Skip to content

Feature A — Verify test-report archives for rejected merge groups (#307 merge-flow) - #1246

Open
oxa-agent[bot] wants to merge 2 commits into
mainfrom
agent/307-test-report-provenance
Open

oxa-agent[bot] wants to merge 2 commits into
mainfrom
agent/307-test-report-provenance

Conversation

@oxa-agent

@oxa-agent oxa-agent Bot commented Nov 18, 2026 •

Copy link
Copy Markdown
Contributor

Feature A · Native work board

Ports the test-report provenance portion of #307 merge-flow → board landing under #1214 and #1017. T2.

An owner can opt in with a private queue-report policy. The existing Board Reader verifies the recorded failed required check IDs against the exact merge-group workflow attempt. It verifies the configured report producers and checks the downloaded SHA-256 digests before appending the archive IDs to the owner journal. Shared reports are fetched once per attempt, with the existing response budget applied to the sample.

This slice records verified archive provenance. Reading failed test names and projecting the author notice follow in the next slice. Landing shadow remains read only with respect to GitHub.

The real CLI fixture checks two failed required checks sharing an archive, a digest mismatch, and an unrelated required check. Selector controls cover mismatched attempts, producer timing, duplicate artifacts, and incomplete listings. The real #607 merge-group run supplied the two-check shared-attempt case, and both published report digests were independently verified. Formatting, whitespace checks, and isolated compilation of oxa-work tests and the work_shadow CLI test passed locally. Both hosted architecture lanes passed. Required shard 2 failed only the sandbox scratch fixture, which timed out in cargo build at 30.026 seconds. The checkout already contains #1201. Required CI recovery and independent review remain pending.

Removes: nothing.

Native shadow sample at the exact head

The x86_64 binary from checksum-matched archive 14381924695 completed an isolated read-only sample at 1a4360b0. The owner journal contains one sample with 57 board entries and verified archive evidence for #496, #538, and #607. For #607, it recorded exact group ace309aea04e2ab94be827a2fa89324326ba2841, run 43815430013 attempt 1, both failed required check IDs, and both independently verified archive IDs and digests. #794 retained its failed required checks without a verified archive entry. The journal SHA-256 was cca3f8cde47820d413897b11b443149fe16ef762ec6c2e93412fc913fc8cc232. The CLI control passed on both hosted architectures. The required-CI scratch-fixture hold remains recorded above.

Reader conflict resolution

Head 40790887 resolves the real conflict with #1206. The existing Reader shares one captured-byte budget and deadline across GraphQL, Actions metadata, and downloads. The CLI control covers that a REST reply which fits by itself can exceed the remaining board-snapshot budget, and that no journal record is appended in that case. Formatting, whitespace checks, and isolated compilation passed. Normal CI ran on the refreshed head. The shadow and scratch results above qualify 1a4360b0 only.

On refreshed head 40790887, the hosted lane passed on both architectures, including the report and shared-budget control. Required shard 1 passed the same control. Its only failure was oxa-parity::keys_settings::key_018, which timed out opening /tree after the UI reported that the session was still busy. Shard 2 passed. Current main contains no verified fix for this failure. Exact-head required CI remains pending recovery.

A fresh isolated read-only sample from checksum-matched artifact 14384586834 at 40790887 completed with the shared Reader limits in effect. The owner journal contains one sample with 55 board entries and archive proofs for #496, #538, and #607. The #607 proof contains the same exact group, two required check IDs, run attempt, and both archive digests as the independent expected values above. #794 retained its raw rejection evidence. Journal SHA-256 4a5d9e2830e285e6d7a886db2e4474b3fe15fc4ef5f876a9cba379d985a25438. This qualifies the read-only report capture on the refreshed head. The required parity-test hold remains open.

@n0rpan

n0rpan commented Nov 18, 2026

Copy link
Copy Markdown
Collaborator

@codex review

@n0rpan

n0rpan commented Nov 18, 2026

Copy link
Copy Markdown
Collaborator

@codex security review

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Nov 18, 2026 •

Copy link
Copy Markdown
Codex Review Summary: ✅ Completed

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-11-18T00:08:32.703748Z 1a4360b Manual request
🔒 Security Review ✅ Completed 2026-11-18T00:09:48.503311Z 1a4360b Manual request
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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 1a4360b071

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".

Comment on lines +114 to +116
let key = (rejection.removal_event().to_owned(), run, attempt);
let proof = match previous.remove(&key) {
Some(saved) if saved.failed_check_runs == checks => Some(saved),

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Share verified reports across PRs in the same workflow attempt

When a rejected merge group contains multiple PRs, their rejection observations reference the same workflow attempt but have distinct removal-event IDs. Keying the cache by removal event, and consulting only the previous sample's cache, downloads and hashes the same archives again for every PR during the first sample. This can also exhaust the shared 4 MiB response budget: two PRs sharing a 2 MiB archive cause the sample to fail once metadata is included, so no rejection evidence is journaled. Cache the verification by workflow run and attempt within the current sample, while retaining the per-PR failed-check validation.

Useful? React with 👍 / 👎.

@chatgpt-codex-connector

Copy link
Copy Markdown

🛡️ Codex Security Review

Security review completed. No security issues were found in this pull request.

Reviewed commit: 1a4360b071

View security finding report

Only the user who started this review can view the report in Codex.

ℹ️ About Codex security reviews in GitHub

This is an experimental Codex feature. Security reviews are triggered when:

  • You comment "@codex security review"
  • A regular code review gets triggered (for example, "@codex review" or when a PR is opened), and you’re opted in so security review runs alongside code review

Once complete, Codex will leave suggestions, or a comment if no findings are found.

This branch has not been deployed

No deployments
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