**** BEGIN LOGGING AT Tue Aug 05 03:00:00 2014 Aug 05 03:40:56 Crofton|work: any chance of dinner this week? Aug 05 05:22:21 Hi! Is there a recipe for Ubuntu's UFW somewhere, or what is the preferred firewall frontend for Yocto? Aug 05 05:29:57 auke-: iptables/netfilter Aug 05 05:54:53 khem: yes, but i'm trying to avoid using iptables directly since i'm used to the simplicity of ufw... Aug 05 05:56:57 I could invest some time in getting to know iptables, but if there is an easy frontend for the simple usecases i would prefer that Aug 05 06:11:09 auke-: I would encourage you to write ufw recipe for OE/Yocto then Aug 05 06:38:06 khem: okay... I hoped there was something available, but i will try to create my own ufw recipe Aug 05 06:43:04 good morning Aug 05 07:14:00 Hai all, Am compiling libhid in yocto and getting following error "http://pastebin.com/ff9XyvkY". Aug 05 07:18:27 I couldn't sort out the issue. Can any one help me solve this? Aug 05 07:25:14 Hi yocto.. Aug 05 07:28:10 while compiling libhid in yocto at poky distro i am getting following error "http://pastebin.com/ff9XyvkY " please help me any one to resolve this issue.. Aug 05 07:29:03 i am struggling for long time to resolve this QA issue in libhid. Aug 05 07:29:20 kindly help us to solve this. Aug 05 07:30:02 can you give pastebin of file /home/elinux1/atmel/poky/build-71xx/tmp/work/armv5te-poky-linux-gnueabi/libhid/0.2.16-r0/temp/log.do_compile Aug 05 07:30:07 embed Aug 05 07:30:24 also /home/elinux1/atmel/poky/build-71xx/tmp/work/armv5te-poky-linux-gnueabi/libhid/0.2.16-r0/temp/log.do_package.1981 Aug 05 07:33:42 embed: there is no libhid_0.2.16.bb in meta-atmel Aug 05 07:33:53 embed: who created it? Aug 05 07:36:58 For my project purpose i have copied libhid into meta-atmel directory. Aug 05 07:37:42 * LetoThe2nd grins Aug 05 07:38:01 like, "this is my code, i just copied it to a random place and now please fix it for me?" Aug 05 07:39:35 LetoThe2nd: even worse, we don't have any coe here, only an error :-D Aug 05 07:39:43 s/coe/code Aug 05 07:39:58 please check this link for do_compile and do_package http://pastebin.com/LEsS4t6G Aug 05 07:40:28 ERROR: QA Issue: libhid: The compile log indicates that host include and/or library paths were used. Aug 05 07:40:46 looks like the packages compilation is b0rk3d Aug 05 07:41:00 respectively, its library/tool detection. Aug 05 07:41:19 so there you go. Aug 05 07:42:01 also check this link http://pastebin.com/bvzBAbyL Aug 05 07:48:17 i'd rather look into its config.log if i were you. Aug 05 08:21:55 morning all Aug 05 08:40:33 hi bluelightning Aug 05 09:01:45 hi mckoan Aug 05 09:47:41 Hi all Aug 05 10:09:02 Hello! Anyone who used yocto to create an image for the i.MX 6 utilite? I read it should be possible... Aug 05 10:27:19 bluelightning: Is there a way to disable multi threaded build for a particular recipe ? Aug 05 10:27:32 lsb_tester: yes, PARALLEL_MAKE = "" Aug 05 10:27:37 (within the recipe) Aug 05 10:27:46 ok ..thanks :) Aug 05 10:29:56 Why all the u-boots in meta-ti? (u-boot, u-boot-am33x, u-boot-glsdk, u-boot-ti-staging) What are the differences? Aug 05 10:45:27 anyone seeing errors like this today? Aug 05 10:45:29 ERROR: Execution of event handler 'sstate_eventhandler' failed Aug 05 10:45:29 Traceback (most recent call last): File "sstate_eventhandler(e)", line 6, in sstate_eventhandler(e=) Aug 05 10:45:32 TypeError: 'NoneType' object has no attribute '__getitem__' Aug 05 10:48:04 JaMa: sounds like that event handler is being sent events it should not be ... are you sure you're using the correct version of bitbake in conjunction with the metadata? Aug 05 10:48:12 hmm maybe there was wrong bitbake version in PATH Aug 05 10:48:23 after re-sourcing init script it seems to parse OK Aug 05 10:54:31 