From 8dfc27b7d305bc7e74c2250fb5080a67071de7ea Mon Sep 17 00:00:00 2001
From: felix <felix.von.s@posteo.de>
Date: Sun, 13 Mar 2016 12:29:15 +0100
Subject: [PATCH 1/2] [utils] js_to_json: various improvements

now JS object literals like { /* " */ 0: ",]\xaa<\/p>", } will be correctly converted to JSON.
---
 test/test_utils.py  | 12 ++++++++++++
 youtube_dl/utils.py | 30 ++++++++++++++++--------------
 2 files changed, 28 insertions(+), 14 deletions(-)

diff --git a/test/test_utils.py b/test/test_utils.py
index a35debfe1..db189f1f6 100644
--- a/test/test_utils.py
+++ b/test/test_utils.py
@@ -631,6 +631,18 @@ class TestUtil(unittest.TestCase):
         on = js_to_json('{"abc": "def",}')
         self.assertEqual(json.loads(on), {'abc': 'def'})
 
+        on = js_to_json('{ 0: /* " \n */ ",]" , }')
+        self.assertEqual(json.loads(on), {'0': ',]'})
+
+        on = js_to_json(r'["<p>x<\/p>"]')
+        self.assertEqual(json.loads(on), ['<p>x</p>'])
+
+        on = js_to_json(r'["\xaa"]')
+        self.assertEqual(json.loads(on), ['\u00aa'])
+
+        on = js_to_json("['a\\\nb']")
+        self.assertEqual(json.loads(on), ['ab'])
+
     def test_extract_attributes(self):
         self.assertEqual(extract_attributes('<e x="y">'), {'x': 'y'})
         self.assertEqual(extract_attributes("<e x='y'>"), {'x': 'y'})
diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index b6e1dc809..1400ff2db 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -1847,24 +1847,26 @@ def js_to_json(code):
         v = m.group(0)
         if v in ('true', 'false', 'null'):
             return v
-        if v.startswith('"'):
-            v = re.sub(r"\\'", "'", v[1:-1])
-        elif v.startswith("'"):
-            v = v[1:-1]
-            v = re.sub(r"\\\\|\\'|\"", lambda m: {
-                '\\\\': '\\\\',
-                "\\'": "'",
+        elif v.startswith('/*') or v == ',':
+            return ""
+
+        if v[0] in ("'", '"'):
+            v = re.sub(r'(?s)\\.|"', lambda m: {
                 '"': '\\"',
-            }[m.group(0)], v)
+                "\\'": "'",
+                '\\\n': '',
+                '\\x': '\\u00',
+            }.get(m.group(0), m.group(0)), v[1:-1])
+
         return '"%s"' % v
 
-    res = re.sub(r'''(?x)
-        "(?:[^"\\]*(?:\\\\|\\['"nu]))*[^"\\]*"|
-        '(?:[^'\\]*(?:\\\\|\\['"nu]))*[^'\\]*'|
-        [a-zA-Z_][.a-zA-Z_0-9]*
+    return re.sub(r'''(?sx)
+        "(?:[^"\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^"\\]*"|
+        '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
+        /\*.*?\*/|,(?=\s*[\]}])|
+        [a-zA-Z_][.a-zA-Z_0-9]*|
+        [0-9]+(?=\s*:)
         ''', fix_kv, code)
-    res = re.sub(r',(\s*[\]}])', lambda m: m.group(1), res)
-    return res
 
 
 def qualities(quality_ids):

From 2f481215d95c738dad530349392f83a488ea145b Mon Sep 17 00:00:00 2001
From: felix <felix.von.s@posteo.de>
Date: Sun, 20 Mar 2016 12:17:57 +0100
Subject: [PATCH 2/2] [ora] minimise fragile regex shenanigans; recognise
 unsafespeech.com URLs

---
 youtube_dl/extractor/ora.py | 20 +++++++++++---------
 1 file changed, 11 insertions(+), 9 deletions(-)

diff --git a/youtube_dl/extractor/ora.py b/youtube_dl/extractor/ora.py
index 8545fb1b8..cfae71bcc 100644
--- a/youtube_dl/extractor/ora.py
+++ b/youtube_dl/extractor/ora.py
@@ -6,13 +6,14 @@ from .common import InfoExtractor
 from ..compat import compat_urlparse
 from ..utils import (
     get_element_by_attribute,
+    js_to_json,
     qualities,
     unescapeHTML,
 )
 
 
 class OraTVIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?ora\.tv/([^/]+/)*(?P<id>[^/\?#]+)'
+    _VALID_URL = r'https?://(?:www\.)?(ora\.tv|unsafespeech\.com)/([^/]+/)*(?P<id>[^/\?#]+)'
     _TEST = {
         'url': 'https://www.ora.tv/larrykingnow/2015/12/16/vine-youtube-stars-zach-king-king-bach-on-their-viral-videos-0_36jupg6090pq',
         'md5': 'fa33717591c631ec93b04b0e330df786',
@@ -28,10 +29,13 @@ class OraTVIE(InfoExtractor):
         display_id = self._match_id(url)
         webpage = self._download_webpage(url, display_id)
 
-        video_data = self._search_regex(
-            r'"(?:video|current)"\s*:\s*({[^}]+?})', webpage, 'current video')
-        m3u8_url = self._search_regex(
-            r'hls_stream"?\s*:\s*"([^"]+)', video_data, 'm3u8 url', None)
+        ora_meta = self._parse_json(self._search_regex(
+            r'(?s);\s*ora_meta = ({.*?});</script>', webpage, 'ora_meta'), display_id,
+            transform_source=lambda data: js_to_json(re.sub('":(document|\().*?(:false|\(\)),', '":null,', data)))
+
+        video_data = ora_meta.get('video', ora_meta.get('current'))
+        m3u8_url = video_data['hls_stream']
+
         if m3u8_url:
             formats = self._extract_m3u8_formats(
                 m3u8_url, display_id, 'mp4', 'm3u8_native',
@@ -60,13 +64,11 @@ class OraTVIE(InfoExtractor):
                 r'"youtube_id"\s*:\s*"([^"]+)', webpage, 'youtube id'), 'Youtube')
 
         return {
-            'id': self._search_regex(
-                r'"id"\s*:\s*(\d+)', video_data, 'video id', default=display_id),
+            'id': video_data.get('id', display_id),
             'display_id': display_id,
             'title': unescapeHTML(self._og_search_title(webpage)),
             'description': get_element_by_attribute(
                 'class', 'video_txt_decription', webpage),
-            'thumbnail': self._proto_relative_url(self._search_regex(
-                r'"thumb"\s*:\s*"([^"]+)', video_data, 'thumbnail', None)),
+            'thumbnail': self._proto_relative_url(video_data.get('thumb')),
             'formats': formats,
         }