Switch back to underline for invalid characters, and make restricted ASCII-only

2 files changed, 10 insertions, 5 deletions

diff --git a/youtube_dl/__init__.py b/youtube_dl/__init__.py
index cbf1dd1a7..c3e0f78e5 100644
--- a/youtube_dl/__init__.py
+++ b/youtube_dl/__init__.py
@@ -274,7 +274,7 @@ def parseOpts():
 			dest='outtmpl', metavar='TEMPLATE', help='output filename template. Use %(title)s to get the title, %(uploader)s for the uploader name, %(autonumber)s to get an automatically incremented number, %(ext)s for the filename extension, %(upload_date)s for the upload date (YYYYMMDD), %(extractor)s for the provider (youtube, metacafe, etc), %(id)s for the video id and %% for a literal percent. Use - to output to stdout.')
 	filesystem.add_option('--restrict-filenames',
 			action='store_true', dest='restrictfilenames',
-			help='Avoid some characters such as "&" and spaces in filenames', default=False)
+			help='Restrict filenames to only ASCII characters, and avoid "&" and spaces in filenames', default=False)
 	filesystem.add_option('-a', '--batch-file',
 			dest='batchfile', metavar='FILE', help='file containing URLs to download (\'-\' for stdin)')
 	filesystem.add_option('-w', '--no-overwrites',
diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index 1f60d34ae..3339f56ec 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -207,15 +207,20 @@ def sanitize_filename(s, restricted=False):
 		elif char == ':':
 			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('_')
+	if not result:
+		result = '_'
+	return result
 
 def orderedSet(iterable):
 	""" Remove all duplicates from the input iterable """