Skip to content

roadmap.md

Latest commit

 

History

History
891 lines (766 loc) · 47.7 KB

roadmap.md

File metadata and controls

891 lines (766 loc) · 47.7 KB

Development roadmap

0. Repository scaffolding

0.1. Workspace foundations

  • 0.1.1. Initialise the Cargo workspace with workspace members and rust-toolchain pin.
  • 0.1.2. Add baseline project metadata (README, LICENSE, contributing guide, CODEOWNERS).
  • 0.1.3. Establish Makefile or justfile targets for fmt, lint, test, and UI test orchestration.

1. Common infrastructure

1.1. Common crate helpers

  • 1.1.1. Implement the common crate helpers for attributes, context detection, spans, and diagnostics.
  • 1.1.2. Wire shared configuration loading with dylint_linting::config_or_default and serde structs.

1.2. UI test harness

  • 1.2.1. Integrate dylint_testing harness boilerplate for UI tests across all lint crates.

1.3. Conditional Dylint expect support

  • 1.3.1. Add the crates/whitaker_support_macros proc-macro crate and implement #[whitaker_support::dylint_expect(...)] with lib, lints, and optional reason arguments. See ADR 002 §Decision outcome / proposed direction and §Functional requirements.
  • 1.3.2. Add the crates/whitaker_support re-export crate, wire both support crates into the workspace, and document the supported attribute surface with API examples. See ADR 002 §Decision outcome / proposed direction and §Migration plan.
  • 1.3.3. Add compatibility coverage that proves the attribute stays warning free in non-Dylint builds, Clippy runs, and Dylint runs with the matching dylint_lib value. See ADR 002 §Technical requirements and §Migration plan. Requires 1.3.1 and 1.3.2.
  • 1.3.4. Document intended usage, narrow-scope review guidance, and pre-expansion limitations for the conditional expect attribute. See ADR 002 §Decision outcome / proposed direction and §Known risks and limitations. Requires 1.3.3.

2. Core lint delivery

2.1. Lint crate template

  • 2.1.1. Establish the lint crate template with shared dependencies and UI test harness boilerplate.

2.2. Core lint implementations

  • 2.2.1. Implement function_attrs_follow_docs with targeted UI scenarios.
  • 2.2.2. Implement no_expect_outside_tests with context-aware diagnostics.
  • 2.2.3. Implement module_must_have_inner_docs ensuring each module opens with an inner doc comment, including UI scenarios for inline, file, and macro-generated modules plus localised diagnostics and behaviour-driven coverage.
  • 2.2.4. Implement public_fn_must_have_docs using effective visibility data.
  • 2.2.5. Implement test_must_not_have_example covering code-fence heuristics.
  • 2.2.6. Implement module_max_lines with configurable thresholds.
  • 2.2.7. Implement conditional_max_n_branches for complex predicates.
  • 2.2.8. Implement no_unwrap_or_else_panic with optional clippy helpers.
  • 2.2.9. Replace ad hoc conditional allow and expect sequences in Whitaker-managed code with #[whitaker_support::dylint_expect(...)], keeping suppressions narrowly scoped to the item they justify. See ADR 002 §Migration plan. Requires 1.3.1, 1.3.2, and 1.3.3.

2.3. Localisation enablement

  • 2.3.1. Add fluent-templates and once_cell to the workspace dependencies and expose a shared common::i18n loader.
  • 2.3.2. Create the locales/ resource tree with an en-GB fallback and secondary cy/gd language samples covering every lint slug.
  • 2.3.3. Refactor lint diagnostics to source primary messages, notes, and help text from Fluent bundles with structured arguments.
  • 2.3.4. Allow locale selection via DYLINT_LOCALE and dylint.toml, and add UI smoke tests that run under at least one non-English locale.

3. Aggregated packaging, installer, and unified CLI

3.1. Suite assembly

  • 3.1.1. Assemble the whitaker_suite cdylib using constituent features and combined lint pass wiring.

3.2. Installer CLI

  • 3.2.1. Implement the installer CLI that builds, links, and stages all lint libraries.
  • 3.2.2. Add test-support Cargo feature exposing StubExecutor and StubMismatch for external test suites. See installer/Cargo.toml feature documentation for usage guidance and caveats.

3.3. Documentation examples

  • 3.3.1. Provide consumer guidance and workspace metadata examples in documentation.

