X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Futils.py;h=40d6823a0f8ddbbf42a54bf4b0aea9e344825600;hb=1a9c655e3b1569f315d4193e877cba0b4a863c63;hp=6f53337d49496561cada56ed149866fa4f949c0e;hpb=51937c086943a3bdbf6f707c75d041ed3b0ba743;p=youtube-dl

diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index 6f53337d4..40d6823a0 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -26,21 +26,24 @@ std_headers = {
 	'Accept-Language': 'en-us,en;q=0.5',
 }
 
+try:
+    compat_str = unicode # Python 2
+except NameError:
+    compat_str = str
+
 def preferredencoding():
 	"""Get preferred encoding.
 
 	Returns the best encoding scheme for the system, based on
 	locale.getpreferredencoding() and some further tweaks.
 	"""
-	def yield_preferredencoding():
-		try:
-			pref = locale.getpreferredencoding()
-			u'TEST'.encode(pref)
-		except:
-			pref = 'UTF-8'
-		while True:
-			yield pref
-	return yield_preferredencoding().next()
+	try:
+		pref = locale.getpreferredencoding()
+		u'TEST'.encode(pref)
+	except:
+		pref = 'UTF-8'
+
+	return pref
 
 
 def htmlentity_transform(matchobj):
@@ -189,24 +192,36 @@ def timeconvert(timestr):
 	if timetuple is not None:
 		timestamp = email.utils.mktime_tz(timetuple)
 	return timestamp
-	
-def sanitize_filename(s):
-	"""Sanitizes a string so it could be used as part of a filename."""
+
+def sanitize_filename(s, restricted=False):
+	"""Sanitizes a string so it could be used as part of a filename.
+	If restricted is set, use a stricter subset of allowed characters.
+	"""
 	def replace_insane(char):
 		if char == '?' or ord(char) < 32 or ord(char) == 127:
 			return ''
 		elif char == '"':
-			return '\''
+			return '' if restricted else '\''
 		elif char == ':':
-			return ' -'
+			return '_-' if restricted else ' -'
 		elif char in '\\/|*<>':
-			return '-'
+			return '_'
+		if restricted and (char in '!&\'' or char.isspace()):
+			return '_'
+		if restricted and ord(char) > 127:
+			return '_'
 		return char
 
 	result = u''.join(map(replace_insane, s))
-	while '--' in result:
-		result = result.replace('--', '-')
-	return result.strip('-')
+	while '__' in result:
+		result = result.replace('__', '_')
+	result = result.strip('_')
+	# Common case of "Foreign band name - English song title"
+	if restricted and result.startswith('-_'):
+		result = result[2:]
+	if not result:
+		result = '_'
+	return result
 
 def orderedSet(iterable):
 	""" Remove all duplicates from the input iterable """
@@ -299,7 +314,7 @@ class ContentTooShortError(Exception):
 
 class Trouble(Exception):
 	"""Trouble helper exception
-	
+
 	This is an exception to be handled with
 	FileDownloader.trouble
 	"""