From 8ee9fbd03bbf351ac0cfde7f67828d78c6c4dedb Mon Sep 17 00:00:00 2001
From: Will Kahn-Greene <willkg@mozilla.com>
Date: Fri, 25 Oct 2024 14:05:48 -0400
Subject: [PATCH] Fix additional < followed by characters and EOF issues (#728)

This fixes these two cases:

* "<some thing thing" where "thing" is repeated twice which kicks up a
  parser error because it thinks it's a duplicated attribute
* "<some thing thing2 " where the space at the end causes a
  expected-end-of-tag-but-got-eof parser error to pop up

In both of these cases, we want the data to be treated as character
data--not a tag.
---
 bleach/html5lib_shim.py | 15 ++++++++++++---
 tests/test_clean.py     |  4 ++++
 2 files changed, 16 insertions(+), 3 deletions(-)

diff --git a/bleach/html5lib_shim.py b/bleach/html5lib_shim.py
index ca1cc8c8..f083db75 100644
--- a/bleach/html5lib_shim.py
+++ b/bleach/html5lib_shim.py
@@ -396,16 +396,25 @@ def __iter__(self):
                 # name that abruptly ends, but we should treat that like
                 # character data
                 yield {"type": TAG_TOKEN_TYPE_CHARACTERS, "data": self.stream.get_tag()}
+
             elif last_error_token["data"] in (
+                "duplicate-attribute",
                 "eof-in-attribute-name",
                 "eof-in-attribute-value-no-quotes",
+                "expected-end-of-tag-but-got-eof",
             ):
                 # Handle the case where the text being parsed ends with <
-                # followed by a series of characters and then space and then
-                # more characters. It's treated as a tag name followed by an
+                # followed by characters and then space and then:
+                #
+                # * more characters
+                # * more characters repeated with a space between (e.g. "abc abc")
+                # * more characters and then a space and then an EOF (e.g. "abc def ")
+                #
+                # These cases are treated as a tag name followed by an
                 # attribute that abruptly ends, but we should treat that like
-                # character data.
+                # character data instead.
                 yield {"type": TAG_TOKEN_TYPE_CHARACTERS, "data": self.stream.get_tag()}
+
             else:
                 yield last_error_token
 
diff --git a/tests/test_clean.py b/tests/test_clean.py
index dc129d0e..da4082a8 100644
--- a/tests/test_clean.py
+++ b/tests/test_clean.py
@@ -167,6 +167,10 @@ def test_bare_entities_get_escaped_correctly(text, expected):
         ("<some thing", "&lt;some thing"),
         # this is an eof-in-attribute-value-no-quotes parser error
         ("<some thing=foo", "&lt;some thing=foo"),
+        # this is a duplicate-attribute parser error
+        ("<some thing thing", "&lt;some thing thing"),
+        # this is an expected-end-of-tag-but-got-eof parser error
+        ("<some thing thing2 ", "&lt;some thing thing2 "),
     ],
 )
 def test_lessthan_escaping(text, expected):