=
Oct 08 12:58:08 <LocutusOfBorg1> neg seems working
Oct 08 12:59:53 <neg> LocutusOfBorg1: ok thanks, I'm trying to get it to work, what did you put in your do_rootfs overrride?
Oct 08 12:59:58 <LocutusOfBorg1> https://paste.debian.net/314946/
Oct 08 13:00:03 <LocutusOfBorg1> no need to override it I guess
Oct 08 13:01:25 <LocutusOfBorg1> well, the resulting file is around ~5 mb, and it has some sort of busybox stuff
Oct 08 13:01:44 <LocutusOfBorg1> I guess because it is used in some sort of dependency
Oct 08 13:03:33 <LocutusOfBorg1> hi folks, any advice about adding qmake to the sdk image? I can add it in the target image, but I'm clueless about the sdk feature adding.
Oct 08 13:06:10 * LocutusOfBorg1 just found a meta-toolchain-qt5
Oct 08 13:06:46 <frsc> otavio: Oh dear! I think I got it. I once copied a customized version of imx-base.inc to my custom layer for testing and forgot to remove it again. And this one is missing the mx6ul definitions...
Oct 08 13:07:07 <frsc> otavio: Thank you for your help and sorry for bothering you...
Oct 08 13:07:35 <LocutusOfBorg1> otavio, sorry for bothering, but it is really nice to see you there
Oct 08 13:07:45 <LocutusOfBorg1> I'm reading your book right now
Oct 08 13:08:34 <otavio> frsc: this explains it :-)
Oct 08 13:08:48 <LocutusOfBorg1> I hope you will write another one too, maybe even more low level
Oct 08 13:08:55 <otavio> LocutusOfBorg1: Oh nice :-D I hope you are enjoying it :-)
Oct 08 13:08:58 <LocutusOfBorg1> or in sync with the new yocto releases
Oct 08 13:09:30 <LocutusOfBorg1> otavio, I'm a yocto user since an year or two, but I discovered *so many* things that I just did manually many times
Oct 08 13:09:40 <ryanstur1er> Actually on closer inspection, I'm not sure my .bbappend is getting bbappended
Oct 08 13:09:42 <otavio> LocutusOfBorg1: we will; but in fact we are finishing a more high level one, to help business decision makers to understand the yocto
Oct 08 13:09:51 <LocutusOfBorg1> so, yes, I found it really helpful
Oct 08 13:10:09 <ryanstur1er> What is the best way to determine if your bbappend file is actually being applied to the recipe you expect it is
Oct 08 13:10:18 <LocutusOfBorg1> otavio, my company will buy it as soon as it is released, I guess I know which book you are talking about
Oct 08 13:10:39 <LocutusOfBorg1> it should be on amazon and you can buy it in preview mode, right?
Oct 08 13:12:54 <ryanstur1er> Curious about this book now....
Oct 08 13:17:45 <otavio> ryanstur1er: use bitbake-layers tool
Oct 08 13:18:20 <otavio> ryanstur1er: LocutusOfBorg1: it is a full open book. Please take a look at: https://github.com/CollaborativeWritersHub/heading-for-the-yocto-project
Oct 08 13:18:46 <otavio> ryanstur1er: LocutusOfBorg1: we will soon announce it; but it is mostly complete. It is a booklet
Oct 08 13:19:14 <LocutusOfBorg1> wow really nice!
Oct 08 13:19:37 <LocutusOfBorg1> you know, the first step is to make the customer aware of the improvements he will have by switching to yocto
Oct 08 13:19:46 <LocutusOfBorg1> this is for sure something that we will use a lot
Oct 08 13:20:03 <otavio> LocutusOfBorg1: please help :)
Oct 08 13:20:05 <LocutusOfBorg1> the second step is to have a good book to train people using yocto on a daily basis
Oct 08 13:20:11 <LocutusOfBorg1> sure otavio
Oct 08 13:20:31 <otavio> LocutusOfBorg1: also if need anything from my side, or O.S. Systems side, let me know.
