Add support for stateful test suites and add chain tests

Makefile

@@ -22,7 +22,7 @@ test:
 	@echo "Running runner unit tests..."
 	go test -v ./runner/...
 	@echo "Running conformance tests with mock handler..."
-	$(RUNNER_BIN) -handler $(MOCK_HANDLER_BIN)
+	$(RUNNER_BIN) --handler $(MOCK_HANDLER_BIN) -vv
 
 clean:
 	@echo "Cleaning build artifacts..."

README.md

@@ -66,6 +66,33 @@ make runner
 
 The runner automatically detects and recovers from crashed/unresponsive handlers, allowing remaining tests to continue.
 
+#### Verbose Flags
+
+- **`-v, --verbose`**: Shows request chains and responses for **failed tests only**
+- **`-vv`**: Shows request chains and responses for **all tests** (passed and failed)
+
+The request chains printed by verbose mode can be directly piped to the handler binary for manual debugging:
+
+```bash
+# Example output from -vv mode:
+# ✓ chain#4 (Get active chain reference from chainstate manager)
+#
+#       Request chain
+#       ────────────────────────────────────────
+# {"id":"chain#1","method":"btck_context_create","params":{"chain_parameters":{"chain_type":"btck_ChainType_REGTEST"}},"ref":"$context_ref"}
+# {"id":"chain#2","method":"btck_chainstate_manager_create","params":{"context":"$context_ref"},"ref":"$chainstate_manager_ref"}
+# {"id":"chain#4","method":"btck_chainstate_manager_get_active_chain","params":{"chainstate_manager":"$chainstate_manager_ref"},"ref":"$chain_ref"}
+#
+#       Response:
+#       ────────────────────────────────────────
+#       {"result":"$chain_ref"}
+
+# Copy the request chain and pipe it to your handler for debugging:
+echo '{"id":"chain#1","method":"btck_context_create","params":{"chain_parameters":{"chain_type":"btck_ChainType_REGTEST"}},"ref":"$context_ref"}
+{"id":"chain#2","method":"btck_chainstate_manager_create","params":{"context":"$context_ref"},"ref":"$chainstate_manager_ref"}
+{"id":"chain#4","method":"btck_chainstate_manager_get_active_chain","params":{"chainstate_manager":"$chainstate_manager_ref"},"ref":"$chain_ref"}' | ./path/to/your/handler
+```
+
 ### Testing the Runner
 
 Build and test the runner:

cmd/runner/main.go

@@ -2,26 +2,36 @@ package main
 
 import (
 	"context"
-	"flag"
 	"fmt"
 	"io/fs"
 	"os"
+	"sort"
 	"strings"
 	"time"
 
+	"github.com/spf13/pflag"
 	"github.com/stringintech/kernel-bindings-tests/runner"
 	"github.com/stringintech/kernel-bindings-tests/testdata"
 )
 
 func main() {
-	handlerPath := flag.String("handler", "", "Path to handler binary")
-	handlerTimeout := flag.Duration("handler-timeout", 10*time.Second, "Max time to wait for handler to respond to each test case (e.g., 10s, 500ms)")
-	timeout := flag.Duration("timeout", 30*time.Second, "Total timeout for executing all test suites (e.g., 30s, 1m)")
-	flag.Parse()
+	handlerPath := pflag.String("handler", "", "Path to handler binary")
+	handlerTimeout := pflag.Duration("handler-timeout", 10*time.Second, "Max time to wait for handler to respond to each test case (e.g., 10s, 500ms)")
+	timeout := pflag.Duration("timeout", 30*time.Second, "Total timeout for executing all test suites (e.g., 30s, 1m)")
+	verboseCount := pflag.CountP("verbose", "v", "Verbose mode: -v shows all requests needed to reproduce failed tests, plus received/expected responses; -vv shows this for all tests (passed and failed)")
+	pflag.Parse()
+
+	// Convert verbose count to verbosity level
+	verbosity := runner.VerbosityQuiet
+	if *verboseCount >= 2 {
+		verbosity = runner.VerbosityAlways
+	} else if *verboseCount == 1 {
+		verbosity = runner.VerbosityOnFailure
+	}
 
 	if *handlerPath == "" {
-		fmt.Fprintf(os.Stderr, "Error: -handler flag is required\n")
-		flag.Usage()
+		fmt.Fprintf(os.Stderr, "Error: --handler flag is required\n")
+		pflag.Usage()
 		os.Exit(1)
 	}
 
@@ -37,6 +47,9 @@ func main() {
 		os.Exit(1)
 	}
 
+	// Sort test files alphabetically for deterministic execution order
+	sort.Strings(testFiles)
+
 	// Create test runner
 	testRunner, err := runner.NewTestRunner(*handlerPath, *handlerTimeout, *timeout)
 	if err != nil {
@@ -65,12 +78,18 @@ func main() {
 		}
 
 		// Run suite
-		result := testRunner.RunTestSuite(ctx, *suite)
+		result := testRunner.RunTestSuite(ctx, *suite, verbosity)
 		printResults(suite, result)
 
 		totalPassed += result.PassedTests
 		totalFailed += result.FailedTests
 		totalTests += result.TotalTests
+
+		// Close handler after stateful suites to prevent state leaks.
+		// A new handler process will be spawned on-demand when the next request is sent.
+		if suite.Stateful {
+			testRunner.CloseHandler()
+		}
 	}
 
 	fmt.Printf("\n" + strings.Repeat("=", 60) + "\n")