chore(#42): roll out ecosystem code-style policy

[gsdali]

What & why

Rolls out the ecosystem's code-style policy to this repo
(ecosystem docs/code-style-policy-proposal-2026-08.md),
following OCCTSwiftScripts#114 /
#115 as the reference implementation
for a repo this size: a full sweep into compliance in one PR and straight to a blocking gate, not
OCCTSwift's gradual "if you touch it, you fix it" exemption manifest (that pattern is for large
repos, not this one). This repo is ~6.6k lines across 30 files, small enough to fully sweep.

Adds:

• .swift-format (100 col, 4-space, the ecosystem's one deliberate divergence from Google's
  2-space default) copied verbatim from OCCTSwiftScripts, plus a full swift-format --in-place
  sweep of Sources/+Tests/.
• .swiftlint.yml, deliberately narrow (only_rules: [orphaned_doc_comment]). SwiftLint's
  default rule set duplicates swift-format's formatting opinions (can disagree with them on the
  same line) and separately opens a large code-quality/complexity surface that overlaps the
  ecosystem's own code-structure policy rather than this one. Scoped down to the one rule that
  catches something swift-format has no equivalent for.
• Scripts/comment-ratio-check.sh: report-only nudge (never fails), bash variant matching
  OCCTSwiftScripts' own convention (not the Python style OCCTSwift/Scripts/ uses).
• .github/workflows/code-style.yml: swift-format lint --strict and swiftlint lint --strict
  as blocking gates, comment-ratio as a non-blocking step. This repo had no CI at all before
  this PR, so this bootstraps it from nothing.
• okf/policies/code-style.md: this repo's 7th policy, linked from okf/index.md.
• Zero first-party C++ bridge files in this repo (confirmed), so clang-format is out of scope,
  same as noted in the tracking issue.

Hand-fixed lint findings (62 total, everything swift-format --in-place couldn't auto-fix)

• 57× BeginDocumentationCommentWithOneLineSummary. Split a doc comment's first sentence from
  its body with a blank /// line, per Google's style guide structure. One genuine swift-format
  quirk found and worked around: a backtick code span immediately followed by a bare 's
  (e.g. `DragGesture`'s) makes the sentence-boundary parser misreport "doesn't terminate with
  a period" even when it does; the fix is moving the possessive inside the span
  (`DragGesture's`), verified empirically against isolated repro cases before applying it
  across every occurrence.
• 1× orphaned_doc_comment (StandardObjects.swift): a file-level summary comment sat before
  // MARK: - Trihedron and Trihedron's own (separately documented) doc comment, describing all
  four types in the file, not just Trihedron. Demoted /// to // since it's file-scoped
  commentary, not documentation of one following declaration, and it doesn't duplicate anything
  already documented.
• 2× AlwaysUseLowerCamelCase (ManipulatorWidget.swift, local to updateRotateDrag):
  R -> rotation, M -> rotateAboutPivot. Local-only, all call sites in the same function
  updated, no public API change.
• 3× EndOfLineComment: moved a trailing // comment that exceeded the line length onto its
  own line above, in SelectionFilterTests.swift (x2) and ManipulatorRotateTests.swift.

Em-dash cleanup

The ecosystem's writing-style policy requires clearing em-dashes "from any file you are already
editing," not only new ones. Every file swift-format --in-place or a hand-fix touched carried
pre-existing em-dashes in doc/line comments (101 occurrences across 26 files); all replaced with
ordinary punctuation (comma, colon, semicolon, or a split sentence), preserving meaning. Also
caught and fixed two misattributed pieces of copied text: the comment-ratio script's header cited
OCCTSwiftScripts#114 (now points at this repo's own #42, crediting #114/#115 as the
reference), and .swiftlint.yml's rationale comment cited OCCTSwiftScripts' own 548 hits measurement
for identifier_name as if it were this repo's number (now attributed correctly).

Verification

Zero behavior change. Formatting, doc comments, and two local-only renames only; no public API,
signature, or logic changed.

• swift-format lint --strict (Sources/+Tests/): 0 violations (was 532 pre-sweep).
• swiftlint lint --strict: 0 violations, 0 serious (was 1: the orphaned_doc_comment above).
• Scripts/comment-ratio-check.sh: no files at or above 1.0x comment:code (Sources/), exit 0.
• swift build: succeeds before and after the sweep (Build complete!).
• swift test: fails to compile both before and after the sweep, with the exact same two
  pre-existing, unrelated compile errors (confirmed via a clean-tree git stash baseline run
  before touching anything):
  • InteractiveContextMutationTests.swift / RemapTests.swift: value of optional type 'SIMD3<Double>?' must be unwrapped (centroids.enumerated().min { ... }?.offset needs a
    force-unwrap or optional chain) — 40 identical error occurrences in both runs.
  • RemapTests.swift: cannot infer contextual base in reference to member 'body' (.body(newObj)
    inside #expect(...)) — 4 identical error occurrences in both runs, just at a shifted line
    number post-sweep since earlier swift-format reflow moved the line down.
  • Not fixed here: pre-existing, unrelated to code style, out of scope for a formatting/docs pass.
    Flagging in case it's worth its own issue; happy to file one if wanted.
• Package.resolved: untouched by this PR (verified no diff from local build/test runs).

Closes #42

Checklist

• New or changed behavior is covered by a unit test in the same PR (not just manual
  verification) — see SecondMouseAU/OCCTReconstruct#397
  for the ecosystem-wide test-coverage standard this is piloting.
  — N/A here: no behavior changed (formatting/docs/CI only), verified via lint tool exit
  codes and the before/after build/test comparison above rather than new tests.

Notes for the reviewer

Per the ecosystem's own gate (no repo implements without a filed issue first), this was filed as
#42 before any of this landed. Holding for review, not to merge.

The pre-existing swift test compile failure (see Verification above) is real and reproducible on
a clean main checkout, unrelated to this change. Flagging rather than fixing, since it's out of
scope for a code-style/formatting PR and touching it would mix an unrelated fix into this diff.

🤖 Generated with Claude Code