{
  "schema": 1,
  "captured_on": "2026-07-31",
  "checkpoint": "spline-subdivision-policy",
  "source_revision": "5a3c9914a4dc722803835916ad72c3cf989d9351",
  "parent_revision": "1fdcd06994577d61e41b0a21e3eebc828a5d67a6",
  "implementation_revisions": [
    "5a3c9914a4dc722803835916ad72c3cf989d9351"
  ],
  "frozen_baseline_revision": "dfc4ec5f9546560765bd33d4a88292581869adb8",
  "baseline": "../baselines/2026-07-31/baseline.json",
  "semantic_evidence": {
    "policy_explicit_public_apis": [
      "Curve2::{split_at,subcurve,clamped_subcurve}",
      "CurveView2::{split_at,subcurve,clamped_subcurve}",
      "NurbsCurve2::{split_at,subcurve,clamped_subcurve}",
      "PolynomialSplineCurve2::{split_at,subcurve,clamped_subcurve}"
    ],
    "return_contract": "ExactCurveResult<CurveOutcome<T>>",
    "internal_composition_uses_raw_policy_kernels": true,
    "complete_public_operation_observes_terminal_consumption_once": true,
    "nurbs_policy_propagation": [
      "parameter-range validation",
      "knot insertion",
      "reconstructed carrier validation",
      "clamped piecewise-Bezier reconstruction"
    ],
    "polynomial_policy_propagation": [
      "unit-weight NURBS subdivision",
      "polynomial carrier reconstruction"
    ],
    "top_level_curve2_preserves_lineage_and_policy": true,
    "terminal_parameter": "1 + ((pi + e) - (e + pi))",
    "strict_terminal_results": [
      "Subdivision blocker",
      "Ordering blocker"
    ],
    "approximate_512_terminal_result": "exact split, subcurve, and clamped spline geometry with Approximate512Consumed",
    "later_strict_query_on_same_source_still_blocks": true,
    "approximate_first_cache_cannot_certify_later_strict": true,
    "polynomial_source_interval_preserves_uncertain_ordering_for_terminal_replay": true,
    "full_domain_structural_subcurve_fast_path_certainty": "Certified",
    "returned_geometry_is_not_sampled_or_rounded": true,
    "compatibility_overload_deref_or_implicit_value_shim_added": false,
    "hypermesh_modified": false
  },
  "layout_and_memory_evidence": {
    "NurbsCurve2_bytes": 8,
    "NurbsData2_historical_synthetic_bytes": 520,
    "NurbsData2_candidate_bytes": 512,
    "NurbsData2_delta_bytes": -8,
    "PolynomialSplineCurve2_bytes": 8,
    "PolynomialSplineData2_parent_bytes": 288,
    "PolynomialSplineData2_candidate_bytes": 288,
    "public_handles_remain_one_word": true,
    "decomposition_native_and_rational_evaluator_caches_use": "PolicyEvaluationCache",
    "strict_fast_path_uses_existing_bounded_knot_refinement_cache": true,
    "approximate_refinement_is_uncached_to_prevent_strict_cache_contamination": true,
    "heap_cache_or_buffer_added": false
  },
  "matched_curve_path_benchmarks": {
    "command": "eleven interleaved parent/candidate executions of matched default-feature release curve_path artifacts pinned to CPU 9",
    "benchmark_harness": "benches/curve_path.rs",
    "parent_revision": "1fdcd06994577d61e41b0a21e3eebc828a5d67a6",
    "candidate_revision": "5a3c9914a4dc722803835916ad72c3cf989d9351",
    "full_domain_subcurve": {
      "iterations_per_process": 20000,
      "parent_process_totals_ns": [
        412992,
        411112,
        413352,
        400133,
        440490,
        492336,
        423940,
        406982,
        454069,
        419071,
        419551
      ],
      "candidate_process_totals_ns": [
        355745,
        436410,
        368834,
        374135,
        359745,
        353565,
        365565,
        351576,
        352056,
        370414,
        375635
      ],
      "parent_median_ns_per_call": 20.95355,
