From 422326bf83e9e64da625ba668f4ee93e364e6930 Mon Sep 17 00:00:00 2001
From: Celes Renata <celes@frameshift.net>
Date: Mon, 20 Apr 2026 04:23:19 +0000
Subject: [PATCH] fix: inttest pipeline timeout and busybox grep compatibility

- Poll job status instead of kubectl wait (catches Failed condition
  immediately instead of waiting 600s for Complete that never comes)
- Replace grep -oP (Perl regex) with POSIX grep -o (BusyBox compat)
---
 infra/inttest/run_pipeline.sh | 35 +++++++++++++++++++++++------------
 1 file changed, 23 insertions(+), 12 deletions(-)

diff --git a/infra/inttest/run_pipeline.sh b/infra/inttest/run_pipeline.sh
index 690df8f..f08a8fd 100755
--- a/infra/inttest/run_pipeline.sh
+++ b/infra/inttest/run_pipeline.sh
@@ -428,15 +428,26 @@ envsubst < "$REPO_ROOT/infra/inttest/runner.yaml" \
   | kubectl apply -n "$NAMESPACE" -f -
 
 log "Waiting for test runner to complete (timeout: 600s) ..."
-if kubectl wait --for=condition=complete job/inttest-runner -n "$NAMESPACE" --timeout=600s; then
-  log "Test runner completed successfully"
-  stage_end "integration_tests" "ok"
-else
-  log "Test runner failed or timed out"
-  # Check if the job failed vs timed out
-  if kubectl wait --for=condition=failed job/inttest-runner -n "$NAMESPACE" --timeout=5s 2>/dev/null; then
-    log "Test runner job reported failure"
+# Wait for either complete or failed — whichever comes first
+JOB_DONE=false
+for i in $(seq 1 120); do
+  STATUS=$(kubectl get job inttest-runner -n "$NAMESPACE" -o jsonpath='{.status.conditions[0].type}' 2>/dev/null || true)
+  if [ "$STATUS" = "Complete" ]; then
+    log "Test runner completed successfully"
+    stage_end "integration_tests" "ok"
+    JOB_DONE=true
+    break
+  elif [ "$STATUS" = "Failed" ]; then
+    log "Test runner job reported failure (tests failed)"
+    stage_fail "integration_tests"
+    PIPELINE_EXIT_CODE=1
+    JOB_DONE=true
+    break
   fi
+  sleep 5
+done
+if [ "$JOB_DONE" = false ]; then
+  log "Test runner timed out after 600s"
   debug_pod_failure "inttest-runner" "app=inttest-runner"
   stage_fail "integration_tests"
   PIPELINE_EXIT_CODE=1
@@ -468,10 +479,10 @@ if [ -n "$RUNNER_POD" ]; then
   # Parse test counts from logs (pytest output format: "X passed, Y failed, Z errors")
   TEST_OUTPUT=$(kubectl logs "$RUNNER_POD" -n "$NAMESPACE" 2>/dev/null || true)
   if [ -n "$TEST_OUTPUT" ]; then
-    # Extract counts from pytest summary line like "41 passed, 2 failed, 1 error"
-    TESTS_PASSED=$(echo "$TEST_OUTPUT" | grep -oP '\d+(?= passed)' | tail -1 || echo "0")
-    TESTS_FAILED=$(echo "$TEST_OUTPUT" | grep -oP '\d+(?= failed)' | tail -1 || echo "0")
-    TESTS_ERRORS=$(echo "$TEST_OUTPUT" | grep -oP '\d+(?= error)' | tail -1 || echo "0")
+    # Extract counts from pytest summary line like "172 passed, 6 failed"
+    TESTS_PASSED=$(echo "$TEST_OUTPUT" | grep -o '[0-9]* passed' | tail -1 | grep -o '[0-9]*' || echo "0")
+    TESTS_FAILED=$(echo "$TEST_OUTPUT" | grep -o '[0-9]* failed' | tail -1 | grep -o '[0-9]*' || echo "0")
+    TESTS_ERRORS=$(echo "$TEST_OUTPUT" | grep -o '[0-9]* error' | tail -1 | grep -o '[0-9]*' || echo "0")
     TESTS_PASSED=${TESTS_PASSED:-0}
     TESTS_FAILED=${TESTS_FAILED:-0}
     TESTS_ERRORS=${TESTS_ERRORS:-0}