**** BEGIN LOGGING AT Thu May 28 03:00:05 2020 May 28 07:57:05 khem thanks, but ill had to add this meta layer manually i guess. Did anyone try dbus-cxx ? May 28 07:58:37 Guest5214: just use sdbus directly. thats way easier than jumping through random hoops only for the sake of ticking a "its a c++-bindung!!!"-checkbox May 28 07:59:57 Letothe2nd In the end I will need to add this layer by myself... I really like the sdbus binding but i really miss some documentation... May 28 08:00:20 Guest5214: sounds like you are mixing up layers and recipes. May 28 08:00:24 I would like to have a good abstraction for that Data itself. This is really painfull May 28 08:00:40 Letothe2nd most likely. I mix this up all the time May 28 08:01:18 A Messagedata Object. But anyway. do you know how i can add a compiled message as argument to a method_call? May 28 08:07:33 nope. May 28 08:18:16 Letothe2nd i will just use it as it is May 28 08:21:08 do you have any advice for me if i want to buy a cheap device which is supported by mainline yocto (without DSP?) Like Raspberry Pi for example... But i would like to use internal flash May 28 08:22:40 beaglebone black, probably. May 28 08:22:48 if you really mean "DSP" May 28 08:22:56 without a "BSP", the answer is "none" May 28 08:23:20 :D Of course i meant BSP. May 28 08:23:32 then "none" May 28 08:23:50 maybe "qemu" May 28 08:24:16 i never actually did DSP on a DSP . I always did that stuff on FPGAs May 28 08:25:29 Hm qemu is nice. But why i like 'embedded' more then 'backend stuff' is that you can actually see LEDs blinking and control real world stuff May 28 08:26:38 then you'll want a BSP May 28 08:27:05 or i need :) Iam not a big fan of that NXP BSP May 28 08:27:25 and yes, i know, watching a video is complicated for you, as it might take an hour *gasp*, but here you go: https://youtu.be/MKpLQ2bGsAk May 28 08:28:12 Letothe2nd sometimes i watch your videos. But not all of them ;) May 28 08:28:45 maybe iam too old, iam more into reading documentations May 28 08:29:20 if you say so. May 28 08:29:55 Oh but maybe not the "mega manual" May 28 08:34:09 it might be just me, but i have seriously increasing problems to take your questions seriously. you know, it all doesn't fit. you hide behind a guest nick, yet you show up each day. you try to use things that are certainly "professional level" (like rauc), then you mess up simple namings and concepts, yet you claim you like reading documentation. and ever so often you just vanish without giving a May 28 08:34:15 response if the support you received was ... May 28 08:34:18 ... helpful, or not. May 28 08:34:46 i might be exaggerating a bit, but thats the impression i got over the last couple of days. May 28 08:36:09 don't get me wrong - if i see a question i can answer right from the top of my head, i will do so. but my ambitions to go beyond that are not very big at the moment. May 28 08:39:29 ok, iam sorry if you feel like this. I will try to do better in the future. You really helped me a lot May 28 08:39:57 :) just speaking my mind. May 28 11:29:54 <_jose-> Hey, is "inherit bluetooth" supported with dunfell May 28 11:30:22 <_jose-> I getting error "Could not inherit file classes/bluetooth.bbclass " May 28 11:30:51 <_jose-> building from dunfell branch May 28 14:17:48 hello, i am trying to debug a Makefile-base recipe on target with gdb. I can set breakpoints and print variables but can't show the source code (gdb-cmd: list ). I tried to install the dev and dbg package May 28 14:18:09 my question is also at stackoverflow: https://stackoverflow.com/questions/62066383/debugging-not-showing-source-files May 28 14:20:51 chris48: wild guess but incorrect makefile which is overriding variables passed by Yocto? Check that you're using lose assignments (?=) in it May 28 14:34:09 ERROR: linux-yocto-dev-5.0-rc++gitAUTOINC+b9a0187fd9_bf089a53f4-r0 do_populate_lic: QA Issue: linux-yocto-dev: The LIC_FILES_CHKSUM does not match for file://COPYING;md5=bbea815ee2795b2f4230826c0c6b8814 May 28 14:34:17 any idea why I get this error? May 28 14:35:02 which branch ? There was a licence change upstream, and most recipes have been updated for it. But what you are using propbably has not. May 28 14:35:09 warrior May 28 14:35:30 qschulz: it is a simple makefile, i think nothing special. What could i override? May 28 14:35:33 short term. just update the recipe with the checksum that is being reported. May 28 14:36:10 otherwise, we'd have to prepare a patch and send it to warrior. May 28 14:36:54 RP: you missed staging the go 1.14.3 upgrade patch May 28 14:41:34 chris48: CC, CFLAGS, LDFLAGS, etc... May 28 14:41:38 pepone: why using 5.0-rc? May 28 14:46:11 We'll be starting in 15 minutes - time to find a drink and get comfy. May 28 14:46:29 qschulz indeed they are explicit used ... but is this the problem why the source files are not coppied when installing dbk packages? May 28 14:46:41 https://www.twitch.tv/letoatreidesthe2nd May 28 14:49:20 chris48: if you ignore the flags passed by Yocto, you're going to struggle with your packages for no reason. May 28 14:49:30 (flags/variables/whatever) May 28 14:50:20 ok, i will try a simple hello makefile May 28 14:50:53 chris48: or as Letothe2nd would say, just use Cmake :) May 28 14:52:25 chris48: just replace the = with ?= for most if not all variables May 28 14:52:29 tgamblin: can you take a look at the meta-python patches on master-next have you looked at all of them ? May 28 14:53:13 chris48: and if that still does not work, maybe worth sharing the makefile so we could check that one and be sure it's not the issue May 28 14:53:54 pepone: why using 5.0-rc? did you forgot to switch from linux-yocto-dev to linux-yocto? May 28 14:53:56 Letothe2nd: welcome back on Twitch :) /me cheers without a drink May 28 14:54:10 qschulz: :) May 28 14:54:32 khem: there should be an ack for each of those May 28 14:54:34 JPEW: `runqemu qemuarm64-secureboot` should work, right? May 28 14:55:25 JPEW: it appears to just sit in BL2 and never boots the kernel May 28 14:55:54 I have a problem with some Git repositories where the location which hosts the repos doesn't let clients through that don't add the '.git' suffix to the URL. Can I somehow tell yocto to prepend '.git' to the URL when cloning from a specific host? May 28 14:56:50 reminder, we'll be live on project setup in 5 minutes: https://www.twitch.tv/letoatreidesthe2nd May 28 14:57:29 qschulz: or just get rid of handwritten makefiles once and for all! May 28 15:01:27 Letothe2nd: 16:50 qschulz| chris48: or as Letothe2nd would say, just use Cmake :) May 28 15:02:48 qschulz: :) May 28 15:04:51 hello gents, I was wondering if there is a best practice to run unit tests for packages built into an image? May 28 15:06:14 qschulz step by step. Cmake is an option and we will work with them in the future May 28 15:06:53 nameclash: throwing names but not knowing what it is... check ptest May 28 15:07:06 JaMa: Thanks seems like that was it May 28 15:07:27 qschulz: thx, will look into it May 28 15:11:57 Letothe2nd: that bottle did not look like beer. Not a real stream. /me looking for a reimbursement form May 28 15:12:56 tgamblin: https://patchwork.openembedded.org/patch/172829/ https://patchwork.openembedded.org/patch/172858/ May 28 15:13:38 khem: ah, missed those. Will build shortly May 28 15:15:11 ok May 28 15:19:12 jonmason: It should work May 28 15:19:55 JPEW: I see from your emails that it works for you. I'm just wondering what I could be doing where its not working for me May 28 15:20:12 jonmason: Try setting TFA_DEBUG = "1" in local.conf May 28 15:20:17 See what is says May 28 15:20:23 sure May 28 15:20:58 jonmason: I usually run `runqemu nographic slirp serialstudio` FWIW May 28 15:21:34 