Reconcile capture text before disposition stamps - #2417
Conversation
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. |
Chris0Jeky
left a comment
There was a problem hiding this comment.
Bounded fresh-context review complete.
Round 1 found one HIGH: already-cancelled retries bypassed the new CAS and could leave a pre-fix masked archived mismatch serving stale durable text. Commit 7062a89a removes that early return and adds a red-first recovery regression.
Round 2 reviewed exact head 7062a89a76159c5d8dcb827529905decbc19ae4f against 468d76dc8e7a7859324c40e6369c3caae57fe647. The prior HIGH is closed and no new CRITICAL/HIGH defect remains. CAS loss rolls back before durable access; a successful idempotent CAS reloads the exact payload and advances queue freshness; archived repair rejection leaves the aggregate unmodified and restores truthful queue fallback.
Verified at this head: Application capture filter 121/121, API capture/queue filter 96/96, focused command-shape test 1/1, docs governance, and diff check. The full solution had one separately tracked #2399 instrumentation failure after 8,725 passes and five declared skips; the exact test passed twice on clean main and once here.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 7062a89a76
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".
Chris0Jeky
left a comment
There was a problem hiding this comment.
Current-base interaction check for pushed head 64a0aace70e629095e5640c3a0fb7056c3be4e2c against 4f0f0525f7e789f3a68a77ebb4fdcad55dac5f96 is clean.
The landed base delta is limited to the Smart CI schema/planner/test files from #2415. The PR diff remains the same ten #2347 files, with no reference or execution-path overlap. Re-proved after the merge: Application 121/121, API 96/96, Smart CI planner 41/41, docs governance, and diff check. The prior HIGH remains closed; no new CRITICAL/HIGH interaction was found.
Summary
LIMIT batchSizethrough onejson_eachparameter.Verification
node scripts/check-docs-governance.mjsgit diff --check origin/main...HEADdotnet test backend/Taskdeck.sln -c Release -m:1 --no-restore: 8,725 passed and five declared skips; one command-shape instrumentation test failed because hosted-worker SQL entered its uncorrelated capture window. The exact test then passed twice on clean main and once on this branch. Existing issue [Backend][Testing] Isolate intermittent batch command-shape samples on Windows #2399 now carries the evidence.Documentation
CaptureService,ContextFabricSettings, anddocs/platform/CONFIGURATION_REFERENCE.mddocs/STATUS.mdis intentionally left to the canonical docs lane; a post-merge Claude sync packet will remove CF-01c: reconcile capture text before stamping dispositions — an out-of-band divergence followed by keep/archive is currently masked forever #2347 from the open CF-01 residualTracking
Risk notes