      "candidate_median_ns_per_call": 18.27825,
      "median_change_percent": -12.7678,
      "paired_geomean_change_percent": -13.4360
    },
    "native_exact_split": {
      "iterations_per_process": 5000,
      "parent_process_totals_ns": [
        12009666,
        11845897,
        11959159,
        12686060,
        12833420,
        11903443,
        11969639,
        11927912,
        11778412,
        12015276,
        12067262
      ],
      "candidate_process_totals_ns": [
        13254801,
        12522071,
        12012766,
        11850267,
        13200305,
        12219892,
        11855766,
        12148756,
        12260609,
        13037596,
        12293726
      ],
      "parent_median_ns_per_call": 2393.9278,
      "candidate_median_ns_per_call": 2452.1218,
      "median_change_percent": 2.4309,
      "paired_geomean_change_percent": 2.7094
    },
    "nurbs_exact_split": {
      "iterations_per_process": 1000,
      "parent_process_totals_ns": [
        7264322,
        7066725,
        7084404,
        7080565,
        7176137,
        7064575,
        7205586,
        7242203,
        7142790,
        7269191,
        7182727
      ],
      "candidate_process_totals_ns": [
        7085244,
        7282750,
        7104652,
        7152809,
        7193947,
        7188287,
        7249022,
        7179817,
        7146020,
        7817133,
        7209705
      ],
      "parent_median_ns_per_call": 7176.137,
      "candidate_median_ns_per_call": 7188.287,
      "median_change_percent": 0.1693,
      "paired_geomean_change_percent": 1.0255
    },
    "interpretation": "full-domain structural extraction is 12.77 percent faster; the real NURBS subdivision path is neutral, while the native split control carries a visible 2.71 percent paired cost that remains a sentinel"
  },
  "matched_competitive_nurbs_evaluation": {
    "command": "eleven interleaved parent/candidate executions of the comparative NURBS evaluation group pinned to CPU 9",
    "hypercurve_parent_ns_per_iteration": [
      1122.6,
      1103.4,
      1109.8,
      1202.7,
      1098.7,
      1110.8,
      1107.5,
      1122.8,
      1181.7,
      1141.3,
      1185.7
    ],
    "hypercurve_candidate_ns_per_iteration": [
      1103.6,
      1114.4,
      1109.5,
      1098.3,
      1133.4,
      1122.6,
      1114.4,
      1116.9,
      1096.3,
      1159.7,
      1124.8
    ],
    "hypercurve_parent_median_ns_per_iteration": 1122.6,
    "hypercurve_candidate_median_ns_per_iteration": 1114.4,
    "hypercurve_median_change_percent": -0.7305,
    "hypercurve_paired_geomean_change_percent": -1.5100,
    "curvo_parent_ns_per_iteration": [
      57.8,
      57.7,
      58.0,
      70.6,
      57.7,
      58.7,
      59.6,
      58.0,
      58.1,
      60.3,
      60.5
    ],
    "curvo_candidate_ns_per_iteration": [
      57.3,
      57.6,
      57.5,
      57.4,
      58.9,
      57.3,
      57.3,
      57.5,
      57.5,
      58.6,
      58.7
    ],
    "curvo_parent_median_ns_per_iteration": 58.1,
    "curvo_candidate_median_ns_per_iteration": 57.5,
    "curvo_median_change_percent": -1.0327,
    "curvo_paired_geomean_change_percent": -3.1037,
    "interpretation": "Hypercurve exact NURBS evaluation is neutral to slightly faster after the policy/cache cut; Curvo remains much faster under a weaker and numerically different evaluation contract and is not an exactness-parity claim"
  },
  "comparative_candidate_snapshot": {
    "command": "one seven-sample comparative release process pinned to CPU 9",
    "rectangles_ns_per_iteration": {
      "hypercurve": 4760.7,
      "cavalier_contours": 950.4,
      "i_overlay": 1016.7,
      "geo": 1263.1
    },
    "star_64_ns_per_iteration": {
      "hypercurve": 35467.0,
      "cavalier_contours": 27576.7,
      "i_overlay": 34427.0,
      "geo": 36659.3
    },
    "star_256_ns_per_iteration": {
      "hypercurve": 431817.5,
      "cavalier_contours": 488948.7,
      "i_overlay": 512953.4,
      "geo": 526117.7
    },
    "star_1024_ns_per_iteration": {
      "hypercurve": 5106945.0,
      "cavalier_contours": 33765776.7,
      "i_overlay": 14778306.2,
      "geo": 15193855.6
    },
    "capsule_inward_offset_ns_per_iteration": {
      "hypercurve": 38398.7,
      "cavalier_contours": 1713.4
    },
    "certified_cubic_offset_ns_per_iteration": {
      "hypercurve_certified": 56171656.5,
      "hypercurve_chord_fallback": 1493590.9,
      "curvo_heuristic": 14969.7
    },
    "interpretation": "the exact 256- and 1024-vertex Boolean lanes remain faster than the measured finite peers; small Boolean, offset, and NURBS rows remain optimization targets, and offset/evaluation peers do not provide Hypercurve's exact topology and terminal-policy guarantees"
  },
  "matched_pathological_curve_region2": {
    "command": "one warm-up pair followed by eight measured interleaved parent/candidate executions of matched default-feature release pathological_regions artifacts pinned to CPU 9 with HYPERCURVE_PATHOLOGICAL_MODE=boolean and HYPERCURVE_PATHOLOGICAL_CELL_LIMIT=67",
    "cells": 67,
    "operations": 268,
    "completed_operations": 268,
    "blockers": 0,
    "checksum": 6,
    "candidate_pairs": 603,
    "retained_fragments": 3248,
    "point_classifications": 134,
    "parent_build_ms": [
      73.520007,
      74.264655,
      79.046837,
      75.576536,
      73.898191,
      74.937499,
      77.287927,
      74.191980
    ],
    "candidate_build_ms": [
      77.052114,
      74.818527,
      79.100754,
      73.782018,
      78.652944,
      74.778870,
      74.894082,
      73.664877
    ],
    "parent_build_median_ms": 74.601077,
    "candidate_build_median_ms": 74.856305,
    "build_median_change_percent": 0.3421,
    "build_paired_geomean_change_percent": 0.6603,
    "parent_boolean_ms": [
      471.668722,
      475.285294,
      492.035115,
      474.790849,
      468.196861,
      471.854610,
      471.069353,
      471.686532
    ],
    "candidate_boolean_ms": [
      468.130055,
      469.608454,
      468.780450,
      468.254726,
      467.713734,
      480.134642,
      467.630110,
      470.746496
    ],
    "parent_boolean_median_ms": 471.770571,
    "candidate_boolean_median_ms": 468.517588,
    "boolean_median_change_percent": -0.6895,
    "boolean_paired_geomean_change_percent": -0.9304,
    "parent_peak_rss_mib": [
      34.5,
      34.5,
      34.4,
      34.5,
      34.3,
      34.4,
      34.4,
      34.4
    ],
    "candidate_peak_rss_mib": [
      34.5,
      34.4,
      34.5,
      34.4,
      34.5,
      34.4,
      34.5,
      34.4
    ],
    "median_peak_rss_mib": 34.4,
    "interpretation": "all topology and checksums are identical; exact Boolean execution improves 0.93 percent by paired geometric mean, construction is within noise, and process-scale memory is unchanged"
  },
  "release_pathological_binary": {
    "measurement": "matched default-feature release executables measured with GNU size and copied then stripped with strip --strip-all",
    "parent_text_bytes": 5233697,
    "candidate_text_bytes": 5234157,
    "text_delta_bytes": 460,
    "text_delta_percent": 0.0088,
    "parent_data_bytes": 266656,