Oct 08 13:20:51 <LocutusOfBorg1> I have a question about sdk
Oct 08 13:21:00 <LocutusOfBorg1> but usually I answer by looking at the code
Oct 08 13:21:02 <LocutusOfBorg1> :)
Oct 08 13:21:06 <otavio> LocutusOfBorg1: if I know, I will happy to help
Oct 08 13:21:23 <LocutusOfBorg1> well, once ago it was possible to install the sdk in a custom location
Oct 08 13:21:41 <otavio> LocutusOfBorg1: yes, it is relocable
Oct 08 13:21:44 <LocutusOfBorg1> but when I changed the location, I found that the cross variables were still looking to the default path
Oct 08 13:22:01 <otavio> LocutusOfBorg1: it is a bug; if this is the case
Oct 08 13:22:15 <LocutusOfBorg1> let me check again
Oct 08 13:22:19 <LocutusOfBorg1> oh well
Oct 08 13:22:25 <otavio> LocutusOfBorg1: but I think it is working fine, at least our customers been not complaining
Oct 08 13:22:29 <LocutusOfBorg1> running with ./ works correctly
Oct 08 13:22:42 <LocutusOfBorg1> damn, sh filename works bad
Oct 08 13:22:49 <otavio> LocutusOfBorg1: ahhh
Oct 08 13:22:54 <LocutusOfBorg1> :s
Oct 08 13:22:57 <otavio> LocutusOfBorg1: maybe your shell is dash
Oct 08 13:23:09 <LocutusOfBorg1> sure, as a Debian Developer I use that one :)
Oct 08 13:23:17 <otavio> LocutusOfBorg1: and I remember it uses bashism
Oct 08 13:23:24 <otavio> LocutusOfBorg1: I am DD as well :)
Oct 08 13:23:41 <LocutusOfBorg1> "#!/bin/bash"
Oct 08 13:23:42 <LocutusOfBorg1> damn
Oct 08 13:23:50 <otavio> LocutusOfBorg1: I used to be the release manager of Debian Installer for a while
Oct 08 13:24:04 <otavio> LocutusOfBorg1: good times ;-D
Oct 08 13:24:05 <otavio> kkk
Oct 08 13:24:48 <LocutusOfBorg1> you still have a nice ddpo
Oct 08 13:25:10 <otavio> LocutusOfBorg1: from old ages; I've been out of Debian development for about 3 years
Oct 08 13:25:33 <LocutusOfBorg1>  Sun, 24 Nov 2013 15:49:13 on debian-devel-changes@lists.debian.org
Oct 08 13:25:33 <LocutusOfBorg1>  Message ID: <E1Vkbve-0001SD-AV@franck.debian.org>
Oct 08 13:25:35 <LocutusOfBorg1> I see
Oct 08 13:25:58 <LocutusOfBorg1> I can see your mobile phone from db.debian.org :)
Oct 08 13:26:03 <otavio> LocutusOfBorg1: I were quite active there when I was involved; I was maintainer of parted, grub, full debian-installer, apt, tasksel, ...
Oct 08 13:26:32 <otavio> LocutusOfBorg1: it is in my email as well :) also there is a US one now :)
Oct 08 13:26:32 <LocutusOfBorg1> nice, I'm quite active too
Oct 08 13:26:51 <otavio> LocutusOfBorg1: I have no time for Debian fun lately ...
Oct 08 13:26:58 <ryanstur1er> I am using the bitbake-layers tool to "show-appends" and I can see that my append is there
Oct 08 13:27:18 <ryanstur1er> When I say "bitbake-layers show-appends" does it show the appends in priority order?
Oct 08 13:27:28 <ryanstur1er> (So the last in the list is the last to be appended?)
Oct 08 13:27:44 <otavio> ryanstur1er: I think so; bluelightning?
Oct 08 13:29:12 <bluelightning> ryanstur1er: it does, yes
Oct 08 13:31:24 <ryanstur1er> Now if I use wildcards in my .bbappend, it seems to change the priority
Oct 08 13:31:49 <ryanstur1er> so connman_%.bbappend is lower priority than connman_1.27.bbappend seemingly invariant of my layer prioritites
Oct 08 13:32:14 <bluelightning> you're still using an older version of the build system right?
Oct 08 13:32:29 <bluelightning> in older versions that was a problem
Oct 08 13:32:41 <otavio> bluelightning: thanks for jumping in :)
Oct 08 13:32:47 <ryanstur1er> Yes thank you indeed
Oct 08 13:32:49 <otavio> bluelightning: are you in ELCE?
