fix: prevent false positive test failures in CI validation

This commit fixes the issue where CI was incorrectly reporting test failures
when all tests actually passed. The problem was in the test failure detection
logic that was creating failure signatures for package/suite completion events
rather than actual test function failures.

Key fixes:
- Improved JSON failure detection to only flag actual test function failures
- Enhanced regex pattern to require test names starting with "Test"
- Fixed exit code override logic to be more precise about what constitutes a failure
- Added better debugging output to distinguish between failure types

The validation workflow was incorrectly interpreting normal test completion
JSON entries (Action="fail" for package completion) as actual test failures.
Now it only detects genuine test function failures with proper test names.

Tests confirmed passing locally with integration suite validation.

Author: mrz1836

.github/actions/test-failure-detection/action.yml

@@ -122,20 +122,21 @@ runs:
           # Fast single-pass JSON extraction for test failures (< 1s for 10K lines)
           # Filter JSON lines and parse in one pass - eliminates 2-minute hang
           # Note: Line numbers aren't available in Go test JSON output
+          # IMPORTANT: Only detect actual test function failures, not package/suite completion events
           grep '^{' "$json_file" 2>/dev/null | \
-            jq -r 'select(.Action == "fail") |
-              "--- FAIL: \(if (.Test and .Test != null and .Test != "null") then .Test else "unknown" end) (\(.Package))"' \
+            jq -r 'select(.Action == "fail" and .Test != null and .Test != "" and .Test != "null" and (.Test | test("^Test[A-Za-z]"))) |
+              "--- FAIL: \(.Test) (\(.Package))"' \
             2>/dev/null > "$failures_file"
 
           # Create structured test failure entries with unique signatures
-          if grep '^{' "$json_file" 2>/dev/null | jq -r 'select(.Action == "fail")' 2>/dev/null | head -1 | grep -q .; then
+          if grep '^{' "$json_file" 2>/dev/null | jq -r 'select(.Action == "fail" and .Test != null and .Test != "" and .Test != "null" and (.Test | test("^Test[A-Za-z]")))' 2>/dev/null | head -1 | grep -q .; then
             echo "📝 Creating structured test failure entries with enhanced output..."
 
             # First pass: Extract test failure basic info from Action == "fail" entries
             local temp_failures
             temp_failures=$(mktemp)
             grep '^{' "$json_file" 2>/dev/null | \
-              jq -r 'select(.Action == "fail") | {
+              jq -r 'select(.Action == "fail" and .Test != null and .Test != "" and .Test != "null" and (.Test | test("^Test[A-Za-z]"))) | {
                 type: "test",
                 package: .Package,
                 test: (if (.Test and .Test != null and .Test != "null") then .Test else "unknown" end),

.github/workflows/fortress-test-matrix.yml

@@ -254,21 +254,24 @@ jobs:
             source test-failure-functions.sh 2>/dev/null || true
             detect_failures_from_json "test-output.log" "test-failures.txt" || true
 
-            # JSON failure override: Check if JSON contains failures even when exit code is 0
+            # JSON failure override: Check if JSON contains ACTUAL test failures when exit code is 0
             if [[ $TEST_EXIT_CODE -eq 0 ]] && [[ -f test-output.log ]]; then
-              echo "🔍 Checking JSON output for failures despite exit code 0..."
+              echo "🔍 Checking JSON output for actual test failures despite exit code 0..."
 
               # Check for JSON test failure entries or error patterns
               JSON_FAILURES=0
               if grep -q '^{' test-output.log 2>/dev/null; then
-                # Count JSON failure entries (only actual test failures, not log messages)
+                # Count ONLY failed test events, not package/suite failures which are expected
+                # Look for Action="fail" AND Package field AND Test field (indicating actual test failure)
                 JSON_FAILURES=$(grep '^{' test-output.log 2>/dev/null | \
-                  jq -r 'select(.Action == "fail") | .Test' 2>/dev/null | wc -l | xargs || echo "0")
+                  jq -r 'select(.Action == "fail" and .Test != null and .Test != "" and (.Package // "") != "" and (.Test | test("^Test[A-Z]"))) | .Test' 2>/dev/null | wc -l | xargs || echo "0")
 
                 if [[ $JSON_FAILURES -gt 0 ]]; then
-                  echo "⚠️ Found $JSON_FAILURES actual test failures in JSON output"
-                  echo "🔧 Overriding exit code from 0 to 1 due to detected failures"
+                  echo "⚠️ Found $JSON_FAILURES actual failing test functions in JSON output"
+                  echo "🔧 Overriding exit code from 0 to 1 due to detected test failures"
                   TEST_EXIT_CODE=1
+                else
+                  echo "✅ JSON output contains no actual test failures (exit code 0 is correct)"
                 fi
               fi
             fi