[Task] Plan individual product releases and verify available prerequisites

Accepted on actual main 570aaa95b09a7043273df2bcd4f533f216804694. Final PR #8692 normally merged after exact-head independent APPROVE + HIGH, Greptile5/5, all nine Actions, two clean lead cycles and six actual plan receipt audits. All seven criteria and source-identical main acceptance. Queue: Done / Accepted / Verification Passed.

Program: #8194
Parent: #8217
Predecessor: #8683, accepted on actual main2a17d4635b75364b6577c21d217fec724ade9ab8.

This Task completes only nonpublishing product planning. Executions remain bound to34cd9576; actual-main tree equality was verified separately. Real selected artifacts/clean consumers, publisher/recovery and operational/archival work remain open. The minor future-trigger path-filter correction is explicitly tracked under #8217 and required in the immediate artifact Task before omitted-input changes. Earlier real failure/rework evidence is retained in #8691, without relabeling old executions.

Outcome

Add a nonpublishing release planner for one explicitly selected product (Core, Studio or Extensions), source line and requested version. Emit the exact selected/excluded package and expected artifact scope, applicable test closure, and read-only prerequisite/version-history observations. Selecting one product must never select another product for publication to satisfy its dependencies.

This is the first bounded deliverable under #8217. Actual selected artifact builds, clean-consumer proofs and controlled artifact-only publisher/recovery are subsequent required deliverables; this plan does not complete #8217.

Source and policy boundaries

Core 3.8/3.9 binds an exact observed Core release-branch commit/tree and separately records published tag/version history. Core is not a [Task] Admit and verify Core-owned Studio and Extensions maintenance commits #8683 admitted product. Refresh actual refs before planning; branch tip and stable tag can differ.
Studio and Extensions 3.8/3.9 bind their matching explicitly admitted Core-owned maintenance continuation, original layout/anchor and source-evaluated inventory/dependencies. Current-main ProjectReferences are not maintenance dependency evidence.
A 3.10 product plan may describe scope, but must be explicitly release-ineligible until the approved first aligned consolidated baseline is satisfied. Preserve the existing finer connector/coupled release-unit model.
The selected product owns its explicit requested version; validate every selected package against fresh published version history for monotonicity within its selected supported release line and no reuse. A newer supported minor line must not by itself prohibit an older line's maintenance patch. Do not allocate versions, tags or refs. Product-qualified tags remain a proposal until separately settled.
Studio selection includes both existing npm IDs, @elsa-workflows/elsa-studio-wasm and @elsa-workflows/elsa-studio-wasm-react, as an atomic same-version/source pair alongside Studio .NET packages. Core-only/Extensions-only plans exclude these Studio artifacts. Original 3.8/3.9 manifests/workflows contain the pair, but checked-in dependency versions are stale: historical same-run tarball/version/dependency semantics require their own later artifact/consumer proof. [Task] Build and prove paired Studio npm archives from Core without publishing #8679 proves current/main only.

Implementation scope

Reuse current release-unit, exact source/evaluated inventory, package-impact, archive metadata and npm-pair primitives where they fit. Preserve the existing Slack unit and sealed consolidated candidate/executor contracts. Avoid a generic release-engine redesign.

For each outgoing prerequisite, identify the exact package/version and applicable TFM/dependency groups and inspect read-only feed metadata. Missing, incompatible, malformed, ambiguous, unreadable or stale evidence makes the plan ineligible with a closed reason; never expand publication scope. Keep affected build/test closure separate from the publication set.

Retain source/controller/tree and inventory identities, selected/excluded package bijections, expected main/symbol/npm artifact names, requested version and per-package history checks, prerequisite observations with origin/time/metadata identity, paired npm intent, eligibility and published:false. No credentials, raw logs, private paths or secret values in public receipts. Expected artifact names are a plan, not byte provenance.

Acceptance criteria

Execute read-only plans for Core, Studio and Extensions on the supported 3.8/3.9 source lines, bound to current exact sources and their evaluated inventories. Selected/excluded package membership is complete and disjoint; no outgoing prerequisite or other product enters the selected publication set.
Verify product-specific source binding, wrong-line/product/source/inventory rejection and explicit 3.10 aligned-baseline ineligibility. Original-layout metadata cannot be replaced by current-main project references.
Verify monotonic/no-reuse requested versions for every selected package using fresh read-only published history; retain unavailable/history ambiguity as explicit ineligibility. No version/tag/ref is allocated or created.
Verify exact prerequisite availability and compatible per-TFM dependency metadata. Contracts cover compatible/available, missing, wrong-version, incompatible framework, malformed/unavailable/stale evidence and duplicate/ambiguous identities; failures do not widen selection.
Studio plans include the exact npm pair atomically with line-specific source/workflow binding and same-version dependency intent. Current/main [Task] Build and prove paired Studio npm archives from Core without publishing #8679 proof is not relabeled as historical artifact proof. Existing finer Slack release-unit/package/test-closure contracts still pass.
Normal and optimized focused contracts and a read-only hosted rehearsal retain exact-head/source/line receipts for all supported product choices. Independently read actual receipts, including blocked outcomes; no synthetic plan is described as a real artifact or consumer proof.
Document the selected-product command, prerequisite failures, source-line rules, proof limits and remaining artifact/publisher steps. Independent exact-head APPROVE + HIGH, applicable CI and Greptile pass, followed by normal shared/main integration and bounded acceptance.