Oct 08 13:32:50 <LocutusOfBorg1> otavio, yes it did the trick :)
Oct 08 13:32:56 <ryanstur1er> I think I read something about that in the lists
Oct 08 13:33:05 <ryanstur1er> I'm using the distro packaged for the intel edison
Oct 08 13:33:10 <bluelightning> otavio: no problem
Oct 08 13:33:14 <ryanstur1er> I would desperately like to be using the new stuff
Oct 08 13:33:15 <bluelightning> otavio: yes, am helping out at dev day at the moment
Oct 08 13:33:17 <ryanstur1er> but I'm not there yet.
Oct 08 13:33:18 <otavio> ryanstur1er: gosh! update it :P
Oct 08 13:33:23 <bluelightning> ryanstur1er: ok, just checking
Oct 08 13:33:30 <otavio> bluelightning: has the Wandboards arrived?
Oct 08 13:33:38 <bluelightning> otavio: I believe so yes
Oct 08 13:33:39 <otavio> bluelightning: I saw Jefro emails
Oct 08 13:33:40 <ryanstur1er> ok to be clear: intel edison != edison branch of thing that was an old thing
Oct 08 13:33:44 <otavio> bluelightning: great!
Oct 08 13:33:49 <bluelightning> ryanstur1er: right, understood
Oct 08 13:33:53 <ryanstur1er> (that has been confusing for folks who are in the know)
Oct 08 13:34:20 <otavio> bluelightning: I wish I could been there :P
Oct 08 13:34:21 <ryanstur1er> I would love to update, but I feel I have to get it working with all their patches and stuff before I jump to the bleeding edge
Oct 08 13:34:31 <ryanstur1er> I can firm up my bbappend version numbers so that the priorities are right
Oct 08 13:36:41 <LocutusOfBorg1> question: I want to use my populate_sdk for image, but I want to add the meta-toolchain-qt5 tools
Oct 08 13:36:44 <LocutusOfBorg1> how to achieve this?
Oct 08 13:37:02 <LocutusOfBorg1> I mean I want to use qmae
Oct 08 13:37:03 <LocutusOfBorg1> qmake
Oct 08 13:37:21 <LocutusOfBorg1> inherit means that I need to copy a bunch of files from meta-qt to my meta
Oct 08 13:42:48 <LocutusOfBorg1> BTW any meta-raspberrypi maintainer here?
Oct 08 13:43:02 <LocutusOfBorg1> I made a patch to update wiringpi, and I would like to share it with you
Oct 08 13:43:30 <LocutusOfBorg1> they changed a lot of stuff, I don't want anyone else loose time to update it
Oct 08 13:45:26 <bluelightning> otavio: would have been nice to see you here definitely - hope we can have a phone bridge / skype set up for tomorrow perhaps?
Oct 08 13:46:00 <ryanstur1er> When the QA barks at me about "recipe depends on /bin/bash or /usr/bin/perl
Oct 08 13:46:07 <ryanstur1er> How do it know??
Oct 08 13:46:26 <otavio> LocutusOfBorg1: inherit populate_sdk_qt5 in the image.
Oct 08 13:46:32 <ryanstur1er> Is it just trolling through the shebang lines of whatever was generated?
Oct 08 13:46:59 <otavio> LocutusOfBorg1: Andrey ... I think you can do a pull request in github for it
Oct 08 13:47:24 <otavio> bluelightning: I will try to attend to the meeting
Oct 08 13:47:41 <otavio> bluelightning: in either case, Richard is proxying me
Oct 08 13:52:41 <LocutusOfBorg1> well otavio it works, sure, but I generate the sdk even without adding -cpopulate_sdk in that way :(
Oct 08 13:53:25 <LocutusOfBorg1> I think I'll split the image in image.bb and image-sdk.bb, and the latter including "inherit populate_sdk populate_sdk_qt5" and the other image
Oct 08 13:53:30 <otavio> LocutusOfBorg1: no. I said inherit
Oct 08 13:53:39 <otavio> LocutusOfBorg1: so populate_sdk work
Oct 08 13:53:55 <otavio> LocutusOfBorg1: I see no need for two images
Oct 08 13:55:01 <LocutusOfBorg1> let me check
Oct 08 13:55:07 <LocutusOfBorg1> so no inherit populate_sdk=
Oct 08 13:55:16 <LocutusOfBorg1> just populate_sdk_qt5
Oct 08 13:55:30 <niteshnarayanlal> any one any idea about packaging libraries via yocto recipe
Oct 08 13:57:04 <LocutusOfBorg1> otavio, not sure why it didn't work the first time
Oct 08 13:57:30 <otavio> LocutusOfBorg1: yes.
Oct 08 13:57:40 <otavio> LocutusOfBorg1: and normal -c populate_sdk
Oct 08 14:01:51 <acidfu> how do I silence a yocto QA error ?
Oct 08 14:03:33 <LocutusOfBorg1> acidfu, by overriding the task
Oct 08 14:03:58 <acidfu> well, there is only one error I want to ignore, not the rest
Oct 08 14:04:05 <acidfu> does overriding the task will ignore all of them ?
Oct 08 14:04:09 <LocutusOfBorg1> sure
Oct 08 14:04:18 <LocutusOfBorg1> another way might be to fix the qa error :p
Oct 08 14:06:26 <acidfu> yup
Oct 08 14:07:10 <gatisp> Hello. I was wondering how to proceed in the following case when:
Oct 08 14:07:10 <gatisp> bitbake A-native
Oct 08 14:07:10 <gatisp> results in:
Oct 08 14:07:11 <gatisp> Nothing PROVIDES 'B-native'
Oct 08 14:07:11 <gatisp> Cross compiling works fine - "bitbake A".
Oct 08 14:07:11 <gatisp> How to I tell bitbake to build also all dependecy recipes as *-native?
Oct 08 14:15:44 <gatisp> ah, so it seems that i need to add "BBCLASSEXTEND = "native nativesdk" to all dependency recipes
Oct 08 14:23:56 <rburton> gatisp: add *and verify the right thing happens*
Oct 08 14:28:56 <gatisp> rburton, trying that now
Oct 08 14:29:04 <gatisp> seems like i will have to bbappend many recipes
Oct 08 14:29:08 <gatisp> just to add that one line
Oct 08 14:29:17 <gatisp> doesn't feel very convenient :)
Oct 08 14:29:24 <rburton> send patches to the origin layer
Oct 08 14:29:58 <gatisp> rburton, is that considered a bug?
Oct 08 14:30:07 <rburton> not so much a bug as a valid extension
Oct 08 14:30:19 <rburton> supporting native builds is opt-in as you need to verify it actually works
Oct 08 14:30:48 <gatisp> ok, lets see if it works first
Oct 08 14:36:33 <LocutusOfBorg1> otavio, how do I get the "debian/developer/LocutusOfBorg" here on irc?
Oct 08 14:36:52 <LocutusOfBorg1> I don't remember how it is called
Oct 08 14:37:02 <LocutusOfBorg1> ident?
Oct 08 14:41:29 <LetoThe2nd> LocutusOfBorg1: its a cloak, talk to the folks in the freenode channel :)
Oct 08 14:48:06 <ohmy> hello everybody
Oct 08 14:48:23 <ohmy> i'm running into a strange errors during packaging and i'd like to have some help please
Oct 08 14:48:53 <ohmy> http://pastebin.com/D3Rc9cts
Oct 08 14:48:57 <ohmy> QA Issue: wayland-ivi-extension: The compile log indicates that host include and/or library paths were used.
Oct 08 14:49:17 <ohmy> Anyone have a begining of idea of what's all about ?
Oct 08 14:50:32 <ohmy> It indicates like i'm including or linking against host files/libraries but i can"t see this anywhere, i've checked all my libraries and they're all ARM and not X86 so how can i investigate please ?
Oct 08 14:51:08 <ryanstur1er> I'd start with the do_compile log that it indicates in your QA fail - can you pastebin that?
Oct 08 14:54:54 <ohmy> ryanstur1er: http://pastebin.com/9ynEn1AM it looks fine
Oct 08 14:57:31 <ohmy> ryanstur1er: compile looks fine, however i saw that the deployment is partial since only some includes files were installed + shared libraries however examples binaries are not copied (in case this details helps)
Oct 08 14:59:10 <joseppc> ohmy: at some point gcc includes what I guess may be a system dir: -I/weston
Oct 08 14:59:55 <ryanstur1er> The compile doesn't look fine, it looks like you've got a reference to your host weston libraries
Oct 08 15:00:06 <ryanstur1er> -L/usr/lib/weston
Oct 08 15:00:15 <ohmy> joseppc: ah damn
Oct 08 15:00:23 <ryanstur1er> compile indicates that it's unsafe for cross.  Did you write this recipe?