Letothe2nd: I am. I heard the pop of the cap :D May 28 15:21:40 JPEW: I tried "MACHINE=qemuarm64-secureboot runqemu nographic serialstdio slirp" and saw the same issue May 28 15:29:56 jonmason: https://github.com/JPEWdev/meta-arm/tree/qemu-secureboot May 28 15:31:22 JPEW: I pushed https://github.com/jonmason/meta-arm/tree/qemu-optee May 28 15:32:00 and they look to be identical May 28 15:32:12 JPEW: are you on dunfell or master branch? May 28 15:33:16 jonmason: master May 28 15:34:01 me too... May 28 15:34:18 It looks like it can't load u-boot for some reason. May 28 15:34:43 where in the target should i find source files when installed correct May 28 15:35:16 JPEW: whats the SHA for your working master. May 28 15:35:41 Currently using 392ba004f42be2fd9dcf121d3dfa5b414ff83f93 of poky May 28 15:37:17 jonmason: BL2 actually boots u-boot, not the kernel. Do you happen to have something that might interfere with that? May 28 15:40:51 JPEW: this is my "build everything from scratch" script. The only odd thing is using SSTATE instead of building it 100% from scratch. I'll try removing that next May 28 15:44:13 jonmason: I just built on your meta-arm commit and it worked :( weird May 28 15:44:27 Letothe2nd: cheers. Docker please :) May 28 15:44:33 JPEW: it's possible my debian env is just crap May 28 15:45:12 but 32 threads building in parallel is just awesome. Super happy with my new-ish PC May 28 15:56:01 Letothe2nd: all hail the beard. May 28 15:59:56 can somebody point me to an example of how to debug a Makefile-based application? May 28 16:02:27 chris48: that's a really vague question. May 28 16:02:31 debug what? May 28 16:02:35 what's the behavior? May 28 16:03:07 i am not able to debug a simple hello(C-code) world example May 28 16:04:46 i start gdb with: `gdb /usr/bin/hello` -> OK May 28 16:05:00 (gdb) b main May 28 16:05:09 (gdb) r May 28 16:05:21 (gdb) list May 28 16:05:32 1 longjmp.c: No such file or directory. May 28 16:06:03 but at least i expected to pause a the main function of my hello-world function May 28 16:09:27 chris48: have you tried 'c' instead of 'r' ? May 28 16:11:01 ? i think i need run to at least start the program. May 28 16:15:20 chris48: yes May 28 16:15:36 chris48: would be better if you explain mode details May 28 16:16:25 i have a makefile based c-helloworld application and i try to debug it on the target May 28 16:17:10 but i can't see the source code in the debugger May 28 16:17:22 chris48: https://pastebin.com/mYEuVAZq May 28 16:18:45 can you list the source code? `(gdb) list 3` May 28 16:22:05 chris48: yes May 28 16:22:20 chris48: please pastebin your sources May 28 16:22:38 ok, give me a second May 28 16:22:42 chris48: I guess you have a problem in the makefile May 28 16:25:04 https://pastebin.com/7XbuhfRj May 28 16:28:25 chris48: you are not using the Makefile because you are calling CC in do_compile May 28 16:28:53 *facepalm* May 28 16:29:26 chris48: have a look at this https://github.com/koansoftware/meta-training/tree/master/recipes-example/example May 28 16:30:08 chris48: (and missing Makefile in SRC_URI as well) May 28 16:32:11 chris48: however your executable should be ok even without makefile May 28 16:34:26 so what do you think went wrong? May 28 16:35:00 missing CFLAGS? May 28 16:36:10 chris48: I don't know I never debug directly on the target, I cross debug May 28 16:37:21 it worked with cross-dbg but i hat to explicit set the source file path with the `dir`-command. May 28 16:37:21 chris48: what's the output of gdb --configuration May 28 16:40:20 https://pastebin.com/QpMyQ279 May 28 16:41:47 with the Makefile it seems to be working better and i am where i stopped with my real application :) The gdb can't find the source code: https://pastebin.com/2g9QXbRS May 28 16:45:58 chris48: so copy hello.c on the target May 28 16:47:14 this is the "yocto"-solution? May 