fix(download): validate torrent bytes before html fallback

app/application/torrent/download.py

@@ -2,7 +2,7 @@ import datetime
 import re
 from functools import lru_cache
 from pathlib import Path
-from typing import Any, Dict, List, Optional, Protocol, Tuple, Union
+from typing import Any, Callable, Dict, List, Optional, Protocol, Tuple, Union
 from urllib.parse import unquote
 
 from torrentool.api import Torrent
@@ -119,6 +119,19 @@ def _parse_filter_size_range(size_range: str) -> Tuple[str, float, Optional[floa
     return "unknown", 0, None
 
 
+def _cache_valid_torrent_content(
+        parser: Callable[[Union[str, bytes]], Tuple[str, List[str]]],
+        cache_backend: Any,
+        cache_path: Path,
+        content: bytes,
+) -> Tuple[str, List[str]]:
+    """解析合法种子并缓存原始二进制内容。"""
+    folder_name, file_list = parser(content)
+    if file_list:
+        cache_backend.set(cache_path.as_posix(), content, region="torrents")
+    return folder_name, file_list
+
+
 class TorrentHelper:
     """
     种子帮助类
@@ -187,6 +200,13 @@ class TorrentHelper:
             if req.content.startswith(b"magnet:"):
                 # 磁力链接
                 return cache_path, req.text, "", [], "获取到磁力链接"
+            # 优先按原始二进制解析，避免合法 torrent 被误判为首次下载 HTML。
+            folder_name, file_list = _cache_valid_torrent_content(
+                self.get_fileinfo_from_torrent_content, cache_backend, cache_path, req.content
+            )
+            if file_list:
+                # 成功拿到种子数据
+                return cache_path, req.content, folder_name, file_list, ""
             if "下载种子文件".encode("utf-8") in req.content:
                 # 首次下载提示页面
                 skip_flag = False
@@ -224,24 +244,17 @@ class TorrentHelper:
                     logger.warn(f"触发了站点首次种子下载，尝试自动跳过时出现错误：{str(err)}")
                 if not skip_flag:
                     return cache_path, None, "", [], "种子数据有误，请确认链接是否正确，如为PT站点则需手工在站点下载一次种子"
+            # 首次下载提示页面完成后，重新按原始二进制解析 POST 响应。
             assert req is not None
-            # 种子内容
             if req.content:
-                # 检查是不是种子文件，如果不是仍然抛出异常
-                try:
-                    # 获取种子目录和文件清单
-                    folder_name, file_list = self.get_fileinfo_from_torrent_content(req.content)
-                    if file_list:
-                        # 保存到缓存
-                        cache_backend.set(cache_path.as_posix(), req.content, region="torrents")
+                folder_name, file_list = _cache_valid_torrent_content(
+                    self.get_fileinfo_from_torrent_content, cache_backend, cache_path, req.content
+                )
+                if file_list:
                     # 成功拿到种子数据
                     return cache_path, req.content, folder_name, file_list, ""
-                except Exception as err:
-                    logger.error(f"种子文件解析失败：{str(err)}")
-                # 种子数据仍然错误
-                return cache_path, None, "", [], "种子数据有误，请确认链接是否正确"
-            # 返回失败
-            return cache_path, None, "", [], ""
+            # 解析失败时终止，避免把 HTML 继续下沉到下载器或缓存。
+            return cache_path, None, "", [], "种子数据有误，请确认链接是否正确"
         elif req is None:
             return cache_path, None, "", [], "无法打开链接"
         elif req.status_code == 429:

tests/test_site_access_ports.py

@@ -125,6 +125,79 @@ def test_torrent_fake_port_preserves_rate_limit_message(monkeypatch) -> None:
     assert result[1:] == (None, "", [], "触发站点流控，请稍后重试")
 
 
+def test_torrent_binary_is_parsed_before_first_download_html_detection(monkeypatch) -> None:
+    """合法二进制种子包含首次下载提示文本时仍应按 torrent 解析。"""
+    marker = "下载种子文件".encode("utf-8")
+    torrent_content = (
+        b"d4:infod7:comment"
+        + str(len(marker)).encode("ascii")
+        + b":"
+        + marker
+        + b"4:name7:foo.mkv6:lengthi1eee"
+    )
+    cached: list[bytes] = []
+
+    class Cache:
+        """记录合法种子缓存写入，不接触真实缓存后端。"""
+
+        @staticmethod
+        def get(*_args, **_kwargs):
+            """固定返回缓存未命中。"""
+            return None
+
+        @staticmethod
+        def set(_key, value, **_kwargs):
+            """记录原始二进制种子内容。"""
+            cached.append(value)
+
+    class HttpPort:
+        """返回包含提示文本的合法二进制种子响应。"""
+
+        @staticmethod
+        def request(**_kwargs):
+            """返回固定 torrent 响应。"""
+            return SimpleNamespace(
+                status_code=200,
+                headers={"Content-Type": "application/x-bittorrent"},
+                content=torrent_content,
+                text=torrent_content.decode("utf-8"),
+                reason="",
+            )
+
+    monkeypatch.setattr(torrent_module, "FileCache", Cache)
+    configure_torrent_port(HttpPort())
+
+    result = TorrentHelper().download_torrent("https://example.test/download/1")
+
+    assert result[1:] == (torrent_content, "", ["foo.mkv"], "")
+    assert cached == [torrent_content]
+
+
+def test_torrent_invalid_http_200_body_is_rejected_before_downloader(monkeypatch) -> None:
+    """HTTP 200 返回非种子内容时应在下载层终止并避免缓存。"""
+
+    class HttpPort:
+        """返回登录页面的种子 HTTP 假端口。"""
+
+        @staticmethod
+        def request(**_kwargs):
+            """返回固定 HTML 响应。"""
+            return SimpleNamespace(
+                status_code=200,
+                headers={"Content-Type": "text/html"},
+                content=b"<html>login</html>",
+                text="<html>login</html>",
+                reason="",
+            )
+
+    monkeypatch.setattr(torrent_module, "FileCache", _EmptyCache)
+    configure_torrent_port(HttpPort())
+
+    result = TorrentHelper().download_torrent("https://example.test/download/2")
+
+    assert result[1:] == (None, "", [], "种子数据有误，请确认链接是否正确")
+
+
 def test_cookie_invalid_parameters_keep_legacy_result_without_ports() -> None:
     """参数错误属于 Application 语义，不应依赖浏览器端口已装配。"""
     assert CookieHelper().get_site_cookie_ua("", "", "") == (