[3.9] gh-135661: Fix CDATA section parsing in HTMLParser (GH-135665) (GH-137774) (GH-139661)

"] ]>" and "]] >" no longer end the CDATA section.

Make CDATA section parsing  context depending.
Add private method HTMLParser._set_support_cdata() to change the context.
If called with True, "<[CDATA[" starts a CDATA section which ends with "]]>".
If called with False, "<[CDATA[" starts a bogus comments which ends with ">".
(cherry picked from commit 0cbbfc4)
(cherry picked from commit dcf2476)

Co-authored-by: Serhiy Storchaka <[email protected]>
Co-authored-by: Łukasz Langa <[email protected]>

Author: 3 people

Lib/html/parser.py

@@ -127,6 +127,7 @@ def reset(self):
         self.lasttag = '???'
         self.interesting = interesting_normal
         self.cdata_elem = None
+        self._support_cdata = True
         self._escapable = True
         _markupbase.ParserBase.reset(self)
 
@@ -164,6 +165,19 @@ def clear_cdata_mode(self):
         self.cdata_elem = None
         self._escapable = True
 
+    def _set_support_cdata(self, flag=True):
+        """Enable or disable support of the CDATA sections.
+        If enabled, "<[CDATA[" starts a CDATA section which ends with "]]>".
+        If disabled, "<[CDATA[" starts a bogus comments which ends with ">".
+
+        This method is not called by default. Its purpose is to be called
+        in custom handle_starttag() and handle_endtag() methods, with
+        value that depends on the adjusted current node.
+        See https://html.spec.whatwg.org/multipage/parsing.html#markup-declaration-open-state
+        for details.
+        """
+        self._support_cdata = flag
+
     # Internal -- handle data as far as reasonable.  May leave state
     # and data to be processed by a subsequent call.  If 'end' is
     # true, force handling all data as if followed by EOF marker.
@@ -238,7 +252,7 @@ def goahead(self, end):
                                 j -= len(suffix)
                                 break
                         self.handle_comment(rawdata[i+4:j])
-                    elif startswith("<![CDATA[", i):
+                    elif startswith("<![CDATA[", i) and self._support_cdata:
                         self.unknown_decl(rawdata[i+3:])
                     elif rawdata[i:i+9].lower() == '<!doctype':
                         self.handle_decl(rawdata[i+2:])
@@ -314,15 +328,28 @@ def parse_html_declaration(self, i):
         if rawdata[i:i+4] == '<!--':
             # this case is actually already handled in goahead()
             return self.parse_comment(i)
-        elif rawdata[i:i+3] == '<![':
-            return self.parse_marked_section(i)
+        elif rawdata[i:i+9] == '<![CDATA[' and self._support_cdata:
+            j = rawdata.find(']]>', i+9)
+            if j < 0:
+                return -1
+            self.unknown_decl(rawdata[i+3: j])
+            return j + 3
         elif rawdata[i:i+9].lower() == '<!doctype':
             # find the closing >
             gtpos = rawdata.find('>', i+9)
             if gtpos == -1:
                 return -1
             self.handle_decl(rawdata[i+2:gtpos])
             return gtpos+1
+        elif rawdata[i:i+3] == '<![':
+            j = rawdata.find('>', i+3)
+            if j < 0:
+                return -1
+            if rawdata[j-1] == ']':
+                self.unknown_decl(rawdata[i+3: j-1])
+            else:
+                self.handle_comment(rawdata[i+2: j])
+            return j + 1
         else:
             return self.parse_bogus_comment(i)
 

Lib/test/test_htmlparser.py

@@ -9,10 +9,13 @@
 
 class EventCollector(html.parser.HTMLParser):
 
-    def __init__(self, *args, **kw):
+    def __init__(self, *args, autocdata=False, **kw):
+        self.autocdata = autocdata
         self.events = []
         self.append = self.events.append
         html.parser.HTMLParser.__init__(self, *args, **kw)
+        if autocdata:
+            self._set_support_cdata(False)
 
     def get_events(self):
         # Normalize the list of events so that buffer artefacts don't
@@ -33,12 +36,16 @@ def get_events(self):
 
     def handle_starttag(self, tag, attrs):
         self.append(("starttag", tag, attrs))
+        if self.autocdata and tag == 'svg':
+            self._set_support_cdata(True)
 
     def handle_startendtag(self, tag, attrs):
         self.append(("startendtag", tag, attrs))
 
     def handle_endtag(self, tag):
         self.append(("endtag", tag))
+        if self.autocdata and tag == 'svg':
+            self._set_support_cdata(False)
 
     # all other markup
 
@@ -739,10 +746,6 @@ def test_eof_in_declarations(self):
             ('<!', [('comment', '')]),
             ('<!-', [('comment', '-')]),
             ('<![', [('comment', '[')]),
-            ('<![CDATA[', [('unknown decl', 'CDATA[')]),
-            ('<![CDATA[x', [('unknown decl', 'CDATA[x')]),
-            ('<![CDATA[x]', [('unknown decl', 'CDATA[x]')]),
-            ('<![CDATA[x]]', [('unknown decl', 'CDATA[x]]')]),
             ('<!DOCTYPE', [('decl', 'DOCTYPE')]),
             ('<!DOCTYPE ', [('decl', 'DOCTYPE ')]),
             ('<!DOCTYPE html', [('decl', 'DOCTYPE html')]),
@@ -755,6 +758,18 @@ def test_eof_in_declarations(self):
         for html, expected in data:
             self._run_check(html, expected)
 
+    @support.subTests('content', ['', 'x', 'x]', 'x]]'])
+    def test_eof_in_cdata(self, content):
+        self._run_check('<![CDATA[' + content,
+                        [('unknown decl', 'CDATA[' + content)])
+        self._run_check('<![CDATA[' + content,
+                        [('comment', '[CDATA[' + content)],
+                        collector=EventCollector(autocdata=True))
+        self._run_check('<svg><text y="100"><![CDATA[' + content,
+                        [('starttag', 'svg', []),
+                         ('starttag', 'text', [('y', '100')]),
+                         ('unknown decl', 'CDATA[' + content)])
+
     def test_bogus_comments(self):
         html = ('<!ELEMENT br EMPTY>'
                 '<! not really a comment >'
@@ -804,7 +819,56 @@ def test_broken_condcoms(self):
             ('startendtag', 'img', [('src', 'mammoth.bmp')]),
             ('unknown decl', 'endif')
         ]
+
+        self._run_check(html, expected)
+
+    @support.subTests('content', [
+        'just some plain text',
+        '<!-- not a comment -->',
+        '&not-an-entity-ref;',
+        "<not a='start tag'>",
+        '',
+        '[[I have many brackets]]',
+        'I have a > in the middle',
+        'I have a ]] in the middle',
+        '] ]>',
+        ']] >',
+        ('\n'
+         '    if (a < b && a > b) {\n'
+         '        printf("[<marquee>How?</marquee>]");\n'
+         '    }\n'),
+    ])
+    def test_cdata_section_content(self, content):
+        # See "13.2.5.42 Markup declaration open state",
+        # "13.2.5.69 CDATA section state", and issue bpo-32876.
+        html = f'<svg><text y="100"><![CDATA[{content}]]></text></svg>'
+        expected = [
+            ('starttag', 'svg', []),
+            ('starttag', 'text', [('y', '100')]),
+            ('unknown decl', 'CDATA[' + content),
+            ('endtag', 'text'),
+            ('endtag', 'svg'),
+        ]
         self._run_check(html, expected)
+        self._run_check(html, expected, collector=EventCollector(autocdata=True))
+
+    def test_cdata_section(self):
+        # See "13.2.5.42 Markup declaration open state".
+        html = ('<![CDATA[foo<br>bar]]>'
+                '<svg><text y="100"><![CDATA[foo<br>bar]]></text></svg>'
+                '<![CDATA[foo<br>bar]]>')
+        expected = [
+            ('comment', '[CDATA[foo<br'),
+            ('data', 'bar]]>'),
+            ('starttag', 'svg', []),
+            ('starttag', 'text', [('y', '100')]),
+            ('unknown decl', 'CDATA[foo<br>bar'),
+            ('endtag', 'text'),
+            ('endtag', 'svg'),
+            ('comment', '[CDATA[foo<br'),
+            ('data', 'bar]]>'),
+        ]
+        self._run_check(html, expected, collector=EventCollector(autocdata=True))
 
     def test_convert_charrefs_dropped_text(self):
         # #23144: make sure that all the events are triggered when
@@ -1041,27 +1105,6 @@ def test_weird_chars_in_unquoted_attribute_values(self):
                             ('starttag', 'form',
                                 [('action', 'bogus|&#()value')])])
 
-    def test_invalid_keyword_error_exception(self):
-        # bpo-34480: check that subclasses that define an
-        # error method that raises an exception work
-        class InvalidMarkupException(Exception):
-            pass
-        class MyHTMLParser(html.parser.HTMLParser):
-            def error(self, message):
-                raise InvalidMarkupException(message)
-        parser = MyHTMLParser()
-        with self.assertRaises(InvalidMarkupException):
-            parser.feed('<![invalid>')
-
-    def test_invalid_keyword_error_pass(self):
-        # bpo-34480: check that subclasses that define an
-        # error method that doesn't raise an exception work
-        class MyHTMLParser(html.parser.HTMLParser):
-            def error(self, message):
-                pass
-        parser = MyHTMLParser()
-        self.assertEqual(parser.feed('<![invalid>'), None)
-
 
 if __name__ == "__main__":
     unittest.main()

Misc/NEWS.d/next/Security/2025-06-18-13-34-55.gh-issue-135661.NZlpWf.rst

@@ -0,0 +1,5 @@
+Fix CDATA section parsing in :class:`html.parser.HTMLParser` according to
+the HTML5 standard: ``] ]>`` and ``]] >`` no longer end the CDATA section.
+Add private method ``_set_support_cdata()`` which can be used to specify
+how to parse ``<[CDATA[`` --- as a CDATA section in foreign content
+(SVG or MathML) or as a bogus comment in the HTML namespace.