feat(format): validated-tensor-v1 + wasmtime-upgrade-v1 + unified-specs-v1 11-gate PARTIAL discharge#1390
Closed
noahgift wants to merge 2 commits into
Closed
feat(format): validated-tensor-v1 + wasmtime-upgrade-v1 + unified-specs-v1 11-gate PARTIAL discharge#1390noahgift wants to merge 2 commits into
noahgift wants to merge 2 commits into
Conversation
…cs-v1 11-gate PARTIAL discharge Bundles three sister contracts in one verdict module: validated-tensor-v1 (FALSIFY-VT-001..004): - VT-001: density gate — non-zero fraction > 0.055 - VT-002: NaN/Inf detection (validation-result-mirrors-input invariant) - VT-003: at least one row has nonzero L2 norm - VT-004: SIMD vs scalar validation — same Err/Ok decision wasmtime-upgrade-v1 (FALSIFY-WASM-001..004): - WASM-001: api_compatibility — `cargo check --features runtime` compiles - WASM-002: behavioral_parity — all runtime tests pass (≥1 ran, 0 failed) - WASM-003: advisory_elimination — zero wasmtime entries in audit/deny - WASM-004: gc feature enables wasm_reference_types unified-specs-v1 (FALSIFY-SPECS-001..003): - SPECS-001: TOC.md ≤ 500 lines (and non-empty) - SPECS-002: zero orphan specs (every .md linked from TOC) - SPECS-003: zero subcrate specs (root-only) ## Five Whys 1. Why bundle these three contracts? They span the validation/runtime/spec coverage band; binding them in one PR captures three sister contracts without duplicate provenance pin overhead. 2. Why does this block ship? Coverage % cannot move while peripheral tensor-validation, wasmtime, and spec-organization gates are unbound at PARTIAL_ALGORITHM_LEVEL. 3. Why mirror-result for VT-002 (input-has-NaN ↔ validation-returns-Err)? The contract says "scanning misses NaN/Inf" is the failure mode. Two-sided check catches both: - false negative (NaN present, validation passed) — privacy/UB risk - false positive (clean input, validation rejected) — usability bug 4. Why fail-on-zero-tests for WASM-002? The contract says "all tests pass." Zero tests run vacuously satisfies that — but means the upgrade was never actually exercised. The gate must Fail rather than vacuously pass for the regression class "test runner broke and skipped everything." 5. Why fail-on-empty-TOC for SPECS-001? Same Fail-on-zero rationale as several earlier discharges (cgp-monorepo-build, decode-hp23, qdot, tensorlayout). Empty TOC silently passes the ≤500 bound but means TOC was deleted. Catching the regression class up-front is cheaper than discovering it later. Adds 37 unit tests including 8-bucket density sweep + 6-bucket TOC size sweep. Realistic-healthy walks the canonical post-upgrade state; pre-fix walks 11 simultaneous regressions across all three contracts (sparse embedding, NaN leak, dead row, SIMD drift, broken API, failing tests, advisory leak, gc disabled, bloated TOC, orphan specs, subcrate spec leftovers). No runtime % shift; algorithm-level coverage advances by 11 gates.
noahgift
force-pushed
the
feat/vt-wasm-specs-001-011-partial-discharge
branch
from
c72ebaa to
68c05b9
Compare
Contributor
Author
auto-merge was automatically disabled
May 12, 2026 09:21
Pull request was closed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Bundles three sister contracts in one verdict module:
validated-tensor-v1(FALSIFY-VT-001..004): density, NaN/Inf rejection, zero-row L2, SIMD paritywasmtime-upgrade-v1(FALSIFY-WASM-001..004): api compat, behavioral parity, advisory clean, gc featureunified-specs-v1(FALSIFY-SPECS-001..003): TOC ≤ 500 lines, no orphan specs, no subcrate specs37 unit tests including 8-bucket density sweep + 6-bucket TOC size sweep.
Algorithm-level coverage advances by 11 gates; runtime ship % unchanged.
Gates bound
> 0.055cargo check --features runtimecompiles.cargo/audit.toml+deny.tomlwasm_reference_typesTOC.md ≤ 500lines AND non-empty.mdfiles in docs/specifications/crates/*/docs/specifications/Five Whys
See commit message — captures mirror-result rationale for VT-002, fail-on-zero-tests for WASM-002, and fail-on-empty-TOC for SPECS-001.
Test plan
cargo test -p aprender-core --lib vt_wasm_specs— 37 passed🤖 Generated with Claude Code