**** BEGIN LOGGING AT Wed Feb 04 02:59:58 2015 Feb 04 05:26:56 What are my networking options with systemd? Feb 04 05:33:43 I think what I can do is IMAGE_INSTALL_append += "connman connman-systemd connman-plugin-loopback connman-plugin-ethernet" Feb 04 07:46:06 In which task does linux-yocto apply patches listed in a .bbappend's SRC_URI? Feb 04 07:47:17 I'd expect that to happen in do_patch, but that doesn't seem to be done here (patch does not show up in log.do_patch, investigation in devshell shows git tree without the patch). Feb 04 07:57:39 good morning Feb 04 08:01:37 pohly: after a -c cleansstate I see the patch listed in run.do_patches patches="...". Feb 04 08:03:05 But the resulting git tree still doesn't show any signs of it. I tried both "git log" and (out of desperation) "git diff" in devshell. Shouldn't it be applied via "git am" (or something like it) and thus show up in the history? Feb 04 08:06:59 And indeed, when I modify the patch that it cannot possibly apply, do_patch succeeds despite that. So the patch really gets ignored. Feb 04 08:12:58 hi, I'm trying to compile an image based on genericx86-64, but one of my recipes complain that it can't link to libGL (linker says "cannot find -lGL"). I tried adding virtual/libgl to the DEPENDS but that doesn't help (the recipe works fine on another architecture already, since it also depends on virtual/egl). Is there something special to do to use openGL on genericx86-64? Feb 04 08:19:12 <_4urele_> Hi everyone Feb 04 08:21:26 Uhh, how subtle. The naming of the patches matters. Call the first patch file "0003-something.patch" and it gets it ignored. Call it "0001-something.patch" and it gets applied. Feb 04 08:25:26 <_4urele_> I'm trying to build an initramfs did anyone try this before Feb 04 08:25:43 <_4urele_> with poky :) Feb 04 08:31:22 _4urele_: i did and got stuck on a number of issues. if you don't have hard requirements for it, an initrd is way easier to build. Feb 04 08:33:46 hi LetoThe2nd, thx to G+ now I know your nick :-D Feb 04 08:34:41 mckoan: hrhr Feb 04 08:34:51 mckoan: you're welcome Feb 04 08:36:17 pohly: that recipe is that weird patch behaviour happening on? Feb 04 08:36:24 oh linux-yocto Feb 04 08:36:28 yep Feb 04 08:36:36 that's special in that it does its own patching, iirc Feb 04 08:37:36 Yes, there's "patchme" and "kgit-forgot-the-name" involved. Feb 04 08:37:44 <_4urele_> LetoThe2nd, I'm not a specialist, I'm don't really see the diffrence... but the goal is to check the filesystem before mounting it "rw" Feb 04 08:38:05 <_4urele_> LetoThe2nd, and have something strong and robust :) Feb 04 08:38:08 Scratch my remark about the naming. It also fails for a full patch series starting with 0001-... Feb 04 08:38:25 It did fail when I only had one such patch. Hmm. Let's double-check that. Feb 04 08:41:56 Yes, it applies a single patch. And it does show up in "git log", as expected. Now why does it not apply any patches when I list 28 out them?! Feb 04 08:42:14 morning all Feb 04 08:42:14 _4urele_: they're fundamentally different in terms of packaging and usage. so if you don't want to dive in deeper, just create an initrd. Feb 04 08:43:03 _4urele_: that is, make and imagetype tar.gz or such and pass the correct parameters to the kernel. after that, make sure you can switch over Feb 04 08:48:24 <_4urele_> LetoThe2nd, I can't find some example on how to build initrd and initramfs with yocto in refmanual Feb 04 08:48:47 <_4urele_> maybe i'm not looking at the right place Feb 04 08:51:56 _4urele_: an initrd is no different from any other rootfs. Feb 04 08:52:01 Yep, about 20 patches work, 28 don't. And no error anywhere in the log. Feb 04 08:52:06 _4urele_: its just how you tell the kernel to get started. Feb 04 08:53:04 <_4urele_> LetoThe2nd, should I create a new Image, should I add features to my current image...? Feb 04 08:53:33 <_4urele_> it is not clear to me how to build the initrd Feb 04 08:53:46 _4urele_: that depends on how your images are structured, there's no way for me to answer this. Feb 04 08:54:22 _4urele_: just one thing is always valid: an "initrd" is just a particular way to treat a standard root file system. Feb 04 08:54:49 so if you want an initrd that does stuff and then