**** BEGIN LOGGING AT Wed Apr 26 03:00:02 2017 Apr 26 07:29:49 I added the layer meta-embedded/python-meta to the layers and now I have this error : ERROR: Layer 'meta-python' depends on layer 'openembedded-layer', but this layer is not enabled in your configuration Apr 26 07:30:01 Can someone help me please ? Apr 26 07:31:03 hamdyaea: what about re-reading the error message and acting accordingly? Apr 26 07:31:33 hamdyaea: hint: if a required layer in not enabled, maybe enbling it helps... ;-) Apr 26 07:32:42 LetoThe2nd: I know but I have an error message too : ERROR: Unable to parse /home/hamdy/poky/meta-openembedded/conf/layer.conf: [Errno 2] file /home/hamdy/poky/meta-openembedded/conf/layer.conf not found Apr 26 07:33:14 LetoThe2nd: meta-openembedded is not a layer it's a directory for the layers Apr 26 07:33:44 hamdyaea: correct. now what might be the solution? if a directory contains severeal layer directories? Apr 26 07:34:34 LetoThe2nd: hamdy@debian:~/poky$ bitbake-layers add-layers meta-openembedded/ Apr 26 07:34:34 The BBPATH variable is not set and bitbake did not find a conf/bblayers.conf file in the expected location. Apr 26 07:34:34 Maybe you accidentally invoked bitbake from the wrong directory? Apr 26 07:34:34 (despite the fact that you probably cloned meta-openembedded into your poky dir, which is bad practise) Apr 26 07:35:25 hamdyaea: look into meta-openembedded. there are several directories. each is a layer. now, what might be the thing to add? Apr 26 07:35:42 LetoThe2nd: I want to add meta-python Apr 26 07:35:55 hamdyaea: ok, and where is meta-python? Apr 26 07:36:24 LetoThe2nd: bitbake-layers add-layers meta-openembedded/meta-python Apr 26 07:36:38 hamdyaea: now that sounds better :) Apr 26 07:36:39 LetoThe2nd: In the directory meta-openembedded Apr 26 07:36:48 LetoThe2nd: no it don't works Apr 26 07:37:08 LetoThe2nd: maybe I have to put the full path, I try Apr 26 07:37:09 because it lacks the main openembedded layer, right? Apr 26 07:37:16 LetoThe2nd: yes Apr 26 07:37:34 well i never use that script, i just write my stuff into conf/bblayers.conf Apr 26 07:37:48 hamdy@debian:~/poky$ bitbake-layers add-layers /home/hamdy/poky/meta-openembedded/meta-python/ Apr 26 07:37:48 The BBPATH variable is not set and bitbake did not find a conf/bblayers.conf file in the expected location. Apr 26 07:37:48 Maybe you accidentally invoked bitbake from the wrong directory? Apr 26 07:37:50 hamdyaea: so why don't you just add the main openembedded layer? :-) Apr 26 07:37:55 LetoThe2nd: this is the result Apr 26 07:38:16 LetoThe2nd: Because I have the same error message Apr 26 07:38:26 hamdyaea: and, did you source oe-init-env? are you already in the build directory? Apr 26 07:38:36 The BBPATH variable is not set and bitbake did not find a conf/bblayers.conf file in the expected location. Apr 26 07:38:42 LetoThe2nd: Maybe that's the problem Apr 26 07:38:56 that pretty sure is there problem. read it again. what does it tell you? Apr 26 07:38:59 LetoThe2nd: Yes they are Apr 26 07:39:41 obviously there is no conf/bblayers,conf in the directory where you are calling bitbake-layers. Apr 26 07:39:52 hamdyaea: so, where might your conf/bblayers.conf be? :-) Apr 26 07:40:13 maybe.... enter the directory where your conf is? Apr 26 07:40:54 LetoThe2nd: yes I try Apr 26 07:42:36 hamdy@debian:~/poky/meta-openembedded/meta-python$ bitbake-layers add-layer meta-python Apr 26 07:42:36 The BBPATH variable is not set and bitbake did not find a conf/bblayers.conf file in the expected location. Apr 26 07:42:42 *sigh* Apr 26 07:42:51 you just are not listening. or not thinking Apr 26 07:42:52 LetoThe2nd: It seems that I miss the BBPATCH variable Apr 26 07:42:56 NOOOOO Apr 26 07:43:10 hamdyaea: where do you want to build? thats the first thing. Apr 26 07:43:43 in tmp/images as usually Apr 26 07:44:09 close. tmp/deploy/images gets created during the build. but where is tmp? Apr 26 07:48:13 LetoThe2nd: in poky/build/ Apr 26 07:48:25 hamdyaea: right. we call that the build directory. Apr 26 07:48:59 hamdyaea: so if you look into your build directory, there is a conf directory. and there is conf/bblayers.conf Apr 26 07:49:00 LetoThe2nd: So where can I add the BBPATH variable ? Apr 26 07:49:29 LetoThe2nd: Ah yes I see now Apr 26 07:49:42 so... if you hav a build directory, that contains a conf/bblayers.conf, and your error message is "bitbake did not find a conf/bblayers.conf" Apr 26 07:49:48 what might that tell you? Apr 26 07:50:36 LetoThe2nd: I'm not so beginner