[core] Decode environment variables with filesystem encoding (Fixes #3854, Fixes #3217, Fixes #2918)

Introduces compat versions of os.getenv and os.path.expanduser

test/test_utils.py

@@ -44,6 +44,9 @@
     limit_length,
     escape_rfc3986,
     escape_url,
+    get_filesystem_encoding,
+    compat_getenv,
+    compat_expanduser,
 )
 
 
@@ -331,5 +334,15 @@ def test_escape_url(self):
         )
         self.assertEqual(escape_url('http://vimeo.com/56015672#at=0'), 'http://vimeo.com/56015672#at=0')
 
+    def test_compat_getenv(self):
+        test_str = 'тест'
+        os.environ['YOUTUBE-DL-TEST'] = test_str.encode(get_filesystem_encoding())
+        self.assertEqual(compat_getenv('YOUTUBE-DL-TEST'), test_str)
+
+    def test_compat_expanduser(self):
+        test_str = 'C:\Documents and Settings\тест\Application Data'
+        os.environ['HOME'] = test_str.encode(get_filesystem_encoding())
+        self.assertEqual(compat_expanduser('~'), test_str)
+
 if __name__ == '__main__':
     unittest.main()

youtube_dl/YoutubeDL.py

@@ -24,6 +24,7 @@
 
 from .utils import (
     compat_cookiejar,
+    compat_expanduser,
     compat_http_client,
     compat_str,
     compat_urllib_error,
@@ -447,7 +448,7 @@ def prepare_filename(self, info_dict):
             template_dict = collections.defaultdict(lambda: 'NA', template_dict)
 
             outtmpl = self.params.get('outtmpl', DEFAULT_OUTTMPL)
-            tmpl = os.path.expanduser(outtmpl)
+            tmpl = compat_expanduser(outtmpl)
             filename = tmpl % template_dict
             return filename
         except ValueError as err:

youtube_dl/__init__.py

@@ -94,6 +94,7 @@
     parseOpts,
 )
 from .utils import (
+    compat_expanduser,
     compat_getpass,
     compat_print,
     DateRange,
@@ -285,7 +286,7 @@ def _real_main(argv=None):
                      u' template'.format(outtmpl))
 
     any_printing = opts.geturl or opts.gettitle or opts.getid or opts.getthumbnail or opts.getdescription or opts.getfilename or opts.getformat or opts.getduration or opts.dumpjson
-    download_archive_fn = os.path.expanduser(opts.download_archive) if opts.download_archive is not None else opts.download_archive
+    download_archive_fn = compat_expanduser(opts.download_archive) if opts.download_archive is not None else opts.download_archive
 
     ydl_opts = {
         'usenetrc': opts.usenetrc,

youtube_dl/cache.py

@@ -9,6 +9,7 @@
 import traceback
 
 from .utils import (
+    compat_expanduser,
     write_json_file,
 )
 
@@ -22,7 +23,7 @@ def _get_root_dir(self):
         if res is None:
             cache_root = os.environ.get('XDG_CACHE_HOME', '~/.cache')
             res = os.path.join(cache_root, 'youtube-dl')
-        return os.path.expanduser(res)
+        return compat_expanduser(res)
 
     def _get_cache_fn(self, section, key, dtype):
         assert re.match(r'^[a-zA-Z0-9_.-]+$', section), \

youtube_dl/options.py

@@ -6,6 +6,8 @@
 import sys
 
 from .utils import (
+    compat_expanduser,
+    compat_getenv,
     get_term_width,
     write_string,
 )
@@ -27,19 +29,19 @@ def _readOptions(filename_bytes, default=[]):
         return res
 