hi, is it OK to add a patch to busybox if it is not integrated into upstream, but fixes an issue with login names? Aug 05 11:24:29 I am trying to use "bitbake busybox", but I do not seem to get any error, nor the busybox source files in the build tree... why may that be happening? Aug 05 11:28:37 lpapp, sstate reuse ? Aug 05 11:29:02 kroon: I do not know... I have done bitbake -c cleanall busybox before that. Aug 05 11:29:18 basically, I have just added my own patch, and trying to build busybox with that. Aug 05 11:29:22 lpapp, try "-c cleansstate", to remove sstate Aug 05 11:29:40 lpapp, ah, adding a patch should trigger a build however.. Aug 05 11:29:48 yeah. Aug 05 11:30:12 also, cleanall oughta include cleansstate, too. Aug 05 11:30:19 I assume that is why it is called "all". Aug 05 11:31:49 hmm, it seems -c cleansstate helped... interesting, so there is at least two issues here then: a) cleanall not being all b) patch not triggering a rebuild. Anyway, I still cannot find the busybox source files in the build tree, so the main issue is there, but why is that? Aug 05 11:35:30 ./tmp/work/armv5te-foo-linux-gnueabi/busybox/1.20.2-r8/busybox-1.20.2/ is empty, so I assume there were some unreported errors? Aug 05 11:37:11 I entered rm -rf ./tmp/work/armv5te-foo-linux-gnueabi/busybox/ and let us see if clean restuff helps... Aug 05 11:38:43 wait, is this because of rm_work? Aug 05 11:42:19 lpapp, yes, rm_work will remove build dir after the output has been stored in sstate cache Aug 05 11:43:12 lpapp, also, I'm not sure it's a good idea to rm -rf stuff in /tmp by hand. there are stamp files elsewhere that rely on a certain state in tmp/ Aug 05 11:45:02 lpapp, I used to have rm_work, nowadays I just wipe tmp/ on a regular basis Aug 05 11:45:35 not sure what is best practise actually Aug 05 11:53:11 kroon: I do not have choice there... rm_work is simply not an option since I do not have 500 TB hard drive for all the builds ;) Aug 05 11:53:25 no rm_work* Aug 05 12:03:20 Hi.. Am compiling qt5 including by meta-qt5, if i compile this package i am getting following error http://pastebin.com/itubPHn8 Aug 05 12:03:44 please help me how to solve this issue. Aug 05 12:06:05 kroon: thank you for your time. Aug 05 12:06:41 lpapp, no problem, good luck Aug 05 12:07:15 -> install: cannot stat `/home/tamilarasi/qt5/poky/build-qt5/tmp/work/cortexa5hf-vfp-poky-linux-gnueabi/qtwayland/5.3.0+gitAUTOINC+98dca3b54f-r0/build/include/QtCompositor/5.3.0/QtCompositor/private/qwayland-server-*.h': No such file or directory Aug 05 12:07:18 JaMa: ^ Aug 05 12:09:35 lpapp: check if you really have 5.3.0 version (directory in QtCompositor) Aug 05 12:09:44 cleanall does call cleansstate Aug 05 12:11:11 embed: ^^ Aug 05 12:12:33 bluelightning: I have been facing a configure error in font-adobe-75dpi for missing fontutil. Aug 05 12:12:51 log extract here: http://pastebin.com/4rEkPBSh Aug 05 12:13:10 lsb_tester: have you tried to add fontutil dependency? Aug 05 12:13:27 it fails with parallel_make = 8 but works when set to 4 Aug 05 12:13:53 JaMa: yes it's there Aug 05 12:14:28 Looks to me like some thread sequencing issue Aug 05 12:14:30 there is only -native one (at least in font-adobe-100dpi_1.0.3.bb) Aug 05 12:15:12 DEPENDS has native one only Aug 05 12:15:17 lsb_tester: I'm not sure how that can be; do_configure isn't parallelised Aug 05 12:15:20 RDEPENDS has target one Aug 05 12:15:54 lsb_tester: RDEPENDS isn't enough, you need target one in DEPENDS Aug 05 12:16:08 But could it be possible that the .pc file is looked for before it is staged? Aug 05 12:16:23 yes, it's possible, because it's missing dependency Aug 05 12:17:21 JaMa: Yes I tried with that approach. It had worked after adding font-util to native in place of font-util-native Aug 