    "candidate_data_bytes": 264936,
    "data_delta_bytes": -1720,
    "parent_bss_bytes": 2320,
    "candidate_bss_bytes": 3560,
    "parent_total_loadable_bytes": 5500353,
    "candidate_total_loadable_bytes": 5499093,
    "total_loadable_delta_bytes": -1260,
    "total_loadable_delta_percent": -0.0229,
    "parent_file_bytes": 6929840,
    "candidate_file_bytes": 6927440,
    "file_delta_bytes": -2400,
    "parent_stripped_bytes": 5504224,
    "candidate_stripped_bytes": 5502984,
    "stripped_delta_bytes": -1240,
    "stripped_delta_percent": -0.0225,
    "frozen_baseline_total_loadable_bytes": 5531293,
    "delta_from_frozen_baseline_total_loadable_bytes": -32200,
    "delta_from_frozen_baseline_total_loadable_percent": -0.5821,
    "frozen_baseline_stripped_bytes": 5532504,
    "delta_from_frozen_baseline_stripped_bytes": -29520,
    "delta_from_frozen_baseline_stripped_percent": -0.5336,
    "interpretation": "the policy-aware spline cache and explicit public outcomes reduce both loadable and stripped artifact size and remain about 30 kilobytes below the frozen consolidation baseline"
  },
  "caller_audit": {
    "hypercurve_tests_benches_and_ui_example": "all migrated and warning-clean",
    "workspace_old_signature_search": "one direct old NurbsCurve2 clamped_subcurve call and accumulated old native-fragment calls remain in Hyperbrep; other similarly named subcurve calls there belong to Hyperbrep's own pcurve and spatial-curve types",
    "hyperbrep_direct_old_spline_calls": 1,
    "caller_cutover_deferred_until_remaining_phase_1_public_policy_surfaces_stabilize": true,
    "compatibility_shim_added": false,
    "hypermesh_modified": false
  },
  "remaining_policy_audit": {
    "public_curve_spline_subdivision_is_a_strict_wrapper": false,
    "public_spline_construction_evaluation_edit_and_decomposition_strict_surfaces_remain": true,
    "next_action": "inventory every remaining hard-coded STRICT public spline path and migrate the next cohesive operation family without introducing policy-free compatibility entry points"
  },
  "static_callgraph": {
    "command": "/home/tim/Documents/GitHub/workspace/tools/hyper-callgraph/target/release/hyper-callgraph --root /home/tim/Documents/GitHub/workspace --out-dir /tmp/hypercurve-spline-policy-callgraph --crate-name hypercurve --include-examples --include-tests --include-bench --include-fuzz --format json",
    "nodes": 22635,
    "edges": 38791,
    "artifact": "/tmp/hypercurve-spline-policy-callgraph/callgraph.json",
    "curve2_policy_subdivision_nodes_present": true,
    "nurbs_raw_policy_subdivision_nodes_present": true,
    "polynomial_raw_policy_subdivision_nodes_present": true,
    "policy_evaluation_cache_nodes_present": true
  },
  "verification": {
    "cargo_test_all_features": "pass",
    "unit_tests": 262,
    "all_integration_tests": "pass except two explicitly ignored release-scale PCB performance corpora",
    "generated_curve_region_boolean_tests": 7,
    "generated_curve_region_boolean_corpus_seconds": 165.86,
    "pathological_operations": "268/268 decided with no blockers",
    "terminal_nurbs_policy_integration_test": "pass",
    "terminal_polynomial_spline_policy_integration_test": "pass",
    "cargo_clippy_all_features_all_targets_deny_warnings": "pass",
    "cargo_clippy_no_default_features_all_targets_deny_warnings": "pass",
    "cargo_check_fuzz_all_targets": "pass",
    "cargo_doc_all_features_deny_warnings": "pass",
    "hypercurve_ui_cargo_clippy_all_targets_deny_warnings": "pass",
    "cargo_fmt_all_check": "pass",
    "git_diff_check": "pass"
  }
}