Oct 08 15:01:05 <ohmy> ryanstur1er:  no i'm using http://git.projects.genivi.org/?p=wayland-ivi-extension.git;a=tree;h=ad3bc5777583996cd61205380cc9e83dc2bcde3d;hb=bac7fc24bb94e37e33f447b58ec74240bb248841
Oct 08 15:01:46 <ohmy> ryanstur1er: for some reason compsitor.h was not found so in two files i changed #include<compositor.h> to #include <weston/compositor.h>
Oct 08 15:02:25 <ohmy> joseppc: thanks i'm trying to find a workarround
Oct 08 15:02:51 <ryanstur1er> ah there you go
Oct 08 15:03:05 <joseppc> ohmy: you're welcome
Oct 08 15:03:50 <ohmy> ryanstur1er: ah yes i wrote the recipe in fact
Oct 08 15:04:28 <ryanstur1er> You could maybe resolve it by adding an include path to your compiler flags in the recipe
Oct 08 15:04:36 <ryanstur1er> leaving the #include alone
Oct 08 15:06:08 <ryanstur1er> Or, if weston isn't a part of the package you're compiling it could be a DEPENDS issue in your recipe
Oct 08 15:11:00 <ohmy> ryanstur1er: weston is available on my sysroot
Oct 08 15:11:48 <ohmy> ryanstur1er: but when i've added message("************************" ${WESTON_LIBRARY_DIRS} ${WESTON_INCLUDE_DIRS}) to the CMakeLists.txt i cant see anything
Oct 08 15:12:09 <ryanstur1er> That's weird
Oct 08 15:12:21 <ryanstur1er> You get the message, but not the directories?
Oct 08 15:12:30 <ohmy> ryanstur1er: really stgrage because after cleanall i can see - checking for module 'weston'  --   found weston, version 1.6.0
Oct 08 15:13:11 <ohmy> ryanstur1er: and the best of the best i dont have weston installed on my host
Oct 08 15:13:52 <ryanstur1er> So that path it claims to be linking against doesn't exist?
Oct 08 15:14:28 <ohmy> ryanstur1er: yes
Oct 08 15:14:52 <ohmy> ryanstur1er: http://git.projects.genivi.org/?p=wayland-ivi-extension.git;a=blob;f=weston-ivi-shell/CMakeLists.txt;h=df91cb93fb0d435e827abb0cb256c85aac3a9e50;hb=bac7fc24bb94e37e33f447b58ec74240bb248841
Oct 08 15:14:57 <joseppc> it could be a truncated path
Oct 08 15:15:07 <ohmy> ryanstur1er: this is the only file where it looks for westing includes and libs
Oct 08 15:16:04 <ohmy> ryanstur1er: weston
Oct 08 15:17:28 <ohmy> ryanstur1er: pkg_check_modules(WESTON weston REQUIRED) succeeds it finds the westong and its version but for some reason WESTON_INCLUDE_DIRS and WESTON_LIBRARY_DIRS are not set
Oct 08 15:19:16 <ryanstur1er> If you do that print statement, but you print ${Weston_FOUND} do you get a 1?
Oct 08 15:20:21 <niteshnarayanlal> ryanstur1er, I am getting an error saying QA Issue: package tpm-emulator contains bad RPATH
Oct 08 15:20:44 <niteshnarayanlal> I tried to look up on internet to resolve it but didn't get any success
Oct 08 15:20:47 <niteshnarayanlal> any idea about it?
Oct 08 15:20:50 <rburton> ohmy: because the include and library paths are on the default search path so get removed.
Oct 08 15:21:02 <ryanstur1er> ooh, actually try printing ${Weston_INCLUDEDIR}
Oct 08 15:21:10 <ryanstur1er> instead of WESTON_INCLUDE_DIRS
Oct 08 15:23:26 <ryanstur1er> @niteshnarayanlal No, not immediately?
Oct 08 15:23:56 <ohmy> ryanstur1er: i foced include path and lib path and everything worked (set(xxxx "xx"))
Oct 08 15:25:21 <ohmy> rburton: thanks
Oct 08 15:25:26 <ohmy> ryanstur1er: from CMakeLists ?
