Blob Blame History Raw
From 843735d9dc129eb34f16c4c08576ac5f742971f6 Mon Sep 17 00:00:00 2001
From: Lumir Balhar <lbalhar@redhat.com>
Date: Tue, 17 May 2022 12:44:01 +0200
Subject: [PATCH] CVE-2019-20916

---
 download.py | 31 +++++++++++++++++++++++++-----
 1 file changed, 26 insertions(+), 5 deletions(-)

diff --git a/download.py b/download.py
index 54d3131..e49130b 100644
--- a/download.py
+++ b/download.py
@@ -54,7 +54,8 @@ __all__ = ['get_file_content',
            'is_url', 'url_to_path', 'path_to_url',
            'is_archive_file', 'unpack_vcs_link',
            'unpack_file_url', 'is_vcs_url', 'is_file_url',
-           'unpack_http_url', 'unpack_url']
+           'unpack_http_url', 'unpack_url',
+           'parse_content_disposition', 'sanitize_content_filename']
 
 
 logger = logging.getLogger(__name__)
@@ -824,6 +825,29 @@ def unpack_url(link, location, download_dir=None,
         write_delete_marker_file(location)
 
 
+def sanitize_content_filename(filename):
+    # type: (str) -> str
+    """
+    Sanitize the "filename" value from a Content-Disposition header.
+    """
+    return os.path.basename(filename)
+
+
+def parse_content_disposition(content_disposition, default_filename):
+    # type: (str, str) -> str
+    """
+    Parse the "filename" value from a Content-Disposition header, and
+    return the default filename if the result is empty.
+    """
+    _type, params = cgi.parse_header(content_disposition)
+    filename = params.get('filename')
+    if filename:
+        # We need to sanitize the filename to prevent directory traversal
+        # in case the filename contains ".." path parts.
+        filename = sanitize_content_filename(filename)
+    return filename or default_filename
+
+
 def _download_http_url(link, session, temp_dir, hashes):
     """Download link url into temp_dir using provided session"""
     target_url = link.url.split('#', 1)[0]
@@ -864,10 +888,7 @@ def _download_http_url(link, session, temp_dir, hashes):
     # Have a look at the Content-Disposition header for a better guess
     content_disposition = resp.headers.get('content-disposition')
     if content_disposition:
-        type, params = cgi.parse_header(content_disposition)
-        # We use ``or`` here because we don't want to use an "empty" value
-        # from the filename param.
-        filename = params.get('filename') or filename
+        filename = parse_content_disposition(content_disposition, filename)
     ext = splitext(filename)[1]
     if not ext:
         ext = mimetypes.guess_extension(content_type)
-- 
2.35.3