# FIX Antenna .NET: original preflight closure before report warmup

The separate report configuration passes eighteen native pairs and independent
cold replay on each accepted host. Each retains 2,480 complete application
operations, 552 captured preflight application frames, 962 primary spans,
576 measured primary values and 480 measured scheduler-delay values. Twelve
explicit preflight cycles per pair (216 total) and 170 warmup values are
excluded from all measured series. HP retains 254 administrative frames; SCAN
retains 252, including original Logon, bilateral TestRequest/Heartbeat and Logout.
Each host refuses fourteen actual invalid CLI invocations and 216 independently
altered copies of actual output. Finite qualification grants no timing or Stage 5
admission and exports zero result cells.

The original SDK public session and shared .NET runner remain unchanged.
Both roles complete twelve prefix cycles, close their application capture,
exchange original SDK TestRequest/Heartbeat closure messages, then begin the
original warmup, per-run prewarm and measured phases. Cold replay checks actual
closure timestamps, peer-request causality and the complete continuous SDK
sequence against the full typed-owner prefixes. No measured application frame
is copied. All administrative frames remain auditable. The earlier qualified
bounded runtime and its immutable deployments stay separately diagnostic.

The report producer alone compiles, without warnings, against the identical
17-file SDK receipt, numeric writer and shared harness. Exact source sections
prove the inherited business/public-send and callback/clock implementations
unchanged; the only task deadline adjustment is the explicit unpaced prefix.
The original reader's prepared-application clock still includes SDK envelope
work inside public SendMessage. This is not a completed-wire tick boundary and
must remain an explicit tick admission limitation. No post-write callback is t0.

The report cold-reader suite passes 37 cases; the separate immutable-generation,
placement/invocation and generic-runtime suite passes 139 with zero skips.
Its new engine identity is `fixantenna-preflight-callback-canonical`, distinct
from the old measured-prefix diagnostic generation. Packaging requires both
independent eighteen-pair reviews and all 216 refusal controls, then binds all
thirteen immutable payload files and zero-copy/closure provenance. Registered
installation, sixteen fresh relocated/fixed-DAC cases and subsequent quiet
physical/report admission remain pending at this checkpoint. Existing root
helpers and sudo registration are unchanged.

| Binding | SHA-256 / source |
|---|---|
| Qualified source | `8f492b0b6a34e7855112ad852ecbef8ea73073e4` |
| Report producer | `c0456bc27ba48c8ca5a20feece424c36206082314bbc22c8bddea468ffca511d` |
| Shared harness | `5ef1c68bfa62b6bf5fe7f92efd5ec6bff4b91b172b46c937e1c7483fd7cbad5e` |
| Original SDK Core | `11b2934a4dd07106292f6d67bf37b65151540efb5e5d2cb0c36bc228d0279f40` |
| Compiler input receipt | `9f0d93258317e3ca0ae154bda7114bf909320697b7a3add4f3929033106e895e` |
| HP native-receipt.json | `68937fac0b9d841e220e14f09e30138cb0d9a62d04bcd6515d38826338ef90cd` |
| HP cold-review.json | `5c3733c4e438b0510966d90509287e505629d524db03ac84444d54b3fea39a61` |
| HP collection-review.json | `873feeff44ab1c167a354ba88ba82f4f8aa26b5e9795098f6c2685fe12d8b47a` |
| SCAN native-receipt.json | `caa56c6887207efb6401760785aad7f5cba533868efffe46874ce7eedc18b045` |
| SCAN cold-review.json | `8d75caaccd7b431bbd9dcef2bbf868ec1b387637e1bbe62a8630e6ecc0fae58d` |
| SCAN collection-review.json | `9bd81a6e0037d4a2215b7a8b58b4648f15b23b5a11a700e175079151d9e874e2` |

Native units are `codex-fullbench-fixantenna-report-20261026-v1.service`,
both exited successfully. Original outputs remain at
`~/build/codex-fullbench/libhft/fixantenna-report-qualification-20261026-v1`
on each host. Independent copies and their complete hash inventories are at
`~/build/codex-fullbench/libhft/fixantenna-report-review-20261026-v1/{hp,scan}`.
These functional outputs are working qualification files; no retained compact
result has been manufactured from them and no old successful result is replaced.


## Installed report configuration

Both separately installed report generations complete sixteen native pairs and
independent replay: eight relocated local pairs and eight fixed-entry DAC pairs
per host, four flows with zero/five warmup. The source is
`87b2eba56337e25a141ce0d619ddd7f55cecd9f4`. All thirteen payload hashes/modes,
original loaded assemblies, UID, CPU masks, measurement affinity, parsed
invocations and namespace identities pass. No producer/SDK/harness rebuild or
root-helper change is made. The original preflight/phase/sample checks pass
through each installed entry point. Physical timing remains pending.

| Binding | SHA-256 |
|---|---|
| HP generation | `e44fd9201c56b3d20612215cc4858cbe3f949220dc1c6277c55f1ff702c1500c` |
| HP runtime-receipt.json | `2426044ae8fc59bd99b81e191974cd97f6895b33907794f4d8254bbfe4304acb` |
| HP cold-review.json | `a827a4f021f9ec7641c37a2f79ed5f6130387951e7400ad5a89199ad515c7cdf` |
| HP collection-review.json | `a763d86bc3074e5ef3c2f5d5881ae7bf845a680c59e77bd1a2f68cbb8a06823f` |
| SCAN generation | `98837482a5e62494f21e0d34aaaa619b93044676b30c622222402ee673565cd6` |
| SCAN runtime-receipt.json | `55347a07106575ebf02e4e82225187c60117a32f106935c6d3fba662f1180ae8` |
| SCAN cold-review.json | `b401db06fc815909b34a53ed38ec5ff83876b755ee62dc33188264d49f8e1beb` |
| SCAN collection-review.json | `42bfa9ed0d342b41275187f87b1c341af38b18a415d92c28f465480301537066` |