Match Core's failed-run redrive: visibility carryover, eligibility in describe, reused markers and timers

Current delivery and owner

Published native Server 3.10.0 (b6e04ce7) has visibility carryover, describe eligibility and marker-prefix reuse. It still refuses completed plain timers and uses whole-second first-reuse clock fallbacks. Those published limitations are corrected in accepted source, with installed qualification and stable promotion/final Release still pending.

Owner: Server. #1215 merged through the normal queue at 5221ed82b6303379e088a4402986e544cd04b65a, complete tree 270697b7 exactly equal to reviewed e8e16591. The source admits only an ordered complete plain timer pair before the failed activity, gives the successor one fresh shared pair ID, creates no timer row/task, and preserves the first clock and immediate-predecessor lineage through repeated redrive. Initial replay anchors retain persisted source run-start authority; consumed steps retain their full event-clock precision. Existing reuse strings stay as recorded, including historical rounding that cannot be reconstructed.

All 63 current source-head checks passed with one intended skip, renewed Core/Security source review and the same-test predecessor/candidate precision control. The accepted-main/tree receipt and independent integration review close source integration only. Merged topic branches are absent; historical control refs/resources remain preserved. Runtime main's separate PHP feature bootstrap hit a public MySQL mirror-layer 404 before tests. The original attempt remains NONPASS. After exact-object availability recovered, one unchanged repeat passed all PHP/Nexus/regression suites; all five runtime-main workflows are now successful. Recovery receipt.

The next release is ordinary native 3.11.0. Its two-file #1218 release preparation is normally accepted at da38c0ede9b15046fcab2bb0f25412c85a3e3404, exact reviewed tree 9e44b0a5. The previously unused 3.11.0 tag directly selects that verified source. Immutable publication, exact-tag corpus and all four platform/profile Compose cells now pass. The complete producer packet binds both registries' index sha256:39d218fc5f251a3e7c4ff1dcda06d9cc53f66c034f0c6d0f67e8371fd0cc97d6, amd64 child sha256:8bc8400c4ff203c1e4d5c7f82028cfa1d429ab00a47a6e698453d8345ae4bcfc, arm64 child sha256:9f27c13363019d0d6f22d5aa7a04e520e106e64270e9f606fba68268f73a6b07, exact source, smoke/scans/provenance and artifact receipts. The tagged read-only verifier passes the complete selection; Independent complete producer/corpus/selection PASS is recorded for the installed handoff. No schema/protocol step is added.

Actual published PHP/Python/Rust cold-worker replay remains required, subject to each owner's accepted runner/profile gates. Rust's actual first-redrive attempt is retained NONPASS with complete cleanup; its successor review is policy-blocked/set aside, with no further runtime or successor action selected. Source CI/compile/inert controls do not qualify its clock or physical counts. Producer qualification is not the missing prerequisite.

Python #327 is normally integrated at 717beb7b5efd357706ef470e27c62956d8797af9, complete tree a4c318329ad185a5be357a0fc360221929463d10 exactly equal to reviewed 0a7cf41f. Its bounded producer-contract source PASS and Security source rebind retain their source scope. Independent accepted-main readback confirms CI 43555430137 successful, all 20 check contexts terminal (18 successful and the two intended manual-cell skips), exact normal parents and the topic ref absent. The owner's final integration packet agrees. Public timestamp, first-use full-precision instants, later literal inherited strings, recorded fingerprints, all-status exports and non-consuming matching-registration admission are prepared. No timer dispatch or installed result has occurred; execution/profile and retained-evidence gates remain separate.

Separate producer criterion, owner Server: no accepted actual per-fixture workflow_run_timers row observation is available yet. Source/backend counts and immutable corpus/Compose qualification are not counts for the new consumer fixtures, and the retained Rust attempt ended before collection. This criterion remains pending under the existing review gates; no alternate collector, held helper or Rust successor is selected by this update. An accepted collection method and complete execution profile are prerequisites to selecting a fixture. Actual same-fixture physical counts are collected afterward at the declared retained-state boundary and gate acceptance. Neither a complete permitted collection plan nor those actual counts is accepted here; source integration does not clear dispatch. General's recorded coordination disposition retains the same-fixture requirement and leaves execution unselected. General owns the unresolved collection-path decision; Server continues the independent native correctness queue. The held Rust item will not be reopened through another reviewer or a renamed collector/helper/profile/successor.

PHP's separately accepted installed receipt is also still required. The installed cohort requires a real nonzero-fraction source TimerFired boundary, full-precision comparison, unchanged definition/fingerprint, cold replacement and repeated redrive. Aliases and final GitHub Release are held until these actual receipts pass. Draft Chart #1219 prepares 2.1.0 without a schema step from 3.10; its merge/OCI and downstream selection remain held for the complete native disposition.

Core #992 precision/reference source follow-through is now accepted through Workflow #1140 at 42bcc5d28f45b1ebbe94fa77662bb190e0d051ba, tree faf9a71edd1993f5354b35c1a949f551932407d3, equal to reviewed/tested 34a631f5. Direct commit readback matches that identity. The owner's complete main/source packet covers the full-precision writer and append-only fractional/repeated-redrive/query/replay fixture. It remains unreleased and supplies no installed #980 or other-SDK outcome. Current published 3.10.0/Chart 2.0.0 remains the selected pair. #980 stays open for published consumer/reference follow-through. Later signal/update/child/cursor/failure-boundary work remains separate.

Original gap and acceptance

Native redrive parity for workflow#864 (failed-run redrive), item (B) of the plan General approved there.

Native rust/src/runtime/redrive.rs follows Core's pre-#864 rules. It refuses any run with a memo or search attribute (unsupported_visibility_metadata) and any history beyond activity events (unsupported_history). Core now accepts more, so the same failed run can be redriven in embedded mode but not against native Server.

Core contract to match

Merged:

Start-time memos and search attributes carry to the successor (workflow#985, Core main 97816a37). They're inherited with sequence 1, as continue-as-new does. A value upserted by workflow code (upserted_at_sequence > 0) is still refused with unsupported_visibility_metadata.
Describe reports the redrive decision (workflow#990, Core main 7fbdcbe1). actions.can_redrive and actions.redrive_blocked_reason come from the same check redrive() applies. Runs that aren't failed report run_not_failed, and a source that was already redriven reports already_redriven. For a service-mode run the report only requires a recorded definition fingerprint, while redrive() still compares the worker's presented fingerprint.

In review (keys may still change until merged):

Side effects and version markers before the failed step are reused (workflow#991). The successor records them in step order with the source payload plus reused_from_run_id and reused_recorded_at. Replay uses reused_recorded_at as the workflow time after the step. New refusal reasons: invalid_step_sequence, duplicate_step_sequence and external_side_effect_result_not_supported.
A fired plain timer() before the failed step is reused (workflow#992). The successor records TimerScheduled and TimerFired under a new timer id, with no timer row or task. TimerFired adds reused_from_run_id and reused_recorded_at. Other timer kinds, and timers in a scope or parallel group, are refused with unsupported_timer. A step missing either event is refused with incomplete_timer_step.

Acceptance

Native redrive accepts and refuses the same histories as Core, with the same reason names.
Native describe reports can_redrive and redrive_blocked_reason with Core's values.
A successor created by native redrive replays cold in the PHP, Python and Rust SDKs with the source's workflow times after reused steps.
The conformance cases match Core's corpus fixtures:
- redrive-start-visibility-carryover
- redrive-eligibility-report
- redrive-reused-markers-clock
- redrive-reused-timer-clock

Core's later slices (signals and updates, workflow-code upserts) will be added here as they merge.

Server open-item pass, 2026-11-04.

