ailang-lang 1.0.1__tar.gz

ailang_lang-1.0.1/AUTHORS.md

@@ -0,0 +1,10 @@
+# AILang Authors
+
+## Maintainer
+
+- Alec Khan
+
+## Contributors
+
+See the full list of contributors at:
+https://github.com/anomalyco/ailang/contributors

ailang_lang-1.0.1/CHANGELOG.md

@@ -0,0 +1,443 @@
+# Changelog
+
+## v1.0.0-M57
+
+### M57 — VS Code Extension Hardening
+
+- **Extension version sync**: Bumped `package.json` to v0.2.0; fixed trailing comma; added icon reference
+- **Code action edits**: Rewrote `compiler/lsp/features/code_actions.py` — quick fixes now generate actual `TextEdit` operations for "import stdlib module" and "remove unused variable" actions
+- **`for` keyword support**: Added to TextMate grammar (`keyword.control.loop.ailang`) and LSP completions (`completion.py`)
+- **Documentation**: `docs/vscode/INSTALLATION.md` (installation guide), `docs/vscode/FEATURES.md` (feature reference), `docs/releases/M57_VSCODE_EXTENSION_REPORT.md`
+- **Extension CHANGELOG**: Updated `extensions/vscode-ailang/CHANGELOG.md` with v0.2.0 entry
+- **Test results**: 103/103 LSP tests passing, zero regressions
+
+## v1.0.0-M56
+
+### M56 — External Adoption Closure (Package Naming + Package Manager Commands)
+
+- **Snake-case package naming**: Resolved naming deadlock — `manifest.py` now accepts `^[a-z][a-z0-9_]*$` (snake_case). Kebab-case accepted with deprecation warning for backward compatibility
+- **Resolver normalization**: `compiler/compilation/resolution.py` tries both kebab-to-underscore and underscore-to-kebab variants when resolving package directories in `lib/`
+- **`ail new` generates `ail.toml`**: Previously only created `main.ail` + README; now creates `ail.toml` and `ail.lock` with snake_case-normalized package name
+- **`ail add`**: Implemented in `tools/ail_package_manager/commands.py` — parses `name@version`, `--path`, `--git`/`--tag`/`--branch`; edits `ail.toml` [dependencies] section
+- **`ail remove`**: Removes dependency lines from `ail.toml` by matching `"<name>"` prefix
+- **`ail update`**: Delegates to `tools.ail_package_manager.installer.install()` for re-resolution
+- **`ail list`**: Reads `ail.toml` dependencies and `lib/` directory to show install status
+- **CLI wiring**: All four commands added to `compiler/cli/main.py` dispatch table, help text, and `tools/ail_package_manager/__main__.py` argparse
+- **Documentation**: `docs/PACKAGE_NAMING_POLICY.md`, `docs/QUICKSTART.md`, `docs/PACKAGES.md`, `docs/research/M56_EXTERNAL_ADOPTION_CLOSURE.md`
+- **Test results**: 32 new tests (19 naming + 13 commands), all passing. Total: 176/176 tests across LSP + naming + commands + CLI
+
+## 0.10.0
+
+### DX-015 — Repository Rename Tool (`ail rename`)
+
+- **`ail rename old_name new_name`**: Repository-wide identifier rename with semantic awareness
+- **Symbol graph scanning**: Parses all `.ail` files via the compiler pipeline to find function declarations, function calls, variable declarations, variable references, parameters, and import path segments matching the old name
+- **`--dry-run`**: Preview changes without modifying files
+- **`--diff`**: Show unified diff of all changes
+- **`--strings`**: Also rename matching string literal values (e.g., map keys)
+- **`--no-verify`**: Skip compiler verification after rename
+- **Atomic file rewriting**: Changes applied from end to start to preserve offsets; temporary `.rename.tmp` file used per file
+- **Rollback bundle**: Every rename creates `.ail/rename/<timestamp>/` with a `manifest.json` and `.orig` backups of every modified file; on failure, all files are automatically restored
+- **Compiler verification**: After successful rename, `ail build` runs on the entry point to detect any compilation errors
+- **Safe identifier validation**: `new_name` must pass `str.isidentifier()` to prevent invalid identifiers
+- **Implementation**: `compiler/rename.py` — ~260 LOC, no modifications to compiler pipeline, parser, or runtime
+
+### DX-016 — Watch Mode (`ail watch`)
+
+- **`ail watch [<entry>]`**: Automatic incremental recompilation on file changes
+- **Filesystem watcher**: Uses `watchdog` library for cross-platform file monitoring (Windows `ReadDirectoryChangesW`, macOS `FSEvents`, Linux `inotify`)
+- **`--poll`**: Polling-mode fallback for network filesystems, Docker, WSL
+- **`--json`**: Machine-readable JSON output for AI tooling integration
+- **`--no-initial`**: Skip the initial full build on startup
+- **Incremental compilation**: `CompilationSession.incremental_recompile()` — on file change, only the changed module and its transitive dependents are re-parsed, re-analyzed, and recompiled. Unchanged modules use their existing ASTs
+- **Dependency invalidation**: `get_transitive_dependents()` computes the full affected set via BFS on the import graph
+- **Cross-module semantic reanalysis**: Only affected modules are deep-analyzed; cross-module export registration for all modules is fast (top-level names only)
+- **SHA-256 change detection**: `FileCache` stores per-file hashes; events that don't change content (editor atomic-save artifacts) are filtered out
+- **Debounce (200ms)**: Threading.Timer-based debounce prevents redundant compiles during AI burst edits
+- **Polling fallback**: Configurable `--poll-interval` (default 500ms), runs in background thread
+- **Implementation**: `compiler/watch.py` — ~370 LOC; `CompilationSession` extended with 6 incremental methods
+
+### Test Results
+
+- **251 total tests** (220 existing + 31 new), all passing
+- DX-015 (ail rename): 19 tests — scan, change computation, apply, rollback, dry-run, diff, string scanning, import scanning, verification
+- DX-016 (ail watch): 12 tests — file hash, cache, incremental compiler initial build, incremental recompile after change, cache update
+- Zero regressions: all existing CLI, session, lexer, parser, AST, IR, semantic, runtime, type-checker tests pass
+
+## 0.8.0
+
+### DX-006 — AILang Dependency Ordering Assistant (`ail order`)
+
+- **Tool implementation**: `tools/ail_order/` — analyzes .ail files for dependency ordering issues
+- **Function discovery**: `discovery.py` extracts function names, calls, and line numbers from source files
+- **Graph analysis**: `graph.py` computes topological levels and detects forward references/cycles
+- **Automatic fix mode**: `fixer.py` reorders functions while preserving comments and formatting
+- **Report generation**: `reporter.py` produces Markdown and JSON reports for CI/LSP integration
+- **CLI integration**: `ail order` command added to compiler CLI with `--json`, `--fix`, `--quiet`, `--stdout` flags
+- **Detection capabilities**: 
+  - Forward reference violations (function called before definition)
+  - Circular function dependencies
+  - Unreachable functions (not reachable from main)
+  - Duplicate function declarations
+- **Project analysis**: Full directory analysis with cross-file insights and report generation
+- **12 acceptance tests**, 8 regression tests, 8 AI validation tests — all passing
+
+## 0.4.0
+
+### DX-007 — AILang Language Server
+
+- **Architecture document**: `docs/architecture/LSP_ARCHITECTURE.md` — JSON-RPC, workspace model, document lifecycle, diagnostics pipeline, feature modules, performance goals, testing strategy
+- **Shared AST utilities**: New `compiler/lsp/utils.py` consolidating 5× duplicated helpers (`walk_ast`, `find_node_at_offset`, `position_to_offset`, `node_range`, `find_references`, `find_enclosing_call`, `callee_name`, `member_access_name`, `find_definition_target`) into a single source of truth
+- **Code duplication eliminated**: `definition.py`, `hover.py`, `references.py`, `rename.py`, `signature_help.py` all refactored to import from `utils.py` — 300+ lines of duplicated code removed
+- **Workspace Symbol Search**: `workspace/symbol` implemented — cross-file symbol search with case-insensitive query matching across functions, variables, and imports
+- **Code Actions foundation**: `textDocument/codeAction` — quick fix scaffolding for import errors, undefined stdlib references, and unused variable warnings
+- **Server capabilities extended**: `workspaceSymbolProvider: true` and `codeActionProvider: true` added to initialize response
+- **103 tests**, all passing (up from 99) — 4 new tests for workspace symbols and code actions
+
+## 0.5.0
+
+### DX-008 — AILang Formatter
+
+- **Architecture document**: `docs/architecture/FORMATTER_ARCHITECTURE.md` — canonical style, CLI flags, token-based formatting, comment handling, idempotency requirement, acceptance criteria
+- **`ail fmt` CLI**: `--diff` (show changes), `--quiet` (errors only), directory-wide formatting, single-file formatting, stdin mode
+- **Token-based formatting**: All 17 token types handled with canonical style rules (4-space indent, space around operators/`:`, newline after `{`, same-line `} else {`)
+- **String-aware comment handling**: Inline comments inside string literals preserved (not misidentified as comments)
+- **Lexer compatibility**: Reuses existing lexer with zero modifications
+- **Repo-wide idempotency**: `--check` mode verifies formatting on all `.ail` files; verified idempotent across the entire repository
+- **82 tests**: 71 formatter tests + 11 CLI tests, all passing
+- **Zero semantic changes**: no modifications to compiler, runtime, or language specification
+
+### M19 — Documentation Canonicalization
+
+- **Consolidated PROJECT_VISION.md + PROJECT_PHILOSOPHY.md** into `VISION_AND_DIFFERENTIATION.md`
+- **Expanded PROJECT_CONSTITUTION.md** with Documentation Rule (§3) and Canonical First Rule (§4)
+- **Created `PRODUCT_ROADMAP.md`** at repository root as the single canonical roadmap
+- **Moved benchmark methodology** from whitepaper section into `ENGINEERING_BENCHMARK_PLAN.md`
+- **Eliminated 7 obsolete/archive files** — ROADMAP.md, PRODUCT_ROADMAP.md (old), etc.
+- **Canonical First Rule** added to AGENTS.md workflow and Constitution
+- **Cross-references updated** across all affected documents
+
+### Quality Gates
+
+- **854 tests**, all passing (772 existing + 82 new formatter)
+- DX-008 (ail fmt): PASS — 82 tests
+- Documentation canonicalization: PASS — no broken links
+
+## 0.7.0
+
+### Engineering Optimization Program (Evidence-Driven Stdlib Evolution)
+
+- **Evidence Analysis**: Reviewed B2–B7 benchmark data. Identified stdlib gaps as #1 friction source (42% of B2 errors).
+- **`file.listdir(path)`**: New stdlib API returning sorted directory entries. Python `os.listdir()` backend. Available via `import file; file.listdir(path)`.
+- **`list.sum(values)`**: New stdlib API returning sum of numeric list items. 3+ independent app pattern threshold met (STDLIB_GAP_ANALYSIS.md).
+- **`list.find_by_key(values, key, value)`**: New stdlib API for finding first map in list with matching property. 3+ app threshold met.
+- **`convert.to_number(value)`**: Fixed — was a no-op identity function. Now correctly converts string/int to integer (same semantics as `to_int`).
+- **`docs/HYPOTHESIS_STATUS.md`**: Created — tracks all 7 engineering hypotheses with evidence status (3 Supported, 1 Partially Supported, 2 Inconclusive, 1 Not Yet Tested).
+- **AGENTS.md updated**: Removed `convert.to_number` no-op pitfall. Updated missing stdlib list to reflect available APIs.
+- **STDLIB_REFERENCE.md updated**: Added documentation for all 4 new/changed APIs. Updated "Known Missing Operations" section.
+
+### Measured Improvement
+
+| Metric | Before (v0.6.x) | After (v0.7.0) | Improvement |
+|--------|:-:|:-:|:-:|
+| B2 L2 (CSV pipeline) | 3 iterations | 1 iteration | **67%** |
+| B2 total | 7 iterations | 5 iterations | **29%** |
+| B2–B6 total | 18 iterations | 16 iterations | **11%** |
+| AILang vs Python ratio | 1.38× | 1.23× | **11% closer** |
+
+### Quality Gates
+
+- All new APIs verified: `file.listdir`, `list.sum`, `list.find_by_key`, `convert.to_number` — build+run confirmed
+- Existing test suite: 18 stdlib tests pass, zero regressions
+- ENGINEERING_EVIDENCE_REPORT.md: updated with before/after comparison
+- HYPOTHESIS_STATUS.md: created with evidence references
+- All P0 APIs implemented (no scope changes)
+
+## 0.6.2
+
+### B2–B7 — Engineering Benchmark Execution
+
+- **B2 Feature Implementation**: 3 levels (sum_even, CSV pipeline, file diff) in AILang + Python — AILang 2.3× more iterations, all compile-time errors
+- **B3 Bug Fix**: 5 bugs (off-by-one, undefined-id, map guard, comparison, infinite recursion) — AILang 1.2× more iterations, 80% first-fix compile rate
+- **B4 Refactoring**: Rename + extract function — parity (1.0×), zero regressions in both languages
+- **B5 Upgrade**: Signature change + CLI conversion — parity (1.0×), zero regressions
+- **B6 Maintenance**: Multi-step feature addition with edge-case handling — parity (1.0×)
+- **B7 AI Context**: Structured guide (AGENTS.md + Playbook) saves 3× iterations (1 vs 3 compile-fix cycles)
+- **15 benchmark artifacts** created: 5 bug-fix files, naive-without-guide comparison, 2 task specs, 10 code modifications
+- **Key stdlib gaps identified**: no `!=` operator, no `listdir`, `convert.to_number` is no-op
+- **Evidence published**: `ENGINEERING_EVIDENCE_REPORT.md` with method, datasets, raw results, summary table, and cost model
+
+### Quality Gates
+
+- **ENGINEERING_EVIDENCE_REPORT.md**: PASS — all 6 benchmarks documented with measurements
+- B2–B7 artifacts: PASS — all build+run verified
+
+## 0.6.1
+
+### B1.1 — AI Provider Integration & Calibration
+
+- **Provider abstraction**: `benchmarks/providers/base.py` — `AIProvider` abstract base class with `complete(prompt)`, `count_tokens(text)`, and `ProviderResult` data model capturing all measurements (request, prompt, response, timing, interaction, cost)
+- **4 provider implementations**:
+  - `OpenAIProvider` — GPT-4o, GPT-4, GPT-3.5 (requires `openai` + `tiktoken`)
+  - `AnthropicProvider` — Claude 3, Claude 3.5 (requires `anthropic`)
+  - `GoogleProvider` — Gemini 1.5, Gemini 2.0 (requires `google-generativeai`)
+  - `LocalProvider` — Ollama, vLLM, llama.cpp via OpenAI-compatible API (requires `openai`)
+- **Provider factory**: `create_provider(name, model, ...)` with registry and auto-detection from environment variables
+- **Calibration module**: `benchmarks/calibration/` — runs identical prompts (short_response, code_understanding, token_counting) across all configured providers to validate measurement infrastructure. Generates immutable calibration reports.
+- **B1 extended**: Three priority modes for token measurement — (1) AIProvider for accurate tokenizer + comprehension prompt, (2) manual `ai_results` dict, (3) approximate 4-char heuristic. B1 version bumped to 0.2.0.
+- **CLI**: `python -m benchmarks calibrate` and `python -m benchmarks list-providers`
+- **Optional dependencies**: Added `[openai]`, `[anthropic]`, `[google]`, `[local]`, `[all]` extras to pyproject.toml
+- **37 new tests**: ProviderResult serialization, interface contract, factory, mocked provider implementations (all 4), token estimation, cost estimation, B1 provider integration, calibration execution, schema stability
+- **81 tests total** in benchmarks/tests/ (44 existing + 37 new)
+
+### Quality Gates
+
+- **935 tests**, all passing (898 existing + 37 new provider/calibration)
+- Provider interface: PASS — all ABC methods implemented, serialization roundtrip verified
+- Calibration: PASS — runs against mock providers, reports errors correctly, serializable output
+
+## 0.6.0
+
+### B1 — Engineering Benchmark Framework
+
+- **Generic measurement framework**: `benchmarks/framework/` — runner, metrics, environ, reporting, dataset modules — reusable by B2-B7
+- **3 canonical datasets**: small (compiler only), medium (stdlib + docs), current_repo (full repository)
+- **Dataset scanning**: Deterministic metadata generation with SHA-256 content hash and repeatability hash
+- **Repository metrics**: Full structural analysis (files, LOC, function count, variable count, import count, dependency depth, symbol density, comment ratio, nesting depth, cyclomatic complexity)
+- **AI comprehension metrics**: Token estimate, context window utilization, comprehension accuracy (placeholder for future AI integration)
+- **Immutable historical results**: Each run stored in `benchmarks/results/<benchmark>/run_<timestamp>/` with measurements, environment snapshot, and human-readable report
+- **Summary aggregation**: `benchmarks/results/summary.json` with all runs across all datasets
+- **CLI**: `python -m benchmarks {setup,b1,list,test}`
+- **44 tests**: 4 runner, 8 dataset, 13 metrics, 9 reporting, 10 B1 integration — all passing
+- **Run IDs**: Timestamp-based, sortable, unique
+- **.gitignore**: `benchmarks/results/` excluded (runtime artifacts); datasets remain version-controlled
+
+### Quality Gates
+
+- **898 tests**, all passing (854 existing + 44 new B1 framework)
+- B1 smoke test: PASS against small, medium, current_repo datasets
+- All 3 datasets scan, validate, and produce metrics
+
+## 0.3.1
+
+### DX-006 — AILang Package Manager
+
+- **Manifest parser**: `ail.toml` parsing via `tomllib` with full validation (package name, semver, dependency format)
+- **`ail init`**: Project initialization — creates `ail.toml`, `main.ail` stub, and `ail.lock`
+- **Local package support**: `path=` dependencies resolved, copied to `lib/<name>/`, full transitive dependency support
+- **Git package support**: `git=` dependencies shallow-cloned, tag/branch/rev checkout, transitive support
+- **Dependency resolver**: Recursive resolution with topological sort and circular dependency detection
+- **Lock file**: `ail.lock` with TOML format, versioned schema, `input_hash` staleness detection, fast replay
+- **Checksum verification**: SHA-256 integrity hashing for all installed packages
+- **Installation engine**: Full orchestration with `--no-lock`, `--offline`, `--frozen-lockfile` flags
+- **Exit codes**: Per TOOLING_ARCHITECTURE.md conventions (0=success, 1=failure, 3=internal error)
+- **Acceptance tests**: 8/8 tests passing covering init, parse, install, and lock file generation
+- **Documentation**: PACKAGE_MANAGER_DESIGN.md approved, tool README created
+
+### M16 — Documentation Architecture Cleanup
+
+- **ADR collision resolved**: Separate ADR files renumbered ADR-001/002/003 → ADR-010/011/012
+- **Status duplication eliminated**: PROJECT_PHASE.md, ROADMAP.md, CURRENT_MILESTONE.md archived; DEVELOPMENT_STATUS.md now canonical
+- **AI guidance consolidated**: MASTER_ENGINEERING_PROMPT.md, FOR_FUTURE_AI.md archived; AGENTS.md canonical
+- **v0.1.0 sprint reports archived**: 21 files moved to `docs/archive/v0.1.0/`
+- **`generated/` added to `.gitignore`**: 9 tracked generated files removed from git tracking
+- **Documentation Ownership Matrix created**: 15 document types with canonical owners
+- **Cross-references updated**: RELEASE_PROCESS.md, AI_MODEL_GUIDE.md, PROJECT_MEMORY.md
+
+## 0.1.0
+
+### Standard Library v1.0
+
+- **json** — `parse(text)`, `stringify(value)` using Python `json` backend
+- **csv** — `parse(text)`, `parse_header(text)`, `stringify(rows)` using Python `csv` backend
+- **time** — `now()`, `timestamp()`, `sleep(ms)`, `format(ts)`
+- **environment** — `get(name)`, `cwd()`, `args()`
+- **random** — `int(min, max)`, `float()`, `choice(collection)`
+- **file** — `exists(path)`, `read(path)`, `write(path, content)`, `append(path, content)`, `remove(path)`
+- **path** — `join(a, b)`, `basename(path)`, `dirname(path)`, `extension(path)`, `normalize(path)`
+- **collections** — `array`, `list`, `map`, `set` with `new`, `get`, `set`, `add`, `contains`, `remove`, `clear`, `keys`, `len`
+- **string** — `concat`, `equals`, `uppercase`, `lowercase`, `length`, `contains`, `starts_with`, `ends_with`, `trim`
+- **math** — `add`, `sub`, `mul`, `div`, `abs`, `min`, `max`
+- **io** — `write`, `writeln`, `println`
+- **convert** — `to_string`, `to_int`, `to_bool`, `to_number`
+- **system** — `exit`
+
+### Performance, Memory & AI Validation (Phase 5B)
+
+- **Stress tests**: 5000 LOC, 10000 LOC, 50 modules, 100 modules
+- **Compile-time benchmarks**: 100/500/1000/5000 LOC with <60s thresholds
+- **Memory benchmarks**: LOC-based with <500MB threshold at 5000 LOC
+- **Determinism**: IR SHA-256 hash verification across 3 compiles
+- **AI validation**: 23 programs generated from public docs, 100% first-pass success
+- **Defect fix**: `convert.to_string` was a no-op; added `__native_to_string` builtin
+
+### Developer Ecosystem Foundation (Phase 6)
+
+- **10 documentation guides** created: Installation, Getting Started, Language Tour, Stdlib Reference, Compiler Architecture, Contributor Guide, Testing Guide, Release Process, Roadmap, Documentation Index
+- **README rewritten** with badges, quick start, full stdlib table, project status
+- **CURRENT_MILESTONE.md** updated to reflect Phase 6 completion
+
+### Quality Gates
+
+- 360 tests, all passing
+- black, ruff, mypy all clean
+
+## 0.1.1
+
+### Documentation Consolidation (Phase 7)
+
+- **Canonical specification**: Created `LANGUAGE_SPEC.md` at repository root — single source of truth (860 lines, 16 sections)
+- **Fixed LANGUAGE_TOUR.md**: Corrected `else` keyword documentation, removed false "no reassignment" claim, removed wildcard import docs, removed `import "string"` path syntax, updated grammar section to reference canonical spec
+- **Archived specifications**: Moved all 9 obsolete design specification files to `archived/specifications/` with ARCHIVED_README.md
+- **Updated cross-references**: INDEX.md, CONTRIBUTING.md, PHASE_5B_REPORT.md, PHASE_6_REPORT.md, MASTER_ENGINEERING_PROMPT.md all now point to canonical spec
+- **CLI improvements**: Fixed PHASE_6_REPORT.md to reference `ail` CLI instead of `python -m compiler`
+- **All quality gates pass**: 374 tests (up from 360), black/ruff/mypy clean
+
+### Validation & Ecosystem Audit (Phase 8)
+
+- **Documentation validation**: 144 code examples tested across 4 documents; all compile and run correctly
+- **Standard Library validation**: All 16 modules tested with edge cases and invalid inputs
+- **Application validation**: 5/5 large applications compile and execute (Expense Tracker, Inventory, Contact Manager, Banking Ledger, Student Management)
+- **Stress validation**: 28/28 tests pass (large LOC, deep nesting, multi-module, determinism)
+- **Benchmark validation**: 25/25 tests pass (determinism, compile time, memory, IR hash)
+- **AI validation**: 23/23 AI-generated programs compile on first pass (100%)
+- **Defect fixes**: Fixed AssertionError crash in AST builder, CLI crash on malformed input, invalid list literal in STDLIB_REFERENCE.md; corrected false float literal and map literal claims in LANGUAGE_SPEC.md
+- **408 tests** (up from 374)
+
+### VS Code Extension (Phase 9)
+
+- **Extension structure**: `extensions/vscode-ailang/` with `package.json`, TextMate grammar, snippets, language configuration
+- **Syntax highlighting**: 15+ token categories (keywords, builtins, stdlib modules, strings, numbers, operators, comments, functions)
+- **Language configuration**: Bracket matching, auto-closing pairs, auto-indentation, folding markers, comment toggling
+- **9 snippets**: `main`, `fn`, `if`, `ifelse`, `import`, `return`, `recur`, `let`, `comment`
+- **Validation**: 12 test `.ail` files covering all syntax features
+
+### Official Formatter (Phase 10)
+
+- **`ail fmt` command**: Format files in-place, `--check` for CI, `--stdin` for pipelines
+- **Single canonical style**: 4-space indentation, same-line braces, spaces around operators, preserved comments
+- **Deterministic and idempotent**: Formatting is stable across multiple runs
+- **27 formatter tests** + 7 CLI tests
+- **Operator precedence parenthesization**: Automatic for correct precedence rendering
+
+### Dogfooding & Real-World Validation (Phase 11)
+
+- **56 AILang applications**: 27 in `apps/`, 29 in `phase11/` — all compile and run
+- **Stdlib addition**: `string.substring(value, start, end)` added
+- **`env_args` fix**: Now returns user args only (strips CLI plumbing)
+- **23 phase11 apps fixed**: Arg index bugs corrected for new `env_args` contract
+- **6 new apps built**: find, payroll, library_mgmt, a_star, expr_eval, markdown_parser
+- **Linter de-chaining**: Fixed chained member-access-on-call patterns in linter
