Commit graph

3,191 commits

Author SHA1 Message Date
Meta Vens
780d8cfa83 Variants map: store authorities after merge-r1 and the Mesh reader cut
One owner per store decision per runtime (header, existence and links, zero-byte,
directory at path, lock, replace, soft write), with pinning tests and open gaps by
follow-up id. Written by a Modeling pass against 705ab669.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 03:00:19 +02:00
Meta Vens
38ead8c4b0 Semver policy: the declaration file owns prefix, changelog and assemblies; first release and measured set are derived
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 02:36:58 +02:00
Meta Vens
23d7beac20 Semver check derives what callers used to declare: first release, tag prefix, changelog and measured set
The checker reads each package's tag prefix, changelog and assemblies directory
from scripts/release-packages.mjs and derives the release record from git tags.
A first release is the absence of any older tag in a checkout that holds other
tags; a malformed same-prefix tag, a checkout without tags and an unreadable
repository refuse. The measured set is every managed assembly (CLI header in
the PE bytes) tracked at the previous tag at any depth; an unbuilt one is a
break, a built assembly with no namesake is an addition.

Deleted with their branches, refusals and tests: --first-release, --api-new,
--tag-prefix, --changelog, --api-baseline-path and the workflow first-release
decision line. Any unknown option is now a usage error.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 02:36:54 +02:00
Meta Vens
3d17254bbf Research: Windows store replace primitives (ReplaceFile vs MoveFileExW vs POSIX rename), with probe counts
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 02:31:11 +02:00
Meta Vens
705ab669b5 Mesh single-file store tests: junction target is one level; the empty-store race test skips Windows
A Starfire run showed the junction helper left its parent directory behind, which the "write went
through the link" assertion mistook for a write. The race test sees nothing-at-the-path windows on
Windows because File.Replace is not atomic to a lock-free reader; that is the store's replace and
is reported as a fork, so the test skips there with that reason.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 02:17:18 +02:00
Meta Vens
b1ff2dcbce Mesh single-file store: tests for dangling link, directory, zero byte, lock, concurrent replace and the older layout; contract sentence
Behavioural tests through Mesh and CultCache for each rule the cut owns, and the persistence
contract's single-readability paragraph states that one runtime has one reader, lock and replace
and that only nothing at the path is missing. Not yet run: verification follows in one job.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 02:02:02 +02:00
Meta Vens
9852fa9367 Mesh single-file helpers read and replace through the CultCache single-file store; a directory at the store path is an error
SingleFileBackingStore gains ReadDurable and ReplaceDurable (internal): CultMesh's read and write
helpers reach the one reader, <file>.lock and the unique-temp atomic replace through them. ReadDisk
decides absence from the path's attributes, so a directory at the path is an IOException, not an
empty store. Includes the directory clause that the spec lists as a separate step.
Builds with the previous commit's deletions; the tree before this commit does not build.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 01:59:42 +02:00
Meta Vens
2e12e9eebb Mesh single-file helpers: delete the second reader, the older catalog layout and the lock-free replace
Deletes CultMesh's ReadSingleFileSnapshot, HasOlderCatalogLayout, ReadLegacySchemaCatalogEntry,
WriteFileAtomically and the File.Exists gate, ReadStore's entry-reader delegate, and the three tests
of the older catalog layout. This commit does not build: the Mesh read and write sites still name
the deleted reader; the next commit routes them through the CultCache single-file store.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 01:59:13 +02:00
Meta Vens
245b0619b4 Declared-range check: find npm's package root on Unix as well as Windows
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 01:35:45 +02:00
Meta Vens
172759fd30 Declared-range check: a clone without release tags is refused rather than passing for want of releases to build
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 01:30:15 +02:00
Meta Vens
4f76bb9391 cultcache-ts 0.15.0; cultnet-ts and cultmesh-ts declare ^0.15.0; a check builds the tagged releases a declared range admits
cultnet-ts imports readSingleFileStore and cultmesh-ts imports schemaIdentityOf and SchemaConflictError, none of which cultcache-ts 0.14.0 exports. The new check builds each tagged release a declared range admits and requires every runtime name the dependent imports to be an export of it.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 01:28:48 +02:00
Meta Vens
2d26c8100e CultCache TS: nothing at the path needs both looks to agree; the second look is a test seam
readStoreBytes reports an absent store only when the read and the lstat both find nothing. Any other lstat failure is thrown, and an entry the lstat finds that is not a link is reported as appearing during the read. The manifest walk also runs a soft pushAll over every vector.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 01:28:47 +02:00
Meta Vens
12805eb32f CultCache Rust: a soft push_all asks the one reader, as TypeScript's does
Absent writes, a store it reads is left byte for byte, and anything the reader refuses (an unreadable file, a dangling link) is refused. The readability manifest walk now runs a soft flush over every vector, so TypeScript and Rust are pinned by the same files.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 01:28:47 +02:00
Meta Vens
133dd9f760 Semver policy: a rebuild is measured against its predecessor, a first release is cross-checked, every tracked assembly counts
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 00:58:34 +02:00
Meta Vens
4c7ae3adc6 Publish workflow: a package's first release passes --first-release, which the checker verifies
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 00:58:33 +02:00
Meta Vens
652b0837d2 Semver check: one rule for an unbuilt tracked assembly, comparisons must run, git failures are refusals, a first release is cross-checked
Every assembly the previous tag tracks that the build does not hand over is a break, whatever its name; a run that compared nothing is refused (the --api-new rename of a lone assembly). The cultlib build now hands over every shipped assembly. --first-release is believed only with no other same-prefix tag and the prefix its changelog directory names. git ls-tree/show failures and an unreadable changelog become refusals. Tests pin the CultMath-shaped rename, the left-reference scenario (a dependency's netstandard base chain changes), the first-release probes and the workflow's own first-release line.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-03 00:58:33 +02:00
Meta Vens
43183276d3 CultMath design.md and README: iv_frustum_ball, the envelope composition, the scenario saving
design.md "Intervals" gains the frustum-ball paragraph and the envelope-then-noise
paragraph: how a consumer composes the bound, the short-circuit, and that the warp
lives on the noise only. r1's slab saving paragraph is replaced by r2's printed
numbers per scenario. README names iv_frustum_ball and the envelope composition.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 00:08:35 +02:00
Meta Vens
1111537b71 CultMath intervals r2: the tile march on cell ranges and the scenario measurement
MarchTile runs the pre-pass once per tile on cell-index ranges. It starts from
the whole grid, skips and doubles on proof, halves on maybe, and runs to the
grid's end. Every ray of the tile integrates exactly the dense cells with the
dense march's Integrate. MarchLod is the void's footprint-aware march inside the
unmasked runs. IntervalSkipHalvesEvaluations prints (a)-(d) at N in {1, 4, 8, 16}
and warp in {0, D}, the void grid (Rh x camera offset x S), and the shipped point
with three marches. It asserts transmittance agreement and skips naming
saving-2x-scenarios when (a), (c) or (e) misses its 2x contract.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 00:04:54 +02:00
Meta Vens
6f2f3a19a3 CultMath intervals r2: cut r1's per-ray march; envelope-then-noise fields and their composed bound
Deleted: SavingField's MarchIntervals (per-ray probes, the StepMax growth cap,
probes attempted where the ball could never prove emptiness), its unconditional
three-snoise sample, and the saving test that skipped naming r1's fork.

Added, test-side only: the scenario model. FogField is Aetheria's nebula over its
gravity-well height map (Volumetric.cginc, Zone.cs, Settings.asset cited by
line): (a) height fog, (b) inside the fog, (c) the zone bowl and four wells.
SlabField is r1's slab, kept as the worst case. VoidField is the shipped cavity
with footprint octave weights. Each composes one bound from existing iv_* ops:
the envelope over the tile slice's box, its fade, and the noise ball from
iv_frustum_ball enlarged by the warp. The pointwise density is envelope-gated.
