**** BEGIN LOGGING AT Wed Nov 18 02:59:59 2015 Nov 18 08:49:09 good morning Nov 18 09:00:24 Anyone know how to disable UART pins on Edison running Yocto Poky? Overriding my scrip on pins GP130 (TX) and GP131 (RX), which I think is because of their UART role. This thread has detail: https://communities.intel.com/inbox Nov 18 09:00:57 JaMa: did you answer my question from yesterday and did I missed while commuting: Nov 18 09:00:57 [17:01] JaMa: can you point me out where in https://github.com/rehsack/meta-openembedded/commit/4fca1c50fceb7a91c2d067475eb47510672856a2 is a PACKAGing mess? Nov 18 09:00:57 [17:02] I tried to understand but didn't get the point Nov 18 09:01:21 Sorry wrong URL it's here - https://communities.intel.com/thread/92519 Nov 18 09:40:50 sno: I meant that current packaging was a mess, what your patch or what kergoth suggested is both ok with me Nov 18 09:45:52 Morning: can someone help me how to use BBMASK correctly? I'm trying to mask to packages by. The first one works, but as soon as I add the second one, the first one stops working Nov 18 09:47:08 are the entries separated by a | ? Nov 18 09:47:24 https://www.yoctoproject.org/docs/2.0/ref-manual/ref-manual.html#var-BBMASK Nov 18 09:48:44 joshuagl: yes the entries are seperated with the |. Like it is mentioned in the manual. Nov 18 09:49:07 no, the . is part of the regex Nov 18 09:50:16 joshuagl: this is how the line looks BBMASK = ".*qmlvideotest|*.ew15demo" Nov 18 09:51:08 The first one works, if it is just BBMASK = "*.qmlvideotest" but as soon as I add the "|.*ew15demo" it fails. Nov 18 09:51:29 is that regex correct? Nov 18 09:52:24 I don't see anything obviously incorrect in the BBMASK assignment Nov 18 09:53:11 hello who can help me with bitbake? Nov 18 09:53:22 I'm not 100% shure. As far as my understanding of regex goes, the line should state: either match .*qmlvideotest or .*ew15demo Nov 18 09:54:53 DatGizmo: just checking: ".*qmlvideotest|*.ew15demo" is incorrect, should be as you written it the second time Nov 18 09:56:15 joseppc: sorry, that was a typo the line is like this: ".*qmlvideotest|.*ew15demo" Nov 18 09:56:23 when i run bitbake i get this error Nov 18 09:56:32 ERROR: OE-core's config sanity checker detected a potential misconfiguration. Nov 18 09:56:32 Either fix the cause of this error or at your own risk disable the checker (see sanity.conf). Nov 18 09:56:32 Following is the list of potential problems / advisories: Your python is not a full install. Please install the module xml.parsers.expat (python-xml on openSUSE and SUSE Linux). Nov 18 09:57:04 i have tried to install several packages but the problem was not solved. I am running ubuntu 14.04 Nov 18 09:58:01 joseppc: joshuagl: Ah sorry, I was simply stupid. I mixed up the git reponame an the recipe name -_- now it's working. Nov 18 09:59:22 DatGizmo: ah cool :) Nov 18 10:00:23 The name of the first packege is qmlvideo not qmlvideotest .. funny that the regex still matched it. Nov 18 10:02:18 DatGizmo: maybe there is another BBMASK being appended somewhere else Nov 18 10:05:12 joseppc: maybe, but for now the BBMASK is working :) thanks for your time. Nov 18 10:08:15 DatGizmo: np ;-) Nov 18 10:25:33 Can anyone help with my UART query above? Nov 18 10:35:55 it looks that SDKs don't work on older host DISTROs anymore :/ Nov 18 10:37:05 sort -V added in b4caa808 and tar --checkpoint (without defining checkpoint action) added in ee4aadd1 so it fails with "unknown checkpoint action" for tar-1.20 Nov 18 10:39:35 JaMa: great, I await kergoth's response for a while and send today or tomorrow more samba cleanups with "disable-patches" :) Nov 18 10:40:51 JaMa: I had bad experience with RPROVIDES - maybe otavio can give you an impression what he had to suffer my rant during openjdk-8 recipe build ;) Nov 18 10:41:50 https://lists.gnu.org/archive/html/bug-tar/2008-10/msg00001.html Nov 18 10:42:53 sno: the biggest problem is because OE-Core does not yet support rprovide resolving for rootfs generation Nov 18 10:43:59 that covers with my problems around talloc - where rootfs creating of xbmc-dev-image failes for debugging galcore driver issues :) Nov 18 10:50:19 sno: the better fix is to prevent that default -dev dependency creeping in Nov 18 10:50:39 but with both changes, the -dev wont be empty and it will be fine Nov 18 11:07:41 Can anyone help with disabling UART GPIO pins on Edison running Yocto? https://communities.intel.com/message/351118#351118 Nov 18 12:35:54 Hi, I'm looking to have a bb append only apply to a specific board. The bb append will contain a single patch file which I believe I could install to the specified board with SRC_URI_append_$machinename. Is there a way I could only apply the bb append say using image_install_append from within the boards image recipe? Nov 18 12:43:52 as the repo in use is used by many boards, with a image recipe with install appends for each board Nov 18 12:51:34 CTtpollard: no, bbappends are always appended Nov 18 12:52:39 joseppc: so i'd have to have the append do SRC_URI_append_$machinename so it only applies it when the specific machine is being built for? Nov 18 12:53:21 CTtpollard: yes, that would do Nov 18 12:53:25 I mean it would apply the append for all, but it would only have a patch apply for that specific machine, all the rest it would just pass Nov 18 12:53:41 correct Nov 18 12:54:20 joseppc: thanks a lot Nov 18 12:55:40 CTtpollard: you're welcome Nov 18 13:29:09 Outside of yocto/poky, I've got an eval board and can build the kernel with the provided defconfig and the manufacturers fork of linux using "make " and "make uImage" However when I create a recipe and use the custom linux URL and defconfig the compilation fails. It seems that the .config file generated in both cases is very different. I've got it compiling now by using the working .config as my defconfig for yocto but t Nov 18 13:57:13 <__b4ub> Hi, got an issue with netfilters and iptables on my build. I want to use iptables to pass stuff to a NFQUEUE, but it doesn't work... Nov 18 13:58:35 <__b4ub> E.g : iptables -A OUTPUT -p udp --dport 4444 -j NFQUEUE --queue-num 0 ## has no effect on my build, while on my desktop pc it's doing what's it's supposed too do. Nov 18 14:01:16 <__b4ub> I have iptables enabled, and netfilters enabled on menuconfig. I also build libnetfilter-queue (think it's in meta-oe). Nov 18 14:22:26 rburton: what about other fixes for thumb1? Nov 18 14:23:17 JaMa: i've a mut2 already, and wanted someone with a vested interest in arm to double check them Nov 18 14:34:12 Hello. Anyone know how to disable UART on TX/RX pins on an Edison running Yocto? This refers https://communities.intel.com/message/351118 Nov 18 14:50:20 hi Nov 18 14:51:49 I have a Makefile that does -I$(includedir)/foo - what's the common way of dealing with that ? I can't find any env var pointing to the target sysroot... Nov 18 14:54:18 well, there is PKG_CONFIG_SYSROOT_DIR but it does not quite look like the right thing to use... Nov 18 14:59:50 hm, nevermind, the problem was somewhere else Nov 18 15:06:56 hm, sorry to contradict myself, but in fact yes, I still do have this problem (should have cleaned the recipe result before concluding that reverting the addition of PKG_CONFIG_SYSROOT_DIR did work...) Nov 18 15:12:20 but hell, gcc --sysroot should do its job... Nov 18 18:27:35 morning all Nov 18 18:27:43 morning Nov 18 19:34:19 morning Nov 18 20:34:06 Hello, I'm trying to build the gstreamer plugins using the git version instead of the default tarballs. I've tried to use PREFERRED_VERSION_gstreamer1.0-plugins-good = "git" in a bbappend file but it seems to still want to use the tarballs Nov 18 20:35:51 you need to set it to actual version, you can use % wildchar at the end Nov 18 20:36:45 so I have the following file name in my layer Nov 18 20:36:46 gstreamer1.0-plugins-good_%.bbappend Nov 18 20:37:51 And in that file I have the particular SRCREV I need and the PREFERRED_VERSION set there. Should I just define the preferred version as git in my image definition instead? Nov 18 20:41:40 all PREFERRED_VERSION does is select between versions of the recipe if there are multiple recipe files Nov 18 20:41:57 and it'll only work at the configuration level; setting it in a bbappend isn't going to do anything Nov 18 20:43:48 deviosity: from how the gstreamer plugins recipes are written it seems like you could just set PV = "git" and SRCREV to the desired value in the bbappend Nov 18 20:44:10 Thanks for clarifying bluelightning. Nov 18 20:44:59 hmm they don't set PV at all, that's wrong Nov 18 20:45:47 PV in the bbappend is definitely doing the trick Nov 18 20:45:57 deviosity: I recommend adding something like PV = "1.6.3+git${SRCPV}" in your bbappend together with required SRCREV Nov 18 20:46:17 then PREFERRED_VERSION_gstreamer1.0-plugins-good = "1.6.3+git%" in distro config (or local.conf etc) Nov 18 20:46:55 it will give you something sortable and upgrade path on target, if you have installed 1.6.2 there already Nov 18 20:46:59 JaMa: the gstreamer inc files have specific logic that checks for PV = "git" Nov 18 20:47:09 JaMa: so I'm not sure that will work in this specific instance Nov 18 20:47:27 if you don't use that you'll have to manually set SRC_URI as well Nov 18 20:48:32 at least AFAICT Nov 18 20:48:54 hmm :/ Nov 18 20:49:18 good that we have our own version of gst*1.0 recipes then Nov 18 20:49:37 well okay it seemed like it was working but has all sorts of issues Nov 18 20:53:41 JaMa: who's we? Nov 18 20:55:57 so setting the PV to git causes it to build the 1.4.5 tarball version agains the git version, which causes all sorts of problems Nov 18 20:56:54 bluelightning: LG Nov 18 20:57:37 deviosity: erm, that's strange because the logic says it should do exactly the opposite of that Nov 18 20:58:15 JaMa: ok... well FWIW these recipes were just sent to the mailing list and I guess nobody complained about how SRC_URI was being handled, if it's undesirable we can always patch it Nov 18 21:17:48 what's this about gstreamer recipes? Nov 18 21:25:42 rburton: there's some rather unusual logic in gstreamer-plugins.inc that sets SRC_URI depending on whether PV is set to "git" or not Nov 18 21:25:59 I can understand the intention, but it is somewhat unorthodox Nov 18 21:26:15 huh, didn't notice that Nov 18 21:44:51 hi all, I'm trying to use Hob with a core-image-base -- when I get to the 2nd part for editing the recipes to add some meta-python ones (python-pip) -- Hob crashes with a "cannot find core-base-layers.bb" error. What's up? Nov 18 21:56:33 Specifically, I see Nov 18 21:56:49 "Hob Exception - Error running command ['matchFile', 'core-image-base.bb']: Nov 18 21:57:00 with the NoSpecificMatch error. Nov 18 21:59:33 It looks like when you edit a recipe, Hob doesn't change the recipe name to the new one. Nov 18 22:01:52 kratsg: Hob is basically deprecated at this point I'm afraid Nov 18 22:02:00 it has lots of limitations such as this one Nov 18 22:26:11 Hi everyone, avoiding the rain here in Atlanta, and doing some Yocto, and I am stuck and hope to get some help from the community. I am trying to add/extend image.bbclass with python functions that require external python modules, specifically pyyaml, so I have included python-pyyaml-native as dependency, but when I include "import yaml" in some python function within the recipes, it complains that no module is found, how does one Nov 18 22:31:47 So I printed the os.sys.path and I see that we have openembedded-core and bitbake libraries there, so I am going to do "os.sys.path.append()" to append my python-pyyaml-native package in... if anyone has any other ways, I think I will try this now. Nov 18 22:33:35 aelmadhoun: does the recipe inherit pythonnative ? Nov 18 22:37:09 bluelightning: thanks! I was able to get it to work. Unfortunately, when I try to add the extra recipes, I get a kernel panic "Failed to execute /init (error -13)" Nov 18 22:37:33 (using bitbake my-custom-recipe with the machine set) Nov 18 22:38:16 kratsg: sounds like maybe IMAGE_INSTALL is being set rather than added to leading to sysvinit no longer being in the image? Nov 18 22:39:03 The line after it: "/bin/sh: earlyprintk: No such file or directory" Nov 18 22:39:13 I'm trying to figure out if it's the extra recipes doing it or not. Nov 18 22:40:31 It looks like image_install is being set Nov 18 22:40:57 checking to see if that fixes it. Nov 18 22:45:26 Hi Ivan, yes it does, it is an extension of image.bbclass Nov 18 22:45:30 bluelightning: that seems to have fixed it! how did you know? Nov 18 22:47:31 bluelightning: yes it does, its not a standard recipe, it is an image.bbclass extension Nov 18 22:47:58 one other question -- I don't know if bitbake can generate a ramdisk image for me. I usually have to use u-boot-tools and then pass in the cpio.gz image through a mkimage command to get my final image for the board. is there an automated way or not? Nov 18 22:48:59 also, unfortunately, it didn't completely work. I added `python-pkgutil` as part of the IMAGE_INSTALL and it's not showing up on the board. :-/ Nov 18 22:51:49 nvm, pkgutil is there! Nov 18 22:51:54 i can't spell. Nov 18 23:01:53 kratsg: re ramdisk images, yes it can, though I am not much of an expert with our u-boot integration Nov 18 23:22:51 ok, so everything looks ok for right now Nov 18 23:23:04 i grabbed pip, and then tried to pip install something -- but it looks like it fails to compile twisted. Nov 18 23:23:16 I'll need to see if I can use a different version of python-twisted (13.2.0 is too old for me) Nov 18 23:23:47 bluelightning: that's what I thought. It looks like it's just a special output format, but it's almost as easy as specifying the command and doing a post-execute thing. Nov 18 23:25:05 gtg, but thanks! Nov 18 23:28:51 can someone point out why STAGING_DIR_NATIVE is not resolved to proper path when I do this in an image.bbclass extension Nov 18 23:28:52 yaml_package = os.path.join(d.getVar('STAGING_DIR_NATIVE', True),d.getVar('PYTHON_SITEPACKAGES_DIR', True), 'yaml') Nov 18 23:29:21 os.sys.path shows => /usr/lib/python2.7/site-packages/yaml Nov 18 23:43:10 Apparently os.path.join in python will strip anything before supplied argument that looks like an absolute path, which PYTHON_SITESPACKAGES_DIR looks like it, so I need to strip the leading "/" Nov 18 23:43:13 bummer Nov 19 00:00:11 anyone setup a slack group for yocto yet? :) **** ENDING LOGGING AT Thu Nov 19 02:59:58 2015