From 13b0e81f17c802320302c5f0490d8bfc44a0faee Mon Sep 17 00:00:00 2001
From: dirkf <fieldhouse@gmx.net>
Date: Sun, 20 Oct 2024 13:22:25 +0100
Subject: [PATCH] [utils] Correctly match class names in
 `get_element[s]_by_class()` * reproduce CSS .classname behaviour ("bar"
 matches "bar", "foo bar baz", etc) * add tests

---
 test/test_utils.py  | 10 ++++++++--
 youtube_dl/utils.py |  2 +-
 2 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/test/test_utils.py b/test/test_utils.py
index 2947cce7e..16523a0c9 100644
--- a/test/test_utils.py
+++ b/test/test_utils.py
@@ -1603,11 +1603,14 @@ Line 1
 
     def test_get_element_by_class(self):
         html = '''
-            <span class="foo bar">nice</span>
+            <span class="foo bar baz-bam">nice</span>
         '''
 
         self.assertEqual(get_element_by_class('foo', html), 'nice')
+        self.assertEqual(get_element_by_class('bar', html), 'nice')
         self.assertEqual(get_element_by_class('no-such-class', html), None)
+        self.assertEqual(get_element_by_class('baz', html), None)
+        self.assertEqual(get_element_by_class('bam', html), None)
 
     def test_get_element_by_attribute(self):
         html = '''
@@ -1626,10 +1629,13 @@ Line 1
 
     def test_get_elements_by_class(self):
         html = '''
-            <span class="foo bar">nice</span><span class="foo bar">also nice</span>
+            <span class="foo bar baz-bam">nice</span><span class="foo bar">also nice</span>
         '''
 
         self.assertEqual(get_elements_by_class('foo', html), ['nice', 'also nice'])
+        self.assertEqual(get_elements_by_class('bar', html), ['nice', 'also nice'])
+        self.assertEqual(get_elements_by_class('baz', html), [])
+        self.assertEqual(get_elements_by_class('bam', html), [])
         self.assertEqual(get_elements_by_class('no-such-class', html), [])
 
     def test_get_elements_by_attribute(self):
diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index ac1e78002..a8a26ee02 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -1960,7 +1960,7 @@ def get_element_by_attribute(attribute, value, html, escape_value=True):
 def get_elements_by_class(class_name, html):
     """Return the content of all tags with the specified class in the passed HTML document as a list"""
     return get_elements_by_attribute(
-        'class', r'[^\'"]*\b%s\b[^\'"]*' % re.escape(class_name),
+        'class', r'[^\'"]*(?<!-)\b%s\b(?!-)[^\'"]*' % re.escape(class_name),
         html, escape_value=False)