EnvelopeBoundEnclosesDensity checks 2,000 slices x 64 warped points with no
tolerance.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 00:04:54 +02:00
Meta Vens
9b9a8b3e05 CultMath intervals r2: iv_frustum_ball, the ball of a screen tile's rays over a depth segment
One new function in C# and both tracked HLSL copies (one blob): the centre on the
tile's central ray at mid-depth, radius ((z1 - z0) / 2) |(m_c, 1)| + z1 N / (sqrt(2) f)
+ warp. TileBallEnclosesEveryRaySegment checks 2,000 seeded tiles x 64 points
(sub-pixel corners at both depth ends with full-length flows) with no tolerance,
and pins the ball to the derivation so a grown ball fails too. The mirror's
compared-function count moves from 52 to 53.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-03 00:04:44 +02:00
Meta Vens
42f05f6681 CultCache C#: a dangling Windows junction at the store path is an IOException, and the tests reach it without a privilege
Opening a junction whose target is gone fails with UnauthorizedAccessException, not the
IOException the dangling-link rule promises. ReadDisk now turns that into the same
IOException. The dangling-link tests build a directory junction (mklink /J), which needs
no privilege, and the file-symlink cases skip where the account cannot create one.
The fixture removes links before its recursive delete.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-02 23:56:53 +02:00
Meta Vens
db4ecb1f8b Readability vectors: a store followed by one stray byte is refused by every runtime
The manifest held only three-byte and whole-store trailers, so a Rust reader that
tolerated a single trailing byte passed every suite while C#, TypeScript and Python
refused the file.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-02 23:53:36 +02:00
Meta Vens
c642bba752 Imagination pass 3d: footprint LOD. The cavity cost as a function of hollow radius, camera offset and ramp width; Q8 retired for a steps-per-pixel contract
The march is footprint-aware: step = max(k z theta_px, min(S/4, lambda_fine/2)),
one smooth weight per octave, the bound enclosing the truncated field. The
cavity cost model is a grid, not a number; the shipped point needs about 6
steps and 12 snoise per pixel against about 35 fixed steps.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
2026-10-02 23:36:17 +02:00
Meta Vens
f218c9e8f3 Tests: pin the authority record shape checks and the dangling-link message
Mutation survivors of the authority reader (record count, schema id, type) and of the
reader's dangling-link error text each get a test that fails when the rule breaks.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-02 23:28:53 +02:00
Meta Vens
210a4f50cd Imagination pass 3c: the true cavity. Unbounded fill carved by one pseudo-Gaussian sphere, the camera re-derived inside the hollow, (e) honestly under 2x, Q8 cavity-contract
Ruling operator-void-true-cavity: the void is a 3D hollow in a cloud volume,
not the 2021 bowl. The outer volume is an unbounded fill so every ray ends in
a wall; Rc grows to 640 so the hollow (198) holds the camera at 186. With no
sky the evaluation ratio for the shipped scene is predicted 1.0-1.2x; the
pre-pass's gain there is sample placement (2.0-2.6x), raised as Q8.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
2026-10-02 23:22:36 +02:00
Meta Vens
9f449fb104 Imagination pass 3b: the void. The old main menu found in Aetheria e6181c11, site-ground r3, the void scenario in interval-ops r2
Ruling operator-ground-void-brush answers Q6 outside the options. The
original main menu's brush, floor, sun, orbit, camera follow, authored light
and stardust rule are read from Aetheria at e6181c11 and carried in their
units; the lighting is the original's analytic pulse made 3D; the TAA
reprojects by camera and flow; the saving test gains the void scenario and
asserts 2x on it.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
2026-10-02 23:04:30 +02:00
Meta Vens
070a58d2aa Imagination pass 3: envelope-then-noise fields, scenarios with empty space, the tile pre-pass, the split Phacelle
Revision 2 specs for the two CultLib cuts of cultmath-tapes, as follow-ups on
hands/cultmath-interval-ops (c9bd003b) and hands/cultmath-glsl-lowering
(5ae207e8), under the rulings operator-tile-amortized-march,
operator-split-phacelle, operator-compiler-downloads and
self-exp-measure-in-site, and the operator's challenge to the march.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
2026-10-02 22:39:33 +02:00
Meta Vens
a23eb8219d TS test: the forged-header case spells its newline as an escape
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-02 22:19:15 +02:00
Meta Vens
8a833730c9 cultnet-ts: Idunn authority files are read by cultcache-ts's one store reader
readAuthorityRecord decoded the store itself and accepted only cultcache.store.v1. It now calls
readSingleFileStore, so v3 reads, a zero-byte file is a StoreUnreadableError, and a refused
header is never echoed. Tests derive their bundles from the Rust-written fixtures.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-02 22:12:15 +02:00
Meta Vens
0122761a74 TS: the one store reader is exported and decides soft pushAll; absent-read and forged-header tests
pushAll({soft}) asked readStore, so it skipped every file the reader refuses. It now asks
#readDisk: absent writes, readable skips, unreadable refuses. The reader is readSingleFileStore,
synchronous and exported, so a consumer reads a store through it instead of decoding one itself.
New tests: soft pushAll over absent, readable and five unreadable files; reading an absent store
writes nothing; a forged header that does not start at the prefix is never echoed.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-02 22:12:15 +02:00
Meta Vens
68c1296b7a Semver check: keep git's stderr out of the refusal (the fatal-message test fails without it)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 21:44:03 +02:00
Meta Vens
c358ef9739 Semver check: drop the unobservable stdio option; the empty-package test really creates the directory
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 21:38:55 +02:00
Meta Vens
df13b502b9 Semver check: built and reference paths are resolved against the process, not the tool's directory
The ApiCompat tool runs from scripts/api-gate, so a relative --api-built or --api-refs
failed as an unreadable assembly. The release scripts pass absolute paths; any other
caller does not have to.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-02 21:30:37 +02:00
Meta Vens
0ac870f132 Semver check tests: pin git's silence, the gate manifest file, netstandard.dll, the first-release hint
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 21:28:26 +02:00
Meta Vens
e7786ffac2 Semver check: a malformed version is refused by name, and refusals are pinned never to repeat input contents
A malformed --version fell into the tag-reading catch and was reported as unreadable
tags. It is now validated first and refused without repeating the value. A canary test
runs seven refusals (measured break, unreadable built DLL, unreadable baseline blob,
missing entry, skipped version, breaking under too small a bump) with a canary in the
changelog and the assemblies and asserts it never appears in the output.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-02 21:27:03 +02:00
Meta Vens
da2b9482e5 Rust: the rename fault fires inside replace_file_atomically, at the call a swallowed error would hide
The seam sat before the call, so a mutant that swallowed the real rename's result was invisible to it. It now makes the rename itself fail, on both platforms.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 21:11:07 +02:00
Meta Vens
7b815e1ca8 TS and Python: pin the refusal text (v3 named, ASCII digits only); TS comments say a zero-byte file is refused
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 21:05:38 +02:00
Meta Vens
9a251fe60c Rust: one store-header reader; bin-encoded headers refused; rename fault seam and the tests the survivors asked for
store_header is the only reader of a store file's leading bytes. It returns the header slot (a MessagePack str only, never bin, as in C#, TS and Python) and separately the proof that the bytes are one whole array, so a newer store cut short after its header is still named unsupported. leading_header, read_array_header and read_string are deleted. decode_store_file takes the header the reader already found.

Vectors bin-header-v1/v3/v9 join the readability manifest as refuses x4. Tests: a rename fault step (WriteStep::Rename) shows a failed rename is NotReplaced, the store intact and no staging file left; every merging writer refuses through a dangling link; the unsupported-format refusal names v1 and v3 and echoes only ASCII-digit headers.

Not yet compiled or run at this commit.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 21:05:37 +02:00
Meta Vens
c4826a39a3 C# tests: the directory refusal shows a header only as store version digits, so the suffixed legacy names are no longer echoed
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 21:00:51 +02:00
Meta Vens
eb79b21e51 C# tests: read "echoed" from how the refusal introduces the header, since it also names the readable formats
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:51:55 +02:00
Meta Vens
3f33445d84 C# tests: assert the header refusal text with plain checks
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:51:00 +02:00
Meta Vens
9e481f59b7 C#: detect a dangling store link by its reparse-point attribute (Caching has no FileInfo.LinkTarget)
Commit 425531d8 does not build.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:49:43 +02:00
Meta Vens
425531d88e C#: a dangling link is an I/O error on read; the directory refusal echoes a header only as store version digits; contract names Rust's real type
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:49:12 +02:00
Meta Vens
75ba6c0f5c Mesh: a zero-byte store file is refused, not read as empty; the alias test is retargeted to one type per schema
CultMesh's single-file reader had its own zero-length-means-empty branch, so a Mesh write
replaced a zero-byte file the cache refuses (merge-r1-precedence). It now asks the cache's
snapshot reader and gets its verdict. The auto-merged test that read a same-schema instance
through MeshNoteAliasDocument cannot be built under retire-schema-alias; it now states the
ownership rule it protected (a database serves only what a shard owns) with the one type.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:49:12 +02:00
Meta Vens
5ae207e82b CultMath design.md, README, notices: the GLSL target
design.md gains "GLSL Target: Source Transformations" (the list, the common-subset rule, the
regeneration command, the derived-file rule, the two parity legs). README says how a WebGL
consumer includes CultMath.glsl. THIRD-PARTY-NOTICES lists CultMath.glsl as MPL-2.0, because
it inlines the Phacelle include.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:09:45 +02:00
Meta Vens
34a9329db5 CultMath tools: get-glslang.ps1 (pinned 16.6.0) and compile-glsl.ps1
compile-glsl.ps1 writes the wrapper a WebGL2 consumer writes (#version 300 es, highp precision,
the library by concatenation, one call of every public function) and compiles it with glslang;
-Body replaces the calls, for the negative control.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:09:45 +02:00
Meta Vens
d2ff33085f CultMath GLSL: shaders/CultMath.glsl generated from the HLSL by one enumerated lowering, pinned
GlslLowering (test project) is the HLSL -> GLSL ES 3.00 transformation as named steps: includes
inlined by the C# mirror's own resolver, guards and texture functions dropped, type names,
static const, casts by a balanced scanner, intrinsic renames, literals passed through, the
CULTMATH_GLSL guard and a provenance line. GlslMirrorTests pins the committed file to the
lowering (CULTMATH_WRITE_GLSL=1 regenerates it), finds no HLSL-only token, finds every public
function but the texture samplers, and pins fixtures/glsl-parity.json (256 seeded cases per
family as float32 bit patterns) to C#.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:09:44 +02:00
Meta Vens
0eeb619b35 CultMath HLSL: write the shader in the subset GLSL ES 3.00 shares, bit for bit
Each edit is legal HLSL that computes the same bits, and closes one dialect gap the GLSL lowering
would otherwise have to special-case:
- snoise and snoise_grad: step(h, 0.0) -> step(h, float4(0.0, 0.0, 0.0, 0.0)); GLSL has no
  step(genType, float).
- pcg: (state >> 28) + 4 -> + 4u; GLSL has no implicit int -> uint.
- cellular and phacelle jitter, cellular id: (float)((uint)hash.x >> 8) before the float
  multiply; GLSL has no implicit uint -> float.
- damp (float2/3/4): the scalar amount is splatted explicitly; GLSL has no lerp(vecN, vecN, float)
  promotion.
- pcg3d(float3) is defined before pcg3d(float2), which calls it; GLSL needs a declaration before use.
Both tracked copies; the C# mirror stays green, FXC cs_5_0 compiles both copies.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:09:44 +02:00
Meta Vens
0f13c7d368 C#: the unreadable-format refusal names every format the runtime reads, v3 included
A mutation survivor on the merge's refusal text: dropping v3 from the list passed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:09:17 +02:00
Meta Vens
e5f06a2a49 Python: fix the payload byte literal in the refusal-text test (the previous commit did not parse)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-02 20:04:16 +02:00