     def _readUserConf():
-        xdg_config_home = os.environ.get('XDG_CONFIG_HOME')
+        xdg_config_home = compat_getenv('XDG_CONFIG_HOME')
         if xdg_config_home:
             userConfFile = os.path.join(xdg_config_home, 'youtube-dl', 'config')
             if not os.path.isfile(userConfFile):
                 userConfFile = os.path.join(xdg_config_home, 'youtube-dl.conf')
         else:
-            userConfFile = os.path.join(os.path.expanduser('~'), '.config', 'youtube-dl', 'config')
+            userConfFile = os.path.join(compat_expanduser('~'), '.config', 'youtube-dl', 'config')
             if not os.path.isfile(userConfFile):
-                userConfFile = os.path.join(os.path.expanduser('~'), '.config', 'youtube-dl.conf')
+                userConfFile = os.path.join(compat_expanduser('~'), '.config', 'youtube-dl.conf')
         userConf = _readOptions(userConfFile, None)
 
         if userConf is None:
-            appdata_dir = os.environ.get('appdata')
+            appdata_dir = compat_getenv('appdata')
             if appdata_dir:
                 userConf = _readOptions(
                     os.path.join(appdata_dir, 'youtube-dl', 'config'),
@@ -51,11 +53,11 @@ def _readUserConf():
 
         if userConf is None:
             userConf = _readOptions(
-                os.path.join(os.path.expanduser('~'), 'youtube-dl.conf'),
+                os.path.join(compat_expanduser('~'), 'youtube-dl.conf'),
                 default=None)
         if userConf is None:
             userConf = _readOptions(
-                os.path.join(os.path.expanduser('~'), 'youtube-dl.conf.txt'),
+                os.path.join(compat_expanduser('~'), 'youtube-dl.conf.txt'),
                 default=None)
 
         if userConf is None:

youtube_dl/utils.py

@@ -203,6 +203,48 @@ def compat_ord(c):
     if type(c) is int: return c
     else: return ord(c)
 
+
+# Environment variables should be decoded with filesystem encoding
+# otherwise this results in issues like #3854 #2918 #3217
+if sys.version_info >= (3, 0):
+    compat_getenv = os.getenv
+    compat_expanduser = os.path.expanduser
+else:
+    def compat_getenv(key, default=None):
+        env = os.getenv(key, default)
+        if env:
+            env = env.decode(get_filesystem_encoding())
+        return env
+
+    def compat_expanduser(path):
+        """Expand ~ and ~user constructs.
+
+        If user or $HOME is unknown, do nothing."""
+        if path[:1] != '~':
+            return path
+        i, n = 1, len(path)
+        while i < n and path[i] not in '/\\':
+            i += 1
+
+        if 'HOME' in os.environ:
+            userhome = compat_getenv('HOME')
+        elif 'USERPROFILE' in os.environ:
+            userhome = compat_getenv('USERPROFILE')
+        elif not 'HOMEPATH' in os.environ:
+            return path
+        else:
+            try:
+                drive = compat_getenv('HOMEDRIVE')
+            except KeyError:
+                drive = ''
+            userhome = os.path.join(drive, compat_getenv('HOMEPATH'))
+
+        if i != 1:  # ~user
+            userhome = os.path.join(os.path.dirname(userhome), path[1:i])
+
+        return userhome + path[i:]
+
+
 # This is not clearly defined otherwise
 compiled_regex_type = type(re.compile(''))
 
@@ -1204,11 +1246,14 @@ def read(self, *args):
         return self.f.read(*args)
 
 
+def get_filesystem_encoding():
+    encoding = sys.getfilesystemencoding()
+    return encoding if encoding is not None else 'utf-8'
+
+
 def shell_quote(args):
     quoted_args = []
-    encoding = sys.getfilesystemencoding()
+    encoding = get_filesystem_encoding()
-    if encoding is None:
-        encoding = 'utf-8'
     for a in args:
         if isinstance(a, bytes):
             # We may get a filename encoded with 'encodeFilename'
@@ -1258,7 +1303,7 @@ def format_bytes(bytes):
 
 
 def get_term_width():
-    columns = os.environ.get('COLUMNS', None)
+    columns = compat_getenv('COLUMNS', None)
     if columns:
         return int(columns)