Skip to content

fix(record): repair the five red preflight record gates to their own contracts - #3369

Merged
lu-zero merged 5 commits into
mudler:mainfrom
lu-zero:row/preflight-gate-repairs
Oct 1, 2026
Merged

lu-zero merged 5 commits into
mudler:mainfrom
lu-zero:row/preflight-gate-repairs

Conversation

@lu-zero

@lu-zero lu-zero commented Oct 1, 2026

Copy link
Copy Markdown
Collaborator

Five preflight record gates are red on main; this branch repairs each to its checker's own contract, reconciled to the true state throughout.

  • check-release-binary-contract and check-windows-release-state: the roadmap's release row had drifted to "no published binary exists" while four other surfaces still carry archive_claims=published-v0.0.2; the row is restored verbatim from 10a64f8 (v0.0.2 published, Windows W14-W16 pending) together with the ENG-RELEASE-WINDOWS state anchor and the v0.0.3-pre.1 pending phrase.
  • check-benchmark-index: sixteen benchmark detail files existed without index rows; the index table gains all sixteen, dispositions read from each file's content. Nothing deleted.
  • check-env-doc: VT_VK_DISABLE and VT_VK_DISABLE_PAGED_ATTN are temporary kernel-internal bisect hooks (vulkan_ops.cpp:982/:1296) and are allowlisted; VT_VK_FENCE_TIMEOUT_MS is an operational watchdog that aborts a wedged dispatch (vulkan_context.cpp:174) and is documented as a Deployment knob in docs/ENVIRONMENT.md.
  • check-oracle-pins: eleven surfaces restated the retired e126687a9a pin; the 2026-09-26 sync advanced the authority to a7c23ac96d / 0.3.0.dev267 but missed them. The restated spans now match the authority in the vllm oracle block, .agents/NOW.md, docs/FEATURES.md and three benchmark docs, with the surrounding since-2026-09-03 prose re-dated to the actual advance.
  • Also gives .agents/NOW.md one line of headroom, fixing a test_record_merge_shape failure that fails on clean origin/main.

Verification: all five gates plus check-agent-record exit 0; record suites 258 passed, gate suites 149 passed. Issue: ISSUE-LOCAL-01M3WMKTBEDAHF226YSATSR7BF, spec committed first.

FOLLOWING_AGENTS_PROTOCOL

Following-Agents-Protocol: true
AI-Assisted: true
Assisted-by: AGENT:zai/glm-5.3-flash [maki]

ISSUE-LOCAL-01M3WMKTBEDAHF226YSATSR7BF: five record/doc gates are red on
origin/main. The spec names each checker's contract and the record each
repair restores; the commit order that follows proves the spec landed
before the implementation it governs.

FOLLOWING_AGENTS_PROTOCOL

Following-Agents-Protocol: true
AI-Assisted: true
Assisted-by: AGENT:zai/glm-5.3-flash [maki]
…anchors

check-release-binary-contract requires the single `| REL | ROAD-V1-RELEASE |`
row to carry `ACTIVE` with the v0.0.2-published text and the pending Windows
publication gate ("Windows W14-W16 are implemented for one PR", "publication
and 32-asset audit remain pending", issue PREFLIGHT-GATE-REPAIRS /
ISSUE-LOCAL-01M3WMKTBEDAHF226YSATSR7BF); check-windows-release-state requires
exactly one `<!-- ENG-RELEASE-WINDOWS: state=ACTIVE publication=pending
artifact=unpublished -->` anchor and the `v0.0.3-pre.1` pending phrase on the
same surface. The record restructure dropped both from .agents/roadmap_v1.md
while engine-matrix, coordination.md, the release spec's machine block and the
state event all still carry the published-v0.0.2 state the checker's
EXACT_MACHINE_FIELDS pin, so the roadmap row was the drifted surface. Both
anchors are restored verbatim from history (10a64f8 for the row, a0a8fed
for the anchor); no expectation was reconciled away, and no published-binary
denial survives to contradict the four sibling surfaces.

FOLLOWING_AGENTS_PROTOCOL

Following-Agents-Protocol: true
AI-Assisted: true
Assisted-by: AGENT:zai/glm-5.3-flash [maki]
check-benchmark-index fails closed on any docs/benchmarks/*.md that no
`| \`id\` | ... | [label](benchmarks/file.md) |` row in the
`## Benchmark detail index` table of docs/BENCHMARKS.md owns, with the ID
required to equal the file stem; 16 legitimate public benchmark detail files
had lost their rows (ISSUE-LOCAL-01M3WMKTBEDAHF226YSATSR7BF). Each is indexed
with a one-line disposition taken from its own content; no file is deleted.

FOLLOWING_AGENTS_PROTOCOL

Following-Agents-Protocol: true
AI-Assisted: true
Assisted-by: AGENT:zai/glm-5.3-flash [maki]
check-env-doc fails closed on any VT_*/VLLM_* name read from src/ or include/
that is neither documented in docs/ENVIRONMENT.md nor listed on
scripts/env-doc-allowlist.txt (ISSUE-LOCAL-01M3WMKTBEDAHF226YSATSR7BF).
VT_VK_DISABLE and VT_VK_DISABLE_PAGED_ATTN are temporary kernel-internal
bisect hooks in src/vt/vulkan/vulkan_ops.cpp (982, 1296) that force ops down
the fallback tier, so they are allowlisted. VT_VK_FENCE_TIMEOUT_MS is a
user-facing operational watchdog (src/vt/vulkan/vulkan_context.cpp:174) that
aborts instead of spinning forever on a wedged dispatch, so it is documented
as a deployment knob with its block-forever default.

FOLLOWING_AGENTS_PROTOCOL

Following-Agents-Protocol: true
AI-Assisted: true
Assisted-by: AGENT:zai/glm-5.3-flash [maki]
…peats it

check-oracle-pins requires each declared pin surface's `<!--pin:commit-->`
span to be a lowercase-hex prefix of upstream-sync.md's vllm_commit and its
`<!--pin:label-->` span to equal the public version of vllm_runtime_version,
because the parity-pin block is the authority and a surface that restates it
must restate the current value; .agents/oracles/vllm.md's oracle-pin block is
held to the same identity (ISSUE-LOCAL-01M3WMKTBEDAHF226YSATSR7BF). The
2026-09-26 sync 4f11dfc advanced the pin to a7c23ac96d / 0.3.0.dev267 but
left the oracle record and five prose surfaces at e126687a9a /
0.28.1rc1.dev132. Every span now restates the current pin, and the
surrounding "since 2026-09-03" claims are re-dated to the advance that
actually produced it. NOW.md also gives back one line so the live digest is
not exactly at its cap.

FOLLOWING_AGENTS_PROTOCOL

Following-Agents-Protocol: true
AI-Assisted: true
Assisted-by: AGENT:zai/glm-5.3-flash [maki]
@lu-zero
lu-zero merged commit 23a90c0 into mudler:main Oct 1, 2026
2 of 18 checks passed
@lu-zero
lu-zero deleted the row/preflight-gate-repairs branch October 1, 2026 21:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant