synclisted

# Syncing a stale branch ## 1. Work out what is stale Read `.keel.json`. Work out the base branch from the current branch's prefix: `integration` for a feature branch, `production` for a hotfix (or, under `trunk` topology, `production` in both cases). Work out which remote holds the canonical repo: `upstream` if you're on a fork (per `keel:start-work`), `origin` otherwise. Call that remote `<remote>` below. git fetch --all --prune git log --oneline HEAD..<remote>/<base> | head -20 If that is empty, the branch is current - say so and stop. ## 2. If this is a fork, fetch upstream's base directly - do not push it git fetch upstream <base> Rebase or merge (step 3) against `upstream/<base>` directly. Do **not** try to "refresh" your fork's own copy of `<base>` by pushing to it (`git push origin <base>`) - keel's guard blocks any push whose destination branch is named `production` or `integration`, and it checks the branch name in the refspec, not which remote or repo owns it. That push will be denied even though it targets your own fork. There's no need for it anyway: always fetch `upstream/<base>` fresh rather than keeping a local mirror in sync. ## 3. Update the working branch Rebase when the branch has not been reviewed yet - it keeps history readable: git checkout <branch> git rebase <remote>/<base> **Merge instead of rebasing if the PR is already under review.** Rewriting history mid-review destroys the reviewer's ability to see what changed: