**** BEGIN LOGGING AT Tue Jul 16 03:02:17 2019
Jul 16 05:05:55 <yocti> New news from stackoverflow: How to deal with symbolic links in Yocto recipe using cmake? <https://stackoverflow.com/questions/57050146/how-to-deal-with-symbolic-links-in-yocto-recipe-using-cmake>
Jul 16 06:09:53 <Chrusel> Good morning everybody! I ask for some expertise regarding integrating Bitbake fetchers. How can I add a customized fetcher (or even a new one) in a separate layer? I have to fetch sources from a ClearCase server (-.-), and somehow, the ccrc fetcher provided by Bitbake does not fulfill the needs. Any ideas appreciated.
Jul 16 06:44:58 <__angelo> hi, preparing a submodule based repo, what is the difference from including separated oe-core (and meta) and bitbake from oe, or poky (with inside meta and bitbake) ?
Jul 16 07:11:22 <ricardocrudo> is ldconfig supposed to work on yocto?
Jul 16 07:52:32 <mcfrisk> hi, trying to update a BSP layer from sumo to master. Any ideas what is causing this: Error, the PACKAGE_ARCHS variable (all any noarch ${PACKAGE_EXTRA_ARCHS_tune-aarch64} my_machine) for DEFAULTTUNE (aarch64) does not contain TUNE_PKGARCH ().Toolchain tunings invalid:
Jul 16 07:53:12 <mcfrisk> Tuning 'aarch64' has no defined features, and cannot be used.
Jul 16 07:53:49 <mcfrisk> to me the BSP layer seems to be quite timple aarch64 one but this sanity check is failing.
Jul 16 08:27:18 <ricardocrudo> I'd like to let yocto install my files using the install rule of the makefile but also install some extra files afterwards. The do_install overwrites the install rule of the makefile. Is it possible to have a do_post_install in my recipe?
Jul 16 08:32:01 <Chrusel> ricardocrudo: does it help when you define a do_install_append() in your recipe?
Jul 16 08:52:26 <ricardocrudo> Chrusel it seems like what I been looking for
Jul 16 08:56:33 <Chrusel> ricardocrudo: (y)
Jul 16 09:26:51 <opennandra> nabokov: got reply from mainling list and it actually works (even I don't know why : )) https://lists.yoctoproject.org/pipermail/yocto/2019-July/046024.html
Jul 16 09:30:47 <rburton> because wks.in files are handled differently
Jul 16 09:36:41 <yocti> New news from stackoverflow: How to Compile SampleApp of Alexa-SDK using bitbake for RaspberryPi <https://stackoverflow.com/questions/57053869/how-to-compile-sampleapp-of-alexa-sdk-using-bitbake-for-raspberrypi>
Jul 16 09:42:52 <mous16> Morning everyone
Jul 16 09:44:48 <opennandra> rburton: ok but when don't add this specific stuff in (for /data handling) then works fine with wks
Jul 16 09:45:03 <opennandra> and there is no documentation for that either (at least I didn't find it)
Jul 16 09:45:22 <opennandra> rburton: can you please shed some light why it is different? thx
Jul 16 09:45:52 <rburton> haven't been following the thread at all, just saying that wks and wks.in are handled differently with regard to variables
Jul 16 09:46:48 <opennandra> ok I see
Jul 16 09:47:04 <opennandra> so probably if variables are not used in wks (use wks) if yes then use wks.in
Jul 16 09:47:06 <opennandra> that make sense
Jul 16 09:47:08 <opennandra> thx
Jul 16 11:38:44 <__angelo> what is the reason why poky/meta of sumo contains some recipes (bb) that are different from openembedded.core/meta (of same sumo tag) ?
Jul 16 12:07:03 <yocti> New news from stackoverflow: How to build yocto with complete gtk+ support <https://stackoverflow.com/questions/57056758/how-to-build-yocto-with-complete-gtk-support>
Jul 16 12:18:30 <Chrusel> __angelo: the higher prioritized layer wins - difficult to give you a precise answer without a specific recipe name - appending and overwriting recipes are basic concepts
Jul 16 13:07:12 <yocti> New news from stackoverflow: aureport is printing -1 instead of the user id <https://stackoverflow.com/questions/57034281/aureport-is-printing-1-instead-of-the-user-id>
Jul 16 13:09:01 <mcfrisk> What's needed to swith a yocto image to utf8 character encoding everywhere? Is it enough set  GLIBC_GENERATE_LOCALES = "en_US.UTF-8" and set IMAGE_LINGUAS = "en_us"?
Jul 16 13:35:20 <rburton> mcfrisk: for historical reasons, yocto's en-us *is* utf8
Jul 16 13:39:39 <mcfrisk> rburton: hmm, I have those in images but various tools fail to print utf8 strings correctly over openssh console and busybox shell..
Jul 16 13:40:18 <mous16> I'm trying to write a recipe for a third-party custom-makefile kernel module. I was able to make it compile with bitbake, but it compile against wrong kernel version. What am I missing?
Jul 16 13:55:36 <freyr69> I don't need meta-yocto-bsp if I use meta-intel, right?
Jul 16 14:59:35 <RP> Engineering meeting: https://zoom.us/j/990892712
Jul 16 15:03:24 <litb> hello folks
Jul 16 15:03:43 <litb> I changed my distribution config file with the PREFERRED_VERSION_linux-yocto = "4.19%"  string
Jul 16 15:03:53 <litb> before it had   "4.9%"
Jul 16 15:04:12 <litb> now when I recompile my kernel module  (i also did -c cleansstate and -c clean), it *still* uses kernel 4.9
Jul 16 15:04:16 <litb> what's going on?
Jul 16 15:05:53 <litb> using warrior and i manually copied 4.9 and 4.14 recipes over from other yocto releases
Jul 16 15:08:23 <litb> actually this makefile is really dumb. it prints "KER: 4.9.0-8-amd64" because that's the kernel of my BUILD host :( I'm assuming it's a false positive
Jul 16 15:09:38 <litb> modinfo shows it uses   vermagic:       4.19.34-yocto-standard SMP preempt mod_unload PENTIUMM   all fine
Jul 16 15:22:12 <rburton> RP: i'm too late to join the engineering call and need to disappear in a bit: did you bring up debian/suse?
Jul 16 15:33:19 <RP> rburton: I did
Jul 16 15:56:32 * kergoth yawns
Jul 16 15:57:10 * kergoth back from a 1 month vacation, road trip across the US, arizona to south carolina. 4-5 day drive one way going 7-8 hours a day... with a 4 year old. so glad to be home
Jul 16 15:59:58 <JPEW> I would never even attempt such a thing with my 4 year old
Jul 16 16:05:54 <kergoth> thank god for his tablet loaded with videos and episodes
Jul 16 16:06:02 <kergoth> and a battery pack
Jul 16 16:45:55 <tlktlk> Hi, Is there a way to change GCCVERSION only for a specific recipe?
Jul 16 16:46:29 <kergoth> not easily, no, as it controls which gcc recipe is built globally via a preferred version
Jul 16 16:47:22 <tlktlk> What about not easily :)
Jul 16 16:47:27 <tlktlk> Even if it is a hackish way
Jul 16 16:47:52 <tlktlk> I would like to try something out, and it requires me to compile a specific with a specific GCC version, while leaving the erst of the system as it is
Jul 16 17:34:30 <CoLa> is it possible to change the #!/bin/sh shebang for the run.do_install task for a single recipe? (background: I got a recipe with bash commands that do not run with Ubuntu's /bin/sh -> dash)
Jul 16 17:44:34 <aehs29> kergoth: nice, haha Im sure that was fun
Jul 16 17:49:26 <kergoth> tlktlk: you'd have to use a different build or configuration, then adjust the recipe to use the other toolchain, not the one built via its dependencies. you can't just copy the recipe, either. you'd have to either build it entirely externally and then reference it, or use multiconfig to try to pull it off
Jul 16 18:25:53 <rburton> CoLa: fix the recipe to use sh, we explicitly say the shell is posix and not bash
Jul 16 18:26:03 <rburton> if you cant then fix your sh symlink
Jul 16 19:12:53 <opennandra> hi it is possible to override bootloader in image recipe?
Jul 16 19:13:15 <opennandra> I'm building 2 mages one for storage medium and second kind of provision one
Jul 16 19:13:25 <opennandra> but would need differencies in bootlaoder
Jul 16 19:13:41 <opennandra> if it possibility to do something like that?
Jul 16 19:19:34 <kergoth> use multiple bootloader recipes, just make the second not provide virtual/bootloader
Jul 16 19:23:02 <opennandra> kergoth: I've tried that but getting then problem with deploy
Jul 16 19:23:20 <opennandra> as first bootloader deploy stuff to deploy dir same as second one
Jul 16 19:23:25 <kergoth> shouldn't be problems unless they're trying to deploy the same files
Jul 16 19:23:31 <kergoth> yeah, don' do that. override do_deploy as necessary
Jul 16 19:23:53 <opennandra> hmm
Jul 16 19:24:03 <opennandra> ok I'll try to play with that but it's nasty :)
Jul 16 19:25:04 <kergoth> not many other options, though. bitbake does *not* like multiple recipes providing the same files or packages anywhere
Jul 16 19:25:44 <opennandra> kergoth: yes I know that, thanks
Jul 16 19:26:42 <kergoth> np, good luck
Jul 16 19:27:57 <robbawebba> I'm trying to build the 5.1.2 kernel within yocto, but I'm running into an issue during do_compile: "make: *** No rule to make target 'bzImage'.  Stop.". Does anyone know if there is something different or new about the build process of the newest stable kernel? I see that the bzImage is still available as a make target in my local clone of the kernel source code, so I'm not quite sure why it's missing.
Jul 16 19:29:08 <robbawebba> One possibility could be that I'm using the Sumo branch of OE, so I'm wondering if the kernel bbclass could have been updated recently to support newer features of the kernel build system?
Jul 16 19:39:37 <kergoth> odds are there's something wrong with your source fetch or unpack steps resulting in $S or $KERNEL_SOURCE not existing, so its' running 'make' in an empty directory, where no targets exist
Jul 16 19:39:46 <kergoth> make won't complain about there being no makefile, since it has implicit rules..
Jul 16 19:52:01 <robbawebba> kergoth: okay, thanks for the input. I'll double-check the work directory to make sure the source directory is unpacked and installed correctly
Jul 16 19:53:18 <robbawebba> on a side not, and I'm not sure how familiar you are with kernel development, but I saw in the logs that a script called mkmakefile is run, and there seems to be a "GEN makefile" step in the kernel build logs. Is there some sort of makefile generation system? I've been working on a 4.9 kernel for some time now, and I don't recall seeing this before.
Jul 16 19:56:35 <kergoth> huh, sadly not familiar with that, i haven't done much kernel development the past few years
**** ENDING LOGGING AT Wed Jul 17 02:59:57 2019