Bright, the global secret would need to be injected at the correct location as part of a build step
KYep
K@j0rge that error might be transitory now that I think about it
Kwe aren't
:latest
Jshould I rerun the checks?
Krerunning them would do nothing
Kwould have to merge
Kor make some not-to-be-merged version that is purely for testing
Jwell, it's already broken so what do we have to lose?
Kyour call
Bspeaking of calls... have a second?
Ksure, I was about to get lunch and get back to real work, I'll call you in a moment
J@Kyle Gospo but if there was an error it'll be localized right?
Kwouldn't build
Btty is fine
Kworst case is the same issue you have now
Jack, merging
Bthis is failing due to no local registry, no?
Jshould we just pin the last known kernel for now?
JI can back this out.
KNot a bad way to go
KBut let's see if this works
KShould be there anyway
Blet's watch it
Jit failed already, same issue
Jdoes anybody remember how to pin a kernel?
EI think we need to fetch it from Koji and do an override replace?
Bthat's how pin/rollback of package works, yes
JI have no idea how to do this, does someone have cycles to help out?
Jlooks like if we had a newer bootupd it might have helped
Rit's probably less jank at that point to pin the entire upstream image
EI'm taking a look but more eyes would be appreciated
Jyeah that might be a better idea
Jjust pin the entire thing lol
KFiguring out why your action can't connect to the local registry in bluefin would also help
KSince bluefin should be signed with the akmod key anyway
EI'll start there
Byep @EyeCantCU, the local registry issue is what blocked signing in bluefin
EAm I looking at the wrong job right now? It looks like it was able to pull and sign the image: https://github.com/ublue-os/bluefin/actions/runs/9571552636/job/26388915091#step:12:478
Bok, well maybe @j0rge simply reverted too quickly
Bit was failing in the PR builds, but you are right, that looks correct
ELet's revert the revert?
Jlol
Jok so am i reverting the revert?
Byeah, the failures i saw on the non-reverted PR merge (getting confusing now) were not the kernel signing
Byes, i think it should work
KSee this is what happens when I go to lunch
Elmao
J