Execution and non-goals

Use isolated worktrees and preserve user changes. Sol 6.1 handles implementation/integration; bounded Luna 6 Extra High audits are suitable. Root retains final integration/QA. Adopt an existing matching PR if one appears.

No registry upload, publisher/credential/OIDC activation, permanent maintenance refs/protections, version allocation, deployment, old-authority retirement or archival. Do not widen the sealed 225-package executor or invoke historical publisher workflows. No Slack provider or new React Studio work.

Definition of ready

Scope/source/npm ambiguities were resolved from actual release refs, historical manifests/workflows and settled owner decisions. No new product-owner choice blocks this nonpublishing slice. The #8683 acceptance dependency is satisfied. Accepted source identities, matching PRs, board and tooling were refreshed before claim.

Claimed #8688 as the sole active delivery Task: In Progress / Assigned / Verification Pending. Native predecessor #8683 is accepted and closed on actual main, commit 2a17d4635b75364b6577c21d217fec724ade9ab8. Complete 153-item board, native blocker/parent, matching open PRs and current issue criteria were refreshed before claim; no matching implementation PR exists.

The merged final PR had removed the temporary shared program branch. codex/elsa-integration-program was recreated at actual main without overwriting an existing ref. Implementation uses a clean isolated worktree and branch codex/elsa-product-release-plans-8688; canonical user changes remain untouched. Sol 6.1 handles the bounded implementation; lead retains final integration and QA.

Scope is the seven existing criteria: six supported product/line read-only plans, exact original-layout source/inventory binding, per-package version history, compatible available prerequisites, atomic Studio npm intent, normal/optimized contracts and exact-head hosted receipts. The first checkpoint is real pre-build SDK metadata evaluation for an accepted Extensions3.8 source. Plans may be ineligible when actual prerequisite/history evidence is missing; selecting dependencies for publication is prohibited.

This does not allocate versions/tags/refs, produce release artifacts, publish packages, activate publisher credentials/settings, retire old authority or archive repositories. Actual selected artifacts/consumers and controlled publisher/recovery remain subsequent required work under #8217/#8220. Full3.10 certification remains separate; Slack and new React Studio remain deferred.

Implementation checkpoint for #8688; acceptance remains pending.

Root independently read the actual generator-bearing pre-build control for Elsa.Caching.Distributed at admitted Extensions 3.8 source 92c27a3dd2c7f1dc107cba34749dd293a5d77743, tree 6b7627924e64d4a6548ff5689000eff2f08561eb, hypothetical version 3.8.5. Both actual generated nuspecs retain net8/net9/net10 dependency groups. Root reconciled their byte hashes, source Git identity, original per-TFM manifest/output policies, actual generator/SDK target hashes and absence of current product DLL/nupkg outputs. The supported metadata projection suppresses compiled file collection without fabricating outputs. Root receipt SHA-256: a17b9fb0a71f5ae058fb6ee2b6a2f5816786c9015093abc7a79a34b137d42794.

This is draft metadata feasibility evidence, not final-head matrix, package, consumer or publisher proof. The earlier inventory run's generator-content blocks are superseded only by fresh evaluation, not relabelled. Root and Luna draft review corrections are being incorporated before the coherent candidate and six-cell hosted rehearsal: same-line version ordering includes higher prereleases; malformed/duplicate metadata must fail safely; unresolved project references must not disappear from test impact. Exact-head independent review, normal/optimized tests, CI, Greptile and ordinary integration remain required.

The complete 153-item board was refreshed. #8688 remains the sole active Task, In Progress / Assigned / Verification Pending. #8636 is a Todo / Agent Ready / Pending buffer after refreshing its bounded corrective scope against actual main, the current core/studio/extensions layout and the owner's resolved coordinated-update policy. Its historical timeout evidence remains intact; no updater run or code change was started, and it is not newly asserted as an archival prerequisite.

