[trissim]

Current394 receiving boundary reviewed

Source-only full-family/dependency review94ee -> exact394 eb2f23e is published at 6aad9f1 in the existing receiving receipt. Complete core/runtime/agent/materialization trees, graph writer/binder/strict reader, Site001 CYX/alignment, publication roles and dependency declarations/gitlinks are unchanged. Exactly four production files differ: merged469 compact sample presentation/shared preview counter and merged474 illumination readiness warmup. Root has the exact delta for confirmation; Dalton received the same boundary. No competing patch, rebuild, reinstall, test replay or native launch. Immutable94ee target remains the parent receiving candidate, NOT current394 sampling/readiness acceptance. WholeRoot R0 stays RED (+1switch/+3arms); this draft is not a waiver or final-main merge request.

Approved private install complete — live gate remains

Published receiving head 6aad9f1; installed verifier/driver remain474759de0. Production remains EXACT Root394 94ee107, carried404 receipt dd3242d. No production, benchmark, packaging configuration or original build-helper changes. This draft targets Root's branch, not a competing394 source implementation.

Parent fully reviewedc5722d937 and independently verified the wheel, then authorized ONE offline/no-deps install. That operation succeeded into this NEW previously absent private target only:

Parent independently rechecked the installed883 payload/metadata/direct-URL identities and retained PARENT-PRIVATE-INSTALL-REVIEW.rst under the same qroot. Updated receiving receipt records that review and removal of only the empty terminal install tmp directory (4096 allocated bytes, no files). Wheel/target/evidence stay retained. No merge into Root without Root coordination; actual Site001/graph validation remains parent-owned after original H003 allowance restoration, no native/scientific launch here.

/home/ts/wt/openhcs-issue-batch-20260929/carrier434-installed-20261002/engineering394404-wheel-94ee-dd324-20261002/target-reviewed-f236ccfb-20261002

Approved wheel SHA256 f236ccfb557c2c08f5ac8c174fd0e5350ad4c9f552fb02b1d302b46d52dd8c60,4224047 bytes. Original wheel/build/source/archive remain unchanged.

• Install PASS:0.90s, aggregate79.3MiB, zero swap. Existing interpreter, --no-deps --no-index --no-compile --no-cache-dir --target; non-overwriting directory claim and wheel-hash admission. No global/canonical package, previous target, dependency, .pth or environment change.
• Installed-byte qualification PASS:0.41s, aggregate20.4MiB. All883 original non-RECORD members byte-match the wheel, including original distribution metadata and native binaries; all814 tracked/792 declared Python members remain covered. Original payload hash/size entries in installed RECORD match; pip's rewritten RECORD/generated metadata are identified, not falsely called byte-equal. Direct URL/wheel SHA matches and target contains no .pth.
• Both operations serial CPU0/CPU1/kernel512MiB/no-swap/60s. Critical resource advisory retained; immediate memory PSI some/full avg10=0.00. No OpenHCS/native/scientific module imported; no application/model/scientific launch, original author replay or broad test rerun.

Complete receipt including original harness/claims/ports. Byte-exact install/member archive, SHA25665a94c07d3fb705f463221486af24b7094d91d7c04a9e22973b83f9577da1adb. Original build/member archive SHA782ffa907f02b1759aa62b1a5135cae85116a22badf699f5e6725d62b8112ead retains first cwd harness failure and full884-member hashes.

Parent's next gate reuses original PublicJourney/McpDevClient owners: public query_plate_files(kind=result) -> exactly inventoried graph path -> stream_plate_files_to_viewer -> owned native source/channel/calibration/geometry/subject readbacks and personally opened raw/result/combined placement. Site001 CYX/order/provenance remains separate actual receiving evidence, not proof from installation. Existing harness and typed contracts are linked/named in the receipt; no new writer/client/registry.

Historical isolated pair5992/5993 DISPLAY91 was TCP-empty at this read-only checkpoint, NOT reserved or IPC-certified. Protected retinal QA6006 is occupied and untouched. Parent rechecks exact owners before launches, admits private target first from outside the source directory, and never defaults to foreign7777. Root owns394 shared source/performance, Singer404/472 receiving, Dalton scientific custody, Dewey H003 handoff. No live/biological/performance or whole394/S3/S4/S6 closure is claimed; all saved evidence/UNKNOWN dispositions remain.

[trissim]

Cross-machine production-freeze handoff (5 Oct)
This DRAFT contains receiving-wheel scripts and retained acceptance evidence, not a new production runtime implementation. Planck owns its reconciliation/closure. Subsequent ordinary receiving16 exists and the determining graph/native fixes have later merged integrations; do not treat this historical wheel branch as an unmerged runtime fix or freeze blocker. Preserve its archived receipts; close associated issues only at the actual accepted scope, not because a package was built.

[trissim]

Owner reconciliation: closing this superseded ordinary-wheel staging draft, not merging it as a runtime patch.

This branch contains receiving scripts and retained qualification evidence, with no outstanding production implementation. Its94ee/f236ccfb wheel remains exactly its historical source qualification, not a current-main or native graph acceptance claim. Root #394 is now merged; later ordinary whole receiving16 is qualified, and #748's automatic point/scalar public path has its own retained acceptance.

Keep all old wheel/source/RECORD receipts and failures. Source-bearing graph inventory-to-native persisted reopen remains explicitly unverified under Singer's #404 receiving ownership; real CZI/Site001 and desktop/long-run obligations are not satisfied by building this wheel. Those separate acceptance boundaries do not make this obsolete packaging branch a production-freeze blocker.