**** BEGIN LOGGING AT Mon Dec 10 03:00:00 2018 Dec 10 09:09:32 aehs29: because the other patch didn't arrive in y inbox until about a day after the first one Dec 10 10:23:05 how can generate an SDK with extra scripts under "environment-setup.d"? Dec 10 11:03:31 mrpelotazo: add a recipe which installs some and add it to the SDK ? Dec 10 11:26:40 RP: should I then install append to the nativesdk class? Dec 10 11:28:42 mrpelotazo: it would be a nativesdk recipe so you'd inherit that class, then add it to list of things to install in the SDK Dec 10 11:34:13 My custom board has separation anxiety. Every now and then, when I stand up to get coffee, it crashes. :p Dec 10 11:35:06 (I think there's a loose connection somewhere and the vibration when I stand up is causing this.) Dec 10 11:40:32 jofr: i orked on several boards that showed that behaviour when jtag conencted. totally reproductibly, rebooted when probe was on and i got up from my chair. Dec 10 11:40:44 so, watch out for ESD related issues. Dec 10 11:41:18 LetoThe2nd: Haha, I will ;) Dec 10 11:57:48 hi guys! i have a strange problem, i am trying to implement bluetooth in my yocto build on my product and it seems to work, but when I try to connect and do something it starts to spam "SMP security requested but not available" endlessly Dec 10 11:58:06 I tried google but cannot find anything helpful Dec 10 12:01:25 hello, I would like to discuss this commit in the toybox recipe: http://cgit.openembedded.org/meta-openembedded/commit/meta-oe/recipes-core/toybox/toybox/OE-path-changes.patch?id=cdd3e9ab99d4ffda673b564ba802b6bd2d40eabf Dec 10 12:01:38 It mentions "OE conventions" Dec 10 12:02:10 where can I read on what motivated OE to use certain paths for certain system tools? Dec 10 12:02:25 is this based on some standard? Dec 10 12:02:42 or on how a certain other distro does this? Dec 10 12:03:50 I am trying to understand whether toybox or OE paths are "more correct" Dec 10 12:12:49 eduardas_m: arguably, neither are more correct Dec 10 12:13:13 that just moves toybox so it matches other paths in oe Dec 10 12:13:42 (mostly because there was a bug in update-alternatives until about a week ago where it didn't like alternatives for commands being in different directories Dec 10 12:15:28 rburton: so now the patch is no longer necessary? Dec 10 12:17:08 if you're using thud, maybe not Dec 10 12:17:26 arguably there's still a good rationale to move them to match Dec 10 12:18:53 rburton: does OE strictly follow FHS specs from Linux foundation? http://refspecs.linuxfoundation.org/FHS_3.0/index.html Dec 10 12:19:02 as much as any other distro does Dec 10 12:19:27 note that eg debian and fedora will have some utilities in different paths Dec 10 12:21:42 Dec 10 12:27:02 Hey all, what's the crack with resurrecting older BSPs? Ive got a few from 1.3,1.4,1.5 that I've inherited but wont build on my ubuntu 16.04 compile machines. Whats the recommended approach? 1) VM 2) additional older build machine 3) extra layer with required fixes in 4) Attempt to bring BSP & tools up to later version 5) something else... Dec 10 12:33:29 no_such_user: I'd say 4... bringing to latest version if possible Dec 10 12:33:50 no_such_user: depends why you're building with the old BSP Dec 10 12:33:55 is this "customer has a bug, i need to fix it? Dec 10 12:33:59 in that case (1) Dec 10 12:34:27 otherwise (4) Dec 10 12:48:57 rburton: They're older boards of ours - no ones built them for a while but I've had ownership passed to me after previous maintainer has left. Initially I just need to know I can build images that work (as documentation is thin) but may need to make some mods. Dec 10 12:49:27 well if mods are for existing customers then they'll want 1, new customers 4 Dec 10 12:52:11 rburton: Am I right in thinking 4) might take a fair bit of effort potentially? I think they're using older meta-ti / meta-freescale layers at their core Dec 10 12:54:23 (It also looks like we've got some yocto 1.1 builds as well but hoping they're forgettable!) Dec 10 13:13:11 no_such_user: yes Dec 10 13:13:35 depends if its just a hardware bsp, or more Dec 10 13:44:35 rburton: I've written on the toybox mailing list about the path patch: http://lists.landley.net/pipermail/toybox-landley.net/2018-December/009969.html Dec 10 13:45:28 rburton: It appears people from OE were trying to get changes for toybox in for this back in 2015 Dec 10 13:45:43 strangely enough busybox does