perf(fts): bound classic conjunction scoring in dense windows

[BubbleCal]

Classic 4/5-term AND evaluation can decode and score every follower even when a partial exact score already proves the candidate cannot compete. Applying score-first evaluation universally is costly when posting overlap is low. This PR enables it only inside a metadata-certified dense range.

• Prove continuous coverage from each current posting block's first/last/count metadata, while retaining actual follower membership checks.
• Build conservative suffix bounds only for eligible live windows and lazily cache exact BM25 norm addends.
• Preserve canonical query-order f32 scores and inclusive tie behavior.
• Reuse an unbounded zero-floor window and invalidate it when local, shared, or compound floors become positive; clear eligibility before phrase evaluation.

The classic path remains used by compound WandCursor and explicit classic searches after #9030 enables standalone wide-query bulk routing.

Most of this diff is regression coverage: approximately 327 production-section additions and 1,334 test additions. Shared fixtures retain the distinct dense/sparse, malformed-metadata, short-tail, high-DocID, score-bit, tie, norm-fallback, seek/shallow, and floor-transition cases. No isolated main A/B result is claimed for this layer.

Validation

• Per-layer inverted-index suite: 603 passed.
• Required all-workspace Clippy and formatting: passed.
• PR metadata/gatekeeper checks: pending. Full hosted test workflows only run for main/release/** bases; this intermediate layer has been tested and linted locally. The full hosted matrix will run after its base is retargeted to main following the parent merge.

Stack, 3/4: #9031 → #9032 → this PR → #9030. Base: #9032. Uses the lazy posting-cursor support introduced in #9031.

Merge in stack order; retarget/restack the remaining PRs after their parent merges.

[lance-gatekeeper]

⚠️ Gate recommendation: approve with a non-blocking risk.

The rebase onto main leaves the score-first implementation unchanged. Its strict eligibility, conservative suffix bounds, canonical score reconstruction, and focused parity/transition coverage continue to support correctness.

The non-blocking risk remains performance: no isolated latency/CPU comparison establishes the net benefit or possible short-tail overhead. Keep the scope narrow until representative classic 4/5-term workloads confirm the trade-off; if the incremental gain is small, progressive scoring after full DocID alignment is the simpler alternative.