1 files changed, 75 insertions, 13 deletions
diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index 3ecd798d7..42ad520f9 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -24,6 +24,7 @@ import socket
 import struct
 import subprocess
 import sys
+import tempfile
 import traceback
 import xml.etree.ElementTree
 import zlib
@@ -228,22 +229,46 @@ else:
         assert type(s) == type(u'')
         print(s)
 
-# In Python 2.x, json.dump expects a bytestream.
-# In Python 3.x, it writes to a character stream
-if sys.version_info < (3,0):
-    def write_json_file(obj, fn):
-        with open(fn, 'wb') as f:
-            json.dump(obj, f)
-else:
-    def write_json_file(obj, fn):
-        with open(fn, 'w', encoding='utf-8') as f:
-            json.dump(obj, f)
 
-if sys.version_info >= (2,7):
+def write_json_file(obj, fn):
+    """ Encode obj as JSON and write it to fn, atomically """
+
+    args = {
+        'suffix': '.tmp',
+        'prefix': os.path.basename(fn) + '.',
+        'dir': os.path.dirname(fn),
+        'delete': False,
+    }
+
+    # In Python 2.x, json.dump expects a bytestream.
+    # In Python 3.x, it writes to a character stream
+    if sys.version_info < (3, 0):
+        args['mode'] = 'wb'
+    else:
+        args.update({
+            'mode': 'w',
+            'encoding': 'utf-8',
+        })
+
+    tf = tempfile.NamedTemporaryFile(**args)
+
+    try:
+        with tf:
+            json.dump(obj, tf)
+        os.rename(tf.name, fn)
+    except:
+        try:
+            os.remove(tf.name)
+        except OSError:
+            pass
+        raise
+
+
+if sys.version_info >= (2, 7):
     def find_xpath_attr(node, xpath, key, val):
         """ Find the xpath xpath[@key=val] """
-        assert re.match(r'^[a-zA-Z]+$', key)
-        assert re.match(r'^[a-zA-Z0-9@\s:._]*$', val)
+        assert re.match(r'^[a-zA-Z-]+$', key)
+        assert re.match(r'^[a-zA-Z0-9@\s:._-]*$', val)
         expr = xpath + u"[@%s='%s']" % (key, val)
         return node.find(expr)
 else:
@@ -827,6 +852,7 @@ def unified_strdate(date_str):
         '%b %dnd %Y %I:%M%p',
         '%b %dth %Y %I:%M%p',
         '%Y-%m-%d',
+        '%Y/%m/%d',
         '%d.%m.%Y',
         '%d/%m/%Y',
         '%Y/%m/%d %H:%M:%S',
@@ -852,6 +878,8 @@ def unified_strdate(date_str):
     return upload_date
 
 def determine_ext(url, default_ext=u'unknown_video'):
+    if url is None:
+        return default_ext
     guess = url.partition(u'?')[0].rpartition(u'.')[2]
     if re.match(r'^[A-Za-z0-9]+$', guess):
         return guess
@@ -1271,9 +1299,15 @@ def int_or_none(v, scale=1, default=None, get_attr=None, invscale=1):
     if get_attr:
         if v is not None:
             v = getattr(v, get_attr, None)
+    if v == '':
+        v = None
     return default if v is None else (int(v) * invscale // scale)
 
 
+def str_or_none(v, default=None):
+    return default if v is None else compat_str(v)
+
+
 def str_to_int(int_str):
     if int_str is None:
         return None
@@ -1440,6 +1474,34 @@ def strip_jsonp(code):
     return re.sub(r'(?s)^[a-zA-Z0-9_]+\s*\(\s*(.*)\);?\s*?\s*$', r'\1', code)
 
 
+def js_to_json(code):
+    def fix_kv(m):
+        key = m.group(2)
+        if key.startswith("'"):
+            assert key.endswith("'")
+            assert '"' not in key
+            key = '"%s"' % key[1:-1]
+        elif not key.startswith('"'):
+            key = '"%s"' % key
+
+        value = m.group(4)
+        if value.startswith("'"):
+            assert value.endswith("'")
+            assert '"' not in value
+            value = '"%s"' % value[1:-1]
+
+        return m.group(1) + key + m.group(3) + value
+
+    res = re.sub(r'''(?x)
+            ([{,]\s*)
+            ("[^"]*"|\'[^\']*\'|[a-z0-9A-Z]+)
+            (:\s*)
+            ([0-9.]+|true|false|"[^"]*"|\'[^\']*\'|\[|\{)
+        ''', fix_kv, code)
+    res = re.sub(r',(\s*\])', lambda m: m.group(1), res)
+    return res
+
+
 def qualities(quality_ids):
     """ Get a numeric quality value out of a list of possible values """
     def q(qid):