05 12:17:21 DEPENDS ensures that font-util.do_populate_sysroot is finished before font-adobe-75dpi.do_configure Aug 05 12:17:25 RDEPENDS doesn Aug 05 12:17:35 t Aug 05 12:17:55 lsb_tester: please send patch with such change Aug 05 12:17:58 hmm, I'd assumed that the target one would not be needed, so I led lsb_tester astray yesterday Aug 05 12:18:06 lsb_tester: my apologies Aug 05 12:18:25 but I think that font-util-native is needed as well Aug 05 12:18:44 IIRC it calls some binary staged by font-util-native Aug 05 12:18:59 and target font-util is there only to make pkg-config happy in do_configure Aug 05 12:19:21 bluelightning: That's fine! Actually it built fine after doing so once. I need to repeatedly test the approach considering the level of inconsistencies I see with results everytime Aug 05 12:19:23 you can also patch configure to look for right font-util Aug 05 12:19:57 JaMa: is this something the inc file should be doing? or is it special to this set of fonts? Aug 05 12:20:23 bluelightning: I don't think that all font recipes are actually using font-util Aug 05 12:20:40 ok, just wondering Aug 05 12:20:53 so doing it in .inc probably won't harm (much) unless we have some font recipes with allarch Aug 05 12:21:53 depends how likely we are to encounter the same issue in other font recipes I guess Aug 05 12:22:35 right, font-adobe-utopia-100dpi_1.0.4.bb seems to have the same problem Aug 05 12:23:36 I came across this issue when in a new msd set up I tried "bitbake font-adobe-75dpi" as first build command. When I used to build it as part of some image recipe I never saw this. Aug 05 12:24:17 May be because all dependency font-util / native etc were already available by then Aug 05 12:24:19 lsb_tester: not surprising Aug 05 12:24:42 the RDEPENDS will build target font-util relatively close to font-* Aug 05 12:25:03 but not always in time before do_configure Aug 05 12:26:07 with high -b this wont be detected even by test-dependencies.sh script Aug 05 12:26:30 Then if we have both font-util & font-util-native in DEPENDS , that should take care of every thing I guess. Aug 05 12:26:46 -b as BB_NUMBER_THREADS Aug 05 12:26:52 yes Aug 05 12:27:41 Mine is set to = 8 although Aug 05 12:46:33 Hai all , Am trying to apply patch to kernel for my custom board. Here is my bbappend file http://pastebin.com/305KWDQP . while compiling am getting the following error "hdo_kernel_configme.26161" ttp://pastebin.com/c4vX6xX8. Aug 05 12:47:42 Can any one help me to fix the problem and compile it with my patch to kernel. Aug 05 12:55:21 sorry ,error link is http://pastebin.com/c4vX6xX8 Aug 05 12:56:04 JaMa: With the said approach it built fine. The patch looks like this: http://pastebin.com/x8E6UTB9 Aug 05 12:58:07 But what worries me now is I have two fontutil.pc files now. The native one looks ok with absolute paths to my bitbake workspace. Aug 05 12:58:55 But the other one has paths resembling to those in my host machine. Aug 05 12:59:57 So just having a doubt whether the build is referring to the system paths indeed. Aug 05 13:00:48 Is there any way to confirm this? Aug 05 13:05:59 there is another problem with bdf and font-* :/ Aug 05 13:05:59 | /OE/build/oe-core/tmp-eglibc/sysroots/x86_64-linux/usr/bin/ucs2any: Can't read mapping file '/usr/share/fonts/X11/util/map-ISO8859-1': No such file or directory! Aug 05 13:06:03 | Makefile:825: recipe for target 'courB10-ISO8859-1.bdf' failed Aug 05 13:06:05 | make: *** [courB10-ISO8859-1.bdf] Error 1 Aug 05 13:06:59 lsb_tester: read what is called from Makefiles Aug 05 13:23:07 hi, can't boot the fsl-community-bsp (daisy) -> core-image-minimal using tftp/nfs. any ideas (experiences)? when I do the same with dora branch everything works fine! Aug 05 13:24:36 when I boot the daisy