复查修复(三): 调度与扫描

- scan: .tmp/.part 过滤 + list_files 分批遍历(消除 10 万条截断误判)
- Scheduler: 持锁只做标记, 锁外异步执行任务(长任务不再冻结调度)
- cron 范围/步进边界校验;FTP 下载后大小校验防静默截断

core/mirror_sync.py

@@ -1308,6 +1308,12 @@ class MirrorSyncManager:
                 else:
                     ftp.retrbinary(f'RETR {remote_filename}', f.write)
 
+            # 校验下载大小与远端一致，防止静默截断
+            actual_size = os.path.getsize(local_path)
+            expected = file_info.get('size')
+            if expected is not None and expected > 0 and actual_size != expected:
+                raise Exception(f"文件大小不匹配: {actual_size} != {expected}")
+            
             return True
         except Exception as e:
             print(f"下载FTP文件失败 {remote_filename}: {e}")

core/scheduler.py

@@ -103,16 +103,27 @@ class ScheduledTask:
                 result.extend(self._parse_cron_part(sub.strip(), min_val, max_val))
             return result
 
-        # 处理范围: "1-5"
+        # 处理范围: "1-5"（校验边界，越界返回空）
         if '-' in part:
+            try:
                 start, end = part.split('-')
-            return list(range(int(start), int(end) + 1))
+                start, end = int(start), int(end)
+                if not (min_val <= start <= end <= max_val):
+                    return []
+                return list(range(start, end + 1))
+            except ValueError:
+                return []
 
         # 处理步进: "*/5"
         if '/' in part:
             base, step = part.split('/')
             base_list = self._parse_cron_part(base or '*', min_val, max_val)
+            try:
                 step = int(step)
+                if step <= 0:
+                    return []
+            except ValueError:
+                return []
             return base_list[::step]
 
         # 单个值
@@ -301,18 +312,23 @@ class Scheduler:
         logger.info("定时任务调度器已停止")
 
     def _run_loop(self):
-        """运行循环"""
+        """运行循环（持锁只做标记，锁外异步执行任务，避免长任务冻结调度）"""
         while self._running:
             try:
                 now = datetime.now()
 
+                due = []
                 with self._lock:
                     for name, task in self.tasks.items():
                         if task.should_run_now():
-                            # 使用线程池执行任务
+                            due.append(task)
-                            from concurrent.futures import ThreadPoolExecutor
+
-                            with ThreadPoolExecutor(max_workers=1) as executor:
+                # 锁外异步执行，不阻塞后续任务的调度
-                                executor.submit(task.run)
+                for task in due:
+                    try:
+                        threading.Thread(target=task.run, daemon=True).start()
+                    except Exception as e:
+                        logger.error(f"调度任务 {task.name} 启动失败: {e}")
 
                 time.sleep(self.check_interval)
 

core/sync_scheduler.py

@@ -259,10 +259,12 @@ class SyncScheduler:
                 self.last_scan_time = time.time()
                 return
 
-            # 扫描文件
+            # 扫描文件（过滤传输中的临时文件）
             scanned_files = []
             for root, dirs, files in os.walk(base_dir):
                 for filename in files:
+                    if filename.endswith(('.tmp', '.part')):
+                        continue
                     full_path = os.path.join(root, filename)
                     rel_path = os.path.relpath(full_path, base_dir).replace("\\", "/")
 
@@ -276,9 +278,16 @@ class SyncScheduler:
                     }
                     scanned_files.append(file_info)
 
-            # 与数据库对比
+            # 与数据库对比（分批遍历，避免 limit 截断导致误判新增/删除）
-            db_files = self.db.list_files(limit=100000)
+            db_paths = set()
-            db_paths = {f.path for f in db_files if not f.is_dir}
+            offset = 0
+            BATCH = 5000
+            while True:
+                batch = self.db.list_files(limit=BATCH, offset=offset)
+                if not batch:
+                    break
+                db_paths.update(f.path for f in batch if not f.is_dir)
+                offset += BATCH
 
             # 检测新增
             local_paths = {f['path'] for f in scanned_files}