perf: accelerate incremental API indexing

[cmilesio]

What

• Add RunCached with content-validated persistent artifacts and incremental typed-schema reuse.
• Parallelize deterministic source parsing and limit package work to selected contracts.
• Publish unchanged-only artifacts under crash-safe platform locks and scavenge only provenance-marked temporary files.
• Add a representative 326-file regression fixture and refresh the recorded framework benchmark snapshot.

Why

Typed schema loading made large API indexes take seconds. This restores a fast development loop without reusing stale output after source, dependency, environment, local-replacement, cache-corruption, or concurrent-publication changes.

Validation

• Root, docs, and examples tests/vet/tidy; root race and coverage passes.
• Compile checks for Darwin, Windows, AIX, Plan 9, JS/Wasm, and WASI.
• Representative medians: 19.9 ms exact hit; 84.8 ms controller edit.
• Fresh seven-sample HTTP benchmark recording and deterministic SVG verification.

Compatibility and operations

RunCached is additive, Run remains available, and the minimum Go version is unchanged. Publication now retains .webindex-artifacts.lock in output directories. Case-only artifact destinations are rejected because they alias on common filesystems; use unambiguously distinct names. Unsafe build configurations fall back to uncached indexing. JS/WASI publication locking is process-local because those targets expose no portable crash-safe cross-process lock.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 89.21569% with 22 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
webindex/index_cache_build_identity.go	92.61%	8 Missing and 5 partials ⚠️
webindex/artifact_lock.go	54.54%	3 Missing and 2 partials ⚠️
webindex/artifact_lock_flock.go	60.00%	2 Missing ⚠️
webindex/index_cache.go	50.00%	1 Missing and 1 partial ⚠️

📢 Thoughts on this report? Let us know!