05:27 <khem> teknix: yes BBB is good, rpi4/rpi5 is good as well.

06:22 <teknix> khem: i have a spare rpi3b, is that good or should i look for a BBB ?

12:45 <vmeson> khem: thanks for the perl ptest info. I think that garbage code was one of the first things I worked on over a long weekend when I started at Wind River 15 years ago!

13:53 <joekale> Not sure if this got lost yesterday when the irc server dropped momentarily. but my scarthgap based builds are failing for an expansion error in the mbedtls recipe.

13:53 <joekale> ERROR: ExpansionError during parsing /builds/oe/oe-builds/workdir/build/../01-meta-oe/meta-networking/recipes-connectivity/mbedtls/mbedtls_3.6.4.bbbb.data_smart.ExpansionError: Failure expanding variable fetcher_hashes_dummyfunc[vardepvalue], expression was ${@bb.fetch.get_hashvalue(d)} which triggered exception FetchError: Fetcher failure: Conflicting revisions (c765c831e5c2a0971410692f92f7a81d6ec65ec2 from SRCREV and v3.6.4 from the ur

13:53 <joekale> I think the SRCREV just needs removed since it specifies the tag?

14:23 <PeterM> joekale: see openembedded-devel maililng list, ton of messages reporting that scarthgap meta-networking layer is broken due to the mbedtls update

14:25 <PeterM> khem: could you take a patch for mbedtls into scarthgap branch? Waiting whole month until Armin opens a new windows for taking patches is too long for a broken layer.

14:26 <PeterM> And this should also fail the scarthgap 5.0.12 which is scheduled to be built in a week or so.

14:29 <NishanthMenon> poky/meta/recipes-connectivity/neard/neard_0.19.bb -> I am a bit confused with iwd trying to build with core-image-sato genericarm64 -> https://gist.github.com/nmenon/811b4422d4fbb435c0b58522e0a08107#file-log local.conf and kas.yml in the link confused why RRECOMMENDS is a problem..

14:39 <khem> teknix:yes pi3b is good too

14:39 <khem> teknix: follow - https://github.com/agherzan/meta-raspberrypi/blob/master/README.md

14:41 <khem> vmeson: I have a hunch the issue is in how quadmath is handled, but have had enough hours to look into it deeply

14:43 <khem> NishanthMenon: what does bitbake-getvar WIRELESS_DAEMON tell you ?

14:44 <khem> vmeson: intel compiler's testsuite was driven using perl as build/driver and I learnt to dislike it there, perhaps it was the way code was written but I never found love for perl again

14:49 <NishanthMenon> khem: iwd https://www.irccloud.com/pastebin/OBp4vRnz/

14:51 <khem> thats your issue

14:51 <khem> default is to use wpa-supplicant thats why others dont see it

14:52 <NishanthMenon> yep.. but i am bit dumb.. I wanted iwd as my daemon... apparently it exposes few more kernel security accelerator driver bugs.. so was trying to get a build with that instead of wpa-supplicant

14:54 <khem> you just need to add meta-oe to your BBLAYERS

14:54 <khem> and it will work

14:54 <khem> iwd is not in core layer

14:54 <NishanthMenon> uggh.. aah, i now understand

14:54 <NishanthMenon> thanks

14:54 <khem> np

14:55 <khem> layers.openembedded.org is alaways a good place to look for recipes that one needs and layers they come from

15:05 <NishanthMenon> khem: thanks for the hint :) I was trying my best to keep it minimalistic :D

15:58 <Crofton> Dinner after OE dev meeting updates on the wiki page: https://www.openembedded.org/wiki/OEDEM2025

15:58 <Crofton> 11:55 AM

15:58 <Crofton> You do not need to attend the meeting to go to dinner

15:59 <Crofton> @NishanthMenon enjoy this definition of minimal: https://github.com/balister/xilinx-minimal

16:00 <NishanthMenon> @Crofton https://github.com/nmenon/poky-imagination/ is what my jenkins builds and tests daily :)

16:01 <Crofton> much more minimal

16:15 <teknix> khem: thanks

17:57 <walter> The amount of information one needs to absorb before getting started is overwhelming. I've read the Introduction section, the Reference Manual, some of the BSP Developer's Guide and now I'm reading the Development Tasks Manual... and it makes me wonder, should I wait for my SBC to arrive and just get my hands dirty? Or should I continue reading? What do you folks recommend? What did you do when you got started? (Hello!)

18:09 <vmeson> walter: minimum steps are: https://docs.yoctoproject.org/brief-yoctoprojectqs/index.html, setup host, git,

18:10 <vmeson> walter: then build and boot core-image-sato (or core-image-minimal) : https://docs.yoctoproject.org/brief-yoctoprojectqs/index.html#building-your-image

18:10 <vmeson> You can learn quite a bit using qemu and if the kernel for your board can run in qemu, you can do initial work there.

18:11 <vmeson> I don't do kernel dev anymore so I'm not sure what public BSPs that aren't explictly qemu* boot using qemu.

18:12 <vmeson> walter: I've been using and working on Yocto for 10 years and I haven't learned everything yet so best to pick a goal and get started.

18:20 <walter> Thanks for the reply! I've got a core-image-sato running with GUI, for both x86_64 and arm64. I've found public BSPs for the board that I've ordererd (an RK3399 powered SBC from Raxda) and I'm thinking of trying to build those, but they probably won't run in qemu.

18:43 <vmeson> walter: Cool. Have you seen / examined: https://layers.openembedded.org/layerindex/branch/master/machines/?q=RK3399&search=1

18:45 <walter> yes, I've already cloned the layer, and I'm trying to use it, but it looks like I need to upgrade from Walnascar to Whinlatter, but I'm wondering, do yocto folks mostly run on "HEAD" instead of actual releases?

18:47 <vmeson> walter: most users building a product pick a release branch. If you're doing development for oe-core, then master. Not sure about the BSP commuity, but my impression is that it's mostly release branches.

18:50 <walter> I see. I'm asking because the branch "whinlatter" doesn't exist in poky, and that layer seems to require it, I'm thinking I need to go with the `master` branch for this to work. I'm now giving it a try.

18:53 <vmeson> Yep. whinlatter is not yet released so master ~= whinlatter now.

18:53 <vmeson> https://wiki.yoctoproject.org/wiki/Releases

19:03 <walter> It seems to work, it's running the tasks for the newly added machine.

19:07 <vmeson> walter: super!