Pull

Workflow

1. Verify git status is clean or commit/stash changes before merging.
2. Ensure rerere is enabled locally:
   • git config rerere.enabled true
   • git config rerere.autoupdate true
3. Confirm remotes and branches:
   • Ensure the origin remote exists.
   • Ensure the current branch is the one to receive the merge.
4. Fetch latest refs:
   • git fetch origin
5. Sync the remote feature branch first:
   • git pull --ff-only origin $(git branch --show-current)
   • This pulls branch updates made remotely (for example, a GitHub auto-commit) before merging origin/main.
6. Merge in order:
   • Prefer git -c merge.conflictstyle=zdiff3 merge origin/main for clearer conflict context.
7. If conflicts appear, resolve them (see conflict guidance below), then:
   • git add <files>
   • git commit (or git merge --continue if the merge is paused)
8. Verify with project checks:
   • Prefer .codex/skills/push/scripts/openase_ci_gate.sh so merge-time validation follows the same local gate as push-time validation.
9. Summarize the merge:
   • Call out the most challenging conflicts/files and how they were resolved.
   • Note any assumptions or follow-ups.

Conflict Resolution Guidance (Best Practices)

• Inspect context before editing:
  • Use git status to list conflicted files.
  • Use git diff or git diff --merge to see conflict hunks.
  • Use git diff :1:path/to/file :2:path/to/file and git diff :1:path/to/file :3:path/to/file to compare base vs ours/theirs for a file-level view of intent.
  • With merge.conflictstyle=zdiff3, conflict markers include:
    • <<<<<<< ours, ||||||| base, ======= split, >>>>>>> theirs.
    • Matching lines near the start/end are trimmed out of the conflict region, so focus on the differing core.
  • Summarize the intent of both changes, decide the semantically correct outcome, then edit:
    • State what each side is trying to achieve (bug fix, refactor, rename, behavior change).
    • Identify the shared goal, if any, and whether one side supersedes the other.
    • Decide the final behavior first; only then craft the code to match that decision.
    • Prefer preserving invariants, API contracts, and user-visible behavior unless the conflict clearly indicates a deliberate change.
  • Open files and understand intent on both sides before choosing a resolution.
• Prefer minimal, intention-preserving edits:
  • Keep behavior consistent with the branch's purpose.
  • Avoid accidental deletions or silent behavior changes.
• Resolve one file at a time and rerun tests after each logical batch.
• Use ours/theirs only when you are certain one side should win entirely.
• For complex conflicts, search for related files or definitions to align with the rest of the codebase.
• For generated files, resolve non-generated conflicts first, then regenerate:
  • Prefer resolving source files and handwritten logic before touching generated artifacts.
  • Run the CLI/tooling command that produced the generated file to recreate it cleanly, then stage the regenerated output.
• For import conflicts where intent is unclear, accept both sides first:
  • Keep all candidate imports temporarily, finish the merge, then run lint/type checks to remove unused or incorrect imports safely.
• After resolving, ensure no conflict markers remain:
  • git diff --check
• When unsure, note assumptions and ask for confirmation before finalizing the merge.

When To Ask The User (Keep To A Minimum)

Do not ask for input unless there is no safe, reversible alternative. Prefer making a best-effort decision, documenting the rationale, and proceeding.

Ask the user only when:

• The correct resolution depends on product intent or behavior not inferable from code, tests, or nearby documentation.
• The conflict crosses a user-visible contract, API surface, or migration where choosing incorrectly could break external consumers.
• A conflict requires selecting between two mutually exclusive designs with equivalent technical merit and no clear local signal.
• The merge introduces data loss, schema changes, or irreversible side effects without an obvious safe default.
• The branch is not the intended target, or the remote/branch names do not exist and cannot be determined locally.

Otherwise, proceed with the merge, explain the decision briefly in notes, and leave a clear, reviewable commit history.