sdcard image it works fine as well, so the issue seem to be related to the combination of daisy and remote file system using nfs !? Aug 05 13:24:54 I only get the following output from the serial: Aug 05 13:25:19 Booting image from tftp... Aug 05 13:25:32 Load address: 0x12000000 Aug 05 13:25:44 Loading: ######### ... Aug 05 13:25:45 JaMa: I think I missed an update somewhere and I've been meaning to ask about it for, like, a week and a half now. Are you planning to handle all merges to meta-networking through the master-next branch now? Aug 05 13:25:55 Using machid 0x10e9 from environment Aug 05 13:26:11 Starting kernel ... Aug 05 13:26:19 and nothing more Aug 05 13:26:43 * joeythesaint has avoided doing any meta-net stuff since the master-next testing with wireshark last month. Aug 05 13:26:53 Just trying to avoid causing any unnecessary conflicts. Aug 05 13:28:35 joeythesaint: please merge all commits you consider OK Aug 05 13:28:56 joeythesaint: I'm including meta-networking changes in master-next just to test them in my jenkins builds Aug 05 13:29:28 but then I usually wait for you to actually merge them (unless they are build fixes etc) Aug 05 13:30:20 joeythesaint: sometimes I merge something just because I don't notice it's actually in meta-networking, let me know if you disagree with some changes I've merged Aug 05 13:31:16 Nope, no concerns, and I know you rebase master-next all the time, I just don't want to cause you extra work. Aug 05 13:33:56 that's OK, feel free to merge (currently there are 4 changes I consider fine for merge): Aug 05 13:33:59 05daddf strongswan: Security Advisory - strongswan - CVE-2014-2891 Aug 05 13:34:01 a5c9311 quagga: Security Advisory - quagga - CVE-2013-2236 Aug 05 13:34:04 a3993a6 znc: Fix for CVE-2013-2130 Aug 05 13:34:06 2a33586 tcpslice: add recipe under tcpdump Aug 05 13:34:15 but you had some comments for tcpslice and I'm not sure if they were completely resolved Aug 05 13:35:06 Yeah, the others are straightforward, I was hoping to hear something back about tcpslice since it's a bit murky. Aug 05 13:35:49 I'll gather the others up now unless you've got them ready to merge now. Aug 05 13:37:54 JaMa: Faced the same error after renaming host machines font directory Aug 05 13:38:36 JaMa: That means we have to force the recipe to pick the native .pc file Aug 05 13:39:09 Other wise it picks all stuffs from host system while building Aug 05 13:43:09 is there a way to set PKG_CONFIG_PATH to the desired path through recipe ? Aug 05 14:40:52 What would be a sane way to remove an RDEPENDS from a packagegroup based on MACHINE Aug 05 14:46:51 WarheadsSE, try using overrides, RDEPENDS__remove_ = "..." Aug 05 14:50:17 bluelightning: My font-adobe-75dpi compile fails because it cannot find fonts/X11/util/map-ISO8859-1. Any missing dependency to add ? Aug 05 14:54:22 lsb_tester: not sure, sorry... Aug 05 16:05:01 bluelightning: I see those files are present in the work dir of font-util. To stage the files into sysroot, under which function in recipe I should write the code? Aug 05 16:05:22 is do_install_append appropriate ? Aug 05 16:09:23 lsb_tester: yes Aug 05 16:09:48 if they should be installed, they should be installed into ${D} from there Aug 05 16:09:49 ok...thanks:) Aug 05 16:11:52 Odd. libnewt blew up at compile since it was looking for gcc 4.9.0 sysroot headers instead of 4.9.1 Aug 05 16:12:32 Figured the gcc uprev would have largely caused a complete nuke and pave of everything... Aug 05 16:19:16 bluelightning: My intention is not to ship the files by rpm. I want to copy them to tmp/sysroot staging location. Aug 05 16:20:53 The font-adobe-* packages require these for compilation Aug 05 16:33:16 logrotate also failed in the same way, needing a forced cleanall to get rid of the gcc 4.9.0 state data. Aug 05 18:43:20 framebuffer fsl-image-machine-test@imx28evk (1/5) rpmdb: BDB1582 Open file handle: .../build-framebuffer/tmp/sysroots/x86_64-linux/var/lib/rpm/__db.001 Aug 05 18:43:23 framebuffer fsl-image-machine-test@imx28evk (1/5) rpmdb: BDB1582 Open file handle: .../build-framebuffer/tmp/sysroots/x86_64-linux/var/lib/rpm/__db.002 Aug 05 18:43:26 framebuffer fsl-image-machine-test@imx28evk (1/5) rpmdb: BDB1582 Open file handle: .../build-framebuffer/tmp/sysroots/x86_64-linux/var/lib/rpm/__db.003 Aug 05 18:43:29 framebuffer fsl-image-machine-test@imx28evk (1/5) Segmentation fault Aug 05 18:43:32 once again :-( Aug 05 18:51:47 I have in gstreamer the PACKAGECONFIG[x264] definition. How do I enable it? Can I do it in my image recipe or only local.conf? How? I looked in the docs and googled but didn't find a good hint (that at least seems to work) Aug 05 18:53:19 volker-: image recipe wouldn't make any sense. it changes how gstreamer is built. you can't modify one recipe from another Aug 05 18:53:27 PACKAGECONFIG_append_pn-gstreamer = " x264" Aug 05 18:53:36 put it somewhere in the config metadata. local.lconf, distro, whatever Aug 05 18:53:58 or in a gstreamer_xx.bbappend file Aug 05 18:54:34 kergoth: is -gstreamer sufficient? or do I need to list the full name like PACKAGECONFIG_pn-gstreamer1.0-plugins-ugly? Aug 05 18:55:00 its a general override. pn- Aug 05 18:55:05 so yes, it'd have to be the latter Aug 05 19:01:20 I am not happy that I can't do it in the image recipe Aug 05 19:01:38 I have all the packages I want to install to an image in this recipe and not in local.conf Aug 05 19:03:00 I have to think about it. I use one distro for two different products which only differ in some packages (and 1-2 features like commercial license + x264 etc) Aug 05 19:04:15 volker-: so it is not same distro Aug 05 19:04:19 volker-: ;-) Aug 05 19:04:49 I would not agree this way. You have ubuntu server edition and desktop edition. They only differ in packages Aug 05 19:05:12 still the same distro, just another "cd image" Aug 05 19:05:28 volker-: but they don't have same packages which different feature sets Aug 05 19:06:04 yocto splits every plugin and kernel module off, anyway, so I don't see there a too big deal. Aug 05 19:06:32 I know how yocto defines it, but I can't always 100% agree with the way to throw it out of the image recipe. Personal opinion :) Aug 05 19:06:33 volker-: ok; go deep in the metadata. I think you'll see what I mean Aug 05 19:10:11 is there an overview which yocto package provides which gstreamer plugin? Aug 05 19:11:06 volker-: the OE-Core follows the GStreamer plugins source recipes so you can mostly map the plugin name to the recipe Aug 05 19:12:08 otavio: I can't follow Aug 05 19:19:52 volker-: so what exactly would happen if you had image one changing how gstreamer builds, and then you "bitbake image1 image2 image3 image4"? Exactly which version of gstreamer packages is built? Aug 05 19:19:58 what you want just doesn't make any sense Aug 05 19:20:26 not with this project, anyway Aug 05 19:20:36 as kergoth indicates.. you can change what is -installed-, but everything is built (configured) once.. Aug 05 19:20:57 so if the configuration needs to be variable, then it's a different project and a different "distribution" configuration Aug 05 19:21:55 I know the build it only once philosophy and it would work here, too. Nevermind. :) Aug 05 19:27:16 python-gst is 0.10 while gstreamer comes with 1.0, too Aug 05 19:36:29 all, i'm having an odd build issue: Aug 05 19:36:32 update-alternatives: Error: not linking /home/justin/development/tc3-detos/detos/builds/tmp-detos-7.3.0-devonit-tc3-eglibc/work/devonit_tc3-oe-linux-gnueabi/detos-squashfs-image/1.0-r12/rootfs//usr/bin/resize