+- **Language freeze declared**: v0.1.x specification frozen — no new keywords, grammar, or syntax changes
+
+### Governance & Philosophy (Phase 11/12)
+
+- **`GOVERNANCE.md`**: Formal proposal process, evidence bars, review process, versioning, backward compatibility, rejected-forever list, freeze policy
+- **`LANGUAGE_EVOLUTION.md`**: Permanent record of all 28 feature requests with dispositions
+- **`PROJECT_CONSTITUTION.md`**: Immutable rules for development
+- **`VISION_AND_DIFFERENTIATION.md`**: Evidence-driven vision, engineering hypothesis, differentiation strategy
+
+### Public Release Preparation (Phase 12)
+
+- **Repository audit**: Removed 9 obsolete/temporary artifacts
+- **GitHub readiness**: LICENSE (MIT), SECURITY.md, CODE_OF_CONDUCT.md, SUPPORT.md, issue templates, PR template
+- **`RELEASE_CHECKLIST.md`**: Formal release checklist for v0.1.1
+- **Documentation audit**: All docs reviewed for broken links, stale examples, version consistency
+- **Example validation**: All examples build, run, and produce documented output
+- **Installation validation**: Verified from clean environment on Windows
+- **VS Code extension**: Verified installation, highlighting, snippets
+- **CI/CD**: All workflows verified
+
+### RC1 Fixes
+
+- **`system.exit()` implemented**: Added `system_exit()` native builtin that calls Python's `sys.exit(code)` to properly terminate the process
+- **`string.substring` documented**: Added to STDLIB_REFERENCE.md and stdlib tables in README.md and LANGUAGE_SPEC.md
+- **Documentation synchronized**: Updated LANGUAGE_TOUR.md to include string module examples and substring workaround
+- **Error code constants**: Centralized PAR001, PAR002, PAR003, LEX001, LEX002, LEX003 constants in diagnostics.py
+- **Test count corrected**: Updated README.md to show 507 passing tests
+
+## 0.3.0
+
+### Developer Experience
+
+- **DX-004 — Benchmark Runner** completed and accepted (auto-discovery, suite modes, configurable repetition, baseline save/compare, regression detection, CI-friendly exit codes, fault-tolerant execution)
+- **DX-005 — Test Generator** completed and accepted (auto-discovery, coverage analysis, three-stage pipeline, intermediate TestCase model, pure Python generators, `--force`/`--dry-run`/`--app` flags, `tests/generated/` separation, dual MD+JSON reports)
+- **tools/common/ — extensions**: Added `hashing.py` (SHA-256 file hashing), `discover_apps()`, `list_py_files()`; `run_ail_build()`/`run_ail_run()` already present
+
+### Quality Gates
+
+- **772 tests**, all passing (up from 658)
+- DX-004 (ail benchmark): PASS — 11 acceptance + 4 regression + 4 AI validation
+- DX-005 (ail testgen): PASS — 9 acceptance + 4 regression + 4 AI validation
+
+## 0.2.1
+
+### Static Analyzer Performance Optimization
+
+- **Replaced 4-pass scan** with single `scan_lines` pass, eliminating redundant function-end detection
+- **Replaced character-by-character** `scan_line_calls_inner` with `collect_calls_inner` using `find_substring` builtin
+- **Replaced `count_calls_file`** (character-by-character scan of all lines) with `build_calls_from_cache` (derives call statistics from per-function cache)
+- **Result:** 3× speedup on hotel_management (252s → 84s), 3× on self-analysis (71s → 24s); self-analysis and large-file analysis now complete (previously never finished)
+
+### Standard Library Additions (ADR-003)
+
+- **`string.find(value, needle)`** — substring search, returns position index or -1
+- **`string.find_from(value, needle, start_pos)`** — substring search with start offset
+- **`string.split(value, delim)`** — split string into list by delimiter
+- Evidence: 8+ independent reimplementations across the app ecosystem satisfied ADR-008 threshold
+
+### Documentation
+
+- **ADR-003** — Decision record for `string.find`/`string.split` stdlib additions
+- **v0.2.1 Release Validation Report** — formal checkpoint with 658 tests, benchmark validation, DX acceptance, before/after performance
+- **PROJECT_PHASE.md** — documents Platform & Developer Experience Engineering phase
+- **PROJECT_CONTEXT.md** regenerated — reflects new stdlib APIs, updated version and test counts
+
+### Quality Gates
+
+- **658 tests**, all passing (up from 624)
+- DX-001 (ail context): PASS — 8 acceptance tests, AI validation
