[thekevinbot]

Fixes #1477

./*/*.md over the depth-bound tree: 32 directories, each tens of thousands of entries wide. A directory's rows reached the SQLite consumer only once its whole listing had been judged. On this tree the consumer waited about 130-190ms for its first chunk, then serialized about 2.7M rows mostly after the walk had already finished.

A directory of more than PART_ENTRIES (8192) entries is now returned from Place::explore as parts. Each part is a Step::Dir holding the shared Arc<Listing> and an entry range. walk_in_order schedules those parts like any directory: workers judge them ahead, and the walk thread hands their rows out in order. Rows start flowing after the first part, and consumption overlaps the walk.

That parallelism replaces judge_all, idle_cores and tree_walk::busy_workers (added in #1522), which are removed. tree_walk.rs goes back to its form before #1522.

Output is unchanged: byte-identical against main for ./*/*.md, ./* and ./**/*.md on the 3.2M-item tree, plus a mixed wide tree.

Measurements

These are paired medians vs fdfind, net of startup, on a host with other agents running.

case	main	this PR
depth-bound ./*/*.md, e2e test (SPEED fastest, 15 pairs)	1.124	0.861
depth-bound ./*/*.md, CLI capture (15 pairs)	1.122	0.910

The capture bench gave these for other shapes: flat ./*.md 0.88 -> 0.71, flat ./** 0.83 -> 0.69, mixed 0.90 -> 0.75, one-level ./* 0.81 -> 0.67.

Expected conflict: #1521 keeps busy_workers in tree_walk.rs, which this PR deletes. The parts also work on top of #1521's ordered heap: about 0.89x fd on that combination.

E2E Verification

• Ran e2e suites locally for every affected SDK
• Python SDK e2e: pass
• TypeScript SDK e2e: pass
• Rust core e2e (if applicable): pass
• packages/python/e2e-attestations/<branch>.json receipt written if packages/python changed (just e2e-attest-python)
• packages/ts/e2e-attestations/<branch>.json receipt written if packages/ts changed (just e2e-attest-ts)
• Command(s) run:
  • cargo test --workspace --features cli
  • cd packages/python && uv run maturin develop --release
  • testing-conventions e2e attest 'uv run python -m pytest tests/e2e/glob_speed_{depth_bound,one_level,recursive,matchers,dotfiles,gitignore,node_modules,symlink}_test.py tests/e2e/query_contract_test.py tests/e2e/fanout_test.py tests/e2e/use_case_join_test.py -q -s'
  • cd packages/ts && testing-conventions e2e attest 'pnpm test:e2e'
• Result summary:
  • Rust: all pass.
  • Python e2e: 28 passed.
  • TS e2e: 18 passed.

Changelog / Migrations

• Changelog fragment added under <root>/<pkg>/changelog.d/ for each changed package (or: skip-changelog trailer on a commit with reason)
• Migration fragment with the same slug added under <root>/<pkg>/migrations.d/ for each changed package (every subsection but Summary _None._ when nothing breaks)