Oct 08 15:27:24 <ohmy> ryanstur1er: from CMakeLists ${Weston_INCLUDEDIR} is empty
Oct 08 15:27:48 <ryanstur1er> Hmm
Oct 08 15:28:03 <ryanstur1er> Saw an issue where this was mis-named, but I guess that's not your issue
Oct 08 15:29:58 <ohmy> ryanstur1er: last thing :) please :) ERROR: QA Issue: wayland-ivi-extension: Files/directories were installed but not shipped hat does this mean ?
Oct 08 15:30:45 <ohmy> ryanstur1er: http://pastebin.com/4VbAh1CA
Oct 08 15:30:47 <rburton> ohmy: your compiler should be using a sysroot and so finding the right headers automatically
Oct 08 15:31:02 <rburton> installed but not shipped means do_install() installed a file, but it wasn't put into a package with FILES_*
Oct 08 15:31:16 <ohmy> rburton: thanks again :)
Oct 08 15:32:04 <ohmy> rburton: well it should i've just created a layer + recipe using yokto tools, haven't wrote any MakeFile by my own
Oct 08 15:38:47 <ohmy> thank you all, for your help and your time really appreciated
Oct 08 15:40:42 <LocutusOfBorg1> thanks LetoThe2nd
Oct 08 15:40:47 <LocutusOfBorg1> #freenode
Oct 08 15:40:58 <LocutusOfBorg1> right'
Oct 08 15:41:00 <LocutusOfBorg1> ^
Oct 08 16:09:20 <LocutusOfBorg1> otavio, how did you get your cloak?
Oct 08 16:35:21 <otavio> LocutusOfBorg1: as said by LetoThe2nd talk to #freenode guys
Oct 08 16:35:28 <otavio> LocutusOfBorg1: but it was looong time ago
Oct 08 16:46:45 <jmesmon> Is there a right way to add entries to ld.so.conf?
Oct 08 16:47:21 <jmesmon> it looks like I'd need to bbappend on the libc?
Oct 08 16:47:45 <otavio> jmesmon: you can use ld.so.conf.d dir, IIRC
Oct 08 16:48:13 <jmesmon> ah, that will work a bit better :) Do all libcs support that?
Oct 08 16:48:32 <otavio> jmesmon: not sure
Oct 08 16:54:23 <kergoth> Hmm, it'd be nice to have a new task that wraps devshell and just changes the devshell PWD from S to B
Oct 08 16:54:26 <kergoth> :)
Oct 08 16:54:33 <kergoth> for one off cases
Oct 08 16:56:04 <ryanstur1er> It seems like most recipes are built around pulling down a source distribution, cross-building t, and installing it
Oct 08 16:56:27 <ryanstur1er> but in the instance I have a recipe that I just want to lay down some files
Oct 08 16:56:38 <ryanstur1er> (configuration files, or docs, or something)
Oct 08 16:56:46 <ryanstur1er> is the re a good example of that somewhere?
Oct 08 16:56:50 <kergoth> bitbake doesn't care what the recipe does :)
Oct 08 16:57:01 <kergoth> a recipe could order a pizza if you really want
Oct 08 16:57:04 <ryanstur1er> I'm getting a little confused by how paths are defined
Oct 08 16:57:05 <kergoth> it just runs tasks
Oct 08 16:57:18 <kergoth> best advice i can give on that: read meta/conf/bitbake.conf
Oct 08 16:57:18 <ryanstur1er> well, for instance, I have to specify a license,
Oct 08 16:57:22 <kergoth> all the most important variables are there
Oct 08 16:57:49 <kergoth> generally, when we have just a recipe or recipe and trivial bits, we'll license those bits as MIT, and point LICENSE to the mit license file that's in oe-core
Oct 08 16:58:00 <kergoth> e.g. packagegroups and images and whatnot usually do that sort of thing
Oct 08 16:58:19 <ryanstur1er> I'm just having a little difficulty knowing where to locate my files, and how to specify them in a do_install step
Oct 08 16:58:28 <ryanstur1er> hmm ok
Oct 08 16:58:50 <kergoth> do_install writes everything relative to ${D}, think of that as /
Oct 08 16:58:59 <kergoth> all the standard paths reltaive to that have variables defined in bitbake.conf.
