17:43 <Ch^W> Yup... confirmed... EdgeOS is, and as far as I can tell always has been, Debian based. TIL...

17:44 <Ch^W> Given all the work they did to hack in Vyatta... it feels like an odd choice.

17:46 <Ch^W> RP: I should be able to do a deep dive on the setup tools stuff, but not until after OSSNA in June. Not sure how the fits with scheduling, but that is the best I can do at the moment.

18:32 <RP> Ch^W: I'd love to think I'd have it resolved by then but we'll see...

18:39 <khem> RP: I am seeing one reproduciblility failure with clang merge - https://autobuilder.yoctoproject.org/valkyrie/api/v2/logs/2145966/raw_inline

18:40 <khem> openmp-src is built with one of clang runtime recipes, is there an easy way for me to look into data to see what could have changed in the package between two runs ?

18:40 <khem> basically diffoscope output

18:40 <RP> khem: if you look at the logs you'll see an path with a url in it

18:40 <RP> valkyrie.yocto.io/pub/repro-fail/oe-reproducible-20250423-7uds3lh5

18:40 <RP> khem: that is diffoscope output

18:41 <khem> ah yes I see there now thanks

18:41 <RP> khem: https://valkyrie.yocto.io/pub/repro-fail/oe-reproducible-20250423-7uds3lh5/packages/diff-html/

18:42 <khem> yeah I see there is a date embedded in there with one of generated files

18:44 <RP> khem: that looks to be the issue

18:48 <khem> RP: I have another issue in nativesdk-clang to fix and then we are good with a-quick

18:56 <RP> khem: sounds good. I've not looked at those patches yet. Struggling with 5.2 :/

19:22 <tgamblin> how does one adjust the size of /var/volatile in a Yocto image? I'm building core-image-minimal and there's only 4.0K available. Is this possible to configure in local.conf?

19:45 <rburton> tgamblin: its a tmpfs isn't it?

19:46 <tgamblin> rburton: what I'm trying to do ultimately is fix the numpy ptests, but it looks like ~100 (or somewhere south of that many) ptests are trying to write to files... but since that's happening in /var/volatile/tmp, there's not enough space

19:46 <rburton> ah fun, probably following a link from /tmp

19:47 <rburton> try creating a temporary directory under cwd and redirecting TMPDIR, if it respects that

19:47 <tgamblin> rburton: thanks, will try it

19:47 <rburton> also fwiw core-image-ptest adds a lot more space to the rootfs for this reason

19:48 <rburton> and you can bitbake core-image-ptest-python3-numpy to get an image with just the numpy tests in

19:50 <tgamblin> rburton: right

20:00 <tgamblin> rburton: looks like it just ends up using 2G of storage created. core-image-ptest.bb suggests that there's a functional upper limit for the image size, so I'm not sure if specifying a much larger IMAGE_ROOTFS_EXTRA_SPACE in there for python3-numpy?

22:03 <rburton> tgamblin: only because it’s trying to fit in specific image type limits. Easily ignored if you use wic images. Surely you don’t need gigabytes?