Merge 8057c858ba into e8d49b1c7f

test/test_utils.py

@@ -1787,6 +1787,9 @@ Line 1
     GET_ELEMENT_BY_ATTRIBUTE_TEST_STRING = '''
         <div itemprop="author" itemscope>foo</div>
     '''
+    VOID_ELEMENT_TEST_STRING = '''
+        <img alt="foo" src="bar.png"><img alt="foobar" src="baz.jpg"><img alt="foo"/>
+    '''
 
     def test_get_element_by_attribute(self):
         html = self.GET_ELEMENT_BY_CLASS_TEST_STRING
@@ -1799,6 +1802,10 @@ Line 1
 
         self.assertEqual(get_element_by_attribute('itemprop', 'author', html), 'foo')
 
+        html = self.VOID_ELEMENT_TEST_STRING
+
+        self.assertEqual(get_element_by_attribute('alt', 'foo', html), '')
+
     def test_get_element_html_by_attribute(self):
         html = self.GET_ELEMENT_BY_CLASS_TEST_STRING
 
@@ -1810,6 +1817,10 @@ Line 1
 
         self.assertEqual(get_element_html_by_attribute('itemprop', 'author', html), html.strip())
 
+        html = self.VOID_ELEMENT_TEST_STRING
+
+        self.assertEqual(get_element_html_by_attribute('alt', 'foo', html), '<img alt="foo" src="bar.png">')
+
     GET_ELEMENTS_BY_CLASS_TEST_STRING = '''
         <span class="foo bar">nice</span><span class="foo bar">also nice</span>
     '''
@@ -1834,6 +1845,10 @@ Line 1
         self.assertEqual(get_elements_by_attribute('class', 'foo', html), [])
         self.assertEqual(get_elements_by_attribute('class', 'no-such-foo', html), [])
 
+        html = self.VOID_ELEMENT_TEST_STRING
+
+        self.assertEqual(get_elements_by_attribute('alt', 'foo', html), ['', ''])
+
     def test_get_elements_html_by_attribute(self):
         html = self.GET_ELEMENTS_BY_CLASS_TEST_STRING
 
@@ -1841,6 +1856,11 @@ Line 1
         self.assertEqual(get_elements_html_by_attribute('class', 'foo', html), [])
         self.assertEqual(get_elements_html_by_attribute('class', 'no-such-foo', html), [])
 
+        html = self.VOID_ELEMENT_TEST_STRING
+
+        self.assertEqual(get_elements_html_by_attribute(
+            'alt', 'foo', html), ['<img alt="foo" src="bar.png">', '<img alt="foo"/>'])
+
     def test_get_elements_text_and_html_by_attribute(self):
         html = self.GET_ELEMENTS_BY_CLASS_TEST_STRING
 
@@ -1853,6 +1873,11 @@ Line 1
         self.assertEqual(list(get_elements_text_and_html_by_attribute(
             'class', 'foo', '<a class="foo">nice</a><span class="foo">nice</span>', tag='a')), [('nice', '<a class="foo">nice</a>')])
 
+        html = self.VOID_ELEMENT_TEST_STRING
+
+        self.assertEqual(list(get_elements_text_and_html_by_attribute(
+            'alt', 'foo', html, tag='img')), [('', '<img alt="foo" src="bar.png">'), ('', '<img alt="foo"/>')])
+
     GET_ELEMENT_BY_TAG_TEST_STRING = '''
     random text lorem ipsum</p>
     <div>
@@ -1881,6 +1906,10 @@ Line 1
             (self.GET_ELEMENT_BY_TAG_RES_INNERSPAN_TEXT, self.GET_ELEMENT_BY_TAG_RES_INNERSPAN_HTML))
         self.assertRaises(compat_HTMLParseError, get_element_text_and_html_by_tag, 'article', html)
 
+        html = self.VOID_ELEMENT_TEST_STRING
+
+        self.assertEqual(get_element_text_and_html_by_tag('img', html), ('', '<img alt="foo" src="bar.png">'))
+
     def test_iri_to_uri(self):
         self.assertEqual(
             iri_to_uri('https://www.google.com/search?q=foo&ie=utf-8&oe=utf-8&client=firefox-b'),

yt_dlp/utils/_utils.py

@@ -167,6 +167,12 @@ JSON_LD_RE = r'(?is)<script[^>]+type=(["\']?)application/ld\+json\1[^>]*>\s*(?P<
 
 NUMBER_RE = r'\d+(?:\.\d+)?'
 
+VOID_ELEMENTS = [
+    'area', 'base', 'br', 'col', 'embed',
+    'hr', 'img', 'input', 'link', 'meta',
+    'param', 'source', 'track', 'wbr',
+]
+
 
 @functools.cache
 def preferredencoding():
@@ -364,15 +370,13 @@ def get_elements_text_and_html_by_attribute(attribute, value, html, *, tag=r'[\w
     if not value:
         return
 
-    quote = '' if re.match(r'''[\s"'`=<>]''', value) else '?'
-
     value = re.escape(value) if escape_value else value
 
     partial_element_re = rf'''(?x)
         <(?P<tag>{tag})
-         (?:\s(?:[^>"']|"[^"]*"|'[^']*')*)?
-         \s{re.escape(attribute)}\s*=\s*(?P<_q>['"]{quote})(?-x:{value})(?P=_q)
-        '''
+        (?:\s[^>"']*|"[^"]*"|'[^']*')*?
+        \s{re.escape(attribute)}\s*=\s*(?P<q>['"])?(?-x:{value})(?(q)(?P=q)|(?=[\s/>]))
+    '''
 
     for m in re.finditer(partial_element_re, html):
         content, whole = get_element_text_and_html_by_tag(m.group('tag'), html[m.start():])
@@ -436,12 +440,17 @@ def get_element_text_and_html_by_tag(tag, html):
             return haystack.index(needle)
         except ValueError:
             raise exc
-    closing_tag = f'</{tag}>'
+
     whole_start = find_or_raise(
         html, f'<{tag}', compat_HTMLParseError(f'opening {tag} tag not found'))
     content_start = find_or_raise(
         html[whole_start:], '>', compat_HTMLParseError(f'malformed opening {tag} tag'))
     content_start += whole_start + 1
+
+    if tag in VOID_ELEMENTS:
+        return '', html[whole_start:content_start]
+
+    closing_tag = f'</{tag}>'
     with HTMLBreakOnClosingTagParser() as parser:
         parser.feed(html[whole_start:content_start])
         if not parser.tagstack or parser.tagstack[0] != tag: