Meta Vens
b941b68085
CultMath tests: pin the FreeViolated record of the real solves, column by column ...
The 30 trace cases' FreeViolated lines (column, gradient magnitude, threshold) are pinned by exact equality beside the base-derived decisions they feed, so a free-set test that is mis-called from SolveTraced (a shifted threshold, a short free count) fails here and not only in the hand-built FreeSetViolated tests.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-11 03:04:25 +02:00
GLSL fixture: small-reach af_snoise and af_fbm families, and the centre leg of the device check [device-check-blind-to-allowance] [device-check-self-referenced]
2026-10-04 11:09:35 +02:00
AffineTests: af_range encloses on every overflowed form, the infinity on the NaN side
2026-10-04 07:20:56 +02:00
Test that the float2 radial falloff overload accumulates, not assigns
2026-09-24 20:13:05 +02:00
CultMath tests: pin the FreeViolated record of the real solves, column by column
2026-10-11 03:04:25 +02:00
CultMath BoundedLeastSquares: the stopping rule's decisions are traced and pinned; Validate is the one owner of n < 0
2026-10-10 22:24:02 +02:00
CultMath tests: a trace case for the cap reached right after a release
2026-10-11 00:07:36 +02:00
CultMath tests: pin the FreeViolated record of the real solves, column by column
2026-10-11 03:04:25 +02:00
tests: fix hashing fixtures, false comments, and gradient tolerance
2026-09-25 23:13:50 +02:00
Compile the real HLSL mirror as C# and with dxc; fix its errors
2026-09-14 13:57:03 +02:00
Move CultMath into the CultLib monorepo
2026-08-26 17:35:51 +02:00
Fix a tolerance-overload bug that made intercept precision asserts inert
2026-09-24 19:38:30 +02:00
CultMath GLSL tests: overloads compared by signature; the two-file header contract pinned
2026-10-03 14:15:32 +02:00
GLSL fixture: small-reach af_snoise and af_fbm families, and the centre leg of the device check [device-check-blind-to-allowance] [device-check-self-referenced]
2026-10-04 11:09:35 +02:00
Affine forms in HLSL: CultMath.Affine.hlsl mirrors math.Affine.cs bit for bit; iv_snoise_ball(float2); the GLSL lowering carries them
2026-10-04 09:47:49 +02:00
Move CultMath into the CultLib monorepo
2026-08-26 17:35:51 +02:00
snoise tie rule: x over y, y over z, x over z (a total order) in C# and both HLSL copies
2026-10-04 03:15:32 +02:00
Make CultMath HLSL-shaped: writable components, bool-vector comparisons
2026-09-14 13:29:59 +02:00
Mirror sweep reaches af_snoise's centred branch: realistic slices, both sides of the choice [mirror-sweep-misses-centred-branch]
2026-10-04 11:09:34 +02:00
Make CultMath HLSL-shaped: writable components, bool-vector comparisons
2026-09-14 13:29:59 +02:00
CultMath intervals: enclosure harness, one-ulp shrink mutants, noise bounds, Lipschitz pin, saving
2026-10-02 19:56:30 +02:00
CultMath: regenerate snoise golden values for the upstream kernel
2026-09-29 12:09:43 +02:00
Move CultMath into the CultLib monorepo
2026-08-26 17:35:51 +02:00
Gate test: restore the affine gate's detail ball and add FogField.InsideFog to the field list [affine-gate-loses-fog-culls] [interval-gate-detail-unpinned]
2026-10-04 12:12:36 +02:00
CultMath 0.4.0: state the true scope of the snoise tie change
2026-10-09 23:07:48 +02:00
CultMath: Phacelle port moves into MPL-2.0 files; prune cut pinned by boundary points
2026-09-30 00:34:27 +02:00
CultMath docs: state the float32 tie band in ulps, as measured (Soul s4)
2026-10-10 00:12:23 +02:00
Pin mtp test-runner for CultMath and document mutation testing
2026-09-22 21:24:05 +02:00
Tie the scalar comparison overloads too, not just vector-vector
2026-09-24 19:49:23 +02:00
b941b68085