It writes the [Key] layout plus non-public keyed members, and registration now refuses any of those that
reaches an object list, so no element id can be dropped. The comparison is made lazily, only after an
ahead resolver returned a non-standard formatter, and an exception from it reads as not equal.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
MessagePack writes a non-public [Key] member only under AllowPrivate, and the element-id walk reads
public members, so the ids such a list holds could never be checked. Graph entries carry Hidden; a hidden
list or a hidden edge whose child reaches adds one problem and nothing to the id path.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
IsLeaf, IsObjectType and DictionaryValueType are deleted. Classify(declared) says Leaf, Map
(IDictionary<,> or IReadOnlyDictionary<,>), Sequence or Owner; a keyed struct is walked as an owner and
is never an element. Plan and Holds take the declared root type and follow declared member types, so a
member declared object or a subclass's extra members under a base-typed member are neither minted into
nor read. Minted path text is unchanged.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Graph edges and lists carry the collection types between member and child; Settle adds them to the path so
[MessagePackFormatter] on a collection class and a resolver claim on List<Elem> are judged. Declares replaces the four
MemberInfo.IsDefined member reads, which ignore inherit for properties. Survey.Judged dedupes the unkeyed-element refusal.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
_readHoldingIds and the pull's Holds walk are removed in the same commit that adds
SingleFileBackingStore.CopiedHoldIds: under the store's lock, while the header on disk is
unmarked, every plain copied record the writer's registry claims and whose type reaches an
object list is decoded, and the first holding an id marks the store. Positive-only, never
refuses a write. CultMesh's typed write attaches the registry its payload was encoded with;
the raw write attaches none. Contract text states the rule and its residual.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Stryker survivors on the id-path walk: dropping the union-subtype edge, dropping the
element base from the path. Both now have a test that fails without them.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The single-file store records, per pull, which plain records decoded holding a non-empty
element id and at what version. A write whose copied record is still at that version
marks the header (v3). The record is positive-only and version-gated, so nothing a load
records can unmark a store. The directory store passes false: only C# writes it.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Registration: a concrete element type of an object list that is not keyed, and
[MessagePackFormatter] on any type or member on the id path or on the id member, are
refused. The id path is a fixed point over the walk's edges. The codec refuses, once per
document type and on every encode and decode path, a resolver ahead of StandardResolver
that formats a type on that path. UnionArmGuardTests moves StubArm out of the list union.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The decode_stack Windows run is at 10a39dac (tests unchanged since 4b55c7d3);
the head after this commit awaits a Soul rerun.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
SpanFoldCoversAccelerationAlone draws spans with zero midpoint velocity and an acceleration along the gradient
that carries n across zero, so the fold's acceleration term and its gradient bound are the only thing holding
the branch departure under fold; two hand mutants survived without it.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Legacy and v1 reads on 64 KiB, redb pull_all and compare-and-swap on 104 KiB,
and one Largest level on 88 KiB, the sizes where a plain from_slice or a missing
outer maybe_grow dies in release as well as debug.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The README and decode_payload claimed the calling thread needs only its own
frames. Measured, the door has a payload-independent floor (release 56 KiB vs
16 for a same-signature non-decoding function; debug 104 vs 88). Say so, and
replace "Windows not yet verified" with the one by-hand Starfire run of
decode_stack on x86_64-pc-windows-msvc.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
SpanStageGradientMatchesDifferences now also compares field.w and field.xyz with an octave sum assembled
in the test from Stage.Point and snoise_grad, so shift, offset, octaveStep, M, lacunarity and gain are pinned
numerically, which a value/gradient self-consistency check cannot do.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The octave stages of a warp chain over a span, in C#: stretched, shifted, offset octave sums whose
field equals fbm_grad and ridged_grad bit for bit at the identity transform and whose curvature
and fold bounds derive from SNOISE_LIPSCHITZ, SNOISE_HESSIAN and SNOISE_THIRD only (local tier from
the midpoint Hessian's Frobenius norm plus the third-derivative term, capped by the global pin).
SpanStageTests encloses the Hessian norm and the branch departure over seeded spans in both regimes
and checks the gradient against differences and the stretch/axis conventions. README and design.md
state the span, the transform and the bound rules. No shader twin: stage-shaders is part 2.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
lo = min(lo, hi) now precedes the b == 0 branch in C# and HLSL (Unity copy byte-identical, GLSL
regenerated), so no clamp ever sees lo above hi; GLSL leaves that undefined. The b == 0 path used to
call clamp(a, lo, hi) with the raw bounds. New tests: the flat and sloped lo > hi values, the
statement-order check on the committed HLSL and GLSL, and an appended glsl-parity.json family of lo > hi
and infinite-bound edge cases (every earlier case byte-identical).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Each held stretch contributes only when its length is positive, so an infinite bound times a zero
length never forms: relu (hi = +inf) and its mirrors give the finite integral instead of NaN, in C#,
HLSL (Unity copy byte-identical) and the generated GLSL. Finite cases are unchanged.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
ThirdTensor is now snoise_hessian's per-corner derivation differentiated once more, in double, on the
kernel's lattice, tie order and hash; central differences of the float Hessian are gone from the
measurement and survive only as the oracle in ThirdTensorMatchesDifferencesOfTheHessian (Richardson
extrapolated differences of the double Hessian). MeasureThird on the closed form: refined 487.0917,
x 1.10 = 535.8009 (the differenced tensor read 488.1047, its roundoff). SNOISE_THIRD, ThirdWitness,
CULTMATH_SNOISE_THIRD (both HLSL copies), the generated GLSL constant and the design.md provenance
paragraph follow.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The compiled mirror declares its static consts without public access; the first version of
EveryShaderConstantEqualsItsCSharpTwin found none and failed. Reflect NonPublic too.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
EveryShaderConstantEqualsItsCSharpTwin reflects each CULTMATH_* literal of the compiled HLSL
mirror and asserts it is bit-equal to the math constant of the same name, so CULTMATH_SNOISE_THIRD
(and every other constant) edited in one runtime fails.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Measured on Yggdrasil: the decode needs 56 KiB (Wide32) and 80 KiB (Largest) in
debug on a thread that returns a bool; returning the record through
std::thread's own frames adds several copies of it, so the tests box it and
size threads at 96 KiB and from 160 KiB.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
