**** BEGIN LOGGING AT Sat May 22 02:59:57 2021 May 22 10:05:09 hello May 22 10:05:29 anyo know is wrong here re2c: /home/damian/kernel/poky/build-qemu/tmp/sysroots-uninative/x86_64-linux/usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.29' not found (required by re2c) May 22 10:05:29 bootstrapping ninja... May 22 10:05:53 with fedora 34 May 22 10:06:34 re2c: /home/damian/kernel/poky/build-qemu/tmp/sysroots-uninative/x86_64-linux/usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.29' not found (required by re2c) May 22 10:06:34 bootstrapping ninja... May 22 10:06:34 Traceback (most recent call last): May 22 12:10:19 dtometzki: almost certainly a uninative issue and need to upgrade to the gcc11 compatible uninative. Patch in master and hardknott, is safe for older releases May 22 12:23:56 RP yaeh with master it runs until now May 22 13:17:32 hey, is host-user-contaminated working properly? (hardknott), I'm getting warnings like "is owned by uid 1000" on files under /home/user/* May 22 13:21:25 I got there symlink for /usr/lib/systemd/user/pulseaudio.service, I think that bitbake might be confused by that May 22 13:28:04 /home/retro/.config/systemd/user/default.target.wants/pulseaudio.service (symlink owned by 1000:1000) points to /usr/lib/systemd/user/pulseaudio.service (0:0) May 22 14:07:33 home like in your system's /home or target May 22 14:18:03 target May 22 18:35:08 moto-timo: I need different images to boot from the eMMC or the SD card, likely using different bootloaders as well. Hence having two foo-emmc and foo-sdcard machines sharing an .inc file and two multiconfigs sharing the same TMPDIR seems the way to go. I could place the bootloader variables and WKS_FILE in the multiconfig directly, but that seems ugly. May 22 18:37:00 v2d: why does it seem ugly? It is the configuration you need May 22 18:38:32 moto-timo: you would define the various UBOOT_* variables and WKS_FILE in the multiconfig? May 22 18:40:06 moto-timo: to be more precise, the eMMC uses a configured barebox. I want to build an image for the SD card (using the default u-boot) to embedded the .wic file for the eMMC image, in order to provision the system. That's why I imagined two machine configurations (one per medium) May 22 18:40:25 v2d: it’s just a config, like any other. If you would have put it in local.conf for one image… it makes sense to put it in multiconfig/foo.conf May 22 18:40:30 hi! after checking the docs, it appears the preferred way for installing on debian is from git? are there packages? May 22 18:40:54 v2d: machine is more about tuning the compiler… May 22 18:41:30 v2d: but there are multiple ways to do this. You can choose what works for you. May 22 18:42:34 moto-timo: hence my question :) Thanks for the hints. I always thought machine configuration was more about "medium and boot strategy", same harware or not. May 22 18:44:05 moto-timo: so you would keep a single machine and two multiconfigs foo-emmc and foo-sdcard, sharing the same TMPDIR, correct? May 22 18:45:39 v2d: li have never done it, but it seems like a cleaner solution when vast amounts of what you are building for the soc (the machine) are going to identical May 22 18:46:25 v2d: in practice, I would actually use different kas targets May 22 18:47:37 v2d: but it seems like a good use case for multiconfig May 22 18:49:59 moto-timo: I'd use different kas files as well, but I think I need multiconfig in order to integrate mc:foo-emmc:image.wic into my wic image for sdcard May 22 18:51:05 something that isn't well covered in OE yet regarding the multiconfig is the paths to the multiconfig built images, but having the same $MACHINE and $TMPDIR at least guarantees that for my case. May 22 20:22:16 moto-timo: sorry, run into a router issue and disconnected :/ May 22 21:54:36 sakoman RP armpit: python3-scons 3.1.2 is marked as insecure by https://pyup.io/changelogs/scons/#3.1.2 ... and 4.1.0 has a number of fixes on top of 4.0.1 ... what do we want to do about dunfell/gatesgarth/hardknott? May 22 22:25:30 moto-timo: is that some general policy for older releases or a specific issue? May 22 22:27:01 RP: just stable releases policy. Now that I know the component is marked as insecure... I should do my due diligence ;) May 22 22:28:51 RP: I'm working on a 4.0.0 recipe for minimum compliance with pyup.io and safety (https://pyup.io/safety/) May 22 22:30:07 moto-timo: upgrading master sounds good, less sure about the other releases May 22 22:30:09 I've found this slide deck about building a xen image for rpi: https://wiki.yoctoproject.org/wiki/images/d/d7/DD10_Xen_Hypervisor_NA20.pdf May 22 22:30:54 are there ready-made images available somewhere? May 22 22:31:04 RP: agreed. I just wanted to raise the flag, in case someone felt strongly. May 22 22:32:07 https://pastebin.com/scm9Bv5J May 22 23:15:23 moto-timo: mktemp isn't a huge problem and scons probably isn't on target that much anyway May 22 23:15:42 RP: good enough :) May 22 23:16:06 RP: I'll push a contrib branch in case anybody wants 4.0.0 on dunfell **** ENDING LOGGING AT Sun May 23 02:59:56 2021