please, It's added in the bblayers.conf from the begining. This is why I ask here Apr 26 07:51:32 hamdyaea: theres nothing wrong with being a beginner. but i belive that just giving you a complete solution without making you think will not give you the understanding that you need the next time. Apr 26 07:52:31 LetoThe2nd: So now it works I don't know why.. Apr 26 07:52:58 LetoThe2nd: I added like before and now it works.. I don't know why.. Apr 26 07:53:17 hamdyaea: :-) Apr 26 07:55:25 LetoThe2nd: Now when I bake I have this error : ERROR: ExpansionError during parsing /home/hamdy/poky/meta-openembedded/meta-python/recipes-extended/python-pyephem/python-pyephem_3.7.6.0.bb Apr 26 07:55:47 hamdyaea: and did you add the base oe layer? Apr 26 07:55:55 LetoThe2nd: I just want to add python-pip Apr 26 07:56:01 LetoThe2nd: Yes I added it Apr 26 07:56:17 hamdyaea: and you did not mix branches? Apr 26 07:59:46 you might put your bblayers.conf into a pastebin so we can look at it. Apr 26 08:00:25 LetoThe2nd: So I removed now the /meta-python because I don't need it. The meta-oe contain python-pip Apr 26 08:01:16 ok, and? Apr 26 08:01:20 GNU nano 2.2.6 File: bblayers.conf Apr 26 08:01:20 # POKY_BBLAYERS_CONF_VERSION is increased each time build/conf/bblayers.conf Apr 26 08:01:20 # changes incompatibly Apr 26 08:01:20 POKY_BBLAYERS_CONF_VERSION = "2" Apr 26 08:01:20 BBPATH = "${TOPDIR}" Apr 26 08:01:20 BBFILES ?= "" Apr 26 08:01:20 BBLAYERS ?= " \ Apr 26 08:01:21 /home/hamdy/poky/meta \ Apr 26 08:01:21 /home/hamdy/poky/meta-poky \ Apr 26 08:01:22 /home/hamdy/poky/meta-yocto-bsp \ Apr 26 08:01:22 /home/hamdy/poky/meta-openembedded/meta-oe \ Apr 26 08:01:23 " Apr 26 08:01:35 i said "put into a pastebin" Apr 26 08:01:59 LetoThe2nd: Sorry I don't know how to do it Apr 26 08:02:29 hamdyaea: well then why don't you ask, or google it if you don't know something? everything is better than just silently ignoring it Apr 26 08:02:49 LetoThe2nd: And I want to create this image : bitbake core-image-x11 Apr 26 08:03:10 the bblayers.conf looks fine to me at a first glance. Apr 26 08:03:37 yes and I did a ls -la on all directories, the owner and the rights are good Apr 26 08:03:50 hamdyaea : https://gist.github.com/ Apr 26 08:04:06 so what is the problem *now*? Apr 26 08:04:14 hamdyaea : if you want to share results and so on. Apr 26 08:04:30 if meta-python is not there, then the recipe error you said above cannot be relevant. Apr 26 08:06:03 LetoThe2nd: No I just saw that python-pip is in meta-python Apr 26 08:06:17 I will add is the layers again Apr 26 08:06:29 i don't get it. Apr 26 08:07:49 LetoThe2nd: so now the Meta-Python layers is there : meta-python /home/hamdy/poky/meta-openembedded/meta-python 7 Apr 26 08:08:03 but I have the error message when I bitbake Apr 26 08:08:44 so you have meta-oe AND meta-python included, right? Apr 26 08:09:20 LetoThe2nd: yes Apr 26 08:09:57 hamdyaea: then please place a full log od the build attempt in a pastebin. so we can see whats going on. Apr 26 08:10:10 hamdyaea: i am pretty sure that you leave out some important bit. Apr 26 08:12:10 and what branch are you on? morty? Apr 26 08:15:33 LetoThe2nd: It's ok I have find the issue. The package meta-openemebdedd was not on the same branch Apr 26 08:16:21 :-) Apr 26 08:16:50 LetoThe2nd: yes morty thanks Apr 26 08:17:07 well if you say it works noow i don't need to reproduce it. Apr 26 08:19:12 I'm trying to switch my init.d system to a systemd system. But when i give uboot custom_params="init=/bin/systemd" to start systemd. but when i boot it returns: Requested init /boot/systemd failed Apr 26 08:19:37 MarcWe: why do you give a specific init command? Apr 26 08:20:20 MarcWe: and please notice: "/bin/systemd" != "/boot/systemd" Apr 26 08:20:26 if i don't give the command it says it can't find init= Apr 26 08:20:46 MarcWe: and you are sure that it is there? Apr 26 08:21:27 MarcWe: e.g. you followed the steps in http://www.yoctoproject.org/docs/2.1/dev-manual/dev-manual.html#selecting-an-initialization-manager Apr 26 08:21:56 yes i dit notice but the /boot/systemd is in the boot log and /bin/systemd is in the uEnv.txt Apr 26 08:22:07 and /bin/systemd exits in the image Apr 26 08:22:41 then you are certainly doing something weird, because i have a systemd image running here and there is no /bin/systemd :-) Apr 26 08:22:43 i don't have that 1 in yet but im using arago where it is selected. Apr 26 08:23:30 there is /sbin/init that does the correct forwarding to systemd. Apr 26 08:23:43 so something in your image is not yet set up correctly. Apr 26 08:24:06 now what. do you have those: DISTRO_FEATURES_append = " systemd" Apr 26 08:24:07 VIRTUAL-RUNTIME_init_manager = "systemd" Apr 26 08:24:09 or not. Apr 26 08:24:39 those are in arago layers but not in the layers of my recepy so im going to try that first Apr 26 08:25:23 maybe check if your image actually gets built with those then. bitbake -e $YOURIMAGE Apr 26 08:25:25 put them in my recipe Apr 26 08:25:47 nono. Apr 26 08:25:55 k Apr 26 08:26:14 it says: "Set the these variables in your distribution configuration" Apr 26 08:26:33 it does not say: "Set these variables in your image recipe" Apr 26 08:26:59 and that is for a reason: it only works in .conf files, because those variables have to be visible during the whole build. Apr 26 08:27:34 things that are set in your image recipe are only visible during the execution of the steps of that recipe. which does not work for setting system-wide things. Apr 26 08:28:17 a oke Apr 26 08:28:42 LetoThe2nd : Can you put that in other words ? I think i'm not understanding and seems important. Apr 26 08:28:50 so i can put it in my conf file Apr 26 08:29:23 ChrysD: in short: things in .conf files are visible everywhere, things in .bb files are only visible for that specific recipe. Apr 26 08:29:37 LetoThe2nd: sort of device and dist variables shout be in .conf Apr 26 08:29:38 LetoThe2nd : Oh yeah it's what i've understood. Apr 26 08:29:48 tnx Apr 26 08:29:49 LetoThe2nd : thanks Apr 26 08:30:13 MarcWe: thats why it sepcifically said: "Set these variables in your distribution configuration" :) Apr 26 08:33:20 LetoThe2nd : THanks for that important information. I missed that. Never checked more in deepth in the conf folder in the layer which there is the image recipe. Apr 26 08:38:12 Slightly off-topic, but most likely the right audience .. anyone here had any experience with using CAN bus on QEMU? Apr 26 08:43:35 My train of thought: It would be SO nice if my devs could have a CAN interface attached to their PCs and make it accessible to the system running under QEMU (whose images happen to be a product of our Yocto setup). Apr 26 08:44:32 JoiF: shouldn't some usb solution do the trick? Apr 26 08:46:18 LetoThe2nd: It might. That would actuallty be pretty great. Google pointed me to someone doing something for PCI interfaces, but just having a USB dongle of sorts would actually be even better. Apr 26 08:46:50 JoiF: maybe soemthing from vector.com can help you? Apr 26 08:47:21 JoiF : you mean that ? https://devel.rtems.org/wiki/Developer/Simulators/QEMU/CANEmulation Apr 26 08:48:04 ChrysD: This is exactly the "someone doing something for PCI interfaces" :) Apr 26 08:48:17 JoiF : have guessed . Apr 26 08:50:04 LetoThe2nd: Perhaps. I'd need to look more closely at their offerings :) Apr 26 08:51:04 LetoThe2nd: They seem to be focused on the automotive industry, but.. a CAN bus is just a CAN bus. ;) Apr 26 08:51:32 JoiF: yeah, but IIRC the pengutronix guys were somehow involved with their drivers, concerning socketcan Apr 26 08:51:45 LetoThe2nd: Ahh, I see. Apr 26 08:53:23 But thanks! Apr 26 08:53:29 I'll have a closer look at it Apr 26 08:53:50 Perhaps it won't make any sense in the long-run since our platform is the Xilinx ZYNQ.. Apr 26 08:54:20 Which means that there is a lot of stuff that's not going to be accessible in a QEMU environment anyway.. Apr 26 08:55:46 JoiF : It's normal that they are focused on the automotive industry as the CAN have been built with the contraints of automotive. Apr 26 08:56:28 JoiF : As a company, it's better to aim on the main market of CAN bus ahah. Apr 26 09:11:51 JoiF: There's a cheapish USB single-speed CAN box out in the wild; don't remember the name, though Apr 26 09:19:58 How can I change the keyboard of a minimal yocto image ? Apr 26 09:21:51 hamdyaea : I'm not sure if IMAGE_LINGUAS change also the keyboard layout. Apr 26 09:23:16 hamdyaea : You should have a recipe for the dedicated language with maybe a service on systemd for the suited languages. Apr 26 09:31:10 hamdyaea : if you are using systemd consider loadkeys for temporary usage or localectl list-keymaps to see if you have the language you want Apr 26 09:31:33 hamdyaea : then change the file /etc/vconsole.conf Apr 26 09:35:18 hamdyaea : but depends it's not sure that you have the vconsole.conf. Could be another file to change permanently. But still, you will need in yocto to patch the recipe wich provide that file. Apr 26 09:38:41 ChrysD: thanks Apr 26 09:39:32 How can I add two extra image features "tools-sdk" and "debug-tweaks" at the same time ? Apr 26 09:40:20 hamdyaea : EXTRA_IMAGE_FEATURES += "tools-sdk debug-tweaks" ? Apr 26 09:42:46 hamdyaea : you have trouble to put more than one extra_image_features ? Apr 26 09:44:32 ChrysD: Thanks it works now Apr 26 10:04:00 ChrysD: Definately. :) Apr 26 10:17:45 "Use a Shared State Cache (sstate) among groups of developers who are on a fast network. The best way to share sstate is through a Network File System (NFS) share. The first user to build a given component for the first time contributes that object to the sstate, while subsequent builds from other developers then reuse the object rather than rebuild it themselves." Apr 26 10:18:08 Does the sstate mechanism provide all the necessary locking etc.? Apr 26 10:18:14 i.e. is this robust? Apr 26 10:21:49 Also, I'm wondering what's appropriate to put under SCM? Apr 26 10:24:05 Currently I have a written documentation telling users/developers to git clone poky and the appropriate layers, source oe-init-build-env and then sed-lines such as: sed -i "/meta-yocto-bsp/a\ $(dirname $PWD)/meta-foss \\\\ " conf/bblayers.conf Apr 26 10:24:48 I'm just wondering if it's appropriate to put the conf-directory under SVM or something like that, when someone wants to set up a build environment, they don't need to be editing files? Apr 26 10:24:55 s/SVM/SCM/ Apr 26 10:47:56 Is it possible to set a default only if another value is missing? E.g. will the following work: SYSTEM_VERSION??="fallback-version" SYSTEM_VERSION?="${APPVERSION}" ? Apr 26 10:49:09 I suspect that SYSTEM_VERSION will be set to "" if ${APPVERSION} is empty an thus it will not use the "fallback-version" string Apr 26 10:57:41 hello guys Apr 26 11:04:00 Harro Apr 26 11:04:30 Ox4: Didn't you priv-msg me a few days ago? Apr 26 11:05:12 JoiF: no Apr 26 11:05:53 Ox4: Ahh, yes you did. I see it in my logs now. Apr 26 11:06:43 Ox4: So what I was suggesting with the bitbake -u thing (where knotty is the default) is to see if there is something that you could use instead of knotty in order to suppress the output. Apr 26 11:07:08 But perhaps you need to create your own (which you could base on knotty) that would print out less information. Apr 26 11:07:13 guys, I have a question regarding net-snmp compilation. I have enabled python-modules in the bbappend file and now I receive an error during compilation. Could you please help? Here is the log: http://0x0.st/2n Apr 26 11:07:50 JoiF: I have only knotty and ncurses there :-( Apr 26 11:08:06 Ox4: Me too. Apr 26 11:08:51 Ox4: That's why it was more of a theoretical suggestion.. Not very helpful, I know.. ;) Apr 26 11:10:15 JoiF: anyway, thank you :) Apr 26 11:10:42 JoiF: do you have any ideas about this one -> http://0x0.st/2n ? Apr 26 11:12:55 Not really Apr 26 11:13:08 But I'm a n00b here.. so.. ;) Apr 26 11:16:02 Except just the obvious .. it's trying to enter a python directory and execute setup.py, neither of which exist.. Apr 26 11:18:34 I need gtk in my image someone know how to add it please ? Apr 26 11:18:59 hamdyaea: gtk2 or gtk3 Apr 26 11:19:16 hamdyaea: by now you should know https://layers.openembedded.org/layerindex/branch/master/recipes/?q=gtk Apr 26 11:26:29 gtk2 Apr 26 11:27:45 hamdyaea : Everytime you search a layer, you should check in https://layers.openembedded.org/layerindex/ Apr 26 11:28:09 hamdyaea : And as you know how to install a package, shouldn't be a problem. Apr 26 11:28:32 JoiF: but it exists Apr 26 11:29:20 Ox4: Under /home/yocto/yocto_varsomam33/tisdk/build/arago-tmp-external-linaro-toolchain/work/cortexa8hf-neon-linux-gnueabi/net-snmp/5.7.3-r0/build ? Apr 26 11:31:05 hamdyaea : If i remember gtk2+ recipes are already in poky under meta layer. Apr 26 11:32:22 https://layers.openembedded.org/layerindex/recipe/5762/ Seems like they're just called gtk+ .. which happens to be version 2.x Apr 26 11:32:53 yes gtk+ in oe-core is gtk2 Apr 26 11:33:48 hamdyaea : poky/meta/recipes-gnome/gtk+ Apr 26 11:36:09 hamdyaea: you know what to do now right. bitbake gtk+, use oe-pkgdata-util to examine what binary packages were generated, add what you want to your image. Apr 26 11:37:57 rburton : i've never checked what oe-pkgdata-util is. I should investigate. Thx. Apr 26 11:39:54 JoiF: no, under /home/yocto/yocto_varsomam33/tisdk/build/arago-tmp-external-linaro-toolchain/work/cortexa8hf-neon-linux-gnueabi/net-snmp/5.7.3-r0/net-snmp-5.7.3 directory Apr 26 11:42:37 Ox4: I may be wrong, but