tomoaki/hakorune
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1,958 Commits 1 Branch 0 Tags
39affbf00d30e994a27d7438b7d1a5199867a1f8
Commit Graph

15 Commits

Author SHA1 Message Date
nyash-codex
b71a18495d feat(joinir): Phase 88 - Pattern4 continue + variable step increment 
Continue Pattern 拡張:
- then側の i=i+const 差分加算 + acc更新を許可
- continue_pattern.rs:193 で可変ステップ検出

Dev Router 許可:
- ast_lowerer/mod.rs:92 で normalized_dev feature時に新パターンを有効化

Fixtures & Tests:
- jsonparser_unescape_string_step2_min fixture追加（submodule）
- normalized_joinir_min.rs に shape テスト追加
- shapes.rs に expected shape 定義

Documentation:
- joinir-architecture-overview.md に Phase 88 到達点を追記

Impact:
- Pattern4 continue + 可変インクリメント（i+=1 or i+=2）対応
- _unescape_string 制御構造の土台確立
- normalized_dev tests PASS

Next: _unescape_string 残り複合ループ対応

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
 2025-12-14 00:29:56 +09:00
nyash-codex
8c2bc45be6 refactor(joinir): Phase 85 - Quick wins: loop_patterns removal, DebugOutputBox, dead_code audit 
Quick Win 1: Remove loop_patterns_old.rs (COMPLETED)
- Deleted obsolete legacy loop pattern dispatcher (914 lines)
- All patterns (Break/Continue/Simple) now in modular loop_patterns/ system
- Moved helper functions (has_break_in_loop_body, has_continue_in_loop_body) to analysis.rs
- Updated loop_frontend_binding.rs to remove fallback
- Verified zero regressions: 974/974 lib tests PASS

Quick Win 2: DebugOutputBox consolidation (COMPLETED)
- New module: src/mir/join_ir/lowering/debug_output_box.rs (170 lines)
- Centralized debug output management with automatic HAKO_JOINIR_DEBUG checking
- Refactored 4 files to use DebugOutputBox:
  - condition_env.rs: 3 scattered checks → 3 Box calls
  - carrier_binding_assigner.rs: 1 check → 1 Box call
  - scope_manager.rs: 3 checks → 3 Box calls
  - analysis.rs: Updated lower_loop_with_if_meta to use new pattern system
- Benefits: Consistent formatting, centralized control, zero runtime cost when disabled
- Added 4 unit tests for DebugOutputBox

Quick Win 3: Dead code directive audit (COMPLETED)
- Audited all 40 #[allow(dead_code)] directives in lowering/
- Findings: All legitimate (Phase utilities, future placeholders, API completeness)
- No unsafe removals needed
- Categories:
  - Phase 192 utilities (whitespace_check, entry_builder): Public API with tests
  - Phase 231 placeholders (expr_lowerer): Explicitly marked future use
  - Const helpers (value_id_ranges): API completeness
  - Loop metadata (loop_update_summary): Future phase fields

Result: Net -858 lines, improved code clarity, zero regressions
Tests: 974/974 PASS (gained 4 from DebugOutputBox tests)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
 2025-12-13 19:25:11 +09:00
nyash-codex
acb6720d9b Phase 61: structural if-sum+break lowering (dev-only) 2025-12-12 22:15:41 +09:00
nyash-codex
80e952b83a feat(joinir): Phase 54 SELFHOST-SHAPE-GROWTH - 構造軸育成 + 偽陽性観測 
Phase 53 成果を踏まえ、構造シグネチャ軸を 5+ に育て、
偽陽性観測テストで name ガード縮小準備を整えた。

方針変更: 新ループ追加 → 構造軸育成 + 偽陽性率測定に焦点変更
- 理由: Phase 53 で selfhost P2/P3 実戦パターン追加済み
- 焦点: 既存ループに対する構造軸拡張 + 精度測定

主な成果:

1. 構造軸 5+ 達成:
   - carrier 数
   - carrier 型
   - Compare パターン
   - branch 構造
   - NEW: Compare op 分布 (count_compare_ops ヘルパー)

2. 偽陽性観測テスト追加:
   - test_phase54_structural_axis_discrimination_p2()
   - test_phase54_structural_axis_discrimination_p3()

3. 重要な発見 - 偽陽性率 ~50%:
   - P2: selfhost P2 が正しく検出されず (name ガード依存)
   - P3: selfhost P3 が Pattern4ContinueMinimal と誤検出 (構造的類似性)
   - 結論: 構造判定のみでは分離不十分、name ガード必須と判明

変更内容:

- shape_guard.rs (+80 lines):
  - count_compare_ops() 構造軸ヘルパー追加
  - detect_shapes() pub 化 (テストから呼び出し可能に)
  - SelfhostVerifySchemaP2/SelfhostDetectFormatP3 enum 追加 (将来用)

