Composable footer opening and file pruning

[joseph-isaacs]

Add FooterOpen and FooterPrune under vortex_file::planning. The first stage reads and validates a footer through the planning IO protocol; the second rejects empty files and files whose statistics prove the predicate cannot match. Surviving files pass to a caller-supplied Next<OpenedFile>.

The stages are separate from existing file-opening and scan implementations and use the file crate's private helpers without widening their visibility. End-to-end tests finish with a diagnostic {start, end} morsel that exists only in test support; this PR does not evaluate file data or switch public scan entry points.

Tests cover cached and large footers, known/unknown size, stable requests, failures, conservative pruning with missing statistics, and footer-only reads. Builds, tests, formatting, and lint checks have not been run.

Third of three draft PRs: traits → simple driver → footer opening/pruning. Based on ji/scan-protocol-driver. Further planners and production scheduling are follow-on work.

Stack order:

1. Scan protocol traits
2. Simple scan planning driver
3. Composable footer opening and file pruning (this PR)

[codspeed]

Merging this PR will degrade performance by 10.75%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

❌ 1 regressed benchmark
✅ 221 untouched benchmarks
⏩ 2403 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	bitpack_blocked_compress_avx2	6.8 µs	7.6 µs	-10.75%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing ji/scan-protocol-footer (de8be02) with ji/scan-protocol-driver (493b7d9)}

Footnotes

1. 2403 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