Oct 08 16:59:02 <kergoth> bindir, sbindir, etc
Oct 08 16:59:16 <nrossi> kergoth: are you subtlety suggesting someone makes pizza-party.bb? (https://github.com/coryarcangel/Pizza-Party-0.1.b/)
Oct 08 16:59:46 <kergoth> hehe. it's a long standing joke, though true, from back when we first started the project
Oct 08 16:59:56 <kergoth> don't think anyone ever got around to making such a recipe though
Oct 08 17:00:33 <ryanstur1er> It's the other side of that I'm wondering about
Oct 08 17:00:48 <ryanstur1er> I know how to tell bitbake where the files go to
Oct 08 17:01:10 <ryanstur1er> but how do I tell it where they come from, if they'r enot coming from a source distribution that was fetched and stored at ${s}
Oct 08 17:01:37 <kergoth> add the files to SRC_URI as file://, all files in file:// will end up in ${WORKDIR} rather than ${S}
Oct 08 17:01:42 <kergoth> there are plenty of examples of this
Oct 08 17:01:48 <ryanstur1er> bam,
Oct 08 17:02:29 <ryanstur1er> ok great, that's it.
Oct 08 17:03:09 <kergoth> Hmm, I think it'd be nice if the 'set -e' that we set for our functions/tasks was set via the sh commandline, rather than in the run. script
Oct 08 17:03:21 <kergoth> then you could source the run script in the devshell and it wouldn't exit your devshell on failure
Oct 08 17:03:26 <kergoth> so you could then run oe_runmake and whatnot directly
Oct 08 17:03:53 <rburton> sounds easy, patches welcome ;)
Oct 08 17:04:02 <nrossi> ryansturler: also for reference you can change "S" in your recipe to make it clearer that the source is actually in WORKDIR
Oct 08 18:12:48 <kergoth> the bitbake ui really needs to handle sigterm and properly clean up the tty state on exit
Oct 08 18:30:35 <kergoth> Hmm, getting a KeyError in the find_chains() process when bitbake is digging into a dependency loop :(
Oct 08 18:30:42 <kergoth> thought those were all fixed
Oct 08 18:58:07 <khem`> I see that nano has a RSUGGESTS on ncurses-terminfo and it gets pulled in by default by opkg. how can I blacklist it
Oct 08 18:58:35 <khem`> RRECOMMENDS can be blacklisted but I guess rsuggests are not
Oct 08 18:58:37 <ryanstur1er> I have made a custom recipe that drops a few files in /etc
Oct 08 18:58:40 <khem`> am I missing something
Oct 08 18:58:43 <ryanstur1er> which seems to "build" fine
Oct 08 18:59:11 <ryanstur1er> but when I do build_rootfs, I get opkg_install_cmd: Cannot install package mypackage
Oct 08 19:09:35 <ryanstur1er> I'm sure what I'm doing wrong is silly as heck:
Oct 08 19:09:36 <ryanstur1er> http://pastebin.com/HwhLjqTQ
Oct 08 19:10:45 <ryanstur1er> Feels silly to have it wrong after actually having got my actual source code to build/install and run :P
Oct 08 19:22:44 <bryan_> hi
Oct 08 19:23:19 <bryan_> can i get ibus recipe?
Oct 08 19:24:50 <rburton> khem: do you care about uclibc?  fails to build in jethro if you just set TCLIBC
Oct 08 19:30:29 <ryanstur1er> clear
Oct 08 19:30:29 <ryanstur1er> ls
Oct 08 20:01:23 <khem`> rburton: I do a bit how does it fail
Oct 08 20:01:34 <khem`> I thought I provided upfates to it
Oct 08 20:01:44 <khem`> and it was building minimal images for all arches
Oct 08 20:01:45 <khem`> on master
Oct 08 20:01:52 <khem`> a month or so ago
Oct 08 20:02:11 <khem`> but use musl :)
Oct 08 20:02:15 <khem`> its far better
Oct 08 20:04:58 <rburton> well, yeah
Oct 08 20:05:05 <rburton> so send patches to remove uclibc ;)
Oct 08 20:08:28 <kergoth> I wish there was a way to deal with dependency loops other than hardcoding the affected deps, for certain cases. E.g. the way we do for BASEDEPENDS & the autotools depends.. only added for everything after the things being added
Oct 08 20:08:54 <kergoth> e.g. we need this thing, but for building its deps, assume it's provided, or something
Oct 08 20:08:58 * kergoth scratches head
Oct 08 21:01:18 <kergoth> Hmmmmm
**** ENDING LOGGING AT Fri Oct 09 02:59:58 2015