No package publication, version allocation, tag, permanent maintenance ref, publisher activation, deployment or archival occurred. #8217, #8220 and the full program remain open.

Release-planning implementation is now in PR #8691, exact head 328cd137a6156219c18d84f476df05cad0e0d32e, into codex/elsa-integration-program.

Standards: the one diagnostic-path finding was corrected and independently rechecked; no remaining actionable findings. Spec: no implementation gaps or scope creep; independent APPROVE + HIGH. Current39 affected contracts pass normally and optimized; prior151 full-suite results retain their actual prior-head attribution. The lead read all seven additions and the complete fix, verified file/evidence hashes, ran the prior full suite and current regression, and checked workflow syntax.

Read-only hosted rehearsal 37967102089, attempt1 is confirmed running on this exact head: all six Core/Studio/Extensions ×3.8/3.9 plan jobs and contracts have started. Actual public receipts and complete history/prerequisite results, CI and Greptile remain pending; a deliberately blocked plan must retain its reasons, not be described as artifact or consumer proof.

The complete153-item board and59 open PRs were refreshed: #8688 is the only active Task, now In Progress / In Review / Verification Running; #8636 remains the small Agent Ready buffer. Main/shared remain2a17d and canonical user changes are preserved. No product compilation/publication, permanent maintenance ref/protection, credential, deployment, retirement or archival activation occurred. All seven acceptance criteria and parent #8217 remain open.

The first hosted rehearsal has a verified execution failure: Studio3.9 job113944141124 returned an incomplete plan_input_or_metadata_unavailable receipt, not a completed plan with an explicit unavailable-prerequisite decision. Its original ZIP artifact11633628286 is retained; provider/download SHA256 5fe058da58b9f52fe2e9569abc6728339edd36bdeed583cd6bcde9540ebaafef matches.

Extensions3.9 completed with a full eligible76-package receipt (artifact11634117565, original ZIP SHA256 3c925e6e270d849ea4a9f026a40db755092375351a33e438ab2c3159eefd57fc); independent receipt QA is active. Other product/line jobs remain in flight. Neither outcome is artifact or consumer proof.

#8688 returns to In Progress / Assigned / Verification Failed while Sol6.1 diagnoses the exact Studio3.9 source path in isolation. No rerun, code workaround, merge or acceptance is authorized by the source-review approval alone. The lead retains integration/QA and will batch any verified correction before a new exact-head review and hosted checkpoint.

Lead rework checkpoint: PR #8691 remains at 328cd13; its initial approval is superseded and Task #8688 remains In Progress / Assigned / Verification Failed.

The first six-cell hosted rehearsal is terminal. Core 3.8, Core 3.9 and Extensions 3.9 produced complete eligible plans. Studio 3.8 and Extensions 3.8 retained unavailable source-metadata members; Studio 3.9 produced an incomplete receipt. These outcomes are retained at their original controller, not accepted or relabeled.

A local correction commit 360a8c2 is frozen but not pushed: the actual distinct Studio 3.8/3.9 npm workflows are now bound separately; a third supported SDK file-collection projection preserves original content policy while collecting real dependency/framework metadata. Forty-six affected tests passed normally and optimized. Root read actual paired SDK nuspecs, source project/assets and both genuine copied asset bytes; dependency/framework metadata matches, while contentFiles/file collections deliberately differ. No package or consumer proof is claimed.

Before the next push, the same worker is addressing newly received review findings: per-TFM reference inventory, overlapping recursive restore writes, Git exception containment, manual-run concurrency and wrong-branch failure reporting, plus small new-file style corrections. The original Extensions 3.8 hosted failure and separate ProtoGenTask CI failure remain unexplained; neither is waived or labeled flaky.

The board was reread: #8688 is the sole active delivery Task; #8636 remains the shallow ready buffer. All seven acceptance criteria remain open. No publication, credential activation, permanent maintenance refs, retirement or archival occurred. Next gate: one coherent corrected head, independent review, then fresh six-cell hosted receipts and applicable exact-head CI/Greptile.

Preserve historical Studio npm recipes and prebuild content metadata (#…

Retain conditional release references and serialize metadata staging (#…

Corrected PR #8691 is now pushed at 855d46561a857729fc13c1b625322d65e16afe72 into the shared integration branch. Independent exact-head review and correction ledger passes; the lead independently passed all49 planner tests. Worker normal/optimized49 suites and actionlint/diff checks pass with their attribution recorded.

Fresh six-source planning rehearsal 37971246567, applicable CI and Greptile/CodeRabbit review are running or queued. #8688 remains the sole active Task: In Progress / In Review / Verification Running. All seven acceptance criteria stay open until actual current-head receipts, CI/review, shared/main integration and bounded acceptance are verified. Previous328 failures and their unresolved cause limits remain preserved. No artifact production, publication, maintenance-ref activation, retirement or archival is claimed.

Planning slice #8691 normally merged after its exact-head review, all seven green workflows and actual six-receipt QA. Verified shared merge 34cd9576e07cbc0170a58ac60bc121f75b99c04b has ordered parents main2a17d / reviewed855 and identical tree cfc65f4e6db94b06eb6794a67acec4f3143531ec (merge-verification SHA256 c6192c9e8820778a27f4e532104b9a4e9eea1abc5babb546f6f22a844cfc793c). Final main integration is #8692, attached and open at34cd9576. Two clean lead cycles are recorded there; fresh exact-head independent review, hosted receipts, CI and automated review are running. Task remains the sole active leaf, In Progress / In Review / Verification Running, with seven criteria unaccepted until actual main verification. No publication, permanent ref/protection, credential activation, retirement or archive action.

Accepted all seven #8688 criteria on actual main 570aaa95b09a7043273df2bcd4f533f216804694 after ordinary final PR8692 merge. Verified ordered parents [2a17d4635b75364b6577c21d217fec724ade9ab8,34cd9576e07cbc0170a58ac60bc121f75b99c04b], exact reviewed tree cfc65f4e6db94b06eb6794a67acec4f3143531ec, clean isolated main checkout and continued containment of all35 advertised 3.8.x/3.9.x source release refs. Main verification SHA-256 10619956d5ee4102ebb248ac86eb327fa33d4a62cbdf70ae719a202c754104e0.

All six actual plans: complete/disjoint original-source inventories, product/ownership scope and outgoing-publication exclusions passed independent record audits.
Exact source/line/inventory rejection, all four admitted descendants and explicit3.10 ineligibility contracts passed.
Fresh per-package monotonic/no-reuse history passed native replay:696 NuGet and4npm decisions; no allocation/tag/publication.
All2,526 outgoing prerequisite edges join source nuspec groups/exact versions and native range/TFM checks; unavailable/missing/malformed/stale/ambiguous/wrong-version/TFM negatives pass without widening scope. This is dependency-metadata proof only.
Both historical Studio lines bind original recipe/manifests and exact atomic npm pair/dependency intent; independent Studio audit passed. Existing Slack/unit/npm contracts remain passing; no historical artifact claim.
Exact34cd hosted contracts passed162 normal +162 optimized without skips. All six original final-head ZIP receipts passed provider/head/run/byte identities, full record audit and native replay. Earlier failed/blocked328 evidence retains its original attribution.
Documentation, author-independent APPROVE + HIGH, two clean lead cycles, Greptile5/5 and all9 Actions passed on exact34cd; normal shared/main integration and actual-main tree acceptance are verified. CodeRabbit's completed advisory review has one acknowledged minor future-trigger correction openly tracked in [Feature] F3.1: Explicit release-unit manifests and scoped versions #8217 and required in the immediate artifact Task; no fix is falsely claimed.

Seven-criterion acceptance SHA-256 f39e8a4641c166cdddb77056f6ef54fb64e1a652da4f439656d873d81185cd97; exact-head merge gate; actual six-plan readback.

Acceptance is for the nonpublishing planner only. Executions remain attributed to34cd; byte-identical main integration does not invent a new-main run. The158 applicable-test closure entries are a plan, not executed product tests. Real selected artifacts/clean consumers, historical npm bytes, controlled publishers/recovery and operational/archival approvals remain required under open parents #8217/#8216/#8220/#8194. Closing this Task and moving it to Done / Accepted / Passed; continuing the program.

Plan independent historical product releases without publishing (elsa…

Keep product plan failure receipt I/O inside diagnostic boundary (els…

Navigation Menu

Repository navigation

[Task] Plan individual product releases and verify available prerequisites #8688

Outcome

Source and policy boundaries

Implementation scope

Acceptance criteria

Execution and non-goals

Definition of ready

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

[Task] Plan individual product releases and verify available prerequisites #8688

Description

Outcome

Source and policy boundaries

Implementation scope

Acceptance criteria

Execution and non-goals

Definition of ready

Activity

sfmskywalker commented on Oct 9, 2026

sfmskywalker commented on Oct 9, 2026

sfmskywalker commented on Oct 9, 2026

sfmskywalker commented on Oct 9, 2026

sfmskywalker commented on Oct 9, 2026

sfmskywalker commented on Oct 9, 2026

sfmskywalker commented on Oct 9, 2026

sfmskywalker commented on Oct 9, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions