From a084dce780ec0bdbe4e1107576b3df8931a18568 Mon Sep 17 00:00:00 2001
From: Marius Ciepluch <11855163+norandom@users.noreply.github.com>
Date: Tue, 22 Sep 2026 21:56:20 +0200
Subject: [PATCH] test(push): a verified source freeze for long measurement
 runs

Every long run mounts src into a container, and mounting the working tree is a live mount: edits
made during a multi-hour run are picked up mid-run, and a measurement whose code changed underneath
it is not a measurement of anything. A five-hour PHP transfer replay was lost that way today.

The export has to be checked too. git archive piped into tar once produced a tree missing the whole
ontology -- every tracked file after groovy alphabetically -- and reported success, because tar
status in a pipeline was never read. Every scan on that tree died at its first fact load and looked
like a slow run for fourteen minutes.

So: archive to a file, extract, count against git, and refuse an export without the ontology.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EmM1j46qgKMh3zqD1QMwEo
---
 benchmarks/push/freeze_source.sh | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)
 create mode 100755 benchmarks/push/freeze_source.sh

diff --git a/benchmarks/push/freeze_source.sh b/benchmarks/push/freeze_source.sh
new file mode 100755
index 0000000..a1802ee
--- /dev/null
+++ b/benchmarks/push/freeze_source.sh
@@ -0,0 +1,28 @@
+#!/usr/bin/env bash
+# Export the COMMITTED source for a long measurement run, and prove the export is whole.
+#
+# Every long run here mounts src into a container. Mounting the working tree is a live mount: edits
+# made during a multi-hour run are picked up mid-run, and a measurement whose code changed underneath
+# it is not a measurement of anything -- the failure looks exactly like the thing being measured
+# failing. A five-hour PHP transfer replay was lost that way on 2026-09-22.
+#
+# And the export has to be checked. `git archive | tar -x` once produced a tree missing the whole
+# ontology (every tracked file after `groovy` alphabetically) and reported success, because tar's
+# status in a pipeline was never read. Every scan on that tree died at its first fact load and looked
+# like a slow run for fourteen minutes. So: archive to a file, extract, count against git.
+#
+# Usage: benchmarks/push/freeze_source.sh <dest-dir>   -> prints the frozen commit on success
+set -euo pipefail
+dest="${1:?destination directory}"
+rm -rf "$dest" && mkdir -p "$dest"
+git archive HEAD src benchmarks > "$dest.tar"
+tar -xf "$dest.tar" -C "$dest"
+rm -f "$dest.tar"
+want=$(git ls-files src benchmarks | wc -l)
+have=$(find "$dest" -type f | wc -l)
+if [ "$want" -ne "$have" ]; then
+  echo "TRUNCATED export: git tracks $want files, export holds $have" >&2
+  exit 1
+fi
+test -d "$dest/src/openultrasast/ruleset/semantic" || { echo "export lacks the ontology" >&2; exit 1; }
+echo "$(git rev-parse --short HEAD) $have files"