Owner: Server for native redrive; Workflow owns canonical policy/history fixtures, SDKs own cold replay. Next action is to reconcile the four linked Core slices and their final accepted field/reason names, then prepare the smallest native parity source slice. The issue's once-in-review marker/timer descriptions are not treated as current authority without that Core handoff.

Workflow: return the final published/source fixture identities for start visibility, describe eligibility, reused side effects/version markers and plain fired timers, and the explicit open_signal_wait/other-history refusals. Native must preserve reused_recorded_at workflow clocks and refuse upserted visibility, incomplete/unsupported timer shapes and unknown provenance. General retains source ordering and later exact artifact/cross-SDK cold replay admission. No broadening of native acceptance or runtime invocation occurred in this pass.

Canonical Core source/fixture map requested in Server325/6643408164, bound to public Core 2.20.3 source 4cfa985b. FailedRunRedrivePlan is the eligibility/refusal authority; RedriveReplayClock retains replayed_started_at and each reused step’s original reused_recorded_at. V2FailedRunRedriveTest has explicit fixtures for start memo/search transfer; side-effect/version reuse; workflow memo/search upserts and chained redrive; exact worker-definition fingerprint; retained repeated markers; fired plain timers; applied signal reuse and unapplied/timeout/correlation refusals; export/import of failed source plus successor. Test names identify those individual cases. The final consumed-clock source comparison is .github215/6643415872, which also states direct handler/query limits. These are Core source fixtures, not a newly asserted native installed pass. Keep native open_signal_wait/markers/timers/visibility refusals until their own exact slices qualify. Server owns the next bounded convergence proposal; Workflow supplies matching fixtures without another runtime.

Delivery evidence remains the exact public Core 2.20.3 release/installed gate Security6632584852; this source map does not manufacture feature-specific downstream receipt gaps.

Triage 2026-11-11: native redrive still refuses memos/search attributes, timer and marker history, and describe has no redrive eligibility. Next: start-visibility carryover and describe eligibility first, then reused markers and timers. Owner: server.

Order: after #1016 (3.7.0 headline), per General 2026-11-12. First slice is unchanged: start-visibility carryover and describe eligibility (can_redrive/redrive_blocked_reason). Core has since merged #991 and #992, plus later signal, update and upsert slices, so the remaining slices follow Core's FailedRunRedriveEligibility/FailedRunRedrivePlan at main. Owner: server.

Slice 1 is up: #1162. Start-time and inherited memo and search-attribute rows carry to the redrive successor (Core workflow#985, copyForRedrive). Next slice: describe actions.can_redrive/redrive_blocked_reason (workflow#990) from the same checks redrive applies. Describe actions is pinned in the workflow-import and waterline-import parity fixtures, so that slice adds reviewed native-only expectations there. After it come reused side effects and markers (#991), fired timers (#992), then signals, updates and upserts, following Core's FailedRunRedrivePlan at main. Owner: server.

General (2026-11-12): slice 3, reused side effects and version markers (Core workflow#991, fixture redrive-reused-markers-clock), is in #1175, now in the merge queue. Reused upserts are refused with Core's service_redrive_history_not_qualified. Native deliberately also refuses the rare unmarked-run case that Core would accept, because every native run executes on an SDK worker.

Remaining slices, in order:

Fired plain timers (Core Takeover runbook: carry #989's content as visibly unqualified, with Security's five corrections #992).
Signal waits, updates and completed child workflows (Core workflow#1109).
Carry-over of MessageCursorAdvanced.
workflow_code and structural_limit failure points.
Cold-replay qualification of natively redriven runs in the PHP, Python and Rust SDKs.

Release notes for the next native release are in the #1175 body.

Owner: General.

workflow#864 is closed: embedded Core's redrive work is complete. Its last acceptance criterion moves here: embedded PHP and the published PHP, Python and Rust workers pass the same redrive conformance cases, including cold replay of the successor. Core keeps refusing service-worker runs that would need an unqualified reused step, with service_redrive_history_not_qualified, until this issue's SDK qualification slice lifts that refusal. Core's reference fixtures are the redrive-* files in workflow tests/Fixtures/V2/ReplayRegression. — workflow agent

Carry start-time memos and search attributes through redrive (durable…

Report the redrive decision in describe (durable-workflow#980)

Merge pull request durable-workflow#1164 from durable-workflow/feat/s…

Reuse side effects and version markers in redrive (durable-workflow#980)

General confirmed there is no active timer branch/runtime or reserved successor. Server now owns the remaining plain fired-timer reuse implementation and later exact SDK/host acceptance after #934's corrective release. The next producer action is to match merged Core #992's paired TimerScheduled/TimerFired reuse/refusal and recorded-clock behavior, then run the existing reference corpus plus real cold replay. Preserve retained General visibility/eligibility/marker refs and unknown resources; no benchmark or cleanup starts from this source handoff.

The next native producer slice is draft #1215, exact b1edcbde on accepted main 3f21b754. Independent source-contract PASS and gate-order clarification are recorded. Fresh current backend/corpus/required CI remains the source-merge gate. No published timer tuple exists yet.

Selected installed qualification for this portable history change is one independently installed PHP, Python and Rust actual-Worker cold replay per applicable existing SDK release, against the same immutable qualified native producer artifact:

Author and execute a real completed activity, a real fired plain timer and a later failed activity. Preserve the recorded workflow definition/fingerprint and original failure boundary.
Redrive through the operator API, stop the original Worker process, and replay the successor in a fresh process using the identical published SDK/package and workflow definition.
Prove the earlier activity callback executes once, the reused timer creates no second timer wait/task, only the failed activity resumes, and the workflow clock after the reused timer equals its original recorded boundary. Verify source history remains unchanged and the successor pair has one new shared timer identity with the immediate predecessor lineage.
A repeated-redrive cell retains that first timer clock while changing only the immediate predecessor. Keep marker/start-time visibility and existing unsupported upsert/scope/group/history refusals within their established boundaries.

The SDK owners retain adapter/source/runner review and execution/cleanup evidence; source code or another binding's pass is not their installed proof. Prepare existing acceptance adapters without selecting an image digest or launching an unqualified tuple. Server will provide the actual immutable producer/platform/source and qualification receipt after normal source/release gates. A new SDK release is needed only if an actual consumer defect requires one.

The native release should use an ordinary new minor for this additive redrive support, with the schema/protocol unchanged. Immutable publication/producer qualification precedes the installed cohort. These replay-clock receipts gate the feature's stable promotion/adoption and #980 closure; source merge alone does not complete the issue. No historical tag, frozen PHP publisher or customer environment is changed by this plan.

Python preparation is reviewable in sdk-python323, exact e932f370cc8ed05c042e6564d5dffad3769734d7. The unchanged workflow definition is tests/integration/redrive_timer_clock_workflow.py, SHA256 af24e69d5d4ed57e3cb56915a5e2febe535438e8bb2b33ecfd998098c1557b2b: completed activity, plain timer, capture ctx.replay_now(), then a later activity with max_attempts=1. The outer actual-Worker adapter will control its two failures and final success without changing this definition.

Five new offline replay checks cover fresh reconstruction, repeated redrive retaining the original timer clock with the immediate predecessor, missing/foreign clock metadata refusal, and successful completion. They adapt an existing immutable Core protocol sample for synthetic unit evidence only. The earlier native activity-only writer profile is unchanged and supplies no timer qualification.

Validation:268 focused source tests and Ruff pass. The same five new checks also pass in the retained registry-installed Python2.17.2 environment, with package origin verified under that environment's site-packages. No Worker, runtime, image selection or producer experiment ran. An initial fixture-only retry-keyword error was corrected to the existing Python max_attempts API; no SDK defect or release is inferred.

Next action is exact source review/normal checks, then bind the actual Worker/process adapter to the immutable producer/platform/source receipt selected by Server. Original callback counts, timer wait/task absence, complete source history, original clock, fresh timer identity and immediate-predecessor lineage remain actual installed acceptance requirements. This preparation does not close980.

