MCP Email Service

# UID优先+回退逻辑修复方案 ## 问题总结 用户在 `fetch_emails()` 和 `get_email_detail()` 中实现了 UID 支持，但其他写操作函数（`mark_email_read`, `delete_email`, `move_email_to_trash` 及其批量版本）只尝试 UID，不回退到序列号，导致向后兼容性问题。 ## 需要修复的函数 1. `mark_email_read` - 行 616 2. `delete_email` - 行 686 3. `move_email_to_trash` - 行 763 4. `batch_move_to_trash` - 行 849 5. `batch_delete_emails` - 行 950 6. `batch_mark_read` - 行 1009 ## 修复模式 对于单个邮件操作： ```python # Try UID first result, data = mail.uid('store', fetch_id, '+FLAGS', '\\Seen') # If UID fails, fallback to sequence number if result != 'OK' or not data or data == [None]: logger.debug(f"UID operation failed for {fetch_id}, trying sequence number") result, data = mail.store(fetch_id, '+FLAGS', '\\Seen') if result != 'OK': raise Exception(f"Failed to mark email as read: {result}") ``` 对于批量操作： ```python for email_id in email_ids: try: # Try UID first result, data = mail.uid('store', email_id, '+FLAGS', '\\Seen') # Fallback to sequence number if UID fails if result != 'OK' or not data or data == [None]: result, data = mail.store(email_id, '+FLAGS', '\\Seen') if result == 'OK': success_count += 1 else: failed_ids.append(email_id) except Exception as e: logger.warning(f"Failed to process email {email_id}: {e}") failed_ids.append(email_id) ``` ## 关键点 1. **优先使用 UID**：UID 更稳定（删除/添加邮件时不变） 2. **自动回退**：UID 失败时自动尝试序列号 3. **向后兼容**：支持旧代码传递序列号 4. **详细日志**：记录回退行为以便调试 5. **保持函数签名**：不破坏现有 API ## 实施步骤 鉴于文件已恢复，需要： 1. 重新应用原有的 IMAP 连接泄漏修复（try/finally 块） 2. 添加 HTML 转文本和正文截断功能 3. 为所有写操作函数添加 UID 回退逻辑