09:03 <walter> Hello, I have a question, how does one fix this error? https://bpa.st/XX6IOM5RGNWTWU76OXEDOUCJ5U -- I've checked the poky source, and the license seems to be defined correctly in `python3-markupsafe_3.0.2.bb`. So it's probably an issue on the python side, right? (thanks)

10:31 <Scorpi> walter: "error: invalid command 'egg_info'" - I guess this is the actual error

14:14 <walter> oh, good point, I didn't see that line

14:40 <walter> I got over the error. I had to install some python3 packages in my OS to fix that error. That was strange, I wasn't expecting to need some OS dependencies.

17:54 <khem> walter: which release are you on ?

17:55 <khem> sometimes versions of setuptools in use can also cause such problems

19:49 <walter> I've on the master version of poky, and I've had to run update setuptools to latest version via PIP, the OS version wasn't recent enough (even though I'm on Fedora 42, and it has recent versions of everything)

20:03 <walter> Question, what are these QA errors? Are they for reproducible builds, and can I disable them? ERROR: QA Issue: File X in package mesa-vulkan-drivers contains reference to TMPDIR [buildpaths] -- https://bpa.st/PIDTYQ5HPRX7AQT5CHG6SLM53E

20:06 <walter> I'm monkeying around with some BSP layers for a board I'll have next week, so I don't really care about reproducible builds.

20:38 <walter> Is there a way to downgrade a package? I'm looking at the `Errors and Warnings` section in the QA section of the Reference Manual, I found my error there, but it recommends finding and fixing the issue! But, I'm not experienced enough to do that, so I'm thinking that maybe downgrading the package would fix the issue.

21:21 <RP> walter: if you want to ignore it, ERROR_QA:remove = "buildpaths" will disable that check, at your own risk

21:47 <walter> thanks! that has worked, but I've ran into other issues, so I've switched away from the poky master branch and I'm now trying my luck with `walnascar`.

23:33 <khem> RP: -march=core2 works -march=x86-64-v3 shows the failure in that perl test

23:35 <khem> source code shows bit of pointer type punning and Pack/unpack uses C structs/unions and copies them as raw memory blobs. structs contain padding or partially written fields.it seems perl is assuming they will be zero initialized

23:36 <khem> I will try building with -fsanitize=memory and see if my assersion is right.

23:37 <khem> It works with clang fine and I guess its because we use -fno-strict-aliasing with clang

23:38 <khem> maybe we should add -march=core2 to cflags in perl recipe and move on