cli-guard GitHub and Forgejo mirrors have diverged lockdown sync commit Forgejo-only #36

New issue

Open

opened 2026-05-28 10:58:04 +00:00 by coilysiren · 0 comments

coilysiren commented

2026-05-28 10:58:04 +00:00

Owner

Symptom

Pushing coilysiren/cli-guard main to its two origin push URLs (GitHub + Forgejo mirror) succeeds on GitHub but is rejected non-fast-forward on Forgejo. The two remotes have diverged:

Forgejo-only: ed36b49 lockdown: sync to coily v2.45.0 [skip ci] (an automated lockdown sync) - never reached GitHub.
GitHub-only: the recent dispatch commits (e.g. surface-split, consult, surface-to-isolation inversion) - never reached Forgejo.

Impact

go get github.com/coilysiren/cli-guard@... resolves against GitHub, so the Go module path is unaffected. But the Forgejo mirror is stale and any future "force-sync GitHub -> Forgejo" would drop the automated ed36b49 lockdown commit; conversely a Forgejo-first push keeps failing.

Likely cause

An automated lockdown: sync ... [skip ci] job committed straight to Forgejo main without the same commit landing on GitHub, so the mirror is no longer a strict ancestor/descendant of GitHub.

Ask

Reconcile the two cli-guard remotes (decide whether ed36b49 should be replayed onto GitHub main, or dropped) and confirm which remote is the write source of truth for cli-guard so the mirror stops drifting. Surfaced while landing coilysiren/cli-guard#35 (consumer coilysiren/coily#145).

## Symptom Pushing `coilysiren/cli-guard` `main` to its two `origin` push URLs (GitHub + Forgejo mirror) succeeds on GitHub but is rejected non-fast-forward on Forgejo. The two remotes have diverged: - Forgejo-only: `ed36b49 lockdown: sync to coily v2.45.0 [skip ci]` (an automated lockdown sync) - never reached GitHub. - GitHub-only: the recent dispatch commits (e.g. surface-split, consult, surface-to-isolation inversion) - never reached Forgejo. ## Impact `go get github.com/coilysiren/cli-guard@...` resolves against GitHub, so the Go module path is unaffected. But the Forgejo mirror is stale and any future "force-sync GitHub -> Forgejo" would drop the automated `ed36b49` lockdown commit; conversely a Forgejo-first push keeps failing. ## Likely cause An automated `lockdown: sync ... [skip ci]` job committed straight to Forgejo `main` without the same commit landing on GitHub, so the mirror is no longer a strict ancestor/descendant of GitHub. ## Ask Reconcile the two cli-guard remotes (decide whether `ed36b49` should be replayed onto GitHub `main`, or dropped) and confirm which remote is the write source of truth for cli-guard so the mirror stops drifting. Surfaced while landing coilysiren/cli-guard#35 (consumer coilysiren/coily#145).

coilysiren referenced this issue from coilyco-bridge/coily

2026-05-28 11:04:07 +00:00

coily GitHub and Forgejo mirrors have forked divergent feature work on both sides #148

coilysiren referenced this issue

2026-05-28 11:04:27 +00:00

dispatch: invert surface-to-isolation mapping main vs worktree+branch #35

coilysiren referenced this issue from coilyco-bridge/coily

2026-05-28 11:11:32 +00:00

coily GitHub and Forgejo mirrors have forked divergent feature work on both sides #148

coilysiren added the

label

2026-05-31 01:54:14 +00:00