From e62b3bfe04d831376c8e9cbb0bca9c857fb9d98e Mon Sep 17 00:00:00 2001
From: Joseph Ferano <joseph@ferano.io>
Date: Mon, 14 Sep 2026 07:24:31 +0700
Subject: [PATCH] The survey did not move, which is what the dev gate is for

---
 HANDOFF-x86-abi-marker.md | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/HANDOFF-x86-abi-marker.md b/HANDOFF-x86-abi-marker.md
index c95e6bb..edade72 100644
--- a/HANDOFF-x86-abi-marker.md
+++ b/HANDOFF-x86-abi-marker.md
@@ -149,11 +149,20 @@ merely intended, and these two crossed tests are what will catch a half-done ver
 
 | | before | after |
 |---|---|---|
-| `spike/x86/survey.sh` | 103 MATCH / 0 DIFFER / 0 REFUSED | see below |
-| skip breakdown | 28 does-not-compile / 8 no-main / 2 runs-forever | see below |
+| `spike/x86/survey.sh` | 103 MATCH / 0 DIFFER / 0 REFUSED / 0 NOX86 | **103 / 0 / 0 / 0** |
+| skip breakdown | 28 does-not-compile / 8 no-main / 2 runs-forever | **28 / 8 / 2** |
 | `spike/x86/cells.sh` | 4/4 ok | **4/4 ok** |
 | `dune test --root .` | exit 0, 232 checks, 0 failures | **exit 0, 232 checks, 0 failures** |
 
+The survey is the measurement that could have moved and did not, which is the point of running it: both
+backends now emit a symbol into every `--dev` build, and the survey's default is a release build on both sides,
+so an unguarded marker would have shown up as 103 identical-but-different objects rather than as a wrong
+answer. It agrees byte-for-byte on what the programs print.
+
+Run it detached — `setsid timeout 2400 spike/x86/survey.sh > log 2>&1 </dev/null` — or a signal to this
+harness's process group comes back as `SURVEY_EXIT=143`, which is not a result. The log is block-buffered
+through the redirect and stays empty until the end; that is not a hang.
+
 ## What remains
 
 - Item 3 of `HANDOFF-x86-redef.md`: `flan dev` and `flan reload` choosing host and module backend together.