ci: improve test failure detection in parallel tests

1. ci-build.sh: Add verify_test_failure_detection() function
   - Tests that single test failures are properly detected
   - Tests that parallel test failures are properly detected
   - Runs at CI start to catch framework issues early

2. Makefile.am: Fix check-functional-parallel exit code handling
   - OLD: used 'wait' without checking individual exit codes
   - NEW: capture each PID and check exit code with 'wait $pid'
   - On failure: print the log file and set failed flag
   - Exit with error if any group failed

This ensures CI correctly fails when:
- Any unit test fails
- Any of the 4 functional test groups fails
- Any parallel build configuration fails

Makefile.am

@@ -311,17 +311,19 @@ tests_functionaltests_functionaltests_LDADD = -lcmocka -lstabber @FORKPTY_LIB@
 check-functional-parallel: tests/functionaltests/functionaltests
 	@echo "Running functional tests in parallel (4 groups)..."
 	@mkdir -p $(builddir)/test-logs
-	@( \
+	@failed=0; \
-		./tests/functionaltests/functionaltests 1 > $(builddir)/test-logs/group1.log 2>&1 & \
+	./tests/functionaltests/functionaltests 1 > $(builddir)/test-logs/group1.log 2>&1 & pid1=$$!; \
-		./tests/functionaltests/functionaltests 2 > $(builddir)/test-logs/group2.log 2>&1 & \
+	./tests/functionaltests/functionaltests 2 > $(builddir)/test-logs/group2.log 2>&1 & pid2=$$!; \
-		./tests/functionaltests/functionaltests 3 > $(builddir)/test-logs/group3.log 2>&1 & \
+	./tests/functionaltests/functionaltests 3 > $(builddir)/test-logs/group3.log 2>&1 & pid3=$$!; \
-		./tests/functionaltests/functionaltests 4 > $(builddir)/test-logs/group4.log 2>&1 & \
+	./tests/functionaltests/functionaltests 4 > $(builddir)/test-logs/group4.log 2>&1 & pid4=$$!; \
-		wait \
+	wait $$pid1 || { echo "Group 1 FAILED (exit $$?)"; cat $(builddir)/test-logs/group1.log; failed=1; }; \
-	) && ( \
+	wait $$pid2 || { echo "Group 2 FAILED (exit $$?)"; cat $(builddir)/test-logs/group2.log; failed=1; }; \
-		echo "=== Test Results ==="; \
+	wait $$pid3 || { echo "Group 3 FAILED (exit $$?)"; cat $(builddir)/test-logs/group3.log; failed=1; }; \
-		grep -E 'PASSED|FAILED|Running' $(builddir)/test-logs/group*.log; \
+	wait $$pid4 || { echo "Group 4 FAILED (exit $$?)"; cat $(builddir)/test-logs/group4.log; failed=1; }; \
-		! grep -q FAILED $(builddir)/test-logs/group*.log \
+	echo "=== Test Results Summary ==="; \
-	)
+	grep -E 'PASSED|FAILED|Running' $(builddir)/test-logs/group*.log || true; \
+	if [ $$failed -ne 0 ]; then echo "FUNCTIONAL TESTS FAILED"; exit 1; fi; \
+	echo "All functional test groups passed!"
 endif
 endif
 

ci-build.sh

@@ -23,6 +23,79 @@ error_handler()
 
 trap error_handler ERR
 
+# Verify that test failures are properly detected
+# This is a meta-test: it runs a deliberately failing test
+# and checks that the test framework reports the failure correctly
+verify_test_failure_detection()
+{
+    echo
+    echo "==> Verifying test failure detection..."
+    
+    # Create a simple failing test
+    cat > /tmp/test_must_fail.c << 'EOF'
+#include <stdarg.h>
+#include <stddef.h>
+#include <setjmp.h>
+#include <cmocka.h>
+
+static void test_that_must_fail(void **state) {
+    (void)state;
+    assert_true(0); // This MUST fail
+}
+
+int main(void) {
+    const struct CMUnitTest tests[] = {
+        cmocka_unit_test(test_that_must_fail),
+    };
+    return cmocka_run_group_tests(tests, NULL, NULL);
+}
+EOF
+    
+    # Compile the failing test
+    if ! gcc -o /tmp/test_must_fail /tmp/test_must_fail.c -lcmocka 2>/dev/null; then
+        echo "Warning: Could not compile test failure verification (cmocka not available?)"
+        echo "Skipping test failure detection verification"
+        return 0
+    fi
+    
+    # Test 1: Single failing test detection
+    echo "  Testing single test failure detection..."
+    if /tmp/test_must_fail > /tmp/test_must_fail.log 2>&1; then
+        echo "ERROR: Test that should fail returned success (exit code 0)"
+        echo "This means the test framework is NOT detecting failures correctly!"
+        echo "--- Test output ---"
+        cat /tmp/test_must_fail.log
+        echo "--- End output ---"
+        rm -f /tmp/test_must_fail /tmp/test_must_fail.c /tmp/test_must_fail.log
+        exit 1
+    fi
+    echo "  ✓ Single test failure correctly detected"
+    
+    # Test 2: Parallel failure detection (simulates check-functional-parallel)
+    echo "  Testing parallel test failure detection..."
+    failed=0
+    /tmp/test_must_fail > /tmp/p1.log 2>&1 & pid1=$!
+    true > /tmp/p2.log 2>&1 & pid2=$!  # This passes
+    /tmp/test_must_fail > /tmp/p3.log 2>&1 & pid3=$!
+    true > /tmp/p4.log 2>&1 & pid4=$!  # This passes
+    
+    wait $pid1 || failed=$((failed + 1))
+    wait $pid2 || failed=$((failed + 1))
+    wait $pid3 || failed=$((failed + 1))
+    wait $pid4 || failed=$((failed + 1))
+    
+    if [ $failed -ne 2 ]; then
+        echo "ERROR: Expected 2 failures in parallel tests, got $failed"
+        echo "Parallel failure detection is broken!"
+        rm -f /tmp/test_must_fail /tmp/test_must_fail.c /tmp/test_must_fail.log /tmp/p?.log
+        exit 1
+    fi
+    echo "  ✓ Parallel test failures correctly detected (2 of 4 failed as expected)"
+    
+    rm -f /tmp/test_must_fail /tmp/test_must_fail.c /tmp/test_must_fail.log /tmp/p?.log
+    echo "✓ Test failure detection verified"
+}
+
 num_cores()
 {
     # Check for cores, for systems with:
@@ -34,6 +107,9 @@ num_cores()
         || getconf _NPROCESSORS_ONLN 2>/dev/null
 }
 
+# Run test failure detection verification first
+verify_test_failure_detection
+
 ./bootstrap.sh
 
 tests=()