Repository navigation
Fix release verification and registration lifecycle defects - #619
Conversation
|
The existing review head The separate formatter workflow fails before formatting at the existing high-severity dependency audit. Local reproduction and the bounded fix are tracked in #620. The draft remains open while those results and licensed gates are pending. This does not change #613's remaining review/integration and OOM-investigation limits. |
Performance NumbersMeasured commit: Benchmark or harness paths changed, so historical deltas are non-comparable:
Historical Standalone deltaDelta direction: + is better; - is worse. Historical delta omitted because benchmark or harness code changed in this pull request. Current TargetMap evidenceThese raw rows are current-run diagnostics. Use fresh bracketed controls before accepting a micro-optimization. |
|
Updated the existing draft to exact head
Original host samples (43 files plus metadata), package sources, and clean active scene are restored and verified; the restored baseline assembly is fresh. All six existing user edits are preserved. Supported-version and stripped IL2CPP validation, OOM root-cause limits, and maintainer review remain separate gates. This PR stays draft; no issue closes and no release-readiness claim follows. |
|
Exact head Unity Tests 37703838396 has admitted its 2021.3.45f1 job through the central organization lock and reached Standalone work. Its mode XML and final cleanup gate are still pending; individual continue-on-error step conclusions do not establish a passing Unity result. Remaining supported-version, performance, stripped-player, integration, and human review gates keep this PR draft. |
|
The supported-floor result is now available: Unity 2021 job 113073678698 fails the cache integration test because compilation remains active throughout its 64-update precondition ( Official UPM Test Framework 1.4.5 source confirms EditModeRunner locks assembly reloads while tests execute. The same job log starts another compilation for changed |
|
Pushed exact head Native evidence at Unity 6000.4.6f1: 10 genuine pending-compilation probes, 50 separate executions of the final fixture, 117/117 affected tests, and 975/975 Editor tests, including 13 sample-quality cases. All accepted results have zero failed/skipped/inconclusive tests and verified framework cleanup. Original host samples and package files are restored by hashes, and the baseline assembly is fresh. Source and raw evidence are detailed in #613. Commit and push hooks pass; the committed fixture SHA-256 matches the tested bytes. The new supported-version, performance, and stripped-player gates remain required. This PR stays draft for those gates, review, and integration. The separate OOM root cause remains unresolved. |
|
DISCLOSURE: LLM-GENERATED TEXT Pushed 5c74fd4 after the supported Unity 2021 job exposed a documentation capture dependency on the uninitialized IMGUI style cache. The compiler specimen now resolves the same font through the public inspector skin API; its diagnostic/source assertions and the verified compilation handoff remain unchanged. Verification: original cold-cache failure reproduced 50/50; candidate passes 50/50 cold-cache probes, 50 distinct native runs with confirmed cleanup, related documentation 9/9, and full ordinary Editor 975/975 including all 13 sample quality cases. Format/spelling/diff and commit/push hooks pass. Original editor files, 43 samples/root metadata and six user edits are restored byte for byte; the baseline assembly is fresh and the temporary fixture is absent. New exact-head Unity run 37711381000 and performance run 37711380137 are queued. New CSharpier, Prettier, documentation deployment and Unity Docs Gate already succeed; CI and devcontainer checks are still running. All licensed jobs use central admission and acquired-scoped return/classify/release/confirmed cleanup. This remains a draft: supported-head CI, shipping verification and human review remain pending. No issue closed. |
|
Commit The independent #690 Windows comparison now proves an observer array-stride defect: the old observer decodes one holder but misdecodes multiple holder records. SDK-matched buffers and structure size fix the owned two/three-holder cases. Production-only PR #619 remains draft. Shipping #685, compiler #687, reclamation, configuration/owner costs, calibration, final imports, copy review and release readiness remain open. No calibrated throughput, parity or native allocation claim. |
|
Current commit Every Editor scope has 1,031 cases: 1,016 pass and the same 15 skip. All three new #690 SDK-layout/two-holder/three-holder regressions pass on all four versions. Every prior outcome and all 372 configuration cases remain exact. Mono/IL2CPP totals are 27,793/27,789 for 2021, 2022 and 6.3, and 27,792/27,788 for 6.5. Default processing takes 38/66/66/66 seconds under the unchanged 120-second cap. Internal performance retains 207 passes. Comparisons retain 470 passes/66 unchanged skips plus five separate profiler controls, exact source/profile/process records, replay and matched DLL/PDB identity. All 48 timing CSV allocation fields remain -1/unmeasured. No calibrated throughput/parity or whole-owner-cost claim follows. The explicit #689/#690 acceptance audit passes on this draft source ( |
|
The existing #684 scale-test and #688 failed-output launch-gate protocols now pass their scoped acceptance audits. Both are resolved in this draft source; their issues are closed. The audit rechecks exact integrated source, all eight scale passes/12 raw archives, actual-function and public Mono2x RED/GREEN/mutation controls, full script results and qualified-parent cleanup. SHA256 Scale verdict and retained unrelated failures. Launch-gate verdict and limits. Fresh73b checks remain pending. This PR remains draft for the open reclamation, attribution, shipping, cost, calibration and release gates; no merged/published claim follows. |
|
Current73b6341d qualification is complete for the scopes that ran: seven standard workflows pass; the Unity workflow is CANCELLED. Unity2021 and2022 preserve all prior outcomes and372configuration cases, including1,021Editor passes/15same skips and all five new file controls. Internal performance207PASS qualifies. Comparison536=470PASS/66unchangedSkip plus five separate PlayMode profiler controls qualifies, with exact profiles/process/source/replay/native DLL-PDB identity and cleanup. Eight qualified archives retain94raw members. The cancelled6.3packet adds one13-byte version-probe log; it has no NUnit results. All five licensed jobs that entered confirm cleanup;6.5never starts. Default processing takes66/35seconds under the unchanged120-second cap. Cancellation detail and explicit correction. Consolidated disposition PR619 remains draft. Two supported Unity versions are incomplete, cancellation intent is pending clarification, and existing manual reclamation, attribution, shipping/compiler, cost, calibration and release gates remain open. No resubmission or new native study/consumer admission occurs. User approval for this PR and future commits remains recorded. |
|
Prepared two local checked-build corrections without changing PR head73b: signature hashing05a8 and registration-handle hashing303ede68 (#659/#691). The combined clean local child is These commits are unpushed/unintegrated and do not qualify current-head Unity or release acceptance. Existing PR/future-commit approval persists. PR remains draft; unchanged current-head Unity run40593459251 is cancelled and the cancellation-intent clarification is still pending. No new Native admission, resubmission or source push follows from these local results. Full local receipt d3d47615ea2732af5b4acf61981d6755316950d4af0a86efd95ec6a17cbb1806. |
|
Prospective local upstream reconciliation: GitHub reports this PR behind master. Current master4e1c53a148bec0601f4cd9384ae62f06304c936f adds only two commits since the common base: reviewed PR686 lock repins/calling-step budgets and its existing automated performance-document refresh. I will prepare an ordinary merge into a separate local child of303ede68, leaving Main73b and all frozen source/evidence worktrees unchanged. Inspect any conflicts and require the final Runtime, Editor, Tests, generators, scripts, package inputs, shipping bytes and workflow tree to remain exact303ede68 unless an actual upstream delta justifies separate review. Retain the upstream generated data with its original source attribution; it is not a measurement of the new candidate or calibrated release qualification. Run affected documentation/static checks and normal hooks before a local commit. Unexpected production or workflow deltas stop this preparation for review. No source push, Native submission, new PR, campaign-data adaptation or acceptance closure. |
|
The prepared local merge is clean, unpushed Before any future target data, I am preparing a separate fff83bc qualification packet using the unchanged73b default classifiers/baseline12 XMLs and the same internal/comparison/profile/process/cleanup gates. Only head and local evidence/worktree paths may change in the perf wrappers/replay; no thresholds, case inventories, outcomes, source controls, profile or cost verdicts change. This packet will contain no run bindings or admission. Main stays73b and the cancellation question still blocks new Native submissions/source pushes. The upstream timing rows retain their measured dd93 source and are not new candidate/calibrated measurements. |
|
Future-head preparation is frozen locally forfff83bcfbabf1f3675ae67d5c75d03aed7fc6992:1889 source-file hashes, the unchanged four default analysis files and12 baseline XML hashes, and perf/replay wrappers changed only by the declared head/evidence/worktree string substitutions. Inverse substitutions recover every original byte. Python AST and PowerShell parsing pass. All existing case/outcome/profile/process/native-input/binary-PDB/cleanup and unmeasured-cost rules remain intact; expected full qualification remains14archives/120members with1036Editor and372configuration cases. Preparation SHA-256 UNADMITTED: no run-bindings file, target data, source push or Native submission. Fresh new-head run/job metadata must precede target qualification; never reuse73b artifacts or IDs. Current-consumer and stopped-research preparations remain unchanged. Main/PR still73b; cancellation-intent clarification remains pending. Existing PR/future-commit approval still applies, and no new review approval is requested. |
|
Final pre-admission control for preparedfff83bcf: all three default/internal/comparison qualifiers stop at the exact missing NEW run-bindings.json before reading target metadata. No qualification file or Native target directory was created; analysis/rules stayed unchanged. Receipt SHA-25675c987ca29eb752bffe9b4238920df323c2e478eadc01c28904a9d822725f300. This verifies the missing-binding guard only, not successful target classification or Native readiness. No source push, admission or resubmission. Cancellation intent remains unanswered; original session cutoff00:50:37UTC is preserved and handoff is current. |
|
Prepared next revision cbd4022 combines the locally verified checked-hash corrections (#691/#659), current master, the local Unity-project metadata correction and qs6.17.0. Standing authorization for this PR and future commits applies to this distinct new source revision. The cancelled73b run remains retained; its cause is unknown and its missing scopes do not qualify the new head. No same-source retry is planned. Local evidence: 1,464 Node and2Python tests pass, as do format/Markdown/spelling/validate:all and normal commit hooks. The hash corrections retain the previous751docs/26,338portable qualification and paired checked/unchecked evidence; the new session's five changed files preserve all Runtime/Editor/Tests/generator/workflow/script/package-manifest Git objects from preparedfff. The local project graph's27entries match retained publisher metadata and the real CI-generated manifest, excluding its two local IDE additions and local package URI. Actual clean local-project resolve/import stays unverified. All362shipping bytes match the existing059b tarball. Serializer acceptance audit. Before target CI data:1889physical source hashes and the existing12baseline XML hashes are frozen. Default analyses are byte-exact to the prior prepared packet; performance/replay analyses change only declared head, paths and labels, with inverse byte-equivalence checks. All21repository comparison source hashes match their retained33-record source roster. New-head preparationSHA256 e70bcced897a3361812e05d63d9c9c5225ae137a3b27f0c2d8e2665d84939651; rules88a354caaf7a16458497071fca221a91564813a897b87c1fd504a4d30f7d8396. No target run bindings or native results exist yet. Require all fresh standard workflows plus all four supported default scopes and both performance scopes, exact old outcomes/new controls, raw archive/member/source identities, profiles/process settings and cleanup. Expected full ordinary evidence remains14archives/120rawmembers. Allocation/owner-cost and calibrated-performance claims stay unmeasured; local-project imports, shipping/compiler/reclamation investigations and wallstop review remain separate. This PR stays draft and is not merge-ready. |
Bugbot needs on-demand usage enabledBugbot uses usage-based billing for this team and requires on-demand usage to be enabled. A team admin can enable on-demand usage in the Cursor dashboard. |
|
Important Review skippedToo many files! This PR contains 350 files, which is 50 over the limit of 300. To get a review, reduce the PR to 300 files or fewer by splitting it into smaller PRs or changing its base branch. Usage-priced reviews support at most 300 files.
 |
Why
Nested callbacks, resets and failed replay can lose subscriptions or invoke retired callbacks. Generated identifiers and checked hash arithmetic also fail valid inputs. Incomplete player builds can pass old output checks. The local test project contains package metadata from a different Editor.
What changed
How we know
All eight workflows for
cbd40222are green. The four ordinary Unity-version scopes and both performance correctness scopes have qualified source, artifact, process and cleanup evidence: 14 archives and 120 raw members. Unity run, performance run. Local checks pass: 1,464 Node tests, two Python tests, formatting, Markdown, spelling and package validation. The hash fixes have 28 actual-source guards and paired 26,338-case checked and unchecked Runtime checks. Prior full documentation checks retain 751 passes.The corrected local graph matches exact Unity 2022 distribution and registry metadata. All 362 shipping files match the prepared tarball. The separate clean local-project experiment found UniRx compilation failures from missing Physics/Physics2D modules; its raw upload was blocked and final cleanup was unconfirmed. The correction is being prepared locally under #692. Actual clean import remains unqualified; the central runner quarantine must be resolved before further licensed admission. New-head evidence protocol.
Remaining verification
This remains a draft. Historical cancellation and failures remain retained. Shipping/compiler, manual reclamation, native costs, calibration, final imports/copy and wallstop review remain open. Standard green checks do not replace those gates.
References #575, #641, #642, #644, #645, #646, #647, #649, #650, #651, #652, #653, #654, #655, #659, #660, #661, #662, #663, #668, #669, #670, #671, #673, #674, #675, #676, #677, #679, #680, #681, #682, #683, #685, #687, #691 and #692.