Add new --encoding option (Fixes #2650)

youtube_dl/YoutubeDL.py

@@ -8,6 +8,7 @@ import datetime
 import errno
 import io
 import json
+import locale
 import os
 import platform
 import re
@@ -159,6 +160,7 @@ class YoutubeDL(object):
     include_ads:       Download ads as well
     default_search:    Prepend this string if an input url is not valid.
                        'auto' for elaborate guessing
+    encoding:          Use this encoding instead of the system-specified.
 
     The following parameters are not used by YoutubeDL itself, they are used by
     the FileDownloader:
@@ -1200,6 +1202,9 @@ class YoutubeDL(object):
     def print_debug_header(self):
         if not self.params.get('verbose'):
             return
+
+        write_string('[debug] Encodings: locale %s, fs %s, out %s, pref %s\n' %
+                 (locale.getpreferredencoding(), sys.getfilesystemencoding(), sys.stdout.encoding, self.get_encoding()))
         write_string('[debug] youtube-dl version ' + __version__ + '\n')
         try:
             sp = subprocess.Popen(
@@ -1264,3 +1269,19 @@ class YoutubeDL(object):
         # (See https://github.com/rg3/youtube-dl/issues/1309 for details)
         opener.addheaders = []
         self._opener = opener
+
+    def encode(self, s):
+        if isinstance(s, bytes):
+            return s  # Already encoded
+
+        try:
+            return s.encode(self.get_encoding())
+        except UnicodeEncodeError as err:
+            err.reason = err.reason + '. Check your system encoding configuration or use the --encoding option.'
+            raise
+
+    def get_encoding(self):
+        encoding = self.params.get('encoding')
+        if encoding is None:
+            encoding = preferredencoding()
+        return encoding

youtube_dl/__init__.py

@@ -263,6 +263,9 @@ def parseOpts(overrideArguments=None):
         '--ignore-config',
         action='store_true',
         help='Do not read configuration files. When given in the global configuration file /etc/youtube-dl.conf: do not read the user configuration in ~/.config/youtube-dl.conf (%APPDATA%/youtube-dl/config.txt on Windows)')
+    general.add_option(
+        '--encoding', dest='encoding', metavar='ENCODING',
+        help='Force the specified encoding (experimental)')
 
     selection.add_option(
         '--playlist-start',
@@ -540,8 +543,6 @@ def parseOpts(overrideArguments=None):
             write_string(u'[debug] System config: ' + repr(_hide_login_info(systemConf)) + '\n')
             write_string(u'[debug] User config: ' + repr(_hide_login_info(userConf)) + '\n')
             write_string(u'[debug] Command-line args: ' + repr(_hide_login_info(commandLineConf)) + '\n')
-            write_string(u'[debug] Encodings: locale %r, fs %r, out %r, pref: %r\n' %
-                         (locale.getpreferredencoding(), sys.getfilesystemencoding(), sys.stdout.encoding, preferredencoding()))
 
     return parser, opts, args
 
@@ -786,6 +787,7 @@ def _real_main(argv=None):
         'include_ads': opts.include_ads,
         'default_search': opts.default_search,
         'youtube_include_dash_manifest': opts.youtube_include_dash_manifest,
+        'encoding': opts.encoding,
     }
 
     with YoutubeDL(ydl_opts) as ydl:

youtube_dl/postprocessor/ffmpeg.py

@@ -53,8 +53,9 @@ class FFmpegPostProcessor(PostProcessor):
 
         if self._downloader.params.get('verbose', False):
             self._downloader.to_screen(u'[debug] ffmpeg command line: %s' % shell_quote(cmd))
-        p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+        bcmd = [self._downloader.encode(c) for c in cmd]
-        stdout,stderr = p.communicate()
+        p = subprocess.Popen(bcmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+        stdout, stderr = p.communicate()
         if p.returncode != 0:
             stderr = stderr.decode('utf-8', 'replace')
             msg = stderr.strip().split('\n')[-1]

youtube_dl/utils.py

@@ -539,7 +539,6 @@ def encodeFilename(s, for_subprocess=False):
         encoding = 'utf-8'
     return s.encode(encoding, 'ignore')
 
-
 def decodeOption(optval):
     if optval is None:
         return optval