- normalized_joinir_min.rs (+110 lines):
  - 偽陽性観測テスト 2 個追加 (P2/P3 各1)
  - canonical shapes vs selfhost shapes 構造判定精度測定

- phase49 doc (+200 lines):
  - Phase 54 節完成版
  - 偽陽性分析結果記録
  - name ガード縮小方針明記

- enum 拡張対応:
  - bridge.rs (+8 lines)
  - normalized.rs (+8 lines)
  - ast_lowerer/mod.rs (+2 lines)

偽陽性観測結果 (2025-12-12):
- P2 構造判定: selfhost P2 検出失敗 → name ガード必須
- P3 構造判定: selfhost P3 が Pattern4 と誤判定 → 構造的類似性問題
- 総合: 偽陽性率 ~50% → 構造軸 5 本では不十分

次フェーズ方針 (Phase 55+):
- Phase 55-A: 条件複雑度軸追加 (BinOp/UnaryOp ネスト深度)
- Phase 55-B: 算術パターン軸追加 (Mul/Sub/Div 出現パターン)
- Phase 56: selfhost 実戦ループ追加 (6 本以上蓄積)
- Phase 57: 誤判定率 < 5% 達成後に name ガード縮小開始

name ガード撤去条件 (Phase 57):
- 構造軸 8+ 本確立
- selfhost P2/P3 各 6 本以上蓄積
- 誤判定率 < 5% 達成
- 複合的特徴量ベース判定実装

回帰テスト: ✅ 939 PASS, 0 FAIL (既存挙動不変)

Files Modified: 8 files
Lines Added: ~408 lines (net)
Implementation: Pure additive (feature-gated)

Phase 54 完了！構造軸育成・偽陽性観測基盤確立！
 2025-12-12 17:12:58 +09:00
nyash-codex
7b0db59100 feat(joinir): Phase 53 - SELFHOST-NORM-DEV-EXPAND implementation 
Expanded selfhost dev Normalized target with 2 practical P2/P3 loop variations,
strengthened structural signature axis, and implemented two-stage detection.

Key Changes:

1. Documentation (phase49-selfhost-joinir-depth2-design.md +128 lines):
   - Added Phase 53 section with candidate selection rationale
   - Documented two-stage detector strategy (structural primary + dev-only name guard)
   - Defined structural axis strengthening (carrier count/type, branch patterns)

2. Fixtures (+210 lines):
   - selfhost_args_parse_p2.program.json (60 lines): P2 with String carrier + conditional branching
   - selfhost_stmt_count_p3.program.json (150 lines): P3 with 5 carriers + multi-branch if-else

3. Structured Builders (fixtures.rs +48 lines):
   - build_selfhost_args_parse_p2_structured_for_normalized_dev()
   - build_selfhost_stmt_count_p3_structured_for_normalized_dev()

4. ShapeGuard Two-Stage Detection (shape_guard.rs +80 lines):
   - Added SelfhostArgsParseP2/SelfhostStmtCountP3 to NormalizedDevShape enum
   - Implemented is_selfhost_args_parse_p2(): P2 core family + name guard
   - Implemented is_selfhost_stmt_count_p3(): 2-10 carrier check + name guard
   - Updated capability_for_shape() mappings

5. Bridge Integration (bridge.rs +8 lines, normalized.rs +10 lines):
   - Added shape handlers delegating to existing normalizers
   - Added roundtrip reconstruction handlers

6. Entry Point Registration (ast_lowerer/mod.rs +2 lines):
   - Registered selfhost_args_parse_p2/selfhost_stmt_count_p3 as LoopFrontend routes

7. Dev VM Comparison Tests (normalized_joinir_min.rs +40 lines):
   - normalized_selfhost_args_parse_p2_vm_bridge_direct_matches_structured()
   - normalized_selfhost_stmt_count_p3_vm_bridge_direct_matches_structured()

8. Test Context Fix (dev_env.rs):
   - Added thread-local test context depth counter
   - Fixed deadlock in nested test_ctx() calls via reentrant with_dev_env_if_unset()

Structural Axis Growth:

P2 family:
- Carrier count: 1-3 (unchanged)
- NEW: Type diversity (Integer/String mixed)
- NEW: Conditional branching patterns (Eq-heavy comparisons)

P3 family:
- NEW: Carrier count upper bound: 2-10 (was 2-4)
- NEW: Multi-branch if-else (5+ branches with nested structure)
- NEW: Complex conditional patterns

Test Results:
- normalized_dev: 40/40 PASS (including 2 new tests)
- lib regression: 939 PASS, 56 ignored
- Existing behavior unchanged (normalized_dev feature-gated)