not need any such fixes Dec 10 13:46:19 so perhaps it would make sense to suggest Rob Landley for toybox to act as a busybox replacement and install stuff where busybox does? Dec 10 13:48:02 if toybox install paths are not changed, I feel like the patch in meta-openembedded should be extended to cover all of the tools as it certainly does not cover all of them at this moment Dec 10 13:51:10 for example, the patch does not currently fix the dd install path Dec 10 13:52:37 New news from stackoverflow: Conflicts between python3-azure packages Dec 10 14:15:29 Hi, can anyone tell me how to know while file in the downloads directory a recipe is looking for, so that I place the zip I've downloaded in there manually? Dec 10 14:26:06 if its a tarball then just put it in there and touch a .done file too Dec 10 14:27:43 I tried that, but it's not finding it for some reason. Dec 10 15:15:20 Hello, I am trying to build python 3.7.1 using the thud branches of oe any poky. I have tried using `devtool upgrade python3` but I'm getting some merge conflicts and errors. I see that there are patches here `https://patchwork.openembedded.org/series/14778/#` Does devtool upgrade apply those patches. I'm pretty new to Yocto and wondering what the best process is to upgrade Python? Dec 10 15:38:17 dresserd: yes devtool will try to apply them. upgrading python isn't easy though, there are patches on the list for 3.7.1 already that need more wor Dec 10 15:38:17 work Dec 10 15:54:35 rburton, hi Dec 10 15:54:53 today strange failure with qemumipsel Dec 10 15:55:24 ERROR: linux-yocto-4.18.14+gitAUTOINC+bf98e195a4_465252ba0e-r0 do_package_qa: QA Issue: Endiannes did not match Dec 10 15:55:31 for kernel modules Dec 10 15:55:47 doh, I built with an older chackout a few days ago... Dec 10 15:59:11 ant_home: if you can replicate then looking at what was built would be good Dec 10 15:59:17 file is sufficient to poke at the binaries Dec 10 16:00:57 ok, btw +DEFAULTTUNE = "mips32r2el" Dec 10 16:02:36 cough patches welcome Dec 10 16:02:40 the ko ELF 32-bit MSB relocatable, MIPS, MIPS32 version 1 (SYSV) Dec 10 16:10:26 RP: haha thas weird, ok, I was just wondering in case you had trouble applying it or something Dec 10 16:10:34 hi! Dec 10 16:10:49 I've warnings with poky on sumo Dec 10 16:10:57 rburton, it's be ../kernel-build-artifacts/anubis.ko: file format elf32-tradbigmips Dec 10 16:11:03 "Intentionally failing postinstall scriptlets of ['busybox'] to defer them to first boot is deprecated. Please place them into pkg_postinst_ontarget_${PN} ()." Dec 10 16:11:15 (objdump from devshell) Dec 10 16:11:17 I think the busybox recipe isn't up to date Dec 10 16:11:45 * ant_home looking at class changes Dec 10 16:11:47 aehs29: I have no idea why they go delayed for so long :( Dec 10 16:12:00 pkg_postinst_${PN} should become pkg_postinst_ontarget_${PN} Dec 10 16:13:19 TBF its probably the smtp sever here Dec 10 16:15:55 https://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-core/busybox/busybox.inc?h=sumo Dec 10 16:24:36 * armpit ugh.. gen RSA keys very long now Dec 10 16:26:09 * armpit maybe 2 min Dec 10 17:19:58 RP: Sorry the persist_data patch wasn't against master... I forgot I still had the atfork() patch locally. Did it fix the problem? Dec 10 17:20:28 JPEW: we've not seen it again so far... :) Dec 10 19:37:49 RP: is there any way to avoid reparsing due to http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=a0db1211caf364ec69502b82e583e074431d642a Dec 10 19:38:01 RP: other than setting the variable manually on my env Dec 10 19:38:54 RP: basically, my question is if we should whitelist that or something, since we are setting it ourselves Dec 10 20:23:51 New news from stackoverflow: How to render to a second screen without being the DRM master? Dec 10 22:42:25 aehs29: I'm not sure I understand what you mean :/ Dec 10 22:56:01 Hello, I've got a quick question: Why is the `install` command preferred over the `mkdir -p` command in do_install tasks? One answer that comes to mind would be proper error handling, but are there more/other benefits? Dec 10 22:56:36 robbawebba: modes and uid/gid Dec 10 23:44:40 neverpanic: ahh makes sense. Thanks! Dec 10 23:56:17 robbawebba: file creation, copying, mode setting , directory creation all in single command Dec 10 23:56:20 makes it small Dec 10 23:56:25 and less buggy Dec 10 23:58:12 khem: nice, thanks for the info! **** ENDING LOGGING AT Tue Dec 11 03:00:01 2018