Merge pull request 'test: auto-generate test files, verify content, CLI args, test_data dir' (#7) from improve-test-py into main

Reviewed-on: #7

.gitignore

@@ -1,2 +1,4 @@
 build
 data_copied
+test_data/
+__pycache__/

src/client/client.c

@@ -207,6 +207,14 @@ int main(int argc, char *argv[]) {
                       config->compression_level);
         }
       }
+    } else if (strcmp(argv[i], "--source-dir") == 0 && i + 1 < argc) {
+      free(config->send_directory);
+      config->send_directory = str_dup(argv[++i]);
+    } else if (strcmp(argv[i], "--dest-dir") == 0 && i + 1 < argc) {
+      free(config->receive_root_directory);
+      config->receive_root_directory = str_dup(argv[++i]);
+    } else if (strcmp(argv[i], "--save-to-disk") == 0) {
+      config->save_to_disk = true;
     } else {
       handle_arg(argv[i], "-m", &config->use_multithreading,
                  "Enabled Multithreading");

src/client/scanner.c

@@ -7,10 +7,14 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
+#include <sys/stat.h>
+#include <unistd.h>
 
 DirectoryScanner *directory_scanner_create(char *root_directory) {
   DirectoryScanner *scanner = malloc(sizeof(DirectoryScanner));
   scanner->directories = queue_create(100, free);
+  scanner->current_dir = NULL;
+  scanner->current_path = NULL;
   queue_enqueue(scanner->directories, str_dup(root_directory));
   return scanner;
 }
@@ -18,11 +22,16 @@ DirectoryScanner *directory_scanner_create(char *root_directory) {
 void directory_scanner_destroy(DirectoryScanner *scanner) {
   if (scanner == NULL)
     return;
+  if (scanner->current_dir) {
+    closedir(scanner->current_dir);
+    scanner->current_dir = NULL;
+  }
+  free(scanner->current_path);
   queue_destroy(scanner->directories);
   free(scanner);
 }
 
-Chunk *chunk_data_to_chunk(ArrayList *chunk_data) {
+static Chunk *chunk_data_to_chunk(ArrayList *chunk_data) {
   void **chunk_items = array_list_to_array(chunk_data);
   Chunk *chunk = chunk_create((File **)chunk_items, chunk_data->size);
   free(chunk_items);
@@ -31,29 +40,57 @@ Chunk *chunk_data_to_chunk(ArrayList *chunk_data) {
   return chunk;
 }
 
+static int open_next_directory(DirectoryScanner *scanner) {
+  if (scanner->current_dir) {
+    closedir(scanner->current_dir);
+    scanner->current_dir = NULL;
+  }
+  free(scanner->current_path);
+
+  if (queue_is_empty(scanner->directories))
+    return 0;
+
+  scanner->current_path = (char *)queue_dequeue(scanner->directories);
+  scanner->current_dir = opendir(scanner->current_path);
+  if (scanner->current_dir == NULL) {
+    perror("Could not open directory!");
+    exit(EXIT_FAILURE);
+  }
+  return 1;
+}
+
 Chunk *directory_scanner_next(DirectoryScanner *scanner) {
   ArrayList *chunk_data = array_list_create(file_destroy);
   unsigned long long chunk_data_size = 0;
 
-  while (!queue_is_empty(scanner->directories)) {
+  while (1) {
-    char *path = (char *)queue_dequeue(scanner->directories);
+    if (scanner->current_dir == NULL) {
-    DIR *dir;
+      if (!open_next_directory(scanner))
-    struct dirent *entry;
+        break;
-    dir = opendir(path);
-    if (dir == NULL) {
-      perror("Could not open directory!");
-      exit(EXIT_FAILURE);
     }
-    while ((entry = readdir(dir)) != NULL) {
+
-      if (strcmp(entry->d_name, ".") == 0 || strcmp(entry->d_name, "..") == 0) {
+    struct dirent *entry = readdir(scanner->current_dir);
+    if (entry == NULL) {
+      closedir(scanner->current_dir);
+      scanner->current_dir = NULL;
+      free(scanner->current_path);
+      scanner->current_path = NULL;
       continue;
     }
-      char *cur_path = path_cat(path, entry->d_name);
+
+    if (strcmp(entry->d_name, ".") == 0 || strcmp(entry->d_name, "..") == 0)
+      continue;
+
+    char *cur_path = path_cat(scanner->current_path, entry->d_name);
     struct stat stats;
-      stat(cur_path, &stats);
+    if (stat(cur_path, &stats) != 0) {
-      if (!S_ISREG(stats.st_mode))
+      free(cur_path);
+      continue;
+    }
+
+    if (!S_ISREG(stats.st_mode)) {
       queue_enqueue(scanner->directories, (void *)cur_path);
-      else {
+    } else {
       File *file = file_create(cur_path, &stats);
       array_list_add(chunk_data, file);
       chunk_data_size += file->stats.st_size;
@@ -62,9 +99,7 @@ Chunk *directory_scanner_next(DirectoryScanner *scanner) {
       free(cur_path);
     }
   }
-    closedir(dir);
+
-    free(path);
-  }
   if (chunk_data->size > 0)
     return chunk_data_to_chunk(chunk_data);
   return NULL;

src/client/scanner.h

@@ -3,8 +3,12 @@
 
 #include "chunk.h"
 #include "queue.h"
+#include <dirent.h>
+
 typedef struct {
   Queue *directories;
+  DIR *current_dir;
+  char *current_path;
 } DirectoryScanner;
 
 DirectoryScanner *directory_scanner_create(char *root_directory);

test.py

@@ -1,27 +1,31 @@
+import argparse
+import filecmp
 import os
+import shutil
 import subprocess
+import sys
+import tempfile
 import time
 
-# --- Configuration ---
+TEST_DIR = os.path.join(os.path.dirname(os.path.abspath(__file__)), "test_data")
+DEFAULT_SOURCE_DIR = os.path.join(TEST_DIR, "source")
+DEFAULT_DEST_DIR = os.path.join(TEST_DIR, "dest")
+
 SERVER_CMD = ["./build/server"]
 base_client_cmd = ["./build/client"]
 
-# --- Resource Limit Configuration ---
+DISK_DEVICE = "/dev/nvme0n1p5"
-# 💾 Disk throttling settings
+READ_BPS_MAX = "15M"
-DISK_DEVICE = (
+WRITE_BPS_MAX = "10M"
-    "/dev/nvme0n1p5"  # IMPORTANT: Change this to your disk (e.g., /dev/nvme0n1)
-)
-READ_BPS_MAX = "15M"  # Max read speed (M for megabytes)
-WRITE_BPS_MAX = "10M"  # Max write speed
 
-# 🐢 Network throttling settings (Linux tc)
 NET_LIMIT = "100mbit"
 NET_DELAY = "100ms"
 NETWORK_INTERFACE = "lo"
-NET_LIMIT_CMD = f"sudo tc qdisc add dev {NETWORK_INTERFACE} root netem rate {NET_LIMIT} delay {NET_DELAY}".split()
+NET_LIMIT_CMD = (
+    f"sudo tc qdisc add dev {NETWORK_INTERFACE} root netem rate {NET_LIMIT} delay {NET_DELAY}".split()
+)
 NET_RESET_CMD = f"sudo tc qdisc del dev {NETWORK_INTERFACE} root".split()
 
-# --- Build the client command prefix with throttling ---
 CLIENT_CMD_PREFIX = [
     "sudo",
     "systemd-run",
@@ -32,11 +36,10 @@ CLIENT_CMD_PREFIX = [
     f"IOWriteBandwidthMax={DISK_DEVICE} {WRITE_BPS_MAX}",
 ]
 
-# --- Test Cases ---
 TEST_CASES = [
     {"name": "Standard (Single-threaded)", "flags": []},
     {"name": "Multithreading (-m)", "flags": ["-m"]},
-    {"name": "Compression (-c 0)", "flags": ["-c 0"]},
+    {"name": "Compression (-c)", "flags": ["-c"]},
     {"name": "Chunk Serialization (-s)", "flags": ["-s"]},
     {"name": "Compression + Chunk Serialization (-c -s)", "flags": ["-c", "-s"]},
     {"name": "Multithreading + Compression (-m -c)", "flags": ["-m", "-c"]},
@@ -48,70 +51,146 @@ TEST_CASES = [
 ]
 
 
-def run_suite(env_name, apply_limits):
+def generate_test_files(source_dir):
+    if os.path.exists(source_dir):
+        shutil.rmtree(source_dir)
+    os.makedirs(source_dir)
+
+    target_total = 50 * 1024 * 1024
+    written = 0
+
+    files = {
+        "small.txt": b"hello world\n",
+        "medium.txt": b"the quick brown fox jumps over the lazy dog\n" * 5000,
+        "binary.bin": bytes(range(256)) * 1000,
+        "nested/subdir/deep.txt": b"deeply nested file\n",
+        "nested/another.txt": b"another nested file\n" * 50,
+    }
+    for rel_path, content in files.items():
+        full_path = os.path.join(source_dir, rel_path)
+        os.makedirs(os.path.dirname(full_path), exist_ok=True)
+        with open(full_path, "wb") as f:
+            f.write(content)
+        written += len(content)
+
+    i = 0
+    while written < target_total:
+        chunk_size = min(5 * 1024 * 1024, target_total - written)
+        rel_path = f"bulk/file_{i}.dat"
+        full_path = os.path.join(source_dir, rel_path)
+        os.makedirs(os.path.dirname(full_path), exist_ok=True)
+        with open(full_path, "wb") as f:
+            f.write(b"0" * chunk_size)
+        written += chunk_size
+        i += 1
+
+    total_mb = written / (1024 * 1024)
+    print(f"  Generated {total_mb:.1f} MB of test data in {source_dir}")
+
+
+def verify_transfer(source_dir, dest_dir):
+    source_dir = os.path.abspath(source_dir)
+    dest_dir = os.path.abspath(dest_dir)
+
+    received_prefix = os.path.join(dest_dir, source_dir.lstrip(os.sep))
+    if not os.path.exists(received_prefix):
+        return [], ["no received files found"]
+
+    mismatches = []
+    missing = []
+
+    for root, dirs, files in os.walk(source_dir):
+        for f in files:
+            src_path = os.path.join(root, f)
+            rel = os.path.relpath(src_path, source_dir)
+            dst_path = os.path.join(received_prefix, rel)
+
+            if not os.path.exists(dst_path):
+                missing.append(rel)
+            elif not filecmp.cmp(src_path, dst_path, shallow=False):
+                mismatches.append(rel)
+
+    return mismatches, missing
+
+
+def run_suite(env_name, apply_limits, source_dir, dest_dir):
     results = []
     print(f"\n{'=' * 60}")
-    print(f"🚀 Starting Suite: {env_name}")
+    print(f"Suite: {env_name}")
     print(f"{'=' * 60}")
 
     if apply_limits:
-        print(
+        print(f"  Disk I/O: Reads <= {READ_BPS_MAX}, Writes <= {WRITE_BPS_MAX}")
-            f"Applying Disk I/O Limits: Reads <= {READ_BPS_MAX}, Writes <= {WRITE_BPS_MAX}"
+        print(f"  Network: {NET_LIMIT}, {NET_DELAY} delay")
-        )
-        print(f"Applying Network Limits: {NET_LIMIT}, {NET_DELAY} delay")
         client_prefix = CLIENT_CMD_PREFIX
     else:
-        print("Running Baseline (No limits applied)")
+        print("  Baseline (no limits)")
-        client_prefix = []  # Run normally without systemd-run/limits
+        client_prefix = []
 
     try:
-        # SETUP: Apply or ensure clean network limits
         if apply_limits:
             subprocess.run(NET_LIMIT_CMD, check=True)
         else:
-            # Silently attempt to clear any leftover rules just to ensure a clean baseline
             subprocess.run(NET_RESET_CMD, capture_output=True)
 
         for case in TEST_CASES:
             name = case["name"]
             flags = case["flags"]
+            print(f"\n  --- {name} ---")
 
-            print(f"\n--- Running: {name} ---")
+            if os.path.exists(dest_dir):
+                shutil.rmtree(dest_dir)
 
             server_process = None
             try:
-                # 1. Start the server
-                print("  Starting server...")
                 server_process = subprocess.Popen(
                     SERVER_CMD, stdout=subprocess.DEVNULL, stderr=None
                 )
-                time.sleep(0.5)  # Allow server to bind to port
+                time.sleep(0.5)
 
-                # 2. Build and run the client
+                env = os.environ.copy()
-                client_cmd = client_prefix + base_client_cmd + flags
+
-                print(f"  Running client: {' '.join(client_cmd)}")
+                client_cmd = (
+                    client_prefix
+                    + base_client_cmd
+                    + ["--source-dir", source_dir, "--dest-dir", dest_dir, "--save-to-disk"]
+                    + flags
+                )
+                print(f"    Running: {' '.join(client_cmd)}")
 
                 start_time = time.monotonic()
                 client_result = subprocess.run(
-                    client_cmd, text=True, capture_output=True
+                    client_cmd, env=env, text=True, capture_output=True
                 )
                 end_time = time.monotonic()
-
                 duration = end_time - start_time
 
+                if server_process:
+                    try:
+                        server_process.wait(timeout=5)
+                    except subprocess.TimeoutExpired:
+                        server_process.kill()
+                        server_process.wait()
+                    server_process = None
+
+                mismatches, missing = [], []
                 if client_result.returncode == 0:
-                    results.append(
+                    mismatches, missing = verify_transfer(source_dir, dest_dir)
-                        {
+
-                            "environment": env_name,
+                entry = {
                     "name": name,
-                            "status": "Success",
+                    "suite": env_name,
-                            "time": f"{duration:.4f}s",
+                    "time": f"{duration:.4f}s" if client_result.returncode == 0 else "N/A",
-                            "error": "",
                 }
-                    )
+
+                if client_result.returncode == 0 and not mismatches and not missing:
+                    entry["status"] = "Success"
+                    entry["error"] = ""
                 else:
-                    print(f"  ⚠️ Failed (code: {client_result.returncode})")
+                    entry["status"] = "Failed"
-                    err_msg = (
+                    errors = []
+                    if client_result.returncode != 0:
+                        err = (
                             client_result.stderr.strip().split("\n")[0]
                             if client_result.stderr
                             else (
@@ -120,107 +199,101 @@ def run_suite(env_name, apply_limits):
                                 else "No output"
                             )
                         )
-                    results.append(
+                        errors.append(f"Exit code {client_result.returncode}: {err[:80]}")
-                        {
+                    if missing:
-                            "environment": env_name,
+                        errors.append(f"Missing ({len(missing)}): {', '.join(missing[:5])}")
-                            "name": name,
+                    if mismatches:
-                            "status": "Failed",
+                        errors.append(f"Mismatch ({len(mismatches)}): {', '.join(mismatches[:3])}")
-                            "time": "N/A",
+                    entry["error"] = " | ".join(errors)
-                            "error": f"Exit code {client_result.returncode}: {err_msg[:40]}",
+
-                        }
+                results.append(entry)
-                    )
 
             except subprocess.TimeoutExpired:
-                print("  ⚠️ Timeout (exceeded 15s)")
                 results.append(
-                    {
+                    {"name": name, "suite": env_name, "status": "Timeout", "time": "N/A", "error": "Exceeded 15s"}
-                        "environment": env_name,
-                        "name": name,
-                        "status": "Timeout",
-                        "time": "N/A",
-                        "error": "Exceeded 15 seconds",
-                    }
                 )
             except Exception as e:
-                print(f"  ❌ Error: {e}")
                 results.append(
-                    {
+                    {"name": name, "suite": env_name, "status": "Error", "time": "N/A", "error": str(e)}
-                        "environment": env_name,
-                        "name": name,
-                        "status": "Error",
-                        "time": "N/A",
-                        "error": str(e),
-                    }
                 )
             finally:
-                # Clean up the server for this test case
                 if server_process:
-                    print("  Stopping server...")
                     try:
-                        server_process.terminate()
                         server_process.wait(timeout=5)
                     except subprocess.TimeoutExpired:
                         server_process.kill()
                         server_process.wait()
 
     except subprocess.CalledProcessError as e:
-        print(f"❌ Error running system limit command: {' '.join(e.cmd)}")
+        print(f"  Error running limit command: {' '.join(e.cmd)}")
-        print("Are you running this script with 'sudo' privileges?")
-
     finally:
-        # TEARDOWN: Remove network limits if they were applied
         if apply_limits:
-            print("\nCleaning up limits for this suite...")
             try:
                 subprocess.run(NET_RESET_CMD, check=True, capture_output=True)
-                print("Network limits removed.")
+            except Exception:
-            except Exception as e:
+                pass
-                print(f"⚠️ Could not reset network settings: {e}")
 
     return results
 
 
-os.system("cmake -B build -S .")
+def main():
-os.system("cd build && make")
+    parser = argparse.ArgumentParser(description="FastSync integration test / benchmark")
+    parser.add_argument("--source-dir", default=DEFAULT_SOURCE_DIR,
+                        help="Source directory for test files (default: %(default)s)")
+    parser.add_argument("--dest-dir", default=DEFAULT_DEST_DIR,
+                        help="Destination directory for received files (default: %(default)s)")
+    parser.add_argument("--keep-data", action="store_true",
+                        help="Keep test_data directory after run")
+    parser.add_argument("--no-throttled", action="store_true",
+                        help="Skip throttled suite (requires sudo)")
+    parser.add_argument("--no-unlimited", action="store_true",
+                        help="Skip unlimited suite")
+    args = parser.parse_args()
 
-# --- Main Execution ---
+    os.system("cmake -B build -S . > /dev/null 2>&1")
-all_results = []
+    ret = os.system("cd build && make -j$(nproc) 2>&1 | tail -3")
+    if ret != 0:
+        print("Build failed")
+        sys.exit(1)
 
-# 1. Run Baseline (No Limits)
+    generate_test_files(args.source_dir)
-all_results.extend(run_suite("Unlimited", apply_limits=False))
+    os.makedirs(args.dest_dir, exist_ok=True)
 
-# # 2. Run Throttled (With Limits)
+    try:
-all_results.extend(run_suite("Throttled", apply_limits=True))
+        all_results = []
 
-# --- Print Comparison Table ---
+        if not args.no_unlimited:
-print("\n" + "=" * 105)
+            all_results.extend(
-print(f"{'fastSync BENCHMARK RESULTS (COMPARISON)':^105}")
+                run_suite("Unlimited", False, args.source_dir, args.dest_dir)
-print("=" * 105)
-print(
-    f"{'Configuration':<45} | {'Environment':<12} | {'Status':<10} | {'Time':<10} | {'Details/Error':<20}"
-)
-print("-" * 105)
-
-# Sort results by test case name first, then environment to easily compare
-# This groups the baseline and throttled results for the same test next to each other
-# sorted_results = sorted(
-#     all_results,
-#     key=lambda x: (
-#         TEST_CASES.index(
-#             next(item for item in TEST_CASES if item["name"] == x["name"])
-#         ),
-#         x["environment"],
-#     ),
-# )
-
-for res in all_results:
-    status_symbol = (
-        "✅"
-        if res["status"] == "Success"
-        else ("⏳" if res["status"] == "Timeout" else "❌")
             )
-    status_str = f"{status_symbol} {res['status']}"
+
+        if not args.no_throttled:
+            all_results.extend(
+                run_suite("Throttled", True, args.source_dir, args.dest_dir)
+            )
+
+        print("\n" + "=" * 110)
+        print(f"{'RESULTS':^110}")
+        print("=" * 110)
+        print(f"{'Configuration':<45} | {'Suite':<12} | {'Status':<8} | {'Time':<10} | {'Details'}")
+        print("-" * 110)
+
+        for res in all_results:
             print(
-        f"{res['name']:<45} | {res['environment']:<12} | {status_str:<10} | {res['time']:<10} | {res['error']:<20}"
+                f"{res['name']:<45} | {res['suite']:<12} | {res['status']:<8} | {res['time']:<10} | {res['error']}"
             )
-print("=" * 105)
+
+        failed = [r for r in all_results if r["status"] != "Success"]
+        if failed:
+            print(f"\n  {len(failed)} test(s) FAILED")
+            sys.exit(1)
+        else:
+            print(f"\n  ALL {len(all_results)} TESTS PASSED")
+
+    finally:
+        if not args.keep_data:
+            shutil.rmtree(TEST_DIR, ignore_errors=True)
+
+
+if __name__ == "__main__":
+    main()