Skip to content

Proof receipt at c31bf1a: record main's run with the Lean legs under lake (31 rows PWG-2ii) - #57

Open
hyperpolymath wants to merge 1 commit into
mainfrom
claude/receipt-main-c31bf1a
Open

hyperpolymath wants to merge 1 commit into
mainfrom
claude/receipt-main-c31bf1a

Conversation

@hyperpolymath

Copy link
Copy Markdown
Owner

Summary

Follow-up to #53. That change moved the Lean leg and the M-2 bridge under lake, which changed the Lean leg's inputs, so the Lean receipt stopped describing the tree and 31 rows fell to PWG-2i. This records the proofs.yml run on the merge commit of #53, so all three legs have a current receipt again.

Grades, derived by the pinned standards reader from the receipt (never asserted): of 32 theory rows, 31 are PWG-2ii and 1 (M-2) is ungraded. No row is PWG-1, and this does not close PN-7.

Changes

  • scripts/proof-receipt.sh: the run ledger holds the three jobs of run 44324353274 on c31bf1a (Agda 134871496782, Rocq 134871496777, Lean 134871496113), written with --run-row. Only the Rocq row names an image (rocq/rocq-prover:9.3.0@sha256:97a68f23…). The Agda and Lean rows name none: their checkers are installed at run time in base images pinned by digest, which the owner ruled on 2026-11-18 is not isolation (PN-7).
  • scripts/proof-receipt.sh: the shared generator is pinned at standards 146fb8f8, the merge of fix(proof-receipt): run scope, claims with no Agda proof, checked-but-unreconciled legs; reader: foundation-diverse needs RECONCILED metadatastician/standards#1671 on standards main. The four pinned files have the same SHA256 as at the branch commit f4e83044, so nothing about the receipt format changes.
  • verification/receipts/resource-algebra-proofs_praxis.deed: regenerated. The Lean receipt is the first from a build under lake, and carries the tropical-types import at f2e465c again.
  • docs/status/PROOF-STATUS.adoc: census, per-row grades, the leg and receipt tables, and the earlier-receipts list (the run at 15fa113 moves there).
  • docs/status/PROOF-NEEDS.adoc, CHANGELOG.adoc: follow the receipt.

RSR Quality Checklist

Required

  • Tests pass: scripts/proof-receipt.sh --check (regenerates the receipt, refuses any difference, runs the §8.5.3 git checks, checks each current receipt against the tree, and runs the reader with a negative control)
  • Linter is clean (shellcheck scripts/proof-receipt.sh)
  • No banned language patterns
  • SPDX license headers present on all new/modified source files (no new files; no header changed)
  • No secrets, credentials, or .env files included

As Applicable

  • Documentation updated (PROOF-STATUS.adoc, PROOF-NEEDS.adoc)
  • CHANGELOG updated

Testing

Run locally on this branch:

  • bash scripts/proof-receipt.sh --write, then --check: every check holds; the negative control (first claim recorded as PWG-1) is refused.
  • bash scripts/check-adoc-renders.sh: passes.
  • The three ledger rows were produced by --run-row from the tree at c31bf1a; the Lean row's import digest is the one the earlier receipt recorded (c176b35a…).

Not run here: the Agda, Rocq and Lean builds themselves. The receipt cites the CI jobs above for those; this change does not touch a leg's inputs.


Generated by Claude Code

…lake (31 rows PWG-2ii)

The ledger of scripts/proof-receipt.sh now holds the three jobs of
proofs.yml run 44324353274 on c31bf1a, the merge of #53. All three legs
have a current receipt again, so the reader derives PWG-2ii for 31 rows
and ungraded for M-2. Only the Rocq row names an image; the Agda and Lean
rows name none (PN-7 stays open), so no row is PWG-1.

The shared generator is pinned at standards 146fb8f8, the merge of
metadatastician/standards#1671; the four pinned files have the same
SHA256 as at the branch commit f4e83044.

PROOF-STATUS, PROOF-NEEDS and CHANGELOG follow the receipt; the run at
15fa113 moves to the earlier receipts.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013CuBToxr5h9catptGiTsFX
@sonarqubecloud

Copy link
Copy Markdown

hyperpolymath commented Nov 18, 2026 •

Copy link
Copy Markdown
Owner Author

Five checks failed on a6f8272, and none of them on this change: each log ends in API rate limit exceeded for installation between 14:41 and 14:43 UTC.

Check Where it stopped
gate-pr / CI gate "could not publish the re-run pending status; NO VERDICT WAS PRODUCED" after the rate-limit error
Proven 3-outcome test & bench suite setup-bun could not list oven-sh/bun tags: 403, rate limit
Hypatia Neurosymbolic Analysis rate limit while reporting
analyze (actions, none) CodeQL finished its queries, then the upload hit the rate limit
Squabble CI chains & required-context gate its 144 tests passed and the required-context check printed OK; the installer step then got "release not found" from the API in the same window

The quota was spent just before: the merge commit of #53 (c31bf1a) has 248 workflow runs, 221 of them CI gate (203 failed with "could not produce a verdict"). The gate starts a run for every other workflow's start and finish, and each run calls the API.

What this PR is about did pass on this commit: Proof receipt, derived grades, git checks, Core proofs and expected rejections, Rocq recheck leg, Lean recheck leg, Proofs verdict and Coverage exemption holds.

No fix exists in this repository; the gate's fan-out lives in the cicd suite. I will re-run the five failed jobs once, spaced apart, after the hourly quota resets.

Update, 16:40 UTC. The five jobs were re-run once each from 15:29 UTC and passed; CI/gate then read "all 16 required checks passed". At 16:27 UTC the gate's scheduled sweep (run 44347146602) got HTTP 502 from GitHub while reading the jobs of run 44326953644 and wrote CI/gate = failure ("could not read the jobs of run 44326953644") over that verdict. No job on this commit failed. The gate writes the verdict again on its next sweep, or when any workflow run on this commit changes; I am not re-running a passing job to force that.


Generated by Claude Code

@hyperpolymath
hyperpolymath marked this pull request as ready for review November 18, 2026 15:03
Repository owner deleted a comment from coderabbitai Bot Nov 18, 2026
@hyperpolymath
hyperpolymath enabled auto-merge (rebase) November 18, 2026 15:06
auto-merge was automatically disabled November 18, 2026 15:32

Branch protection rule check failed

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.

3 participants