diff --git a/SOURCES/CVE-2021-43818.patch b/SOURCES/CVE-2021-43818.patch
new file mode 100644
index 0000000..7413abc
--- /dev/null
+++ b/SOURCES/CVE-2021-43818.patch
@@ -0,0 +1,127 @@
+diff --git a/src/lxml/html/clean.py b/src/lxml/html/clean.py
+index 6f3f7de..da5af16 100644
+--- a/src/lxml/html/clean.py
++++ b/src/lxml/html/clean.py
+@@ -73,18 +73,25 @@ _looks_like_tag_content = re.compile(
+
+ # All kinds of schemes besides just javascript: that can cause
+ # execution:
+-_is_image_dataurl = re.compile(
+- r'^data:image/.+;base64', re.I).search
++_find_image_dataurls = re.compile(
++ r'^data:image/(.+);base64,', re.I).findall
+ _is_possibly_malicious_scheme = re.compile(
+- r'(?:javascript|jscript|livescript|vbscript|data|about|mocha):',
+- re.I).search
++ r'(javascript|jscript|livescript|vbscript|data|about|mocha):',
++ re.I).findall
++# SVG images can contain script content
++_is_unsafe_image_type = re.compile(r"(xml|svg)", re.I).findall
++
+ def _is_javascript_scheme(s):
+- if _is_image_dataurl(s):
+- return None
+- return _is_possibly_malicious_scheme(s)
++ is_image_url = False
++ for image_type in _find_image_dataurls(s):
++ is_image_url = True
++ if _is_unsafe_image_type(image_type):
++ return True
++ if is_image_url:
++ return False
++ return bool(_is_possibly_malicious_scheme(s))
+
+ _substitute_whitespace = re.compile(r'[\s\x00-\x08\x0B\x0C\x0E-\x19]+').sub
+-# FIXME: should data: be blocked?
+
+ # FIXME: check against: http://msdn2.microsoft.com/en-us/library/ms537512.aspx
+ _conditional_comment_re = re.compile(
+@@ -512,6 +519,8 @@ class Cleaner(object):
+ return True
+ if 'expression(' in style:
+ return True
++ if '@import' in style:
++ return True
+ if '',
+ lxml.html.tostring(clean_html(s)))
+
++ def test_sneaky_import_in_style(self):
++ # Prevent "@@importimport" -> "@import" replacement.
++ style_codes = [
++ "@@importimport(extstyle.css)",
++ "@ @ import import(extstyle.css)",
++ "@ @ importimport(extstyle.css)",
++ "@@ import import(extstyle.css)",
++ "@ @import import(extstyle.css)",
++ "@@importimport()",
++ ]
++ for style_code in style_codes:
++ html = '' % style_code
++ s = lxml.html.fragment_fromstring(html)
++
++ cleaned = lxml.html.tostring(clean_html(s))
++ self.assertEqual(
++ b'',
++ cleaned,
++ "%s -> %s" % (style_code, cleaned))
++
++ def test_svg_data_links(self):
++ # Remove SVG images with potentially insecure content.
++ svg = b''
++ svgz = gzip.compress(svg)
++ svg_b64 = base64.b64encode(svg).decode('ASCII')
++ svgz_b64 = base64.b64encode(svgz).decode('ASCII')
++ urls = [
++ "data:image/svg+xml;base64," + svg_b64,
++ "data:image/svg+xml-compressed;base64," + svgz_b64,
++ ]
++ for url in urls:
++ html = '' % url
++ s = lxml.html.fragment_fromstring(html)
++
++ cleaned = lxml.html.tostring(clean_html(s))
++ self.assertEqual(
++ b'',
++ cleaned,
++ "%s -> %s" % (url, cleaned))
++
++ def test_image_data_links(self):
++ data = b'123'
++ data_b64 = base64.b64encode(data).decode('ASCII')
++ urls = [
++ "data:image/jpeg;base64," + data_b64,
++ "data:image/apng;base64," + data_b64,
++ "data:image/png;base64," + data_b64,
++ "data:image/gif;base64," + data_b64,
++ "data:image/webp;base64," + data_b64,
++ "data:image/bmp;base64," + data_b64,
++ "data:image/tiff;base64," + data_b64,
++ "data:image/x-icon;base64," + data_b64,
++ ]
++ for url in urls:
++ html = '' % url
++ s = lxml.html.fragment_fromstring(html)
++
++ cleaned = lxml.html.tostring(clean_html(s))
++ self.assertEqual(
++ html.encode("UTF-8"),
++ cleaned,
++ "%s -> %s" % (url, cleaned))
++
+ def test_formaction_attribute_in_button_input(self):
+ # The formaction attribute overrides the form's action and should be
+ # treated as a malicious link attribute
diff --git a/SOURCES/fix-threading-tests.patch b/SOURCES/fix-threading-tests.patch
new file mode 100644
index 0000000..27a5052
--- /dev/null
+++ b/SOURCES/fix-threading-tests.patch
@@ -0,0 +1,26 @@
+diff --git a/src/lxml/tests/test_threading.py b/src/lxml/tests/test_threading.py
+index 8948c3e..5ede3f8 100644
+--- a/src/lxml/tests/test_threading.py
++++ b/src/lxml/tests/test_threading.py
+@@ -130,7 +130,7 @@ class ThreadingTestCase(HelperTestCase):
+
+
+ ''' + '\n'.join('' % i for i in range(200)) + '''
+-
++
+ ''')
+ self.assertRaises(etree.XSLTParseError,
+ etree.XSLT, style)
+@@ -153,9 +153,10 @@ class ThreadingTestCase(HelperTestCase):
+ self.assertTrue(len(log))
+ if last_log is not None:
+ self.assertEqual(len(last_log), len(log))
+- self.assertEqual(4, len(log))
++ self.assertTrue(len(log) >= 2, len(log))
+ for error in log:
+- self.assertTrue(':ERROR:XSLT:' in str(error))
++ self.assertTrue(':ERROR:XSLT:' in str(error), str(error))
++ self.assertTrue(any('UnExpectedElement' in str(error) for error in log), log)
+ last_log = log
+
+ def test_thread_xslt_apply_error_log(self):
diff --git a/SPECS/python-lxml.spec b/SPECS/python-lxml.spec
index cc5da76..3247621 100644
--- a/SPECS/python-lxml.spec
+++ b/SPECS/python-lxml.spec
@@ -1,15 +1,8 @@
-%if 0%{?rhel} > 7
-# Disable python2 build by default
-%bcond_with python2
-%else
-%bcond_without python2
-%endif
-
%global modname lxml
Name: python-%{modname}
Version: 4.2.3
-Release: 3%{?dist}
+Release: 4%{?dist}
Summary: XML processing library combining libxml2/libxslt with the ElementTree API
License: BSD
@@ -28,6 +21,17 @@ Patch0: CVE-2020-27783.patch
# Fixed upstream: https://github.com/lxml/lxml/commit/2d01a1ba8984e0483ce6619b972832377f208a0d
Patch1: CVE-2021-28957.patch
+# Fix for CVE-2021-43818: HTML Cleaner allows crafted
+# and SVG embedded scripts to pass through
+# Fixed upstream:
+# https://github.com/lxml/lxml/commit/12fa9669007180a7bb87d990c375cf91ca5b664a
+# https://github.com/lxml/lxml/commit/f2330237440df7e8f39c3ad1b1aa8852be3b27c0
+Patch2: CVE-2021-43818.patch
+
+# Make test more resilient against changes in latest libxslt releases
+# Fixed upstream: https://github.com/lxml/lxml/commit/acef361ca80ff9afd828d91c98ea91c92f9d09af
+Patch3: fix-threading-tests.patch
+
BuildRequires: gcc
BuildRequires: libxml2-devel
BuildRequires: libxslt-devel
@@ -41,22 +45,6 @@ home page < or see our bug tracker at case you want to use the current ...
%description %{_description}
-%if %{with python2}
-%package -n python2-%{modname}
-Summary: %{summary}
-BuildRequires: python2-devel
-BuildRequires: python2-setuptools
-BuildRequires: python2-Cython
-Recommends: python2-cssselect
-Recommends: python2-html5lib
-Recommends: python2-beautifulsoup4
-%{?python_provide:%python_provide python2-%{modname}}
-
-%description -n python2-%{modname} %{_description}
-
-Python 2 version.
-%endif # with python2
-
%package -n python3-%{modname}
Summary: %{summary}
BuildRequires: python3-devel
@@ -76,30 +64,17 @@ Python 3 version.
%build
export WITH_CYTHON=true
-%if %{with python2}
-%py2_build
-%endif # with python2
%py3_build
%install
-%if %{with python2}
-%py2_install
-%endif # with python2
%py3_install
%check
-%if %{with python2}
-%{__python2} setup.py test
-%endif # with python2
-%{__python3} setup.py test
-
-%if %{with python2}
-%files -n python2-%{modname}
-%license doc/licenses/ZopePublicLicense.txt LICENSES.txt
-%doc README.rst src/lxml/isoschematron/resources/xsl/iso-schematron-xslt1/readme.txt
-%{python2_sitearch}/%{modname}/
-%{python2_sitearch}/%{modname}-*.egg-info/
-%endif # with python2
+# The tests assume inplace build, so we copy the built library to source-dir.
+# If not done that, Python can either import the tests or the extension modules, but not both.
+cp -a build/lib.%{python3_platform}-%{python3_version}/* src/
+# The options are: verbose, unit, functional
+%{python3} test.py -vuf
%files -n python3-%{modname}
%license doc/licenses/ZopePublicLicense.txt LICENSES.txt
@@ -108,6 +83,10 @@ export WITH_CYTHON=true
%{python3_sitearch}/%{modname}-*.egg-info/
%changelog
+* Thu Jan 06 2022 Charalampos Stratakis - 4.2.3-4
+- Security fix for CVE-2021-43818
+Resolves: rhbz#2032569
+
* Wed Mar 24 2021 Charalampos Stratakis - 4.2.3-3
- Security fix for CVE-2021-28957
Resolves: rhbz#1941534