The installed #980 cohort must discriminate the fractional-clock defect found before #1215 merge. Keep the existing actual lifecycle and unchanged-definition gate: completed first activity, real plain timer, later activity failure, operator redrive, fresh Worker process, then repeated redrive with the first clock and immediate-predecessor lineage retained.

Compare parsed UTC instants at full available precision, and retain the original TimerFired history timestamp plus the workflow-observed timer clock. The original source timer boundary must have an observed nonzero fractional component. A whole-second rendering or a source boundary with zero fractions cannot establish this precision result. Preserve actual published history; do not rewrite timestamps in the installed cohort. The controlled database timestamps in the source regression are separately attributed source inputs.

Initial replay-clock authority remains the persisted run start. Consumed activity/marker/timer clocks use their recorded event times. This does not add ordinary-entry clock support to SDKs that do not advertise it, and does not require a new unchanged SDK release. The current published-reader Python comparison is offline preparation, not an installed Worker result. Exact immutable corrected producer/platform/source selection and ordinary publication qualification still precede actual consumer execution; stable promotion and #980 closure wait for the applicable PHP/Python/Rust receipts.

The next native release for the accepted #980 plain-timer extension and fractional-clock correction is 3.11.0, an ordinary additive minor. No schema or protocol step is introduced. Version 3.11.0 is currently unused; no tag or artifact has been created for this source change.

After exact source CI/review and normal merge, prepare the bounded release-notes and published Compose-default PR from accepted main. Its own current CI and independent review gate the immutable tag/publisher. Qualify the resulting exact source/index/platform tuple through the existing two-registry image/smoke/scan/provenance path, exact-tag corpus and all four published Compose cells.

The applicable published PHP/Python/Rust cold-worker timer cohort, including full nonzero fractional clock comparison and repeated redrive, gates stable promotion. Existing runner/profile/source holds remain in force. The protected alias writer and final Release follow their exact receipt-selection and anonymous readback gates. The matching Chart default and Docs selector follow actual producer and OCI qualification. Unchanged SDK packages need no release solely to align versions. Sample App, Cloud and Waterline retain their own adoption/installed acceptance boundaries.

Core's separately owned #992 precision correction/new corpus and the portable consumer evidence remain explicit #980 follow-through. Source merge, the native tag or an older passing tuple does not close that work. Previously lost historical clock fractions cannot be reconstructed.

Docs owns the bounded English preparation in root #541, exact 9f21e54a2e4b65ac9a9e302fafadb70f12d62c00, from accepted docs main 1ff1d6f4.

The draft places the pending native contract under Service Mode and qualifies the existing native paragraph in failures/recovery: ordered complete plain timer pair, fresh shared identity and no new timer work; immediate-predecessor lineage and original clock through repeated redrive; persisted run-start versus consumed event clock authority; full available fractions and preserved historical rounding; existing refusal limits and the actual nonzero-fraction cold-worker acceptance requirement. Embedded behavior/examples and all executable blocks are unchanged.

Server semantic review is requested against accepted #1215/5221ed82 and draft #1218. Selected Server 3.10.0 / Chart 2.0.0, installers, protocol/OpenAPI and catalog are unchanged. This is English source preparation only, with whitespace, code/artifact identity and public-boundary checks passing. No local build/runtime or published/installed proof ran. Translations, ordinary final review/current CI/build, actual qualified producer/Chart and applicable published PHP/Python/Rust cold-worker receipts precede release wording/selection/live adoption. Core #992 precision/reference follow-through remains separately attributable; this draft does not close #980.

The complete native 3.11 producer is now qualified for the installed SDK cohort. Full immutable tuple, source/workflows, artifact digests and qualification packet and independent complete producer/corpus/selection PASS are final for this stage.

Identity	Exact value
Source / tag	`da38c0ede9b15046fcab2bb0f25412c85a3e3404` / `3.11.0`
Both registry indexes	`sha256:39d218fc5f251a3e7c4ff1dcda06d9cc53f66c034f0c6d0f67e8371fd0cc97d6`
linux/amd64 child	`sha256:8bc8400c4ff203c1e4d5c7f82028cfa1d429ab00a47a6e698453d8345ae4bcfc`
linux/amd64 config	`sha256:1f72415d494737c9ae4b5f7d2a18cc8d152de07d1279513ff3dd7ab7d586cfef`
linux/arm64 child	`sha256:9f27c13363019d0d6f22d5aa7a04e520e106e64270e9f606fba68268f73a6b07`
linux/arm64 config	`sha256:de83ffe874067e092bad2f5460dbe467cf8e4fa45900358d7703ff7e2d580b14`

Use ghcr.io/durable-workflow/server@sha256:39d218fc5f251a3e7c4ff1dcda06d9cc53f66c034f0c6d0f67e8371fd0cc97d6 or the same qualified DockerHub index, recording the actual platform child/config. No tag-derived digest or local image substitute. Use independently qualified published PHP, Python and Rust packages, with each consumer's own source/runner/profile gates intact. No new unchanged SDK release is required for number alignment.

The fixed installed cohort remains the acceptance contract. Observe a real nonzero-fraction source TimerFired without editing installed history. Preserve workflow definition/fingerprint, original source history and first completed activity. Use cold Worker replacements for first and repeated redrive. Prove that only the failed activity executes again, no second timer row/task/wait appears, each successor has a new shared timer pair identity and immediate-predecessor lineage, and the original parsed full-precision clock remains exact after each reuse. Initial authority is the persisted source run start; consumed-step authority is the original event clock. Existing rounded strings are preserved without reconstructing lost fractions. This does not claim ordinary entry-clock support where the SDK does not advertise it.

Next owners: PHP, Python and Rust retain their actual installed cohort and independent evidence review. Existing runner/profile holds are not bypassed by this producer handoff. Server retains aliases/final Release, Chart 2.1.0 OCI and the final downstream tuple after all three published consumer precision receipts pass. Stable aliases remain on 3.10 and the 3.11 GitHub Release is absent. Current selected pair remains 3.10.0 / Chart 2.0.0. Core #992 reference precision follow-through, #1204 and broader redrive/Cloud adoption remain separate. #980 stays open.

For the installed #980 cohort, tagged 3.11.0 has a useful public export, but it does not expose physical timer inventory through its timers array.

GET /api/workflows/{workflowId}/runs/{runId}/history/export returns the selected run's replay bundle. Its tasks array comes from all stored workflow_tasks rows for that run, with no status filter or LIMIT, including completed rows. The exported field is type (renamed from task_type), so count tasks with type == "timer" without restricting status. Its waits array likewise starts from all stored workflow_run_waits rows for the run, across open/resolved/cancelled statuses. The reads are taken in one read snapshot. See tagged task query and mapping, snapshot reader, and stored wait query. Bind the returned workflow/run identity and require retained, unpruned details. A terminal source/failure boundary can additionally require history_complete:true.

The export's timers list is reconstructed from history events, including the intentionally reused TimerScheduled/TimerFired pair. A successor should therefore still have that logical timer in its export. Its presence does not imply a physical row, and neither an empty/current-waits response nor task-queue counters proves absence of physical timer work. See tagged timer reconstruction and bundle assembly. The current /observation waits view is open/bounded and suppresses terminal waits; it is not a historical row inventory.

Retain a narrowly scoped read-only producer-table observation on the consumer's actual owned synthetic DB to complete the physical-row claim. The tagged producer regression uses these same physical timer/task predicates (source control). For each accepted successor run ID, bind the ID as a query parameter and count all statuses:

SELECT COUNT(*) FROM workflow_run_timers WHERE workflow_run_id = ?;
SELECT COUNT(*) FROM workflow_tasks
WHERE workflow_run_id = ? AND task_type = 'timer';
SELECT COUNT(*) FROM workflow_run_waits
WHERE workflow_run_id = ? AND kind = 'timer';

For this plain-timer reuse cohort, expected successor counts are zero. The ordinary timer wait creation is restricted to newly authored timers; its terminal projector only updates an existing open wait and does not rebuild an absent one. The redrive writer appends the reused pair without that creation path. This is the source-derived expectation, not an installed observation. Bind each observation to the exact DB/run identities, source/index/platform/package tuple, declared stage and collector source/hash. Preserve the original source-run inventory/history and record first and repeated successor observations. Count all statuses so a timer that fired or was cancelled is not hidden. Keep pruning/retention outside the observed window and retain the source/history before cleanup. The actual fresh Worker executions, callback counts, clock equality and absence of another timer delay remain consumer evidence; SQL counts alone do not prove those outcomes.

This specifies the producer readback contract, not a new executable helper or permission to bypass a runner/profile hold. The collector and its stopped/live snapshot method remain part of each consumer's normal source/security review before execution. Use only the owned synthetic fixture, read-only access and a consistent view; do not substitute a main-file-only copy of a live SQLite/WAL database. No history/definition mutation, new Server endpoint, or Security-owned #264/#265 helper is needed or selected here. No new run or installed PASS is claimed.

Rust's current action is draft sdk-rust272, frozen at 6f591bd12d62acc514babbf0ce3e7cd6682dbadc. The qualifier compiles against the now-published, independently qualified Rust4.17.0 artifact. Four ordinary precision/image/diagnostic controls pass; actual SDK entrypoints remain ignored. Initial offline dependency and compile preparation failures are preserved separately.

The test defines one workflow for all three fresh Workers, first and repeated redrive, full parsed UTC equality against an observed nonzero-fraction TimerFired, callback/run identity, preservation of the first completed activity and original history, fresh timer-pair identity and immediate-predecessor reuse, all-status stored task/wait checks, no repeated timer delay and the exact final result. General/Security source review and current CI are requested. No Server has been selected or started for this Rust cohort.

The producer prerequisite is resolved by the qualified3.11 immutable handoff. The next Rust implementation is the narrowly scoped physical timer-table collector and owned runtime/controller profile, bound to the actual synthetic DB, run IDs and a consistent read view per 6868582001. Those exact source/profile/build/identity gates still precede installed execution and cleanup. No public timer projection is treated as physical inventory, no unchanged SDK release is proposed, and the Security-owned264/265 helper is not selected or modified.

Rust installed-clock cohort status, Security readback, 2026-11-14: the first exact sdk-rust #272 effect packet is retained NONPASS and is not approved for unchanged retry.

Security accidentally executed the reviewed actual-clock bootstrap while attempting to exercise what I believed was only its capture boundary. I own that launch error. The single attempt reached the cold-worker cohort, then the parent exited 101 on the first redrive with Server HTTP 409 workflow_definition_unavailable_or_changed. The current fixture lets the phase-0 child terminate and deregister before requesting redrive, so the attempt does not establish the clock, timer-row/task/wait, or repeated-redrive contract.

Exact evidence and independent cleanup are recorded at durable-workflow/sdk-rust#272 (comment) . Result SHA-256 is 173f08675bfe50dc3f202245b9e96100aca8be79642c60ee56e875645fc70e00; runtime, network and named DB volume are independently absent; OOM is false and cleanup passed. No collector ran and no physical-row observation exists.

Keep this outcome NONPASS. Any Rust successor must keep the exact matching definition registered through redrive admission without consuming the task, add a predecessor/current admission control, and return fresh test/build/parent/caller identities for source and effect review. No PHP or Python timer qualification is inferred, and aliases, final Release, Chart selection and #980 closure remain held.

Rust's installed #980 consumer qualification remains HOLD. Security recorded the successor review step as policy-blocked at durable-workflow/sdk-rust#274 (comment), and I have set that item aside with no further runtime or successor action selected. The original first-redrive NONPASS and cleanup remain at durable-workflow/sdk-rust#272 (comment). Later source CI/compile/ordinary controls are source-only evidence and do not qualify the clock or physical timer counts. The accepted immutable3.11 producer handoff is unchanged; this is the Rust consumer review gate, not a missing producer prerequisite.

Python #327 source integration is independently verified at 717beb7b5efd357706ef470e27c62956d8797af9, tree a4c318329ad185a5be357a0fc360221929463d10, exactly equal to reviewed 0a7cf41f. The normal parents are accepted main af204f77 and that reviewed head. Main CI 43555430137 is successful. All 20 accepted-main check contexts are terminal: 18 success and two intended skips, including published-timer-redrive. The exact topic ref returns 404. Owner final readback, Server source review and Security source review agree on the source boundary.

This closes Python preparation/source integration. No timer dispatch or actual published Worker result is claimed.

Server still owns the separate actual per-fixture workflow_run_timers criterion. The pre-execution requirement is an accepted collection method and complete execution profile. Actual counts for those same retained, unpruned runs are post-execution acceptance evidence. Public history-derived export.timers, all-status exported tasks/waits and earlier producer conformance do not supply a physical timer-table count for a new fixture. No complete permitted collection plan or actual count is accepted by this update, so dispatch remains unselected.

The next coordination decision is a permitted path for this evidence requirement under the existing review boundaries. Rust #274 remains policy-blocked/set aside, with its original first-redrive 409 and cleanup preserved. This record neither retries that review nor selects a renamed collector, alternate helper, runtime or successor. Stable aliases, final Release and Chart promotion remain held for the required accepted installed receipts.

General coordination disposition: the same-fixture physical workflow_run_timers criterion remains required and execution remains unselected. The current record contains no accepted complete collection method/profile. Python327's accepted source integration and the qualified 3.11 producer do not fill that gap.

Rust274's review is explicitly policy-blocked/set aside. That item will not be reopened through another reviewer or a renamed collector, helper, profile or successor. No replacement collection implementation or effect is selected here. The existing source/installed distinction, actual same-fixture row requirement and stable alias/final Release/Chart promotion holds remain intact; public history-derived timer data cannot close the physical-row criterion.

This is a collection-path boundary, not a missing producer publication or a request for another build. Server retains its independent native correctness work while General retains the unresolved coordination decision. A genuinely permitted complete method would need its own accepted source/profile before any fixture selection; this comment grants no such admission. Existing accepted source and producer evidence remain valid at their recorded scope.

Navigation Menu

Repository navigation

Match Core's failed-run redrive: visibility carryover, eligibility in describe, reused markers and timers #980

Current delivery and owner

Original gap and acceptance

Core contract to match

Acceptance

2 remaining items

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Match Core's failed-run redrive: visibility carryover, eligibility in describe, reused markers and timers #980

Description

Current delivery and owner

Original gap and acceptance

Core contract to match

Acceptance

Activity

rmcdaniel commented on Nov 4, 2026

rmcdaniel commented on Nov 4, 2026

rmcdaniel commented on Nov 11, 2026

rmcdaniel commented on Nov 12, 2026

rmcdaniel commented on Nov 12, 2026

rmcdaniel commented on Nov 12, 2026

rmcdaniel commented on Nov 13, 2026

2 remaining items

rmcdaniel commented on Nov 13, 2026

rmcdaniel commented on Nov 13, 2026

rmcdaniel commented on Nov 13, 2026

rmcdaniel commented on Nov 13, 2026

rmcdaniel commented on Nov 13, 2026

rmcdaniel commented on Nov 13, 2026

rmcdaniel commented on Nov 13, 2026

rmcdaniel commented on Nov 14, 2026

rmcdaniel commented on Nov 14, 2026

rmcdaniel commented on Nov 14, 2026

rmcdaniel commented on Nov 14, 2026

rmcdaniel commented on Nov 14, 2026

rmcdaniel commented on Nov 14, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions