AILANG Implementation Status
Current Version
Check GitHub Releases for the current stable version and CHANGELOG.md for detailed release notes.
Feature Documentation
All implemented features are documented in Design Documents, which is auto-generated from design_docs/implemented/ and always lists the latest versions — no hand-maintained list here.
Component Status
Core Language (Complete)
| Component | Status | Notes |
|---|---|---|
| Lexer | Complete | Unicode, all token types, ~550 LOC |
| Parser | Complete | Recursive descent + Pratt parsing, ~1,200 LOC |
| Type System | Complete | Hindley-Milner, row polymorphism, type classes |
| Evaluator | Complete | Tree-walking interpreter with modules |
| REPL | Complete | History, completion, type checking |
| Effects | Complete | IO, FS, Net, Clock, Rand, DB, AI, Process |
| Modules | Complete | Imports, exports, cross-module calls |
| Bytecode VM | Partial | run --bytecode, ~40× faster than the evaluator; 142 of 205 runnable examples run fully on the VM, the rest fall back to the evaluator (--strict-bytecode fails instead) |
| Go Codegen | Partial | compile --emit-go: values boxed as interface{}, no tail-call elimination; --emit-go-v2 is frozen |
AI-First Features (Complete)
| Feature | Status | Notes |
|---|---|---|
| Inline Tests | Complete | tests [(input, expected)] syntax |
| Structured Errors | Complete | JSON output, error codes |
| Schema Registry | Complete | Versioned JSON schemas |
| Teaching Prompts | Complete | ailang prompt command |
| Multi-Model Evals | Complete | Claude, GPT, Gemini support |
| Semantic Caching | Complete | SimHash + neural embeddings |
Verification (v0.6.1+)
| Feature | Status | Notes |
|---|---|---|
| Contracts | Complete | requires/ensures clauses |
| Policy Mode | Complete | Redundant verification |
| SMT Backend | Complete | Z3 integration, cross-function inlining |
| Bounded Recursion | Complete | --verify-recursive-depth N (Dafny-style unrolling) |
Known Limitations
See Limitations for the full list.
Active bugs:
- Polymorphic arithmetic in lambdas panics (use named functions as workaround)
- Pattern guards parsed but not evaluated
Not yet implemented:
?error propagation operator- Typed quasiquotes
- CSP concurrency (deferred)
Testing & Quality
CI Pipeline
Every push runs comprehensive checks:
| Check | Command | Purpose |
|---|---|---|
| Unit tests | go test ./... | 50+ packages with timeout protection |
| Parser tests | make test-parser | Golden file comparison |
| Coverage gates | make gate-all-packages | Per-package minimum thresholds |
| Golden drift | make check-golden-drift | Detect unintended output changes |
| Fuzzing | make fuzz-parser | Random input testing |
| Import system | make test-imports | Module loading verification |
Run Locally
# Quick test
make test
# Full coverage report
make test-coverage
# Coverage percentage
make test-coverage-badge
# Specific test suites
make test-parser # Parser golden tests
make test-imports # Import system
make test-stdlib-canaries # Standard library health
Example Validation
Example files serve as integration tests. Check current status in README or run:
make verify-examples # Verify all examples
make update-readme # Update pass/fail counts
See Also
- Design Documents — Auto-generated feature documentation
- Roadmap — Planned features
- CHANGELOG — Release history
- Limitations — Known issues