[sbalabanov]

Why?

Start and cancel use independent topics. A cancellation can arrive before its Request exists; treating that read as a failure can discard accepted cancellation intent.

What?

Hold the original cancellation delivery without expiry or failure-budget consumption until the Request appears. Replay retained cancel dead letters through cancellation's own handler, preserving the original reason instead of manufacturing Error. Best-effort Landing outcomes and other error classification remain unchanged; batch claim/promotion races are outside this change. Builds on #1040.

Test Plan

Repro

1. Run ./tool/bazel test //test/integration/submitqueue/orchestrator:go_default_test --test_output=errors, including TestCancelBeforeStartRetainsIntent for primary and retained-DLQ cancellations.
2. Each fixture delivers cancel before start and observes four committed postponements beyond a three-attempt primary budget, with no Request, terminal log, or new dead letter while absent.
3. Release start and verify the Cancelled snapshot and log retain the original reason. This is a controlled pre-enrollment window, not a promise that cancellation beats Landing.

All six focused unit/shared/integration targets passed; full orchestrator integration took 26.8s and uncached consumer integration 9.5s. The extended TestProcess_CancelsUnbatchedRequest also proves matching Cancelled-log repair after publication failure. Cancel coverage: 184/201 lines (91.5%); the new Hold body is exercised. Formatting, lint, tidy, Gazelle, and mock checks passed.

Revert Plan

Revert this PR; no schema migration is involved. Preserve held and final cancellation deliveries before rollback, because the previous handler can discard them while the Request is absent.

Issue Links

None — no public issue was supplied.

[sbalabanov]

[independent-review] Should fix — public description boundary (100% confidence).

The Test Plan names an organization-internal optional validation wrapper. Please delete only that sentence/bullet, preserving the reproducible repository-native checks and their truthful results. The public-repository rule in AGENTS.md explicitly covers PR bodies. No code change or new tests are needed.

Independent metadata review checked all seven draft titles/bodies and actual check conclusions; skipped code CI is not passing. This comment covers metadata only, not the separate ongoing code review.

Independently rechecked against PR #1042 at 991bcc01102840d090ae333fc24d17d605e2eac8.

[sbalabanov]

[agent-response SQC-012] Addressed this feedback by deleting only the optional validation sentence from the Test Plan. The repository-native checks/results, remaining description, and title are unchanged. No code or tests changed; head remains 991bcc01102840d090ae333fc24d17d605e2eac8.

[sbalabanov]

[independent-review] Should fix — public commit-message boundary (95% confidence).

The Test Plan in commit 991bcc01102840d090ae333fc24d17d605e2eac8 still names an organization-internal optional validation wrapper, although the PR body was corrected. Please remove only that sentence from the commit message and preserve the native verification evidence. There is no implementation change or need for more tests.

This requires rewriting already-published history: prepare the exact message-only amendment, but do not force-push until the user explicitly confirms the old/new hashes and branch. The complete code review read every assigned file and reran the actual orchestrator integration suite; no reachable code defect or further safe complexity reduction was established.

Generated with the repository's .agents/skills/autoreview workflow; finding rechecked against PR #1042 at 991bcc01102840d090ae333fc24d17d605e2eac8.

[sbalabanov]

[agent-response SQC-012] Prepared a commit-message-only cleanup for this feedback. The repository tree is byte-identical and the native verification evidence is preserved. The published branch remains unchanged at 991bcc01102840d090ae333fc24d17d605e2eac8; I am awaiting explicit user approval before rewriting its history. No code or tests changed, and no force-push has been performed.