it seems to me that it's expecting it under "build", because the make has only "left" the mibs directory. But I have never built this package, so perhaps this is a red herring. Apr 26 11:42:52 s/the make/make/ Apr 26 11:43:05 or: s/the make/the make process/ Apr 26 11:43:08 whichever you prefer ;) Apr 26 11:49:58 Hello, is it possible to share variables from one recipe to an other ? Apr 26 11:51:05 no Apr 26 11:51:15 share a variable at the distro level so all recipes can see it Apr 26 11:51:50 JoiF: ok, thanks Apr 26 11:52:03 Hunk_ : The variable should be set into a ".conf" file which defines the distro variables. Should be located in the folder conf in the meta where the image is defined. Apr 26 11:52:33 I need the WORKDIR of my downloader recipe Apr 26 11:53:04 is it possible to get the WORKDIR of a specific recipe? Apr 26 11:55:32 HUnk: you can't expect there to be anything inside it Apr 26 11:55:53 HUnk: if rm_work is enabled, or the recipe was extracted from sstate, it will be empty Apr 26 11:56:19 if you want to share files between builds they need to go into the sysrot Apr 26 11:56:21 sysroot Apr 26 11:57:15 rburton: ++ for the sysrot typo ;-) Apr 26 11:59:28 Hunk_ : located in poky/meta/conf/bitbake.conf Apr 26 12:00:26 Hunk_ : there is defined WORKDIR, you can do " cat bitbake.conf | grep WORKDIR " and you will see how it's made . Apr 26 12:01:02 Can I source a another file and use its variables in a shell function to a recipe? Apr 26 12:01:18 ChrysD: but the point is you can't rely on there being anything in there unless you're actually doing the build of that recipe Apr 26 12:01:43 rburton : +1 Apr 26 12:01:46 sveinse: shell functions are executed by the shell so yes, just use . Apr 26 12:01:54 E.g source ${IMAGE_ROOTFS}/etc/os-release and later echo "$PRETTY_NAME" Apr 26 12:02:01 rburton: thanks Apr 26 12:02:07 ok ty Apr 26 12:02:16 anyone has experience with gitsm fetcher? Apr 26 12:02:18 sveinse: . not source as source is a bashism Apr 26 12:02:39 rburton: yeah, I understood this from your reply, thanks Apr 26 12:02:42 the unpack step of gitsm: is really slow Apr 26 12:08:19 I need python-dev but I don't find it in the recipes. Is there another way to install it ? Apr 26 12:09:27 hamdyaea: if a -dev package is provided by python it should be automatically built. check your packages folder if its already there. Apr 26 12:09:34 then you can rdepend on it Apr 26 12:10:39 its called - wait for it - python-dev Apr 26 12:10:51 * LetoThe2nd waits... and waits... and waits... Apr 26 12:10:53 python-dev: Apr 26 12:10:53 /usr/include/python2.7/Python-ast.h Apr 26 12:10:53 /usr/include/python2.7/Python.h Apr 26 12:10:54 etc Apr 26 12:11:09 rburton: ok I look at it Apr 26 12:11:45 hamdyaea: to me it sounds like you are trying to tinker some specific software or thing into your board. you're asking us about a lot of dependencies and such, and runtime installation things. Apr 26 12:12:09 hamdyaea: generally, it is a vastly superior way to just properly package that software up. Apr 26 12:12:10 LetoThe2nd: I try to run a python software Apr 26 12:12:26 hamdyaea: well then properly package that thing up. Apr 26 12:13:03 hamdyaea: all this runtime stuff is wasted time, once you get the second board. Apr 26 12:13:10 or the first one breaks. Apr 26 12:13:20 or you break it because of tinkering. Apr 26 12:13:23 or, or, or. Apr 26 12:14:40 so why not start with something like http://git.openembedded.org/meta-openembedded/tree/meta-python/recipes-connectivity/python-txws/python-txws_0.9.1.bb for example, and do it right from the beginning. Apr 26 12:18:09 LetoThe2nd: Twisted Web Socket it's usefull for me ?? Apr 26 12:19:02 hamdyaea: It's an example on how I might package a python-whatever Apr 26 12:19:06 you* Apr 26 12:19:14 hamdyaea: not twisted websocker per se. but its a recipe that shows you how to package a pip-based software package for oe Apr 26 12:19:44 hamdyaea: as an example for the rdepends, inherits, etc. Apr 26 12:29:15 Hello! Anyone working with the raspberry pi layer https://github.com/agherzan/meta-raspberrypi? I was trying to build for arm64 but it seems that video core userspace libs like libbcm_host are not included. Is this normal? Maybe that’s because no 64bit version is available? Apr 26 12:30:44 I need to have pygtk, I have some error when I compile it from the source.. can you help me please ? Apr 26 12:31:05 "i have some error" Apr 26 12:31:10 oh, *that* error. Apr 26 12:31:28 hamdyaea: you are really making it hard for people to help you. Apr 26 12:32:27 LetoThe2nd : relax. Ahah. Apr 26 12:32:35 hamdyaea: generally, whenever you have a problem, it works like this 1) review the problem. often the error message already tell you what to do 2) assemble a concise, and detailed description of the problem. 3) google it. 4) ask. Apr 26 12:33:04 hamdyaea: the source on the target, or are you bitbaking it? Apr 26 12:35:10 I can't install it from the source on the target. I have an error message that tells me that my version of python is older than 2.3.5 but my version of python is 2.7. And I have try all possibilites from there.. I want to try to bitbake it now Apr 26 12:35:45 hamdyaea : seems a python problem and not a yocto one. Apr 26 12:36:01 hamdyaea: the goal should be to bitbake everything Apr 26 12:36:15 especially stuff like pygtk for which there is a recipe Apr 26 12:36:21 partially, it sounds like its a python problem by doing things at runtime, in-target one more time. Apr 26 12:36:37 hamdyaea : python -v tells you what ? Apr 26 12:37:26 ChrysD: 2.7.12 Apr 26 12:38:16 in google i see that there is a python-pygtk with recipes-devtools but I don't find it anymore on the git Apr 26 12:39:35 http://git.yoctoproject.org/cgit.cgi/poky/plain/meta/recipes-devtools/python/python-pygtk/?h=laverne Apr 26 12:40:04 https://layers.openembedded.org/layerindex/recipe/40399/ <— last existed in jethro Apr 26 12:40:07 rburton: i can only see the one in meta-debian Apr 26 12:40:23 so i'd say there is no pygtk recipe at the moment. Apr 26 12:41:29 if you want pygtk2 the you can grab that recipe i posted Apr 26 12:41:31 might work Apr 26 12:41:39 it was removed on the grounds that its ancient Apr 26 12:41:46 but if you want ancient software that's your problem Apr 26 12:41:47 rburton: thanks I try Apr 26 12:52:55 rburton: How can I change it now from the branch jethro to the branch morty ? I will not put them in the same directory so it will be no clonflicts Apr 26 12:53:50 'git checkout ' Apr 26 12:54:12 rburton: I just find it but thanks for the answer Apr 26 12:55:15 Image bb's seem to use "DESCRIPTION" and "SUMMARY" quite interchangeably, and according to the documentation, if no DESCRIPTION is defined, then what ever is in SUMMARY is used for DESCRIPTION. Can I therefore deduce that SUMMARY is the one I should use (as I get 2 for 1)? Documentation doesn't specify. Apr 26 12:56:42 yes Apr 26 12:56:58 Thanks :) Apr 26 13:17:02 rburton: When i change the branch it modify the file localy too. It means that I loose the python-pygtk... Apr 26 13:17:17 rburton: I try to copy the files from one to another Apr 26 13:19:58 rburton: How can I add it to my bitbake ? Apr 26 13:22:46 LetoThe2nd: Do you know how I can install python-pygtk from jethro to morty ? Apr 26 13:24:30 hamdyaea: just copy it over? Apr 26 13:24:39 hamdyaea: to get one file from another branch (jethro): "git checkout jethro -- meta/recipes-devtools/python/python-pygtk_2.24.0.bb" Apr 26 13:24:59 hamdyaea: in general i would say, create an empty layer for your custom recipes and put it in there. Apr 26 13:25:39 ^ yeah that would be the reasonable thing to do Apr 26 13:26:09 yocto-layer create should do that quite nice. just go to the directory that you want to contain the layer directory before calling it, as it will use the currend pwd. Apr 26 13:26:59 hamdyaea: the single file is unlikely to help a lot thought , I'm sure there are patches and such Apr 26 13:27:31 If you do any changes, the best is to have your own meta because if for one reason you git your meta, you will loose all your modification before. So to know what you have change, you should always do your own meta using the same structure as the recipe you use. Apr 26 13:27:50 and if you have a own layer, you can happily stuff the recipe and everything related into it. Apr 26 13:28:14 ChrysD: i guess you mean "layer" when you say "meta" Apr 26 13:28:17 And you can even put your meta on git to never loose it and share it with other people =D Apr 26 13:28:25 LetoThe2nd : Yeah sorry. Apr 26 13:28:48 OK I think I have to create my own layer now :-$ Apr 26 13:29:23 hamdyaea : It's very easy and better Apr 26 13:30:18 hamdyaea : be used of doing patch and bbappend and save it into your layer so that you know the changes you made from layer already made. Apr 26 13:32:17 ChrysD: Yes I try it now Apr 26 13:32:43 hamdyaea : changed my life tbh ahah Apr 26 13:35:10 I have a question because of the git fetcher, if the git fetcher has downloaded the current state and i update my repo. When I get this updates on bitbake? Apr 26 13:35:35 Hunk_: when ever you set a new srcrev in your recipe. Apr 26 13:35:58 Hunk_: usually the recipe sets a specific branch and commit that the fetcher is meant to get. Apr 26 13:37:24 i point on a branch with now specific commit Apr 26 13:37:27 always the neweset Apr 26 13:38:23 bitbake uses whats in your recipe. so if you update the recipe with a new srcrev, then all is well Apr 26 13:39:19 but when the recipe shows to dev Apr 26 13:39:34 the recipe shows to dev? Apr 26 13:39:41 to an feature branch Apr 26 13:39:51 not specific commit Apr 26 13:40:47 SRC_URI"git://url;branch=dev;destsuffix=name \ git://url2;destsuffix=name_url2" Apr 26 13:40:56 no you got me wrong. the key point is that the recipe gets modified. if the recipe is not modified, bitbake detects it is still the same and hence says "ok, i can recycle what i have." Apr 26 13:41:17 in short: just pointing to branches does not work right and predictably. Apr 26 13:41:48 hm ok Apr 26 13:42:00 i have to point to a specific commit or tag? Apr 26 13:42:31 either manually set a srcrev (the good and reproductible way) or use AUTOREV (read up on it) for the time of development, and get rid of it as quick as possible again, as it breaks reproductible builds and tries to fetch every time. Apr 26 13:42:53 for a recipe to be sane, it basically has to point to a specific tag/commit, yes. Apr 26 13:43:05 I'm using AUTOREV Apr 26 13:44:13 *sigh* thanks for mentioning that once i spent ten minutes explaining the opposite. Apr 26 13:44:50 then i suggest re-reading http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#platdev-appdev-srcrev Apr 26 13:45:09 and verifiying with bitbake -e that the autorev actually gets used. Apr 26 13:48:27 If you do not want to update your local configuration file, you can add the following directly to the recipe to finish enabling the feature: SRCREV = "${AUTOREV}" Apr 26 13:49:02 exactly. Apr 26 13:49:15 I did this ... I have the problem that under download/git2/my_repo is my repo and this repo isn#t updated Apr 26 13:49:25 because this repo is downloaded a while ago Apr 26 13:49:40 15:45 < LetoThe2nd> and verifiying with bitbake -e that the autorev actually gets used. Apr 26 13:51:38 @letoThe2nd Apr 26 13:51:55 ok the problem ist that I use multiple SCMs Apr 26 13:53:50 a recipe that pulls from multiple upstream repos? Apr 26 13:57:38 rburton. I'm at a loss as to how that missing tools error popped up for you. Apr 26 13:57:56 I changed do_validate_branches[depends] = "kern-tools-native:do_populate_sysroot" Apr 26 13:58:10 and the meta data task has: addtask kernel_metadata after do_validate_branches Apr 26 13:58:14 * zeddiii is confused Apr 26 14:13:56 zeddii, can't help on this issue but I have a pointer for the other one on the ML (savedefconfig) Apr 26 14:15:43 I've had a bug for it for ages, which I 95% implemented and then didn't send out for testing, so I'm waiting until the next cycle now. Apr 26 14:16:02 but if you have a layer and kernel + defconfig I can test against, that is helpful for sure. Apr 26 14:16:05 could somebody help me with net-snmp with python modules? Apr 26 14:16:12 zeddii, https://tinyurl.com/mzm626j Apr 26 14:16:14 I cannot build it at all :-( Apr 26 14:16:21 we already were there Apr 26 14:16:46 ah yes :D Apr 26 14:17:26 I will use that. thanks. Apr 26 14:17:44 I hope it does still apply Apr 26 14:18:03 I also did this patch last september: Apr 26 14:18:05 https://pastebin.com/ Apr 26 14:18:14 and clearly have sat on it for months :D Apr 26 14:18:52 might as well convince people to use emacs or vi, if they can agree on anything kernel related :P Apr 26 14:19:34 heh Apr 26 14:20:19 zeddii: that link needs work Apr 26 14:20:32 pff Apr 26 14:21:23 details Apr 26 14:21:25 :P Apr 26 14:21:30 https://pastebin.com/2yQD8zLe Apr 26 14:21:38 * zeddii wonders what ate the / Apr 26 14:21:54 who says doing 6 things at once causes mistakes! ;) Apr 26 14:57:44 nrossi: So. I have created a layer which is (for now, to get me started) a stripped down+modified version of meta-xilinx, and now I want to add some kernel configuration in order to get Netfilter and the rest of it. recipes-kernel is untouched (except for those modifications I'm currently unsuccessfully trying to do). Here's a short writeup: https://gist.github.com/johannfr/248ee0f8e3ac22affa653c652fa40fa0 Apr 26 14:58:59 nrossi: I'm not getting any errors, but those modules/features are not getting included in my kernel. It not in my modules-.tgz either. Apr 26 14:59:34 Auto-installing Xilinx. Brave man. Apr 26 14:59:59 nrossi: Do you see what I'm doing wrong? Like, is there something stupid/PEBKACy jumping out at you? Apr 26 15:01:39 JoiF: do you want them as modules or built-in? Apr 26 15:01:54 preferably built in Apr 26 15:02:16 JoiF: also, .scc files aren't for "CONFIG_*" thats the .cfg files Apr 26 15:03:54 JoiF: e.g. the linux-yocto ones for netfilter -> http://git.yoctoproject.org/cgit/cgit.cgi/yocto-kernel-cache/tree/features/netfilter/netfilter.scc http://git.yoctoproject.org/cgit/cgit.cgi/yocto-kernel-cache/tree/features/netfilter/netfilter.cfg Apr 26 15:04:13 nrossi: Now that you mention it, I can see that now.. :p I was sure I was reading some .scc files with CONFIG_ in them, but apparently my brain has farted here. Apr 26 15:12:44 nrossi: Woohoo, it works! Thanks! :D Apr 26 15:17:20 nrossi: Also, I completely mis-interpreted "non-hardware" in the config. What I thought this meant was that this would only be built for "non-hardware" images, i.e. qemu and such. But now that I think about it again, that assumption didn't make any sense at all. Apr 26 18:01:36 Does anyone have any experience getting realtek8188 wifi dongles working? Apr 26 20:48:29 how can such a difference in whitespce screw-up the hash ? : Apr 26 20:48:29 bitbake-diffsigs ./x86-64/universal-4.9/6e/sstate:quilt-native:x86_64-linux:0.64:r0:x86_64:3:6e6c86f1c00d85d0deddd6b325a5d331_configure.tgz.siginfo ./x86-64/universal-4.9/2d/sstate:quilt-native:x86_64-linux:0.64:r0:x86_64:3:2dc1b3bd198baf22f595a0616315ab62_configure.tgz.siginfo Apr 26 20:48:29 basehash changed from f75ad8790c8372b85a5754c08194f48a to 75a36a448e9b4970817512193259563c Apr 26 20:48:29 Variable DEPENDS value changed from '${@autotools_dep_prepend(d)} ${BASEDEPENDS} ' to '${@autotools_dep_prepend(d)} ${BASEDEPENDS}' Apr 26 20:48:46 (whitespace after {BASEDEPEND} ) Apr 26 21:02:42 dl9pf: whitespace is considered material unfortunately... the question is how did it get in there? Apr 26 21:03:09 or perhaps more importantly how did it end up changing? Apr 26 21:09:32 yes, thats the question Apr 26 21:46:22 how can I narrow that down ? Apr 26 21:47:18 find 2 build with such an effect and look at the bitbake -e quilt-native | grep ^DEPENDS ? Apr 26 21:55:56 found another one ... bitbake-diffsigs `find ~/.yocto/sstate-cache/defaulttune/ -name *texinfo-dummy-native*configure*siginfo` Apr 26 21:55:56 basehash changed from 3acfe1b0fadbb7489c5f15229c48d12f to 3f8689b3dcf5d168a107271c1d4d72d4 Apr 26 21:55:56 Variable DEPENDS value changed from '${BASEDEPENDS}' to '${BASEDEPENDS} ' Apr 26 22:06:03 halstead: can't seem to acccess the new ab? Apr 26 22:07:11 rburton: Firewall upgrade issue. Working on it now. Apr 26 22:07:28 cool Apr 26 22:08:56 rburton: sorry about the delay. it was supposed to be 0 downtime. Apr 26 22:11:44 halstead: np, bed time now anyway :) Apr 26 22:51:05 Firewall issues resolved. Apr 26 22:51:40 \o/ Apr 26 22:53:17 Thanks clsulliv :) Apr 26 23:13:58 Is anyone here familiar with mender? Apr 26 23:14:07 i'm having trouble with it on a raspberrypi3 Apr 26 23:14:10 specifically uboot Apr 27 01:08:31 my pi 3 running yocto with mender keeps stopping at the U-Boot prompt even though no key is pressed Apr 27 01:08:34 it's driving me crazy Apr 27 01:08:41 I can't seem to skip the prompt Apr 27 01:10:03 I've set 'setenv bootdelay 0' in boot.cmd Apr 27 01:10:05 no dice Apr 27 01:24:31 :( Apr 27 01:26:29 is something hooked up to the console uart? Apr 27 01:27:50 i would try disconnecting the rx pin on the pi if so. Apr 27 01:28:09 m4t: I'm pretty new to this. There aren't any pins connected on the pi Apr 27 01:28:18 oh, not sure then. sorry. Apr 27 01:28:46 is it possible that it's trying to read from serial and it thinks a key has been pressed? Apr 27 01:29:44 m4t: Apr 27 01:30:21 well, that's what i thought could be happening but you said nothing is hooked up to the uart Apr 27 01:31:07 m4t: I am writing enable_uart=1 to config.txt Apr 27 01:32:15 not sure how that would interrupt u-boot Apr 27 01:32:20 you might try #u-boot Apr 27 01:32:22 yeah, idk Apr 27 01:32:32 #u-boot ? Apr 27 01:32:44 irc channel... Apr 27 01:32:55 k Apr 27 01:32:59 thanks Apr 27 01:33:03 np Apr 27 01:35:21 m4t: do you think bluetooth could be interfering? Apr 27 01:39:33 Here is what I'm seeing: https://pastebin.com/3ezrigzu Apr 27 01:49:47 not sure, sry Apr 27 01:57:52 ok Apr 27 01:57:59 thanks anyway **** ENDING LOGGING AT Thu Apr 27 03:00:00 2017