From 0f8bbc7c1e2c2f6fec0b17753ac0d9c4a3ad4317 Mon Sep 17 00:00:00 2001
From: Craig Jennings <c@cjennings.net>
Date: Fri, 22 May 2026 20:12:14 -0500
Subject: fix(test): fail clearly when the VM forward port is taken

A test run launched qemu without first checking the SSH forward port, so a collision with another VM already holding it surfaced only as an opaque "Failed to start VM," with qemu unable to bind and no hint why. I added a port_in_use check in run_test before the launch: it errors with the port number and the SSH_PORT override to set, records the failure, and moves on.

The check lives in run_test, not start_vm, because start_vm runs in a command substitution (vm_pid=$(start_vm ...)) where this harness's non-exiting error() would be captured as the PID instead of failing the run. The pure half, port_listening_in, takes an `ss -tln` snapshot as a string so it's unit-testable.
---
 scripts/test-install.sh | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)

(limited to 'scripts')

diff --git a/scripts/test-install.sh b/scripts/test-install.sh
index 8cfe2ec..4220d20 100755
--- a/scripts/test-install.sh
+++ b/scripts/test-install.sh
@@ -142,6 +142,20 @@ cleanup_disks() {
     rm -f "$VM_DIR"/test-"${test_name}"-disk*.qcow2
 }
 
+# Pure predicate: does this `ss -tln` snapshot show <port> already
+# listening? Kept separate from the live ss query (port_in_use) so it's
+# unit-testable with fixture output. The ":<port> " match anchors on the
+# colon and trailing space so 2222 doesn't match 12222 or 22222.
+port_listening_in() {
+    local port="$1" ss_output="$2"
+    grep -q ":${port} " <<<"$ss_output"
+}
+
+# Live check: is <port> currently bound by a listening socket?
+port_in_use() {
+    port_listening_in "$1" "$(ss -tln 2>/dev/null)"
+}
+
 # Start VM and return PID
 start_vm() {
     local test_name="$1"
@@ -863,6 +877,19 @@ run_test() {
     step "Creating VM disks..."
     create_disks "$disk_count" "$config_name"
 
+    # Fail clearly if the forward port is taken (another VM?) before launching,
+    # rather than letting qemu fail to bind and report an opaque "Failed to
+    # start VM". Runs here, not inside start_vm — start_vm is command-
+    # substituted (vm_pid=$(...)), where error()'s output would be captured as
+    # the PID instead of failing the run.
+    if port_in_use "$SSH_PORT"; then
+        error "Port $SSH_PORT already in use (another VM?). Set SSH_PORT=<free port> and retry."
+        cleanup_disks "$config_name"
+        TESTS_FAILED=$((TESTS_FAILED + 1))
+        FAILED_TESTS+=("$config_name")
+        return 1
+    fi
+
     step "Starting VM..."
     local vm_pid
     vm_pid=$(start_vm "$config_name" "$disk_count")
-- 
cgit v1.2.3