Phase 53 Achievements:
✅ P2/P3 each gained 1 practical variation (2 total)
✅ Two-stage detection: structural primary + dev-only name guard
✅ Structural axis expanded: 4 axes (carrier count/type/Compare/branch patterns)
✅ All tests PASS, no regressions
✅ Test context deadlock fixed (0.04s for 29 tests)

Files Modified: 14 files
Lines Added: ~516 lines (net)
Implementation: Pure additive (feature-gated)

Next Phase (54+):
- Accumulate 6+ loops per P2/P3 family
- Achieve 5+ stable structural axes
- Target < 5% false positive rate
- Then shrink/remove name guard scope
 2025-12-12 16:40:20 +09:00
nyash-codex
386cbc1915 Phase47-B/C: extend P3 normalized shapes and quiet dev warnings 2025-12-12 07:13:34 +09:00
nyash-codex
7200309cc3 feat(joinir): Phase 48-A - P4 (continue) Normalized minimal implementation 
Pattern4 (continue) integration into Normalized JoinIR pipeline complete.

Key changes:
- P4 minimal fixture: skip i==2 pattern, single carrier (acc)
- ShapeGuard: Pattern4ContinueMinimal detector (structure-based)
- StepScheduleBox: ContinueCheck step (eval order: HeaderCond → ContinueCheck → Updates → Tail)
- normalize_pattern4_continue_minimal(): Delegates to P2 (95% infrastructure reuse)
- Tests: 4 integration tests (normalization/runner/VM Bridge comparison×2)

Design validation:
- P4 (continue) = reverse control flow of P2 (break)
- Same loop_step(env, k_exit) skeleton
- Same EnvLayout/ConditionEnv/CarrierInfo infrastructure
- Only difference: evaluation order and control flow direction

Architecture proof:
- Normalized JoinIR successfully handles P1/P2/P3/P4 uniformly
- Infrastructure reuse rate: 95%+ as designed

Tests: 939/939 PASS (+1 from baseline 938, target exceeded!)

Files modified: 10 files (~305 lines added, pure additive)
- pattern4_continue_min.program.json (NEW +126 lines) - P4 fixture
- fixtures.rs (+31 lines) - P4 fixture loader
- shape_guard.rs (+60 lines) - Shape detection
- step_schedule.rs (+18 lines) - Schedule + test
- normalized.rs (+35 lines) - Normalization function
- loop_with_break_minimal.rs (+4 lines) - ContinueCheck handler
- bridge.rs (+5 lines) - VM bridge routing
- ast_lowerer/mod.rs (+2 lines) - Function registration
- normalized_joinir_min.rs (+84 lines) - Integration tests
- CURRENT_TASK.md (+13 lines) - Phase 48-A completion

Next steps:
- Phase 48-B: Extended P4 (multi-carrier, complex continue)
- Phase 48-C: Canonical promotion (always use Normalized for P4)
 2025-12-12 06:31:13 +09:00
nyash-codex
ed8e2d3142 feat(joinir): Phase 248 - Normalized JoinIR infrastructure 
Major refactoring of JoinIR normalization pipeline:

Key changes:
- Structured→Normalized→MIR(direct) pipeline established
- ShapeGuard enhanced with Pattern2 loop validation
- dev_env.rs: New development fixtures and env control
- fixtures.rs: jsonparser_parse_number_real fixture
- normalized_bridge/direct.rs: Direct MIR generation from Normalized
- pattern2_step_schedule.rs: Extracted step scheduling logic

Files changed:
- normalized.rs: Enhanced NormalizedJoinModule with DevEnv support
- shape_guard.rs: Pattern2-specific validation (+300 lines)
- normalized_bridge.rs: Unified bridge with direct path
- loop_with_break_minimal.rs: Integrated step scheduling
- Deleted: step_schedule.rs (moved to pattern2_step_schedule.rs)

New files:
- param_guess.rs: Loop parameter inference
- pattern2_step_schedule.rs: Step scheduling for Pattern2
- phase43-norm-canon-p2-mid.md: Design doc

Tests: 937/937 PASS (+6 from baseline 931)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2025-12-12 03:15:45 +09:00
nyash-codex
a4756f3ce1 Add dev normalized→MIR bridge for P1/P2 mini and JP _atoi 2025-12-11 22:12:46 +09:00
nyash-codex
af6f95cd4b Phase 33 NORM canon test: enforce normalized dev route for P1/P2/JP mini 2025-12-11 20:54:33 +09:00
nyash-codex
a3d5bacc55 Phase 30.1 & 73: Stage-3 features env and JoinIR flag cleanup 2025-11-30 14:30:28 +09:00
nyash-codex
3af98964ed feat(joinir): Phase P2-P4 Loop Pattern boxification 
- P2: Create loop_patterns/ module structure
  - common.rs: shared helpers (ParsedProgram, LoopContext, etc.)
  - simple.rs: generic loop lowering
  - filter.rs, print_tokens.rs: delegate to simple
  - break_pattern.rs, continue_pattern.rs: new modules

- P3: Create LoopFrontendBinding layer
  - Separate function name detection from pattern lowering
  - detect_loop_pattern() for pattern classification

- P4: Break/Continue pattern migration
  - Move Break/Continue handling from loop_patterns_old.rs
  - Add LoopPattern::Break and LoopPattern::Continue variants

- Design: 3-layer architecture
  - LoopFrontendBinding: function name → LoopPattern
  - loop_patterns: LoopPattern → JoinIR
  - Bridge/VM: Box semantics

All 56 JoinIR tests pass.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
 2025-11-29 09:04:18 +09:00
nyash-codex
fd83903f87 feat(joinir): Phase P1 If Handler boxification - 40% code reduction 
## Summary
Refactored loop-internal If statement handling into a boxified module
structure, achieving 154-line reduction (40%) from stmt_handlers.rs
with zero regression.

## Implementation
- Created if_in_loop/ module (9 files, ~480 lines)
  - pattern.rs: IfInLoopPattern enum (5 variants)
  - lowering/{empty,single_var_then,single_var_both,conditional_effect,unsupported}.rs
- Replaced lower_if_stmt_in_loop() (154 lines) with lower_if_stmt_in_loop_boxified()
- Made StatementEffect pub(crate) for module visibility

## Pattern Classification
1. Empty: condition-only check (no side effects)
2. SingleVarThen: if { x = a } → x = cond ? a : x
3. SingleVarBoth: if { x = a } else { x = b } → x = cond ? a : b
4. ConditionalEffect: if pred(v) { acc.push(v) } (filter pattern)
5. Unsupported: complex cases (Phase 54+)

## Test Results
✅ 56 JoinIR tests PASSED (0 failed, 0 regression)
✅ Build successful (1m 02s)
✅ Improved maintainability (easier to add new patterns)

## Files Changed
- src/mir/join_ir/frontend/ast_lowerer/
  - if_in_loop/ (9 new files)
  - mod.rs (+2 lines: module + pub use)
  - stmt_handlers.rs (-154 lines: 40% reduction)
- CURRENT_TASK.md (+5 lines: Phase P1 記録)
- docs/development/refactoring/p1-if-handler-boxification-plan.md (new)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
 2025-11-29 07:20:56 +09:00
nyash-codex
e27934d91a feat(joinir): Phase 52-53 LoopFrontendBinding JSON + Statement Handlers 
Phase 52: LoopFrontendBinding JSON generation fixes
- Add receiver_to_json() for Field node structure (me.tokens)
- Add needs_me_receiver() for instance method detection
- Fix "condition" → "cond" key for JoinIR Frontend
- Add me parameter propagation in loop_patterns.rs
- Add JoinIR-compatible type fields in ast_json.rs
  - Variable → "type": "Var"
  - Literal → "type": "Int"/"Bool" (literal_to_joinir_json)
  - BinaryOp → "type": "Binary"/"Compare" (is_compare_op)
  - MethodCall → "type": "Method"

Phase 53: Statement Handler module for loop body
- NEW: stmt_handlers.rs with StatementEffect type
- Support: Local, Assignment, Print, Method, If statements
- If lowering: single variable update → Select instruction
- Remove hardcoded assert in loop_patterns.rs
- Replace with generic lower_statement() calls

Test results: 56 JoinIR tests PASS, 7 loop_frontend_binding tests PASS

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
 2025-11-29 04:42:16 +09:00
nyash-codex
447bbec998 refactor(joinir): Split ast_lowerer and join_ir_vm_bridge into modules 
ast_lowerer.rs → ast_lowerer/ (10 files):
- mod.rs: public surface + entry dispatch
- context.rs: ExtractCtx helpers
- expr.rs: expression-to-JoinIR extraction
- if_return.rs: simple if→Select lowering
- loop_patterns.rs: loop variants (simple/break/continue)
- read_quoted.rs: read_quoted_from lowering (Phase 45-46)
- nested_if.rs: NestedIfMerge lowering
- analysis.rs: loop if-var analysis + metadata helpers
- tests.rs: frontend lowering tests
- README.md: module documentation

join_ir_vm_bridge.rs → join_ir_vm_bridge/ (5 files):
- mod.rs: public surface + shared helpers
- convert.rs: JoinIR→MIR lowering
- runner.rs: VM execution entry (run_joinir_via_vm)
- meta.rs: experimental metadata-aware hooks
- tests.rs: bridge-specific unit tests
- README.md: module documentation

Benefits:
- Clear separation of concerns per pattern
- Easier navigation and maintenance
- Each file has single responsibility
- README documents module boundaries

Co-authored-by: ChatGPT <noreply@openai.com>

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
 2025-11-28 17:42:19 +09:00