feat(cli): auto-seed git worktrees from their base checkout

[RaghavChamadiya]

What

Makes git worktrees seamless: repowise init and repowise update inside a linked worktree now detect the base checkout automatically (via git rev-parse --git-common-dir), seed the worktree's index from it, and catch up incrementally. No flags, no paths. Follow-up to #655, which introduced the manual --seed-from flag.

cd my-feature-worktree
repowise update
[worktree] Unindexed linked worktree of /path/to/base; seeding its index.

• --seed-from <path> stays as an explicit override for unusual layouts
• new --no-seed forces a cold init inside a worktree
• update auto-seeding means post-commit hooks and coding agents running in a fresh worktree get the fast path with zero setup

Bug fixes along the way

Seeded indexes no longer split in two. upsert_repository matches by local_path, so a seeded wiki.db still pointed at the base checkout; the first update in the worktree minted a second repository row named after the worktree dir and regenerated everything under it, while the seeded pages stayed orphaned under the old row (breaking prior-page reuse and repo-scoped queries). Seeding now adopts the copied row: name and local_path are re-pointed at the worktree and the repo-level pages are retargeted.

[workspace] notices render again. [dim][workspace][/dim] was being parsed as console markup and printed without its prefix. Both the workspace notice and the new [worktree] notice are escaped.

The #655 test suite is green again. The seed tests were failing on main's CI (integration tests only run on push to main, so the PR run never exercised them) and failed differently on Windows:

• raw git commit calls without identity env (exit 128 on CI runners); now routed through the identity-injecting _git helper
• git checkout main assumed the default branch name; now captured via rev-parse
• with sqlite3.connect(...) never closes the handle, which broke worktree cleanup on Windows; replaced with closed-connection helpers
• the git_work_repo fixture pins REPOWISE_DB_URL to the base repo's db, silently routing the worktree's delegated update into the wrong database; the seed tests now drop the env var and use repo-local DBs
• substring assertions broke when Rich line-wrapped long Windows temp paths; whitespace is collapsed before matching
• cleanup order (rmtree before git worktree remove) left git bookkeeping pointing at a missing dir

The index-only survival assertion moved from wiki_pages (empty in index-only runs) to graph_nodes, and the "new file gets a page" assertion is dropped: update-time coverage budgeting prevents it on every platform, tracked as #746.

Docs

New docs/WORKTREES.md (behavior, skip conditions, overrides, troubleshooting); WORKSPACES FAQ, CLI_REFERENCE, README, and QUICKSTART updated; changelog entries added under Unreleased (and the #655 entry moved out of the 0.25.0 section it had landed in).

Testing

• 5 new unit tests for worktree detection and seed preconditions
• 4 new integration tests (init auto-seed, --no-seed, update auto-seed, unindexed-base fallback) plus the 5 repaired feat(cli): allow seeding worktree index from base branch #655 tests
• full tests/integration/test_cli.py (28) and tests/unit/cli + tests/unit/workspace (1162) green on Windows; end-to-end verified manually with the mock provider (single repository row, seeded pages reused, state advanced to worktree HEAD)

[repowise-bot]

✅ Health: 7.7 (unchanged)

📋 At a glance
4 hotspots touched · 4 new findings introduced · 5 co-change pairs left out · 4 files with recent fix history · 2 dead-code findings.

Files & modules (2)

• packages (3 files)
  • .../init_cmd/command.py
  • .../update_cmd/command.py
  • .../cli/helpers.py
• tests (1 file)
  • tests/integration/test_cli.py

⚠️ Change risk: moderate _{(riskier than 64% of this repo's commits · raw 9.3/10)}
This change's risk is driven by:

• more lines added than baseline
• more scattered than baseline

🩹 Review priority _{(files here with the most recent bug-fix history — defects cluster, so review these first)}

• .../cli/helpers.py — fixed 10× in the last ~6 months
• tests/integration/test_cli.py — fixed 5× in the last ~6 months
• .../init_cmd/command.py — fixed 3× in the last ~6 months
• .../update_cmd/command.py — fixed 3× in the last ~6 months

🔎 More signals (3)

🔥 Hotspots touched (4)

• .../init_cmd/command.py — 14 commits/90d, 1 dependents · primary owner: Raghav Chamadiya (82%)
• .../update_cmd/command.py — 10 commits/90d, 3 dependents · primary owner: Raghav Chamadiya (90%)
• .../cli/helpers.py — 23 commits/90d, 68 dependents · primary owner: Raghav Chamadiya (72%)

1 more

• tests/integration/test_cli.py — 8 commits/90d, 1 dependents · primary owner: Paarth SIkka (36%)

🔗 Hidden coupling (3 files)

• .../cli/helpers.py co-changes with these files (not in this PR):
  • .../cli/test_helpers.py (9× — 🟢 routine)
  • docs/CLI_REFERENCE.md (5× — 🟢 routine)
• .../init_cmd/command.py co-changes with these files (not in this PR):
  • docs/CLI_REFERENCE.md (7× — 🟢 routine)
  • .../init_cmd/workspace.py (6× — 🟢 routine)
• .../update_cmd/command.py co-changes with .../update_cmd/persistence.py (5× — 🟢 routine) — not in this PR.

💀 Dead code (2 findings)

• 💀 .../cli/helpers.py ensure_db (confidence 1.00)
• 💀 .../cli/helpers.py is_interactive_session (confidence 1.00)

---

📊 Full report · ⭐ Star Repowise · 📥 Install bot · Last updated 2026-07-10 10:32 UTC
_{Silence on a single PR with [skip repowise] in the title · Per-repo toggle on repowise.dev/settings?tab=bot}

[swati510]

LGTM