v1.31.0

Two additions this release: the FluentSchema builder — an instance-based mirror
of the FluentRule factory that lets you drop the FluentRule:: prefix — and a fifth
automatic optimization, worker-global memoization of Laravel's rule-string parsing.
Both are backward-compatible; existing rules() methods and call sites are unchanged.

Added

• FluentSchema builder. Define a schema(FluentSchema $rules) method on a form
  request and chain field starters off the injected $rules builder instead of
  repeating the FluentRule:: prefix on every line. When a request defines both
  schema() and rules(), schema() wins — detection keys off the
  FluentSchema-typed parameter, so an unrelated schema() method is never hijacked.
  For inline use, RuleSet::define(fn (FluentSchema $rules) => [...]) builds a
  RuleSet from the same builder. FluentSchema exposes typed delegates for every
  FluentRule factory (single source of truth on FluentRule) and forwards macros,
  and mirrors the schema(JsonSchema) shape Laravel's AI SDK uses for structured
  output.

• MemoizingValidator. A Validator subclass that memoizes
  ValidationRuleParser::parse() for string rules in a bounded, worker-global static.
  Laravel re-parses each string rule on every internal probe — and, on a validator
  reused across array items, once per item; the memo collapses those repeats to a
  single hash lookup. OptimizedValidator now extends it, so the residual slow path
  of the optimized entry points (HasFluentRules, FluentValidator,
  RuleSet::validate(), HasFluentValidation) picks it up automatically. Output stays
  byte-identical to Laravel; only string rules are cached. The cache is soft-capped and
  reset on overflow, and holds only rule-string → parse-result pairs (never request
  data), so it stays Octane-safe.

Performance

• Per-item slow-rule validation reuses parsed rules across an array.
  RuleSet::validate() and HasFluentRules route per-item slow-rule validators
  through MemoizingValidator, so a large array whose items fall through to Laravel
  (custom Rule objects, cross-field references) parses each rule string once for the
  whole array rather than re-parsing per item. The gain is largest on conditional-heavy
  and slow-rule-heavy payloads; simple fast-checkable arrays are unaffected — they
  already bypass Laravel entirely.
• Custom validator resolvers are preserved. When an application registers a custom
  Validator::resolver(), the per-item path uses that validator unchanged, so any
  resolver-provided behaviour continues to apply.

Internal

• Extracted the factory-state reflection copy shared by the optimized validator paths
  into ValidatorStateCopier, and taught it to carry wildcard implicitAttributes so
  nested-wildcard expansion metadata pairs with the copied rules.

What's Changed

• chore(deps): bump actions/checkout from 6 to 7 by @dependabot[bot] in #18

Full Changelog: 1.30.1...1.31.0

---

Benchmark results

Scenario	Optimizations	Native Laravel	Optimized	Speedup
Product import — 500 items, simple rules	Wildcard, fast-check	232.4ms	3.4ms	~69x
Nested order lines — 1000 orders × 5 line items	Wildcard, fast-check (nested)	2885.6ms	21.9ms	~132x
Event scheduling — 100 items, field-ref dates	Wildcard, partial fast-check	33.3ms	1.0ms	~32x
Article submission — 50 items, custom Rule objects	Wildcard only	10.7ms	3.0ms	~4x
Conditional import — 100 items, 47 conditional fields	Wildcard, pre-evaluation	3449.9ms	51.5ms	~67x
Login form — 3 fields, no wildcards	Fast-check (flat)	0.2ms	0.0ms	~15x