to /bin/busybox.nosuid since /home/justin/development/tc3-detos/detos/builds/tmp-detos-7.3.0-devonit-tc3-eglibc/work/devonit_tc3-oe-linux-gnueabi/detos-squashfs-image/1.0-r12/rootfs//usr/bin/resize exists and is not a link Aug 05 19:37:12 This happens occasionally, and I usually have to clean out my build directory completely to resolve it. I'm hoping someone here might have some insight on a less invasive solution! Aug 05 19:37:51 theres's a recipe installing resize but not using update-alternatives for it, whereas busybox uses update-alternaitves for it. the fix is to fix the recipe providing resize to use update-alternatives for it as well Aug 05 19:38:43 kergoth, I am sure that is correct, but I am not familiar with update-alternatives (i.e. how to do what you just said lol) Aug 05 19:39:06 i believe the packages involved are e2fsprogs and xterm Aug 05 19:56:31 in a recipe common to both native and target, how to write a function (sysroot_stage_all_append() in my case ) only for native ? Aug 05 20:01:23 JBD1986: I had a similar case by having install_append in an bbappejnd Aug 05 20:03:38 volker-, I don't think I have a related install_append Aug 05 20:08:55 can anyone explain to me how to get the xterm_293 recipe to use update-alternatives? Aug 05 20:20:17 lsb_tester: be warned, you almost never want to touch sysroot_stage_all, unless you need to put files in paths that aren't included in the default behavior (e.g. unusual ones) Aug 05 20:20:20 * kergoth yawns Aug 05 20:20:37 lsb_tester: but you can always use the _class-native and _class-target overrides. foo_append_class-native Aug 05 20:24:30 I have a problem where the gcc compiler does not find a header file which in fact is in the sysroot Aug 05 20:24:58 Any hints on how to troubleshoot that? Aug 05 20:57:41 does someone know why python-pygobject uses --disable-introspection? Aug 05 20:59:40 volker-, introspection cannot be cross compiled iirc Aug 05 21:00:12 radhus: do you have an error message? Aug 05 21:00:31 | tapdisk-queue.h:32:20: fatal error: libaio.h: No such file or directory Aug 05 21:00:31 | #include Aug 05 21:00:48 is it looking for cross but seeing the native tree? ir is the include path missing something? Aug 05 21:01:36 Daemon404: thanks. You seem to be right, using the correct buzzword popped the following up in google: https://github.com/Guacamayo/meta-gir Aug 05 21:01:38 is libaio.h under usr/include or somewhere else? Aug 05 21:02:35 mr_science: yeah. seems like this particular x86_64-poky-linux-gcc does not have a --sysroot flag though Aug 05 21:02:36 and which sysroot is it in? Aug 05 21:02:40 i actually thought so Aug 05 21:03:22 afaik x86_64-poky-linux-gcc shouldn't need it Aug 05 21:03:27 exists here: tmp/sysroots/std-core2/usr/include/libaio.h Aug 05 21:03:30 std-core2 is my $MACHINE Aug 05 21:04:09 so your target arch is also x86_64? Aug 05 21:05:06 yes Aug 05 21:05:54 seems a little weird using bitbake to build the same arch as your host Aug 05 21:06:05 can't say i've ever tried that Aug 05 21:06:19 really? Aug 05 21:06:25 should work, though, as long as a vendor is set, to ensure BUILD_SYS != TARGET_SYS Aug 05 21:06:26 hehe, never thought it was weird :) Aug 05 21:06:26 afaik Aug 05 21:06:30 seems like it should still work Aug 05 21:06:38 * kergoth shrugs Aug 05 21:07:38 maybe I should buy a ARM board to build my x86_64 images ;) Aug 05 21:08:18 radhus: i'm a gentoo guy Aug 05 21:09:09 native builds and all that... Aug 05 21:09:20 yeah, I kinda miss gentoo Aug 05 21:09:33 * mr_science has a built a *lot* of native arm/ppc/sparc/x86 Aug 05 21:10:21 but i had to come back to oe/yocto, mostly to bug kergoth ... Aug 05 21:11:46 * mr_science currently building for beaglebonewhite **** ENDING LOGGING AT Wed Aug 06 02:59:59 2014