3.4. Prebuilt lint libraries

  • 3.4.1. Define artefact naming, manifest schema, and verification policy. See docs/adr-001-prebuilt-dylint-libraries.md.
  • 3.4.2. Build CI automation to compile lint libraries for the supported target matrix and publish .tar.zst assets to the rolling release. See docs/adr-001-prebuilt-dylint-libraries.md §Decision outcome / proposed direction.
  • 3.4.3. Emit manifest.json for each artefact with git SHA, toolchain, target triple, build time, and SHA256. See docs/adr-001-prebuilt-dylint-libraries.md §Decision outcome / proposed direction.
  • 3.4.4. Extend the installer to download and verify prebuilt artefacts before local compilation, falling back on failure. See docs/adr-001-prebuilt-dylint-libraries.md §Decision outcome / proposed direction.
  • 3.4.5. Extract libraries to ~/.local/share/whitaker/lints/<toolchain>/<target>/lib and set DYLINT_LIBRARY_PATH. See docs/adr-001-prebuilt-dylint-libraries.md §Decision outcome / proposed direction.
  • 3.4.6. Record download-versus-build rates and total installation time. See docs/adr-001-prebuilt-dylint-libraries.md §Migration plan.

3.5. Unified CLI foundation

  • 3.5.1. Add a root whitaker binary and move the current installer behaviour behind an internal library boundary. See Whitaker CLI design §Public CLI surface and §Compatibility and migration. Requires 3.2.1.
  • 3.5.2. Implement whitaker check as the default linting path, including lazy dependency repair, --no-install, forwarded cargo arguments after --, and operational error handling that preserves lint-failure exit semantics. See Whitaker CLI design §Public CLI surface and §whitaker check. Requires 3.6.3, 3.6.4, and 3.7.1.
  • 3.5.3. Publish whitaker release artefacts with cargo-binstall metadata that mirror the existing installer packaging flow. See Whitaker CLI design §Public CLI surface and §Compatibility and migration. Requires 3.5.1 and 4.3.1.

3.6. Rule selection and configuration model

  • 3.6.1. Assign stable rule codes, family selectors, and DEFAULT/ALL semantics for the core lint suite. See Whitaker CLI design §Rule identifiers and selection model. Requires 3.5.1.
  • 3.6.2. Implement selector precedence across configuration and CLI flags, including the curated --experimental behaviour for implicit and explicit selections. See Whitaker CLI design §Rule identifiers and selection model. Requires 3.6.1.
  • 3.6.3. Adopt ortho_config for CLI, environment, and whitaker.toml loading, while retaining one-release compatibility with dylint.toml and legacy environment keys. See Whitaker CLI design §Configuration model. Requires 3.5.1.
  • 3.6.4. Add shared --locale, --colour, and --progress controls, and ensure the merged config surface remains localizable and accessible. See Whitaker CLI design §Accessibility and localization requirements and §Configuration model. Requires 3.6.3 and 2.3.4.

3.7. Unified installation and bundle state

  • 3.7.1. Move dependency repair, toolchain provisioning, and bundle installation into whitaker install, including --offline, --build-from-source, and --toolchain flows. See Whitaker CLI design §whitaker install. Requires 3.5.1, 3.4.5, and 4.3.2.
  • 3.7.2. Resolve source builds against the CLI release version or an explicit revision instead of cloning main, and record source provenance for installed bundles. See Whitaker CLI design §whitaker install. Requires 3.7.1.
  • 3.7.3. Add per-bundle manifest files with schema version, Whitaker version, source SHA, build date, toolchain, target, origin, and bundle-kind metadata. See Whitaker CLI design §Bundle manifests and whitaker ls. Requires 3.7.1.

3.8. Status, diagnostics, and repair commands

  • 3.8.1. Implement whitaker ls text output with installed bundle metadata, the effective config path, and per-rule enablement states. See Whitaker CLI design §Bundle manifests and whitaker ls. Requires 3.6.2 and 3.7.3.
  • 3.8.2. Add --json to whitaker ls and whitaker doctor, using stable machine-readable fields that remain untranslated across locales. See Whitaker CLI design §Accessibility and localization requirements and §whitaker doctor. Requires 3.8.1.
  • 3.8.3. Record structured dependency-install and bundle-build failures with timestamps, phase metadata, stderr tails, advice, and log-path references. See Whitaker CLI design §Failure recording. Requires 3.7.1.
  • 3.8.4. Implement whitaker doctor to summarize configuration, toolchains, dependencies, prebuilt availability, bundle-version drift, selected-lint coverage, and recent failures. See Whitaker CLI design §whitaker doctor and §Failure recording. Requires 3.7.3 and 3.8.3.
  • 3.8.5. Add behaviour coverage for locale overrides, plain-progress output, and JSON parity between text and machine-readable status commands. See Whitaker CLI design §Accessibility and localization requirements. Requires 3.6.4, 3.8.2, and 3.8.4.

3.9. Compatibility release and documentation migration

  • 3.9.1. Ship a compatibility release where whitaker-installer dispatches to whitaker install with a deprecation notice and whitaker ls accepts list as an alias. See Whitaker CLI design §Public CLI surface and §Compatibility and migration. Requires 3.7.1 and 3.8.1.
  • 3.9.2. Update docs/users-guide.md, docs/developers-guide.md, docs/publishing.md, and installer-facing workflow documentation to point to the unified CLI and the new configuration surface. See Whitaker CLI design §Configuration model, §Compatibility and migration, and §Expected outcomes. Requires 3.6.3 and 3.9.1.
  • 3.9.3. Remove wrapper script generation, whitaker-ls, and installer-first references once the compatibility release window closes. See Whitaker CLI design §Public CLI surface and §Compatibility and migration. Requires 3.9.1 and 3.9.2.

4. Quality gates and automation

4.1. Continuous integration

  • 4.1.1. Configure CI workflows for fmt, clippy, dylint runs, and per-crate UI tests on a multi-OS matrix.
  • 4.1.2. Add markdownlint, nixie, and other doc/tooling checks to the pipeline.
  • 4.1.3. Enforce lint-level deny rules and fail builds on warnings across the workspace.

4.2. Release metadata

  • 4.2.1. Add cargo-binstall metadata to installer/Cargo.toml for published release artefacts.

4.3. Release workflow

  • 4.3.1. Implement a release workflow that builds whitaker-installer for each supported target, packages .tgz/.zip archives, and uploads them to GitHub Releases tagged v<version>.
  • 4.3.2. Publish repository-hosted dependency binaries for cargo-dylint and dylint-link, and teach the installer to prefer them before Cargo-based installation. (design.execplan.dependency-binary, docs/execplans/install-dependency-binaries.md Stage C-D)

5. Experimental Bumpy Road lint

5.1. Signal and detection pipeline

  • 5.1.1. Implement the per-line complexity signal builder and smoothing window logic.
  • 5.1.2. Detect bump intervals, surface diagnostics with labelled spans, and add configuration options.
  • 5.1.3. Ship UI coverage for positive and negative Bumpy Road scenarios and gate behind a feature flag.

6. Brain trust lints (brain_type and brain_trait)

6.1. Shared cohesion analysis

  • 6.1.1. Add a shared LCOM4 helper in common that builds a method graph and returns connected component counts. See brain trust lints design §Cohesion analysis (LCOM4). Requires 1.1.1.
  • 6.1.2. Define method metadata extraction for field access and method calls, including macro-span filtering. See brain trust lints design §Cohesion analysis (LCOM4) and §Implementation approach.

6.2. brain_type lint

  • 6.2.1. Implement metric collection for WMC (cognitive complexity), brain method detection (CC + LOC), LCOM4, and foreign reach. See brain trust lints design §brain_type signals. Requires 6.1.1.
  • 6.2.2. Implement threshold evaluation and escalation rules, and surface measured values in diagnostics. See brain trust lints design §brain_type rule set (initial defaults). Requires 6.2.1.
  • 6.2.3. Ensure macro-expanded spans are excluded or capped during CC calculation. See brain trust lints design §Metric collection.

6.3. brain_trait lint

  • 6.3.1. Implement trait item counting, default method CC aggregation, and implementor burden metrics. See brain trust lints design §brain_trait signals.
  • 6.3.2. Apply warning and escalation thresholds, and surface measured values in diagnostics. See brain trust lints design §brain_trait rule set (initial defaults).

6.4. Decomposition advice

  • 6.4.1. Build feature vectors for methods and cluster with community detection to form decomposition suggestions. See brain trust lints design §Decomposition advice. Requires 6.2.1.
  • 6.4.2. Emit concise diagnostic notes mapping clusters to extraction suggestions, capped for large types. See brain trust lints design §Decomposition advice.
  • 6.4.3. Use Verus to prove cosine_threshold_met's cross-multiplied threshold check is equivalent to cosine >= 0.20 for non-zero norms and cannot divide by zero. See brain trust lints design §Decomposition advice. Requires 6.4.1.
  • 6.4.4. Use Verus to prove dot_product and norm_squared algebraic properties, including commutativity, non-negativity, and zero-result behaviour when vectors have no overlapping positive features. See brain trust lints design §Decomposition advice. Requires 6.4.1.
  • 6.4.5. Use Kani to verify build_adjacency preserves similarity edges, keeps neighbour indices in bounds, and produces symmetric adjacency lists. See brain trust lints design §Decomposition advice. Requires 6.4.1.
  • 6.4.6. Use Kani to verify propagate_labels preserves valid label indices, returns one label per input vector, and terminates within the supplied iteration bound. Verification harnesses are complete modulo the CBMC state-explosion blocker documented in the design notes. See brain trust lints design §Decomposition advice. Requires 6.4.1.

6.5. SARIF output

  • 6.5.1. Collect brain trust diagnostics into a SARIF 2.1.0 emitter that is opt-in and English-only for tool ingestion. See brain trust lints design §SARIF output.

6.6. Configuration, localization, and tests

  • 6.6.1. Add brain_type and brain_trait configuration sections to whitaker.toml with documented defaults. See brain trust lints design §Configuration, localization, and testing.
  • 6.6.2. Add Fluent localization entries for both lints. See brain trust lints design §Configuration, localization, and testing.
  • 6.6.3. Add UI tests for positive and negative cases under crates/brain_type/ui/ and crates/brain_trait/ui/. See brain trust lints design §Configuration, localization, and testing.

6.7. Documentation

  • 6.7.1. Update docs/users-guide.md with lint descriptions, configuration keys, and SARIF usage. See brain trust lints design §Configuration, localization, and testing.

7. Clone detector pipeline

7.1. SARIF model and shared types

  • 7.1.1. Create the whitaker_sarif crate with SARIF 2.1.0 models, builders, and merge logic. See clone detector design §Crate responsibilities and §SARIF schema and mapping.

7.2. Token pass (Type-1/Type-2)

  • 7.2.1. Implement rustc_lexer normalization, k-shingling, winnowing, and Rabin-Karp hashing. See clone detector design §Pass A: token engine (rustc_lexer).
  • 7.2.2. Implement MinHash + LSH candidate generation with configurable bands and rows. See clone detector design §MinHash and LSH.
  • 7.2.3. Emit SARIF run 0 for accepted Type-1 and Type-2 pairs with stable fingerprints and spans. See clone detector design §SARIF emission (Run 0). Requires 7.1.1.
  • 7.2.4. Add sidecar proof workflows and Makefile targets for clone-detector Verus and Kani checks. Requires 7.2.2. See ADR 003.
  • 7.2.5. Use Verus to prove LshConfig::new rejects zero bands and rows, and enforces bands * rows == MINHASH_SIZE. Requires 7.2.4. See ADR 003 and clone detector design §MinHash and LSH.
  • 7.2.6. Use Verus to prove CandidatePair::new canonicalizes fragment ordering and suppresses self-pairs. Requires 7.2.4. See ADR 003 and clone detector design §MinHash and LSH.
  • 7.2.7. Use Kani to verify bounded MinHasher::sketch invariants, including deterministic output, duplicate-hash insensitivity, and empty-input failure. Requires 7.2.4. See ADR 003 and clone detector design §MinHash and LSH.
  • 7.2.8. Use Kani to verify bounded LshIndex invariants, including no self-pairs, canonical pair ordering, repeated-band deduplication, and insertion-order independence. Requires 7.2.4. See ADR 003 and clone detector design §MinHash and LSH.

7.3. AST refinement (Type-3)

  • 7.3.1. Map candidate spans to ra_ap_syntax nodes and extract AST feature vectors. See clone detector design §Pass B: AST engine (ra_ap_syntax).
  • 7.3.2. Score Type-3 similarity and update SARIF run 1 with cosine and AST hash metadata. See clone detector design §Scoring and acceptance (Type-3) and §SARIF update (Run 1).

7.4. CLI surface

  • 7.4.1. Implement cargo whitaker clones with scan, refine, report, and clean subcommands. See clone detector design §CLI surface.
  • 7.4.2. Emit optional HTML reports and anchor deep links in SARIF results. See clone detector design §Grouping and reporting.

7.5. Dylint integration

  • 7.5.1. Implement the clone_detected lint to load SARIF results, filter to current crate files, and emit diagnostics with help guidance. See clone detector design §Dylint integration (clone_detected lint).
  • 7.5.2. Honour #[allow(whitaker::clone_detected)] and per-file allowlists in configuration. See clone detector design §Dylint integration (clone_detected lint).

7.6. Incrementality and caching

  • 7.6.1. Implement the per-file cache in target/whitaker/clones-cache.bin with config hashing and shard indexing. See clone detector design §Incrementality and caching.

7.7. Testing and validation

  • 7.7.1. Add SARIF golden tests with deterministic ordering. See clone detector design §Testing strategy.
  • 7.7.2. Add Dylint UI tests that consume pre-baked SARIF results. See clone detector design §Testing strategy.
  • 7.7.3. Add property tests for normalization invariants. See clone detector design §Testing strategy.

7.8. Continuous integration

  • 7.8.1. Add CI jobs to run token and AST passes and verify stable SARIF output. See clone detector design §Acceptance criteria.

8. rstest fixture and test hygiene lints

8.1. Shared foundations

  • 8.1.1. Add shared rstest test and fixture detection helpers in common for attribute-based detection and optional expansion-trace fallback. See rstest fixture and test hygiene lints §Lint A: call-site fixture extraction and §Integration constraints. Requires 1.1.1.
  • 8.1.2. Add shared user-editable span recovery helpers for macro-heavy test code paths, and use them to avoid diagnostics on macro-only glue. See rstest fixture and test hygiene lints §Integration constraints and §Lint A: call-site fixture extraction. Requires 1.1.1.
  • 8.1.3. Add shared argument and paragraph fingerprint data models for deterministic grouping across tests. See rstest fixture and test hygiene lints §Lint A: call-site fixture extraction and §Lint C: repeated fixture paragraph detection. Requires 8.1.1.

8.2. rstest_helper_should_be_fixture lint

  • 8.2.1. Create the rstest_helper_should_be_fixture lint crate, register RSTEST_HELPER_SHOULD_BE_FIXTURE, and wire configuration loading defaults. See rstest fixture and test hygiene lints §Lint A: call-site fixture extraction. Requires 8.1.1 and 8.1.3.
  • 8.2.2. Implement call-site collection in #[rstest] tests, including fixture-local classification and constant-aware argument fingerprinting. See rstest fixture and test hygiene lints §Lint A: call-site fixture extraction. Requires 8.2.1.
  • 8.2.3. Implement crate-post aggregation thresholds and actionable diagnostics with span_lint_hir_and_then. See rstest fixture and test hygiene lints §Lint A: call-site fixture extraction. Requires 8.2.2 and 8.1.2.
  • 8.2.4. Add UI pass/fail coverage for repeated no-arg helpers, repeated fixture/constant argument helpers, and non-trigger cases. See rstest fixture and test hygiene lints §Lint A: call-site fixture extraction. Requires 8.2.3 and 1.2.1.

8.3. single_binding_paragraph lint

  • 8.3.1. Create the single_binding_paragraph lint crate and implement the local statement I/O model (defs, uses, muts, and control-flow guards). See rstest fixture and test hygiene lints §Lint B: single-binding paragraph detection. Requires 8.1.2.
  • 8.3.2. Implement the contiguous backward-slice algorithm, single-output checks, and configurable limits for paragraph length and external inputs. See rstest fixture and test hygiene lints §Lint B: single-binding paragraph detection. Requires 8.3.1.
  • 8.3.3. Emit diagnostics with bounded spans and explanatory notes for extraction candidates. See rstest fixture and test hygiene lints §Lint B: single-binding paragraph detection. Requires 8.3.2.
  • 8.3.4. Add UI pass/fail coverage for simple paragraphs, too-short candidates, control-flow-containing blocks, and intermediate reuse cases. See rstest fixture and test hygiene lints §Lint B: single-binding paragraph detection. Requires 8.3.3 and 1.2.1.

8.4. rstest_paragraph_should_be_fixture lint

  • 8.4.1. Create the rstest_paragraph_should_be_fixture lint crate and register RSTEST_PARAGRAPH_SHOULD_BE_FIXTURE with configurable assertion and input constraints. See rstest fixture and test hygiene lints §Lint C: repeated fixture paragraph detection. Requires 8.1.1 and 8.1.3.
  • 8.4.2. Reuse lint B candidate generation and implement assertion-free filtering plus fixture-or-constant input validation. See rstest fixture and test hygiene lints §Lint C: repeated fixture paragraph detection. Requires 8.4.1 and 8.3.2.
  • 8.4.3. Implement cross-test paragraph grouping, identical input fingerprint checks, and crate-post emission controls (emit_once_per_group). See rstest fixture and test hygiene lints §Lint C: repeated fixture paragraph detection. Requires 8.4.2.
  • 8.4.4. Add UI pass/fail coverage for repeated setup paragraphs, assertion presence, differing inputs, and non-fixture-derived inputs. See rstest fixture and test hygiene lints §Lint C: repeated fixture paragraph detection. Requires 8.4.3 and 1.2.1.

8.5. Integration, documentation, and promotion

  • 8.5.1. Add all three lints to the experimental set with feature-gated suite wiring and default configuration stanzas. See rstest fixture and test hygiene lints §Integration constraints and §Comparison and rollout guidance. Requires 8.2.4, 8.3.4, and 8.4.4.
  • 8.5.2. Add Fluent localization entries and diagnostic argument mappings for all three lint slugs. See rstest fixture and test hygiene lints §Integration constraints. Requires 8.5.1 and 2.3.3.
  • 8.5.3. Update user documentation with lint intent, configuration keys, and fixture-extraction remediation guidance. See rstest fixture and test hygiene lints §Comparison and rollout guidance. Requires 8.5.1.
  • 8.5.4. Define promotion criteria from experimental to standard based on UI stability and false-positive tuning across internal repositories. See rstest fixture and test hygiene lints §Integration constraints. Requires 8.5.1.

9. Ownership shape lints

9.1. Shared foundations

  • 9.1.1. Add common::ownership_shape model types, evaluation helpers, and diagnostic argument builders. See ownership shape lints design §11 Shared implementation approach and §16 Diagnostics and localization. Requires 1.1.1.
  • 9.1.2. Add resolved-path classifiers for clone-like operations and wrapper constructors, following the def_path_str plus parsed-segment pattern used by no_std_fs_operations. See ownership shape lints design §11.4 Resolution strategy and §15 Common helper requirements. Requires 9.1.1.
  • 9.1.3. Add exact borrow mappings for the initial concrete type set. See ownership shape lints design §13 Lint 2: owned_param_causes_clone and §15 Common helper requirements. Requires 9.1.1.
  • 9.1.4. Add crate-local MIR summary helpers for local uses, escapes, and boundary classification. See ownership shape lints design §11.5 High-level Intermediate Representation (HIR) prefilter, Mid-level Intermediate Representation (MIR) confirmation and §10 Worked exception model: Servo-style code must stay quiet. Requires 9.1.1 and 9.1.2.

9.2. clone_only_used_by_borrow

  • 9.2.1. Create the lint crate and register CLONE_ONLY_USED_BY_BORROW. See ownership shape lints design §12 Lint 1: clone_only_used_by_borrow. Requires 9.1.1.
  • 9.2.2. Implement HIR candidate prefiltering for clone and to_owned forms. See ownership shape lints design §12 Lint 1: clone_only_used_by_borrow and §11.5 High-level Intermediate Representation (HIR) prefilter, Mid-level Intermediate Representation (MIR) confirmation. Requires 9.2.1 and 9.1.2.
  • 9.2.3. Implement MIR use classification and original-place conflict checks. See ownership shape lints design §12 Lint 1: clone_only_used_by_borrow. Requires 9.2.2 and 9.1.4.
  • 9.2.4. Add machine-applicable suggestions for direct-call and simple-let forms. See ownership shape lints design §12.5 Suggestion policy. Requires 9.2.3.
  • 9.2.5. Add UI pass and fail coverage, including macro and mutable-conflict cases. See ownership shape lints design §18 Testing strategy. Requires 9.2.3 and 1.2.1.

9.3. owned_param_causes_clone

  • 9.3.1. Create the lint crate and register OWNED_PARAM_CAUSES_CLONE. See ownership shape lints design §13 Lint 2: owned_param_causes_clone. Requires 9.1.1.
  • 9.3.2. Implement local call-site clone-pressure collection that records callee identity, argument index, source shape, and retained-source evidence. See ownership shape lints design §13.2.1 Pass A: collect clone-pressure evidence at call sites. Requires 9.3.1, 9.1.2, and 9.1.4.
  • 9.3.3. Implement callee parameter summaries and exported, trait, foreign function interface (FFI), and async suppression rules. See ownership shape lints design §13.2.2 Pass B: summarize callee parameter usage and §13.3 Exemptions. Requires 9.3.2.
  • 9.3.4. Add exact borrow-type help for the initial mapping set without rewriting unsupported signatures. See ownership shape lints design §13.4 Exact borrow mappings and §13.5 Diagnostics. Requires 9.3.3 and 9.1.3.
  • 9.3.5. Add UI coverage for private, exported, trait, async, and non-trigger scenarios. See ownership shape lints design §18 Testing strategy. Requires 9.3.4 and 1.2.1.

9.4. local_shared_ownership

  • 9.4.1. Create the lint crate and register LOCAL_SHARED_OWNERSHIP behind an experimental feature. See ownership shape lints design §14 Lint 3: local_shared_ownership and §19 Rollout plan. Requires 9.1.1.
  • 9.4.2. Implement wrapper-construction detection and non-escape classification for Rc, Arc, interior mutability, and shared-mutable combinations. See ownership shape lints design §14.2 Scope and §14.3 Detection model. Requires 9.4.1, 9.1.2, and 9.1.4.
  • 9.4.3. Implement callback, async, thread, trait-surface, and external-API suppression. See ownership shape lints design §14.6 Servo-style exemptions and §14.7 local_shared_ownership false-positive controls. Requires 9.4.2.
  • 9.4.4. Add diagnostic classes for interior-mutability-only, shared-handle-only, and shared-mutable-wrapper cases. See ownership shape lints design §14.4 Diagnostic classes and §14.5 local_shared_ownership suggestion policy. Requires 9.4.3.
  • 9.4.5. Add Servo-style regression fixtures and framework-boundary negatives. See ownership shape lints design §10 Worked exception model: Servo-style code must stay quiet and §18 Testing strategy. Requires 9.4.4 and 1.2.1.

