Graduate EPAC to its immutable public release - #37
Conversation
|
@codex review Please review exact retirement source |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 89ed78e68c
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review Please review exact head |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: c81d807142
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review Please review exact head 1c6ec9f. The four preceding findings are repaired: current/historical UCNS boundaries are checked independently; byte-exact before/after manifest snapshots carry source commit, Git blob and graph digests; graduation history is independent of future current pins; an unpopulated libs/epac rejects files and symlinks. Actual evidence and a synthetic future-pin fixture pass; 20 corruption fixtures reject. Incoming main f037 is merged with all Python/English Gonol code/data/docs unchanged. Confirmed FALSIFIED standings moved from hmmm to preserved evidence. Current public consumer also binds producer UCNS lock and unchanged consumer-lock bytes. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 1c6ec9fb85
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review Please review exact head 5f47f3a. All five preceding findings are repaired: snapshot bytes now compare directly with git show at their claimed source commits, with full history fetched in every CI consumer of the checker; all three historical receipts match the exact 14-name standings mapping extracted from the committed retirement verifier; historical BASE successor and non-transfer fields are cross-checked; symlinks beneath the retired research path reject; the portfolio report includes merged Python/English Gonol standing/frontier and pins the post-merge documentation commit. Actual transition and future-pin fixture pass; 24 corruption fixtures reject, including a coordinated historical snapshot/hash rewrite, renamed standing, false successor, and directory symlink. The shared authority receipt is unchanged SHA20c33f747e41a571a1b142bb1a3a70c3ef3ae0967973b5025b9ae9b1b2361646, so EPAC8 retains its clean reviewed source5b681 and referenced owning receipt at the now-ancestor1c6. Current-head public replay and all checks gate merge. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 5f47f3a53b
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review Please review exact head 533623a. The four preceding findings are repaired. Every consumer phase now resolves its declared Git tree and verifier bytes. All seven original evidence files and the historical transition facts are also compared directly with the original committed record atc81, preventing coordinated rewrites of receipts, inventories, or their hashes. The inventory requires all28 unique historical paths. Stack CI invokes the canonical portfolio_plan.py validator; its malformed-source regression rejects. The launcher requires the manifest to bind exactly one wheel and one source archive, the complete four-asset inventory, canonical complete SHA256SUMS, unique JSON keys and matching producer UCNS provenance. Actualtransition+futurepin PASS;26 corrupted evidence fixtures reject; valid local release provenance reaches the pre-install boundary while7 malformed launcher cases reject, and the reference portfolio malformed-source case rejects. Current public replay will validate the new launcher against the unchanged published release. EPAC5b681 has clean exactreview5649633599 and all6x209 current-source installs independentlyacceptedSHA9e69e4895a3e795571232629ed01d9b46ba89854eb20aef7d95271b0d28a9583. Sharedreceipt remainsbyteidentical20c33; noEPACsource/releasechange needed. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 533623ac50
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review Please review exact head 233385d. The sole preceding finding is repaired: EPAC_TRANSITION_COMMIT is independently fixed in the checker to c81d807. The receipt must name that anchor, and original receipt/evidence lookups use the constant directly. A disposable two-commit fixture reproduced acceptance by the prior 533623 checker and rejection by this fix; real repository refs and source were unchanged by the probe. The actual transition, future-current-pin fixture and all 26 existing negative fixtures still pass. This commit changes only the checker (5 insertions, 2 deletions). The shared receipt, launcher, manifests, published assets and fully accepted EPAC documentation head are unchanged. Prior 533623 public CI passed and was independently accepted; current CI and this exact-head review gate merge. |
|
Codex Review: Didn't find any major issues. Nice work! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
2 participants
Stack retained a second EPAC implementation after extraction. This change retires its 37 Python implementation/test files, preserves 28 historical files with their actual Stack BASE, and makes the immutable MPL-2.0 EPAC v0.1.0 release the supported consumer and CI input.
The release at EPAC source
949cb1cb304927942966c9fb396caf6227120e7fpassed exact-source review, reproducible four-asset builds, six clean 209-test installs on Python 3.10–3.12, pre-publication Stack verification, public download verification, and reconsumption after forge retirement. Only EPAC implementation/public-contract ownership transfers; all 14 FALSIFIED standings and unresolved scientific questions remain explicit.The manifests, README, AGENTS, CI, release lock, historical BASE, portfolio projection and scoped receipt agree. The checker binds the original graduation to byte-exact before/after manifest snapshots and an archived release lock, while independently checking the current release and UCNS pins. Historical snapshot bytes are checked directly against their Git source commits; CI fetches the required history. Exact standing names, historical BASE successor, and symlink-free retirement are enforced. Later participant or release updates can preserve graduation history. The launcher isolates Python import paths, uses the native archive data filter, and checks producer/consumer UCNS lock agreement. Stack’s other canonical pins are unchanged. Incoming main is reconciled with every Python/English Gonol source and data byte preserved.
Validation: actual transition and a future-pin fixture pass; 26 corrupted-evidence fixtures reject. Local psychsocio 6/6, Python Gonol 11/11, and English Gonol 60/60 pass without skips using its exact historical lexical producer. Portfolio schema, historical Git blob/source identities and current graph digest pass. Current-head public replay, full CI and exact-source review gate merge.
Immutable release · Release qualification matrix · Current Stack public replay
Every historical consumer receipt resolves its Git tree and verifier, and all seven original evidence files are checked against the originally committed transition. The launcher requires the complete wheel/sdist mapping and checksum set and rejects duplicate JSON keys. CI invokes the reference portfolio validator. Nine local launcher/portfolio probes passed.
The original completed transition commit is fixed independently in the checker as c81d807. A disposable two-commit regression reproduces the former redirect acceptance and verifies rejection by the fixed checker.