fix(schema): compare inline reference results

[gaborbernat]

The schema differentials in #1016 need XSLT result comparison and seed-tree mutation. This adds inline XSLT and RELAX NG families alongside the existing XSD runner, using libxslt and libxml2 as references.

Subtree swaps repair references against recipient declarations. Broken references check rejection at compilation or application. Dictionary seeds inherit the specified namespace bindings. Text output compares exact strings; XML output compares expanded names and tree content.

The reference cases exposed accepted undefined RELAX NG references and rejected foreign annotations. RELAX NG sections 4.1 and 4.18 require annotation removal and reference resolution before unused-definition removal. Compilation now follows those rules on its private tree, preserving caller-supplied nodes.

The added compiler checks increase final-head TOTAL instruction counts by 16.56% on the large compile/reuse workload and 13.55% on the small one against the reported base. A fused annotation scan increased both compiler costs, so this retains the original correction. Validation increases remain below 0.4% across the measured RNG families.

The import-root canary and optional evaluation counters remain outstanding in #1016.

[codspeed]

Merging this PR will regress 3 benchmarks

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 1 improved benchmark
❌ 3 regressed benchmarks
✅ 577 untouched benchmarks
⏩ 32 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	test_feature[compile-rng-reuse]	992.4 µs	1,118.5 µs	-11.28%
❌	test_feature[select-relative-sibling]	42.3 µs	45.4 µs	-6.92%
❌	test_feature[compile-rng-reuse-small]	95.7 µs	101.6 µs	-5.75%
⚡	test_feature[shadow-slot-comments]	111.4 µs	83.5 µs	+33.46%

Tip

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

---

_{Comparing gaborbernat:fix/xslt-reference-1016 (2ec126a) with main (83bd497)}

Footnotes

1. 32 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. ↩