Commit 120f32d
fix(submodule): re-pin absolute-zero to a commit that exists
The gitlink pointed at ad085baa7d25de23fd9cf4de3e88e5896e35c708, which does not
exist in hyperpolymath/absolute-zero (GitHub API returns 422) and is absent from
the GitLab and Codeberg mirrors. It was orphaned by an upstream history rewrite
during the estate-wide Ddraig SSG rollout.
Consequences: Dependabot has failed since 2026-07-20 with "upload-pack: not our
ref ad085baa…"; any `submodules: recursive` checkout fails; the local submodule
could not be initialised; and the CNO reference implementation was unreachable.
Re-pinned to current upstream main, 87902bb770e767c10e065d9ac75d111e80a01be1.
NOTE FOR REVIEW: this is a semantic bump, not a restore. The content of the old
pin is unrecoverable, so it cannot be verified that 87902bb7 is the intended CNO
state — please confirm. Per .gitmodules, pointer bumps are a deliberate act.
Re-pinning may also not be durable on its own: the pin was orphaned by an upstream
force-push, so the next sweep can re-orphan it. This repo has already oscillated
three times (#89 convert to submodule, bef4c92 remove it, #117 restore .gitmodules).
Worth settling submodule-vs-vendor, or barring force-push on absolute-zero main.
Verified: `git submodule update --init absolute-zero` now succeeds and checks out
87902bb7 cleanly.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>1 parent a60751f commit 120f32d
1 file changed
Lines changed: 1 addition & 1 deletion
Submodule absolute-zero updated from ad085ba to 87902bb
0 commit comments