9.5. Localization, documentation, and promotion

  • 9.5.1. Add Fluent entries and diagnostic argument mappings for all three lints. See ownership shape lints design §16 Diagnostics and localization. Requires 9.2.5, 9.3.5, 9.4.5, and 2.3.3.
  • 9.5.2. Add Welsh and Gaelic smoke coverage for the ownership-shape diagnostics. See ownership shape lints design §16 Diagnostics and localization and §18 Testing strategy. Requires 9.5.1.
  • 9.5.3. Update docs/users-guide.md and docs/developers-guide.md with lint intent, configuration, and rollout guidance for the ownership-shape suite. See ownership shape lints design §17 Configuration and §19 Rollout plan. Requires 9.5.1.
  • 9.5.4. Define promotion criteria from experimental to standard based on UI stability and false-positive tuning across representative repositories. See ownership shape lints design §19 Rollout plan. Requires 9.5.1 and 9.5.2.

10. Async-trait architecture hygiene lints

10.1. Shared family analysis and configuration

  • 10.1.1. Add whitaker::async_trait_hygiene modules behind the dylint-driver feature and implement the AsyncTraitFamilyIndex data model. See async-trait architecture hygiene Dylint suite design §Shared helper placement in Whitaker and §AsyncTraitFamilyIndex fields. Requires 1.1.1.
  • 10.1.2. Implement dyn-use closure, supertrait closure, boxed-future alias detection, sibling lookup, and adapter discovery with deterministic ordering. See async-trait architecture hygiene Dylint suite design §Algorithms and §AsyncTraitFamilyIndex fields. Requires 10.1.1.
  • 10.1.3. Add shared configuration types and loaders for suite defaults, allowlists, alias policy, and Send policy overrides. See async-trait architecture hygiene Dylint suite design §Configuration model. Requires 10.1.1 and 3.6.3.

10.2. Inventory and family-completeness lints

  • 10.2.1. Create the async_trait_concrete_only and async_trait_signature_markers_present lint crates with diagnostics wired to the shared family index. See async-trait architecture hygiene Dylint suite design §Proposed lint table. Requires 10.1.2 and 10.1.3.
  • 10.2.2. Create the async_dyn_family_requires_native_sibling and async_dyn_family_requires_blanket_adapter lint crates, including family-scoped allowlist support. See async-trait architecture hygiene Dylint suite design §Proposed lint table and §Migration waves and default levels. Requires 10.2.1.
  • 10.2.3. Add UI coverage for dyn-required supertrait cases, high-confidence async-trait marker detection, missing sibling traits, and missing blanket adapters. See async-trait architecture hygiene Dylint suite design §Proposed lint table. Requires 10.2.2 and 1.2.1.

10.3. Signature-shape and migration-ratchet lints

  • 10.3.1. Create the async_dyn_direct_impl_prefers_native and async_dyn_boxed_future_alias_required lint crates and implement their family-shape checks. See async-trait architecture hygiene Dylint suite design §Proposed lint table. Requires 10.1.2 and 10.1.3.
  • 10.3.2. Create the native_async_future_must_be_send and native_async_multi_borrow_requires_named_lifetime lint crates, including per-family non-Send policy overrides. See async-trait architecture hygiene Dylint suite design §Proposed lint table and §Pending architectural decisions and trade-offs. Requires 10.3.1.
  • 10.3.3. Add UI coverage for direct dyn impls, raw boxed-future returns, missing + Send, allowed non-Send families, and multi-borrow lifetime fixes. See async-trait architecture hygiene Dylint suite design §Proposed lint table. Requires 10.3.2 and 1.2.1.

10.4. Suite wiring, rollout, and documentation

  • 10.4.1. Add the async-trait hygiene lints either to a specialist suite or to whitaker_suite behind an explicit opt-in feature, and expose selector defaults for inventory versus ratchet waves. See async-trait architecture hygiene Dylint suite design §Suite plumbing and §Migration waves and default levels. Requires 10.2.3, 10.3.3, and 3.6.2.
  • 10.4.2. Add CI and maintainer guidance that capture timing evidence outside lint execution and document crate-local truth versus workspace-level audit limits. See async-trait architecture hygiene Dylint suite design §CI and xtask responsibilities versus lints. Requires 10.4.1 and 4.1.1.
  • 10.4.3. Update docs/users-guide.md and docs/developers-guide.md with the migration-wave model, configuration keys, and suppression guidance for async-trait hygiene findings. See async-trait architecture hygiene Dylint suite design §Configuration model and §Migration waves and default levels. Requires 10.4.1.

11. Test-support dead-code and masked-expectation analysis

11.1. Detector lints and module-graph discovery

11.2. Overlay builds and workspace analysis

11.3. Reporting, whitaker check integration, and validation

12. Architecture and compile-time hygiene enforcement

12.1. Shared policy schema and lint taxonomy

  • 12.1.1. Add shared [workspace.metadata.whitaker] policy parsing for layers, forbidden dependencies, feature islands, compile-hygiene thresholds, and allowlists. See Whitaker and cargo-compile-hygiene technical design §Policy storage and shared configuration model. Requires 3.6.3.
  • 12.1.2. Register Whitaker lint groups and naming conventions for arch_*, hygiene_*, tests_*, and advisory_* rules so CI and users can enable coherent policy sets. See Whitaker and cargo-compile-hygiene technical design §Whitaker lint namespaces and naming conventions and §Severity levels and build gating. Requires 12.1.1.
  • 12.1.3. Document suppression guidance for early, late, and pre-expansion Whitaker lints, including cfg_attr(dylint_lib = "whitaker", ...) and #[allow(unknown_lints)] escape hatches. See Whitaker and cargo-compile-hygiene technical design §Feature-matrix and “unknown lint” ergonomics. Requires 12.1.2.

12.2. Whitaker lint MVP for architecture and source-level hygiene

  • 12.2.1. Implement whitaker::arch_hexagonal_layer_boundary with resolution-aware internal and external dependency checks, and port the Wildside-style regression fixtures into Dylint UI tests. See Whitaker and cargo-compile-hygiene technical design §whitaker::arch_hexagonal_layer_boundary. Requires 12.1.1 and 1.2.1.
  • 12.2.2. Implement whitaker::hygiene_public_api_leaks_optional_dep and whitaker::hygiene_feature_island_breach, including feature-matrix coverage for guarded and unguarded heavy-dependency paths. See Whitaker and cargo-compile-hygiene technical design §whitaker::hygiene_public_api_leaks_optional_dep and §whitaker::hygiene_feature_island_breach. Requires 12.1.1 and 1.2.1.
  • 12.2.3. Implement whitaker::tests_ui_test_macro_outside_app as a pre-expansion lint and keep whitaker::advisory_async_trait_clear_misuse as an advisory-only rule with a conservative mode. See Whitaker and cargo-compile-hygiene technical design §whitaker::tests_ui_test_macro_outside_app and §whitaker::advisory_async_trait_clear_misuse. Requires 12.1.3 and 1.2.1.

12.3. cargo-compile-hygiene tool and checks

  • 12.3.1. Create the cargo-compile-hygiene Cargo subcommand with check, check --json, explain, and baseline --write entrypoints, using cargo metadata --format-version 1 as the primary graph input. See Whitaker and cargo-compile-hygiene technical design §Tool shape and why it is a Cargo subcommand and §Commands and CLI UX. Requires 12.1.1.
  • 12.3.2. Implement the integration_target_budget, heavy_dependency_not_optional, and duplicate_major_version_hotspots checks with shortest-path or hotspot explanations in reports. See Whitaker and cargo-compile-hygiene technical design §Checks. Requires 12.3.1.
  • 12.3.3. Implement the tls_backend_multiplicity and package_boundary_purity checks, including per-configuration graph analysis for feature-matrix runs. See Whitaker and cargo-compile-hygiene technical design §Checks and §Invocation patterns and feature-matrix handling. Requires 12.3.1.
  • 12.3.4. Add human-readable and JSON reporting with finding IDs, metrics, remediation guidance, and CI-oriented exit semantics. See Whitaker and cargo-compile-hygiene technical design §Output formats and example reports. Requires 12.3.2 and 12.3.3.

12.4. CI, rollout, and migration guidance

  • 12.4.1. Add default, minimal, and all-features CI jobs that run both the Whitaker lint set and cargo-compile-hygiene across the same feature matrix. See Whitaker and cargo-compile-hygiene technical design §Invocation patterns and feature-matrix handling. Requires 12.2.3, 12.3.4, and 4.1.1.
  • 12.4.2. Update docs/users-guide.md, docs/developers-guide.md, and CLI-facing workflow documentation with shared policy examples, report interpretation guidance, and migration sequencing for downstream adopters. See Whitaker and cargo-compile-hygiene technical design §Comparison table: what belongs where and §Migration guidance for Axinite and Gauss. Requires 12.2.3 and 12.3.4.
  • 12.4.3. Define staged rollout criteria for promoting selected architecture and compile-hygiene findings from report-only or warn status to CI-failing policy. See Whitaker and cargo-compile-hygiene technical design §Severity levels and build gating and §Integration, CI, and rollout. Requires 12.4.1.