Jujutsu support

[atimmer]

I've managing a fairly complicated local jj setup with an octopus merge at the top. Agents want to run CI on that top revision, but crabbox has a ton of git assumptions. It just synced the wrong diff 3 times, which burns a lot of unnecessary token. Having jj support would fix that.

[clawsweeper]

Codex review: keeping this open for maintainer follow-up; there is still a little grit to resolve. Reviewed September 3, 2026, 6:51 PM ET / 22:51 UTC.

Summary
Keep open: current main safely prevents the reported wrong-diff sync, but native Jujutsu bookmark/change/commit mapping remains intentionally unsupported pending a minimized fixture-defined contract.

Reproducibility: not applicable. as a bug reproduction: this is a feature request, and current main deliberately rejects native Jujutsu workspaces instead of reproducing the earlier wrong-diff behavior.

Root-cause cluster
Relationship: canonical
Canonical: #1081
Summary: This issue remains the canonical tracker for native Jujutsu revision mapping; the merged work only provides safety boundaries around the unsupported path.

Members:

• partial_overlap: fix(sync): reject native Jujutsu workspaces #1331 - Merged the safe rejection of native Jujutsu workspaces but explicitly left native revision mapping unresolved.
• partial_overlap: fix(sync): validate Jujutsu before delegated provisioning #1366 - Moved the existing rejection before delegated provider calls without adding native revision mapping.
• adjacent_distinct: feat(sync): support explicit synchronization from non-Git workspaces #1334 - Tracks explicit synchronization of intentionally non-Git directories, not Jujutsu revision semantics.
• adjacent_distinct: feat: add opt-in Git overlay sync #1453 - Adds opt-in Git overlay synchronization and still relies on Git-derived manifests.
• adjacent_distinct: feat(sync): seed remote Git metadata from local objects #1593 - Tracks sanitized Git-object seeding for unreachable Git origins, not native Jujutsu mapping.

Proposal only: this assessment does not dispatch repair, suppress jobs, mutate sibling items, close, or merge anything.

Ways to help us reproduce this

• Add a screenshot or short recording showing the behavior.
• Add expected vs actual behavior.
• Include redacted logs or terminal output.

Next step
The maintainer has already set a fixture-first boundary; native mapping is not safe to assign until a minimized octopus-merge fixture supplies the required revision semantics.

Review details

Best possible solution:

First add a minimized octopus-merge fixture that defines required bookmark/change/commit semantics, then implement native mapping with source-blind regression coverage while preserving colocated-Git and --no-sync behavior.

Do we have a high-confidence way to reproduce the issue?

Not applicable as a bug reproduction: this is a feature request, and current main deliberately rejects native Jujutsu workspaces instead of reproducing the earlier wrong-diff behavior.

Is this the best way to solve the issue?

No: heuristic mapping would not be a safe solution. The existing maintainer-approved fixture-first path is the narrowest way to establish a durable native revision contract.

AGENTS.md: found and applied where relevant.

Remaining risk / open question:

• Implementing native mapping without a minimized fixture could select the wrong Jujutsu bookmark, change, or commit and reintroduce incorrect source synchronization.

Codex review notes: model internal, reasoning high; reviewed against 9d0730b5fa72.

Label changes

Label justifications:

• P2: Native Jujutsu support remains a bounded but unsponsored synchronization capability whose safe contract is not yet defined.
• impact:other: Choosing the wrong source revision wastes remote execution work without fitting a more specific owned impact class.

Evidence reviewed

What I checked:

• Current implementation: Current main detects the nearest native Jujutsu boundary and rejects Git-manifest generation with an actionable diagnostic rather than allowing Git discovery to select an outer checkout. No native revision adapter is implemented. (internal/cli/repo.go:945, 9d0730b5fa72)
• Regression coverage for the safe boundary: Tests cover native Jujutsu rejection, colocated Git/Jujutsu manifest generation, explicit Git routing, and rejection before delegated-provider calls. (internal/cli/repo_test.go:796, 9d0730b5fa72)
• Documented supported boundary: The current user documentation says native Jujutsu revision mapping is not supported, while same-root colocated Git metadata and --no-sync remain supported paths. (docs/features/sync.md:82, 9d0730b5fa72)
• Merged safeguard provenance: The safe rejection originated in merged pull request fix(sync): reject native Jujutsu workspaces #1331; a later merged safeguard ensures delegated archive-sync providers are not called for a rejected native workspace. Both changes remain reachable from main. (internal/cli/repo.go:42, 51521d5bc229)
• Released safety fix, not native support: The rejection safeguard is contained in v0.42.0 and later tags, while the delegated-provider preflight is contained in v0.44.0 and later tags. These releases solve the unsafe fallback but do not implement the requested native mapping. (CHANGELOG.md:317, 1b045b57a435)
• Maintainer scope decision: The discussion explicitly keeps this issue open for fixture-backed native revision mapping and rejects heuristic mapping until a minimized octopus-merge fixture defines bookmark/change/commit behavior.

Likely related people:

• steipete: Suggested for follow-up; no historical authorship or introduction is verified. (role: unverified routing candidate; confidence: low)

How this review workflow works

• ClawSweeper keeps one durable marker-backed review comment per issue or PR.
• Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
• A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
• PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
• Maintainers can also comment @clawsweeper review to request a fresh review only.
• Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
• Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
• Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

[steipete]

Maintainer triage: PRs welcome on this one.

Scope guidance for anyone picking it up: sync today is git-manifest-based, so a jj workspace can silently produce the wrong diff (as reported). A good first slice is jj-workspace detection with a clear error pointing at the colocated-git workflow, plus docs — a full native jj revision adapter is a much bigger change and would need a minimized repro to design against. Marking help wanted.