continues towards another filesystem, you will need two distinct images. Feb 04 08:57:59 <_4urele_> LetoThe2nd thanks for the advice ;) Feb 04 08:59:40 _4urele_: have fun. Feb 04 09:09:45 does yocto systemd generate logs in journals/ by default? or its probably part of my BSP ? Feb 04 09:25:04 hello, core-image-weston installs neard and ofono, but i do not want them. so i added them to IMAGE_INSTALL_remove. but how to do a re-build of the image now? Feb 04 09:26:39 ericbutters: I strongly suspect that won't make a difference because those packages aren't added directly to IMAGE_INSTALL, they come in via a packagegroup Feb 04 09:27:30 bluelightning: how to remove them? Feb 04 09:27:57 <_4urele_> ericbutters, I had the same issue, if you want to cleanup without any change in distro, you have to tweak distro feautre Feb 04 09:28:01 or how to find the packagegroup? Feb 04 09:28:45 _4urele_: i also added to DISTRO_FEATURES_remove Feb 04 09:29:08 <_4urele_> ericbutters, I have DISTRO_FEATURES_remove = "nfc apm zeroconf pcmcia bluetooth irda x11 3g wifi dbus ptest usbgadget pci nfs" in my machine.conf Feb 04 09:29:26 ah okay, i try Feb 04 09:29:52 nfc will be neard, but what is ofono here? Feb 04 09:30:06 <_4urele_> ericbutters, 3g to my mind Feb 04 09:30:09 ok Feb 04 09:31:14 _4urele_: how to re-build now? Feb 04 09:31:31 <_4urele_> ericbutters, as far a I remember it comes with a packagegroup (base if i'm right) and it is added if 3g is part of distro feature or machine feature Feb 04 09:31:58 <_4urele_> ericbutters, just "bitbake your-image" again Feb 04 09:33:18 <_4urele_> ericbutters, ofono will be removed from the filesystem but the rpm will be still available Feb 04 09:33:33 that is fine Feb 04 09:33:53 thanks Feb 04 09:34:08 you should probaly consider making your own ditro instead of using DISTRO_FEATURES_remove Feb 04 09:34:17 <_4urele_> ericbutters, you're welcome Feb 04 09:34:58 abelloni: sounds good, how? where do i find the information i need to do so? Feb 04 09:35:23 <_4urele_> abelloni, of course ;), but for now it is enough for me Feb 04 09:35:46 <_4urele_> ericbutters, there is a chapter in the yocto reference manual :) Feb 04 09:36:11 ok good :) i take a look Feb 04 09:36:48 one more thing: how to add a script to be executed on startup in etc/init.d? Feb 04 09:39:35 I usually create a recipe that inherit update-rc.d Feb 04 09:39:49 grep for that, you'll find plenty of examples Feb 04 09:40:03 ok Feb 04 09:42:41 abelloni: i find ../sources/poky/meta/recipes-core/update-rc.d/update-rc.d_0.7.bb -- could you pls explain how to inherit, so that bitbake uses my receipe? sorry but i am new to yocto :) Feb 04 09:43:07 i already use my own layer Feb 04 09:46:43 inherit update-rc.d Feb 04 09:46:46 then http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#var-INITSCRIPT_NAME Feb 04 10:30:04 How do I make sure a recipe is re-packaged when the hash of do_package of one of its RDEPENDS changes? Feb 04 10:30:31 I'm trying do_package[rdeptask] = "do_package", but it doesn't seem to work for me Feb 04 10:31:35 Use case is a packagegroup which depends on eglibc-utils, but I made a change that causes debian.bbclass to rename the package to libc-utils, so the packagegroup needs to be re-packaged with the new name Feb 04 10:47:35 rburton, bluelightning_: I figured out what the linux-yocto issue is. Suppose I have three patches in SRC_URI: a b c. If "b" is already included in the base recipe, then kgit-s2q (called via patchme -> kgit-meta) assumes that this patch and everything preceeding it is already applied, thus skipping patch a even though it is not applied. Feb 04 10:48:07 pohly: worth filing a bug for zedd Feb 04 10:48:37 I ended up in that situation because I added all potentially relevant patches to a .bbappend file. One could argue that I should have been more careful, but the same could also have happened if the SRC_URI list is okay at some point, and then later the base recipe adds one of the patches. Feb 04 10:49:02 Where do I file that bug? Feb 04 10:51:01 bugzilla.yoctoproject.org, for the kernel Feb 04 11:16:38 https://bugzilla.yoctoproject.org/show_bug.cgi?id=7282 "linux-yocto: accidentally skips patches" Feb 04 11:16:40 Bug 7282: normal, Undecided, ---, dvhart, NEW , linux-yocto: accidentally skips patches Feb 04 11:16:58 oh, thanks yocti :-) Feb 04 11:17:08 Yes, I know that this is a robot. Feb 04 11:18:01 I just like to be friendly to machines, lest they get angry/sullen/whatever and start to misbehave... Feb 04 11:28:37 yocti's no fun, with ibot at least it talked back to you Feb 04 11:30:09 hey.. i have a binary that i want to be installed to usr/bin -- i have a receipe but i dont know how to install the binary Feb 04 11:39:21 ericbutters: do_install() containing: Feb 04 11:39:25 install -d ${D}${bindir} Feb 04 11:39:27 install -m 0755 filename ${D}${bindir}/filename Feb 04 12:03:19 bluelightning: i do: http://paste.ubuntu.com/10052391/ and get error: http://paste.ubuntu.com/10052413/ Feb 04 12:04:55 without install -m 0755 ${WORKDIR}/mini-cis ${D}${bindir}/mini-cis it works Feb 04 12:11:59 ericbutters: hmm, that is definitely a bug Feb 04 12:12:35 ericbutters: you may be able to work around it by adding INSANE_SKIP_${PN} += "arch" Feb 04 12:16:28 bluelightning: yes that worked! thanks. what is ${PN}? Feb 04 12:38:59 ericbutters: package name Feb 04 12:42:14 ls Feb 04 14:23:21 morning Feb 04 14:51:53 morning kergoth Feb 04 15:28:16 how to save/distribute local.conf and bblayers.conf ? Feb 04 15:28:42 what is best way to do? Feb 04 15:49:50 <_4urele_> ericbutters, I have no way to do this but I would advice you to look at the bblayers.conf.sample and the local.conf.sample. if you are using freescales scripts, maybe you should also take a look at the "setup-environnement" script Feb 04 15:50:35 <_4urele_> ericbutters, (no way because i'm just new as you are but I searched this way... at this moment I don't need this) Feb 04 15:53:37 _4urele_: yes, setup-environment seems to be a good place to put some copy/install commands for my own confs. thanks Feb 04 15:55:50 <_4urele_> ericbutters, but it erase some configuration if I remember well so I would say that you have to use your own local.conf.sample Feb 04 15:56:03 ericbutters: the idea is to minimise what you put in local.conf such that it doesn't really need to be distributed Feb 04 15:56:54 ericbutters: i.e. a lot of things one might add to local.conf would be better placed in a custom distro config Feb 04 18:22:55 does yacto use the 3.14 ltsi kernel? Feb 04 18:23:44 hi ulf` Feb 04 18:58:29 Does anyone know if there is a way to remove busybox from an yocto image Feb 04 18:59:12 you can disable about 90% of it successfully.. but there are a couple of things that are needed that don't appear to have replacement pieces (last time I looked) Feb 04 19:00:44 you can experiment by disabling busybox though.. Set 'PACKAGE_EXCLUDE = "busybox"' Feb 04 19:00:47 fray: I see.. thanks... have you ever tried ? Feb 04 19:00:56 that will prevent the image creation from allowing busybox to be installed Feb 04 19:01:22 I helped someone else try.. they ended up replacing the couple of remaining items and got it working, but I don't think it was anything more then a test project at the time Feb 04 19:01:48 fray: I did this.. but i got compilation error.... like it was dependency for packagegroup-core-boot Feb 04 19:01:50 i see Feb 04 19:05:44 yup.. that means packagegroup-core-boot needs it's own patch to remove the busybox dep.. Feb 04 19:05:56 there really should be a "no-busybox" way of building the system.. Feb 04 19:08:38 fray: indeed... I will see if i can remove this dep.. or try to figure out how to do it Feb 04 21:42:57 otavio: cough uboot cough :) Feb 04 21:43:45 rburton: oh my. Feb 04 21:43:57 rburton: I am hurrying up. Please excuse me. Feb 04 21:44:07 rburton: I do the test today, I promise. Feb 04 21:44:20 rburton: even if I work along during night :) Feb 04 21:44:23 otavio: no rush really, just thought i'd see if you'd tested it already. sure you're busy! Feb 04 21:44:41 rburton: I need to check the uboot code itself Feb 04 21:44:57 rburton: and alsocheck how we will 'inform' the sdcard class abot the binary Feb 05 00:19:15 tomz: any reason why IMAGE_OVERHEAD_FACTOR and IMAGE_EXTRA_SPACE are in scripts/lib/wic/utils/oe/misc.py instead of using IMAGE_OVERHEAD_FACTOR and IMAGE_ROOTFS_EXTRA_SPACE from the environment ? **** ENDING LOGGING AT Thu Feb 05 02:59:59 2015