Fix memory leak in SpecValidator.iter_errors by caching on instance

[GruffElixir]

Summary

Fixes #523.

SpecValidator.iter_errors was decorated with @lru_cache(maxsize=None). Because @lru_cache was applied to an instance method, the class-level cache dictionary held a strong reference to self (SpecValidator instance) for every validation performed. This prevented garbage collection of the validator and its loaded schemas, causing unbounded memory growth in long-running applications that validate specifications repeatedly.

Solution

1. Removed @lru_cache from SpecValidator.iter_errors.
2. Cached the CachedIterable lazily on the instance (self._cached_errors). This preserves error caching across multiple iterations on the same SpecValidator instance while ensuring instances and schemas are freed once callers release them.
3. Added regression tests verifying both instance-level error caching and validator garbage collection.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.