Roadmap

What We Hope to Achieve

The ultimate goal of LaiRu is a strict, complete, drop-in replacement for the Lasso 9 server on modern Linux — one that Lasso 9 applications can run against without needing code changes or workarounds for missing features. That means:

• Full fidelity to the Lasso 9 language specification, including every edge case of multiple dispatch, trait composition, data-member accessor rules, and continuation semantics.
• The complete Lasso 9 standard library — regex, broader bytes and crypto coverage, XML, mail, network APIs, and every historical alias.
• A full admin and security model matching the historical Lasso 9 admin console behaviour.
• A FileMaker connector, executing through the historical XML gateway and/or the modern FileMaker Data API.
• Production-grade FastCGI process management with graceful reload, worker recycling, and deployment telemetry.
• A compatibility report that can accurately describe which Lasso 9 features a given application relies on and whether LaiRu covers them.

We are building this incrementally and dependency-first. LaiRu 0.1 is the foundation; each subsequent pass adds a verifiable slice with test fixtures before moving to the layer above.

---

Current Priorities

The active queue is ordered from lowest to highest dependency:

1. Value graph consumers Active

Broaden historical DB row/result serialisation APIs beyond the bounded row-map helpers, sessions, and security policies already covered. The foundation exists; coverage needs to grow to match real Lasso applications.

2. Data-member edge cases Active

Finish remaining trait interaction, pure map/array temporary assignment roots, exact accessor/setter naming edge cases, and visibility/precedence rules beyond the current generated accessor, listing, and hasMethod slices.

3. Dispatch and trait semantics Active

Extend ranking for remaining default/rest edge cases and ambiguity behaviour beyond duplicate-count diagnostics. Finish historical trait + conflict, provided-method, and requirement compatibility rules.

4. Source and capture diagnostics Active

Broaden token and parser sub-spans beyond the landed operator/member/invocation failures. Attach exact historical capture-frame source context. Improve remapping for cached template and code-fragment spans.

5. Lazy sequences and query runtime Planned

Introduce public lazy sequence and query result objects. Implement the full eacher/queriable protocol beyond the current callback-style each(handler). Lazy barrier-aware query stages, exact group objects, and null-aware aggregate behaviour.

6. Service and FastCGI hardening Active

Finish remaining upload handoff controls. Grow multipart binary boundary and root-policy coverage. Add graceful FPM lifecycle and reload behaviour (worker recycling, deployment telemetry) on top of the landed config, pool, and multiplexing controls.

7. DBAL completion Active

Implement remaining inline result helpers beyond the current map/runtime helper slice. Transaction lifecycle helpers beyond request/inline flag representation. Layout helpers, charset/data conversion edge cases, and older MySQL/MariaDB reset-policy guidance.

8. Runtime lifecycle and tooling Planned

Startup/shutdown/module/app-state hooks. Thread and scheduler primitives when fixtures require them. Packaging hardening, migration documentation, and richer per-feature compatibility reports.

---

Standard Library

The following standard library areas are in scope but not yet implemented:

• Regex — LaiRu currently fails loudly on regex entry points while a regex engine is selected. This is one of the most commonly needed missing pieces.
• XML — parsing and generation via the historical Lasso XML API.
• Mail — SMTP sending via the historical mail methods.
• Network and process APIs — as real application fixtures require them.
• Broader bytes and crypto coverage — additional digest algorithms, HMAC, symmetric encryption, and broader byte-manipulation methods.
• Full date/time coverage — complete format pattern support, time zone database, and all historical date methods.
• All historical aliases — every documented Lasso 9 method spelling.

---

Language Completeness

• Full error object model — complete historical error object and every handler edge case beyond the current typed-failure/rethrow slice.
• Threading primitives — thread creation, synchronisation, and the thread-variable model across threads.
• Complete multiple dispatch — exact historical precedence for every default/rest/type/trait/parent combination.
• Full callback semantics — all Lasso 9 callback forms beyond the current onCreate slice.

---

Parked Items

These are valid Lasso 9 features that are out of the active queue because they need external infrastructure, product decisions, or a separate implementation pass:

• FileMaker connector — execution via the historical XML gateway and/or modern FileMaker Data API. The DBAL boundary is defined and ready; the connector awaits a fixture server and connector strategy decision.
• MySQL 8 RSA auth CI — live CI coverage for caching_sha2_password and sha256_password RSA users against a real MySQL 8 server (MariaDB does not ship these plugins). The code path is unit-tested; live fixtures are parked.
• Admin/security model — the complete Lasso 9 admin console behaviour, credential storage, audit hooks, and role/permission management beyond the current bounded security policy slice.
• SCGI frontend — superseded by the now-active FastCGI responder.

---

Contributing

LaiRu is fixture-first: a feature is only considered done when it has passing test fixtures. The best way to contribute is to:

1. Find a Lasso 9 behaviour that LaiRu doesn’t yet cover correctly.
2. Write a minimal fixture under lasso-rust/tests/fixtures/ that demonstrates the expected behaviour.
3. Implement the feature or fix until the fixture passes.
4. Reference the LassoGuide section that describes the behaviour in your commit message.

The fixture index at lasso-rust/tests/fixtures/INDEX.md tags existing cases by implementation-plan layer and compatibility area.