28 16:50:35 chris48: I don't understand May 28 16:51:42 chris48: your problem is related to gdb usage, not Yocto May 28 16:53:35 hmm okay, thought yocto is providing an automatic solution. Thx for your time May 28 18:39:58 bluelightning: around? May 28 18:41:20 bluelightning: can you check 6 devtool changes at the top of https://git.openembedded.org/openembedded-core-contrib/log/?h=jansa/devtool (especially the revert of your change)? I don't use devtool much and probably cannot imaging all use-cases people use it for, still waiting for whole selftest to finish locally, then I'll send it to ML for review May 28 18:41:39 kanavin_home: ^ please check them as well May 28 19:35:59 JaMa: these look ok to me but I am not really qualified :) the real stress test is when AUH will operate on the entire oe-core with these patches added in a few weeks - I expect that that it at least won't cause upgrade failures where previously they would succeed May 28 20:01:11 JPEW: I have it working on an Ubuntu VM. So, another weird Debian issue :( Anyway, I'll ack the patches and pull them in. Thanks for all this work! May 28 20:07:00 kanavin_home: ok, thanks May 28 20:23:04 jonmason: Good. Thanks! May 28 20:33:51 JaMa: will do May 28 21:00:29 hi, trying to update from sumo to thud right now and I am getting "automake: not found" from some recipes. Does anyone have any idea what this could be? Using 18.04 as build system May 28 21:09:12 ulfang___: Are these recipes in your own layers? You may need to add automake-native to the dependencies May 28 21:14:30 yes, they are from openembedded-core actually, I have tried a clean build but still get the problem May 28 21:18:47 ulfang___: I need to go so won't be able to look myself but the best way to get assistance is to post the bitbake command and the full bitbake output to a pastebin and link that here May 28 21:31:39 paulbarker Thanks, here the pastbin if anyone can have a look https://pastebin.com/PnCTbSyc May 28 21:55:44 RP: looks like my AB run came out green. will send that kernel pull request later tonight, so it’ll be around in your morning tomorrow. May 28 21:57:07 zeddii: thanks! May 28 22:10:44 ulfang___: Looks like you're missing automake from your host OS. Make sure you've got all the packages listed in the quick start document installed May 28 22:23:26 I'm seeing repo-hooks script not executing issue as part of repo sync when ran locally on my VM. Any idea/pointers? May 28 23:46:22 Hello May 28 23:47:28 I'm trying to build an out of tree driver which is conflicting with an in-kernel driver and I can't seem to figure out how linux-yocto builds it's feature config files. May 28 23:48:13 I have a recipe for an out of tree i40e network driver, but it has conflicting files with kernel-module-i40e, naturally, and I'm trying to disable that from the kernel config. May 28 23:49:48 I see a file, bsp/intel-x86/intel-x86.scc, in the build tree which has the line "include features/i40e/i40e.scc" and I think I just need to remove that somehow to avoid the in-kernel i40e from being built, but that file seems to be auto-generated and I can't find the mechanism that controls it. May 29 00:27:57 OK, it looks like it's not autogenerated, but comes from the yocto-kernel-cache repo AKA kernel-meta, however, I still haven't found if there's a mechanism to just disable that one feature fragment. May 29 00:28:21 It looks like it might be on automatically for all intel-x86 builds. May 29 00:58:46 Maybe the best approach is to add a patch to linux-yocto that patches out that line from the kernel-meta repo. May 29 01:59:38 RP: seeinf https://errors.yoctoproject.org/Errors/Build/103770/ with master-next I wonder if its related to "remove unused DEPLOY_DIR_TOOLS variable from bitbake.conf" May 29 02:20:42 sent a patch to fix it in dfu-util-native **** ENDING LOGGING AT Fri May 29 02:59:57 2020