From: Waylan Limberg <waylan.limberg@icloud.com>
Date: Sat, 10 May 2025 17:36:06 +0000 (-0400)
Subject: gh-86155: Fix data loss after unclosed script or style tag in HTMLParser (GH-22658)
X-Git-Tag: v3.15.0a1~1771
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=53383e90e4df7029f792b7aa81aa2e4cff348ed0;p=thirdparty%2FPython%2Fcpython.git

gh-86155: Fix data loss after unclosed script or style tag in HTMLParser (GH-22658)

When calling .close() the HTMLParser should flush all remaining content,
even when that content is in an unclosed script or style tag.
---

diff --git a/Lib/html/parser.py b/Lib/html/parser.py
index 1b8b6ea0e5ab..1e30956fe24f 100644
--- a/Lib/html/parser.py
+++ b/Lib/html/parser.py
@@ -260,7 +260,7 @@ class HTMLParser(_markupbase.ParserBase):
             else:
                 assert 0, "interesting.search() lied"
         # end while
-        if end and i < n and not self.cdata_elem:
+        if end and i < n:
             if self.convert_charrefs and not self.cdata_elem:
                 self.handle_data(unescape(rawdata[i:n]))
             else:
diff --git a/Lib/test/test_htmlparser.py b/Lib/test/test_htmlparser.py
index 68649e9d6d5e..61fa24fab574 100644
--- a/Lib/test/test_htmlparser.py
+++ b/Lib/test/test_htmlparser.py
@@ -317,6 +317,16 @@ text
                                 ("endtag", element_lower)],
                             collector=Collector(convert_charrefs=False))
 
+    def test_EOF_in_cdata(self):
+        content = """<!-- not a comment --> &not-an-entity-ref;
+                  <a href="" /> </p><p> <span></span></style>
+                  '</script' + '>'"""
+        s = f'<script>{content}'
+        self._run_check(s, [
+            ("starttag", 'script', []),
+            ("data", content)
+        ])
+
     def test_comments(self):
         html = ("<!-- I'm a valid comment -->"
                 '<!--me too!-->'
diff --git a/Misc/NEWS.d/next/Library/2023-02-13-21-41-34.gh-issue-86155.ppIGSC.rst b/Misc/NEWS.d/next/Library/2023-02-13-21-41-34.gh-issue-86155.ppIGSC.rst
new file mode 100644
index 000000000000..bb85481b2296
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2023-02-13-21-41-34.gh-issue-86155.ppIGSC.rst
@@ -0,0 +1,2 @@
+:meth:`html.parser.HTMLParser.close` no longer loses data when the
+``<script>`` tag is not closed. Patch by Waylan Limberg.