Please stabilize dev-python/protobuf-python-4.21.12 (currently ~) It needs to be paired with protobuf 21.12 (already stable). It has existed in Tree since Dec 15th unchanged. Reproducible: Always Actual Results: The stable/~arch mismatch was causing some really weird slot operator conflicts until I figured out I needed to add a p.a.k entry for =dev-python/protobuf-python-4.21.12 - that specific version - Expected Results: emerge -avuDN @world on stable should not try to pull in new N ~ dependencies --autounmask was adding ~amd64 wildcard entries, which was pulling in a way too new version since theres others in tree, which then pulled a way too new version of protobuf just because autounmask just allowed it. Without autounmask, it was confused due to backtracking 6/20 and other packages requiring different slots, yet printing no useful output. Output finally showed with --backtrack=0.
Created attachment 884252 [details] emerge world update but slot conflict with ~amd64 protobuf / protobuf-python when I saw this I was very confused. theres about 4 more logs of me changing options to figure out what it meant.
Created attachment 884253 [details] reduced case - python-axolotl pulls protobuf-python but cant resolve mix of ~
amd64 done
ppc64 done
arm64 done
(In reply to Sam James from comment #3) > amd64 done https://forums.gentoo.org/viewtopic-p-8819415.html
Maybe not the correct bug to notice this, but is there a specific reason why dev-python/protobuf-python-4.21.9 has Python 3.10 through 3.12 enabled, but dev-python/protobuf-python-4.21.12 only has Python 3.10 and 3.11? It's causing trouble with my Portage insisting on downgrading dev-python/protobuf-python if I want to have Python 3.12 support in a few reverse dependencies..
(In reply to waldolemmer from comment #6) > (In reply to Sam James from comment #3) > > amd64 done > > https://forums.gentoo.org/viewtopic-p-8819415.html bugs should be filed separately, not in comments on arch testing bugs where they're easy to miss.
arm done
x86 done all arches done