youtube-dl

commit 448ef1f31c8bcc1550cf907fd46e31026ec981b2
parent 49941c4e4f6e33785a3be1e0d103bd81657d8a0d
Author: Sergey M․ <dstftw@gmail.com>
Date:   Sun, 18 Oct 2015 09:11:02 +0600

[extractor/common] Allow angle brackets in attributes in _og_regexes (#7215)

Diffstat:
M
test/test_InfoExtractor.py
 | 
4
++++
M
youtube_dl/extractor/common.py
 | 
2
+-

2 files changed, 5 insertions(+), 1 deletion(-)
diff --git a/test/test_InfoExtractor.py b/test/test_InfoExtractor.py
@@ -37,12 +37,16 @@ class TestInfoExtractor(unittest.TestCase):
             <meta property='og:image' content='http://domain.com/pic.jpg?key1=val1&amp;key2=val2'/>
             <meta content='application/x-shockwave-flash' property='og:video:type'>
             <meta content='Foo' property=og:foobar>
+            <meta name="og:test1" content='foo > < bar'/>
+            <meta name="og:test2" content="foo >//< bar"/>
             '''
         self.assertEqual(ie._og_search_title(html), 'Foo')
         self.assertEqual(ie._og_search_description(html), 'Some video\'s description ')
         self.assertEqual(ie._og_search_thumbnail(html), 'http://domain.com/pic.jpg?key1=val1&key2=val2')
         self.assertEqual(ie._og_search_video_url(html, default=None), None)
         self.assertEqual(ie._og_search_property('foobar', html), 'Foo')
+        self.assertEqual(ie._og_search_property('test1', html), 'foo > < bar')
+        self.assertEqual(ie._og_search_property('test2', html), 'foo >//< bar')
 
     def test_html_search_meta(self):
         ie = self.ie
diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py
@@ -645,7 +645,7 @@ class InfoExtractor(object):
     # Helper functions for extracting OpenGraph info
     @staticmethod
     def _og_regexes(prop):
-        content_re = r'content=(?:"([^>]+?)"|\'([^>]+?)\'|\s*([^\s"\'=<>`]+?))'
+        content_re = r'content=(?:"([^"]+?)"|\'([^\']+?)\'|\s*([^\s"\'=<>`]+?))'
         property_re = (r'(?:name|property)=(?:\'og:%(prop)s\'|"og:%(prop)s"|\s*og:%(prop)s\b)'
                        % {'prop': re.escape(prop)})
         template = r'<meta[^>]+?%s[^>]+?%s'