Backport upstream PR#863 (“use only lxml for XML handling”)

- Fixes CVE-2023-27476 https://nvd.nist.gov/vuln/detail/CVE-2023-27476
- Fixes GHSA-8h9c-r582-mggc https://github.com/geopython/OWSLib/security/advisories/GHSA-8h9c-r582-mggc
- Fixes RHBZ#2176417 https://bugzilla.redhat.com/show_bug.cgi?id=2176417
This commit is contained in:
Benjamin A. Beasley 2023-03-08 09:09:29 -05:00
commit 3a8551e706
2 changed files with 244 additions and 1 deletions

View file

@ -0,0 +1,232 @@
From 2f81f2b36c01b2a32e8dcf6b2c695b36b1eb0b8f Mon Sep 17 00:00:00 2001
From: Tom Kralidis <tomkralidis@gmail.com>
Date: Fri, 24 Feb 2023 10:15:35 -0500
Subject: [PATCH] use only lxml for XML handling
Backport of PR#863 (https://github.com/geopython/OWSLib/pull/863) to 0.9.2
---
owslib/etree.py | 58 ++++++++---------------------
owslib/util.py | 96 ++++++++++++++++++------------------------------
requirements.txt | 1 +
3 files changed, 52 insertions(+), 103 deletions(-)
diff --git a/owslib/etree.py b/owslib/etree.py
index b8babe8..88ae7ab 100644
--- a/owslib/etree.py
+++ b/owslib/etree.py
@@ -6,9 +6,15 @@
from __future__ import (absolute_import, division, print_function)
import six
-import inspect
-def patch_well_known_namespaces(etree_module):
+from lxml import etree
+from lxml.etree import ParseError
+ElementType = etree._Element
+
+from owslib.namespaces import Namespaces
+
+def patch_well_known_namespaces():
+ """Monkey patches lxml.etree to add some well-known namespaces."""
import warnings
from owslib.namespaces import Namespaces
@@ -17,50 +23,18 @@ def patch_well_known_namespaces(etree_module):
"""Monkey patches the etree module to add some well-known namespaces."""
try:
- register_namespace = etree_module.register_namespace
+ register_namespace = etree.register_namespace
except AttributeError:
- try:
- etree_module._namespace_map
+ etree._namespace_map
- def register_namespace(prefix, uri):
- etree_module._namespace_map[uri] = prefix
- except AttributeError:
- def register_namespace(prefix, uri):
- pass
- warnings.warn("Only 'lxml.etree' >= 2.3 and 'xml.etree.ElementTree' >= 1.3 are fully supported!")
+ def register_namespace(prefix, uri):
+ etree._namespace_map[uri] = prefix
for k, v in six.iteritems(ns.get_namespaces()):
register_namespace(k, v)
-# try to find lxml or elementtree
-try:
- from lxml import etree
- from lxml.etree import ParseError
- ElementType = etree._Element
-except ImportError:
- try:
- # Python 2.x/3.x with ElementTree included
- import xml.etree.ElementTree as etree
-
- try:
- from xml.etree.ElementTree import ParseError
- except ImportError:
- from xml.parsers.expat import ExpatError as ParseError
-
- if hasattr(etree, 'Element') and inspect.isclass(etree.Element):
- # python 3.4, 3.3, 2.7
- ElementType = etree.Element
- else:
- # python 2.6
- ElementType = etree._ElementInterface
-
- except ImportError:
- try:
- # Python < 2.5 with ElementTree installed
- import elementtree.ElementTree as etree
- ParseError = StandardError # i can't find a ParseError related item in elementtree docs!
- ElementType = etree.Element
- except ImportError:
- raise RuntimeError('You need either lxml or ElementTree to use OWSLib!')
+ etree.set_default_parser(
+ parser=etree.XMLParser(resolve_entities=False)
+ )
-patch_well_known_namespaces(etree)
+patch_well_known_namespaces()
diff --git a/owslib/util.py b/owslib/util.py
index 22bcaa7..51ce6d0 100644
--- a/owslib/util.py
+++ b/owslib/util.py
@@ -239,11 +239,8 @@ def nspath_eval(xpath, namespaces):
def cleanup_namespaces(element):
""" Remove unused namespaces from an element """
- if etree.__name__ == 'lxml.etree':
- etree.cleanup_namespaces(element)
- return element
- else:
- return etree.fromstring(etree.tostring(element))
+ etree.cleanup_namespaces(element)
+ return element
def add_namespaces(root, ns_keys):
@@ -254,35 +251,34 @@ def add_namespaces(root, ns_keys):
ns_keys = [(x, namespaces.get_namespace(x)) for x in ns_keys]
- if etree.__name__ != 'lxml.etree':
- # We can just add more namespaces when not using lxml.
- # We can't re-add an existing namespaces. Get a list of current
- # namespaces in use
- existing_namespaces = set()
- for elem in root.getiterator():
- if elem.tag[0] == "{":
- uri, tag = elem.tag[1:].split("}")
- existing_namespaces.add(namespaces.get_namespace_from_url(uri))
- for key, link in ns_keys:
- if link is not None and key not in existing_namespaces:
- root.set("xmlns:%s" % key, link)
- return root
- else:
- # lxml does not support setting xmlns attributes
- # Update the elements nsmap with new namespaces
- new_map = root.nsmap
- for key, link in ns_keys:
- if link is not None:
- new_map[key] = link
- # Recreate the root element with updated nsmap
- new_root = etree.Element(root.tag, nsmap=new_map)
- # Carry over attributes
- for a, v in list(root.items()):
- new_root.set(a, v)
- # Carry over children
- for child in root:
- new_root.append(deepcopy(child))
- return new_root
+ # lxml does not support setting xmlns attributes
+ # Update the elements nsmap with new namespaces
+ new_map = root.nsmap
+ for key, link in ns_keys:
+ if link is not None:
+ new_map[key] = link
+ # Recreate the root element with updated nsmap
+ new_root = etree.Element(root.tag, nsmap=new_map)
+ # Carry over attributes
+ for a, v in list(root.items()):
+ new_root.set(a, v)
+ # Carry over children
+ for child in root:
+ new_root.append(deepcopy(child))
+ return new_root
+
+ # We can just add more namespaces when not using lxml.
+ # We can't re-add an existing namespaces. Get a list of current
+ # namespaces in use
+ existing_namespaces = set()
+ for elem in root.iter():
+ if elem.tag[0] == "{":
+ uri, tag = elem.tag[1:].split("}")
+ existing_namespaces.add(namespaces.get_namespace_from_url(uri))
+ for key, link in ns_keys:
+ if link is not None and key not in existing_namespaces:
+ root.set("xmlns:%s" % key, link)
+ return root
def getXMLInteger(elem, tag):
@@ -401,20 +397,13 @@ def element_to_string(element, encoding=None, xml_declaration=False):
if encoding is None:
encoding = "ISO-8859-1"
- if etree.__name__ == 'lxml.etree':
- if xml_declaration:
- if encoding in ['unicode', 'utf-8']:
- output = '<?xml version="1.0" encoding="utf-8" standalone="no"?>\n%s' % \
- etree.tostring(element, encoding='unicode')
- else:
- output = etree.tostring(element, encoding=encoding, xml_declaration=True)
+ if xml_declaration:
+ if encoding in ['unicode', 'utf-8']:
+ output = '<?xml version="1.0" encoding="utf-8" standalone="no"?>\n%s' % \
+ etree.tostring(element, encoding='unicode')
else:
- output = etree.tostring(element)
+ output = etree.tostring(element, encoding=encoding, xml_declaration=True)
else:
- if xml_declaration:
- output = '<?xml version="1.0" encoding="%s" standalone="no"?>\n%s' % (encoding,
- etree.tostring(element, encoding=encoding))
- else:
output = etree.tostring(element)
return output
@@ -581,18 +570,3 @@ try: # 2.7
from collections import OrderedDict
except: # 2.6
from ordereddict import OrderedDict
-
-
-def which_etree():
- """decipher which etree library is being used by OWSLib"""
-
- which_etree = None
-
- if 'lxml' in etree.__file__:
- which_etree = 'lxml.etree'
- elif 'xml/etree' in etree.__file__:
- which_etree = 'xml.etree'
- elif 'elementree' in etree.__file__:
- which_etree = 'elementtree.ElementTree'
-
- return which_etree
diff --git a/requirements.txt b/requirements.txt
index 7689a61..10ad5db 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,4 @@
+lxml
python-dateutil>=1.5
pytz
requests>=1.0
--
2.39.2

View file

@ -13,6 +13,13 @@ URL: http://geopython.github.io/OWSLib
%global forgeurl https://github.com/geopython/OWSLib
Source0: %{forgeurl}/archive/%{version}/OWSLib-%{version}.tar.gz
# use only lxml for XML handling
# Backport of PR#863 (https://github.com/geopython/OWSLib/pull/863) to 0.9.2
# Fixes: CVE-2023-27476 https://nvd.nist.gov/vuln/detail/CVE-2023-27476
# GHSA-8h9c-r582-mggc https://github.com/geopython/OWSLib/security/advisories/GHSA-8h9c-r582-mggc
# RHBZ#2176417 https://bugzilla.redhat.com/show_bug.cgi?id=2176417
Patch0: 0001-use-only-lxml-for-XML-handling.patch
BuildArch: noarch
%global common_description %{expand:
@ -35,11 +42,13 @@ BuildRequires: python2-dateutil >= 1.5
BuildRequires: python2-requests >= 1.0
BuildRequires: python2-pytz
BuildRequires: pyproj
BuildRequires: python-lxml
Requires: python2-dateutil >= 1.5
Requires: python2-requests >= 1.0
Requires: python2-pytz
Requires: python2-pyproj
Requires: python-lxml
%description -n python2-OWSLib %{common_description}
@ -58,18 +67,20 @@ BuildRequires: python%{python3_pkgversion}-dateutil >= 1.5
BuildRequires: python%{python3_pkgversion}-requests >= 1.0
BuildRequires: python%{python3_pkgversion}-pytz
BuildRequires: python%{python3_pkgversion}-pyproj
BuildRequires: python%{python3_pkgversion}-lxml
Requires: python%{python3_pkgversion}-dateutil >= 1.5
Requires: python%{python3_pkgversion}-requests >= 1.0
Requires: python%{python3_pkgversion}-pytz
Requires: python%{python3_pkgversion}-pyproj
Requires: python%{python3_pkgversion}-lxml
%description -n python%{python3_pkgversion}-OWSLib %{common_description}
%endif
%prep
%autosetup -n OWSLib-%{version}
%autosetup -n OWSLib-%{version} -p1
rm -rf OWSLib.egg-info
# Remove shebangs from non-script sources. The find-then-modify pattern