fix(test): drop a stale benchmark handoff before setup can fail

benchmark-index.sh removed the runtime-root.txt of an earlier run only
after the repository path, the private runtime and the temp directory had
been set up. A run that failed in one of those steps exited non-zero but left
the previous handoff in the reused results directory, and section 7 of
the evaluation plan sources that file unconditionally, so the graph
session would be pointed at a root that step 8 has already removed.

The results directory is now created and the stale handoff removed right
after argument parsing, before any setup step that can fail.

Signed-off-by: Martin Vogel <martin.vogel.tech@gmail.com>

scripts/benchmark-index.sh

@@ -9,6 +9,14 @@ LANG="${2:?}"
 REPO="${3:?}"
 RESULTS_DIR="${4:?}"
 
+OUT="$RESULTS_DIR/$LANG"
+mkdir -p "$OUT"
+# A handoff names this run's index or none: the evaluation loop reuses the
+# results directory, and a stale one would point the graph session at a root
+# step 8 has already removed. It goes before any setup step that can fail, so
+# a run that never reaches the index cannot leave an earlier one behind.
+rm -f -- "$OUT/runtime-root.txt"
+
 # The index must run against a daemon rendezvous and cache this run owns: only
 # CBM_RUNTIME_DIR moves the rendezvous, and without a private cache the
 # benchmark repository was indexed into the operator's live store (#1696).
@@ -59,13 +67,6 @@ REPO_JSON=$(python3 -c 'import json,sys; print(json.dumps(sys.argv[1]))' "$REPO"
 # even though each comes from its own process.
 bench_now_ms() { python3 -c "import time; print(time.monotonic_ns() // 1000000)"; }
 
-OUT="$RESULTS_DIR/$LANG"
-mkdir -p "$OUT"
-# A handoff names this run's index or none: the evaluation loop reuses the
-# results directory, and a stale one would point the graph session at a root
-# step 8 has already removed.
-rm -f -- "$OUT/runtime-root.txt"
-
 echo "INDEX: $LANG ($REPO)"
 
 # Count source files and LOC (exclude .git, vendor, node_modules, build dirs)