flake fix in 700: allow 137 status code from timeout command

Signed-off-by: Ed Santiago <[email protected]>

test/system/700-play.bats

@@ -662,13 +662,20 @@ spec:
       - top
 " > $fname
 
-    # force a timeout to happen so that the kube play command is killed
-    # and expect the timeout code 124 to happen so that we can clean up
+    # force a timeout to happen so that the kube play command is killed.
+    # We *expect* a timeout exit status (124), but as of September 2024,
+    # when run in parallel, we often see 137.
     local t0=$SECONDS
-    PODMAN_TIMEOUT=2 run_podman 124 kube play --wait $fname
+    PODMAN_TIMEOUT=2 run_podman '?' kube play --wait $fname
     local t1=$SECONDS
     local delta_t=$((t1 - t0))
 
+    if [[ $status -eq 137 ]]; then
+        echo "# FIXME-someday: timeout command exited $status" >&3
+    else
+        assert "$status" -eq 124 "Exit status from podman"
+    fi
+
     # Expectation (in seconds) of when we should time out. When running
     # parallel, allow 4 more seconds due to system load
     local expect=4

test/system/helpers.bash

@@ -561,7 +561,7 @@ function run_podman() {
     if [ "$status" -eq 124 ]; then
         if expr "$output" : ".*timeout: sending" >/dev/null; then
             # It's possible for a subtest to _want_ a timeout
-            if [[ "$expected_rc" != "124" ]]; then
+            if [[ -n "$expected_rc" ]] && [[ "$expected_rc" != "124" ]]; then
                 echo "*** TIMED OUT ***"
                 false
             fi