From 448ef1f31c8bcc1550cf907fd46e31026ec981b2 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Sergey=20M=E2=80=A4?= Date: Sun, 18 Oct 2015 09:11:02 +0600 Subject: [PATCH] [extractor/common] Allow angle brackets in attributes in _og_regexes (#7215) --- test/test_InfoExtractor.py | 4 ++++ 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 index 2a00d09a5..938466a80 100644 --- a/test/test_InfoExtractor.py +++ b/test/test_InfoExtractor.py @@ -37,12 +37,16 @@ def test_opengraph(self): + + ''' 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 index a0c4af92f..4365077f1 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -645,7 +645,7 @@ def _get_tfa_info(self, note='two-factor verification code'): # 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']+?%s[^>]+?%s'