[azimafroozeh]

What & why

The first scheduled run of tests-slow-nightly.yml went red on three nightly: tests of the CLI crate that round-trip a fixture receipt through the client and compare the serialized result against the fixture JSON.

• CommitOutput gained published in feat: add unpublished graph commits #942 and kind plus root_commit_id in feat(branching): add unattached graph branches #956; its Deserialize fills the three from defaults when a receipt lacks them and its Serialize always writes them, so a fixture without them never round-trips equal.
• Both PRs updated the gate tests that use this pattern and left the three #[ignore]d nightly owners behind: no PR check runs them, so the drift surfaced on the nightly's first run (two reds before feat(branching): add unattached graph branches #956, three on main since).
• The three fixture objects gain "published": true, "kind": "published", "root_commit_id": null, the values the deserializer derives for a plain published receipt: client::tests::managed_mutations_use_thirty_second_deadline_without_retrying, client::tests::managed_reads_use_thirty_second_deadline_without_retrying, managed::data::tests::managed_load_sends_exact_ndjson_and_preserves_the_server_receipt. No production change.

Backing issue / RFC

• Trivial fast-lane: no issue or RFC, test fixtures only with no production change, motivated by the first red scheduled run of tests-slow-nightly.yml

Checklist

• Focused: one logical change, three nightly fixture receipts carry every field CommitOutput serializes
• Tests: the three edited tests are the regression tests; they run on the nightly's --run-ignored only step, not on Test Workspace
• Public docs: none, no user-facing surface changed
• Release note: N/A, test-only
• Invariants: reviewed against docs/dev/invariants.md, no Hard Invariant weakened, no deny-list item hit; nearest one touched is none

Local verification

• cargo nextest run --workspace --exclude omnigraph-gqt --exclude omnigraph-gqt-served --exclude omnigraph-dst --locked --profile ci --no-fail-fast --run-ignored only --features omnigraph-engine/failpoints,omnigraph-cluster/failpoints -E '<the workflow's 16 enrolled names>' on 0a7bb8c0 before the fix: 13 passed, 3 failed, exit code 100, the three tests above, each assertion left == right failed on the receipt
• the same command after the fix: 16 passed, exit code 0, 41.3 s wall; scripts/check-workspace-test-owners.py target/nextest/ci/junit.xml --passed <the 16 names>: workspace owners OK (16 passed)
• cargo fmt --all --check, cargo clippy -p omnigraph-cli --all-targets --locked -- -D warnings -W clippy::dbg_macro: clean
• the workflow's property-test step at PROPTEST_CASES=48: not run locally; the step and the engine crate are untouched by this change

Notes for reviewers

• A nightly: test changes only when its author remembers it, since no PR check runs it; the same drift recurs with the next CommitOutput field until the gate runs these tests under a settable deadline.

[chatgpt-codex-connector]

Codex Review Summary

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

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-13T12:59:33.503408Z	53c8550	PR opened

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.