+- DX-002 (ail doctor): PASS — 9 acceptance tests
+- DX-003 (ail static_analyzer): PASS — 8/8 acceptance tests (directory analysis timeout resolved with configurable `--timeout`)
+
+## 0.2.0
+
+### DX Tool #001 — ail context Developer Experience Tool
+
+- **Standalone tool**: `tools/ail_context/` generates `generated/PROJECT_CONTEXT.md` for AI consumption
+- **Single-file context**: All project knowledge consolidated into one LLM-optimized document (~6.5KB)
+- **15 sections**: Project overview, philosophy, architecture, constraints, milestone, ADRs, stdlib, benchmarks, testing, frozen components
+- **Acceptance tested**: 9 functional tests, 6 performance tests, 6 content validation tests, 3 AI validation tests
+- **Zero changes**: No modifications to compiler, runtime, or language specification
+
+### Runtime Optimization #001 — Lexical Variable Lookup Cache
+
+- **`Environment.resolve()`** now caches binding locations per-environment in `_resolve_cache: dict[str, Environment]`, eliminating recursive chain walks on repeated variable lookups
+- **~6× speedup** on the static analyzer benchmark (373s → 19.5s, the primary bottleneck workload identified by Python profiling)
+- **52–64% cache hit rate** across all 5 benchmark apps (dice_roller, hangman_game, inventory_mgmt, kanban, static_analyzer)
+- **Negative caching removed**: initial design cached `NameError` sentinels, but `assign` can create new bindings in ancestor environments, making negative entries stale. Only positive results are cached
+- **`get_cache_info()` introspection** added to `Environment` and `Runtime` for testing
+- **`_CacheStats` instrumentation** added to `Environment` (hits/misses/negative_hits counters) — no semantic effect, used exclusively for profiling
+- **102 regression tests** in `tests/test_scope_cache.py` covering basic resolution, shadowing, recursion, reassignment, modules, edge cases, cache introspection, stress, and correctness invariants
+- **624 tests total** (522 existing + 102 new), all passing
+- **Memory overhead**: ~11 KB for the static analyzer workload (98 cache entries across 18 environments)
+- **No semantic changes** — all existing AILang programs remain fully compatible
+- **Runtime frozen** after this release. No further optimizations planned until community feedback identifies new bottlenecks
+
+## 0.1.2
+
+### Compiler QA — Bug Fix Sprint #001
+
+- **BUG-001 fixed**: Empty `return;` no longer crashes with `AssertionError` — produces "Return statement requires an expression" diagnostic instead
+- **BUG-002 fixed**: Missing initializer `let x = ;` no longer crashes with `AssertionError` — produces "Variable declaration requires an initializer expression" diagnostic instead
+- **BUG-003 fixed**: Module name resolution for bare identifiers (`map.set` as a value, not just in call position)
+  - `_resolve_name` now checks `self._modules` for bare module names before falling through to `NameError`
+  - Module functions are now registered in module environments at initialization time
+- **BUG-004 fixed**: Float literal `3.14` no longer produces cryptic "Identifier node missing token" — emits a clear "Float literals are not supported. Use integer division, e.g. 22 / 7" diagnostic at the lexer level (LEX004)
+- **Regression tests added**: `test_ast_rejects_empty_return`, `test_ast_rejects_missing_initializer` in `test_ast_builder.py`; `test_regression_module_function_as_value` in `test_validation.py`; `test_lexer_rejects_float_literal` in `test_lexer.py`
+- **BUG-005 fixed**: Block-level variable shadowing now works — `_execute_block` creates a new `StackFrame` for each block scope, so `let x` inside an `if`/`else` block properly scopes to that block instead of leaking to the enclosing function
+- **BUG-006 fixed**: Deep recursion `count(1000)` no longer crashes — `Runtime.__init__` raises Python's recursion limit from 1000 to 10000 to accommodate AILang function call overhead
+
+### Quality Gates
+
+- **522 tests**, all passing (up from 521)

ailang_lang-1.0.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 AILang Contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

ailang_lang-1.0.1/MANIFEST.in

@@ -0,0 +1,5 @@
+include LICENSE
+include README.md
+include CHANGELOG.md
+recursive-include stdlib *.ail
+recursive-include examples *.ail