Raise OpcError for a package part with invalid XML instead of a bare lxml error

src/docx/opc/oxml.py

@@ -15,6 +15,7 @@
 
 from docx.opc.constants import NAMESPACE as NS
 from docx.opc.constants import RELATIONSHIP_TARGET_MODE as RTM
+from docx.opc.exceptions import OpcError
 
 # configure XML parser
 element_class_lookup = etree.ElementNamespaceClassLookup()
@@ -35,7 +36,12 @@
 
 def parse_xml(text: str) -> etree._Element:
     """`etree.fromstring()` replacement that uses oxml parser."""
-    return etree.fromstring(text, oxml_parser)
+    try:
+        return etree.fromstring(text, oxml_parser)
+    except etree.XMLSyntaxError as exc:
+        # A malformed [Content_Types].xml or .rels part would otherwise raise a
+        # bare lxml XMLSyntaxError while opening the package.
+        raise OpcError("package contains invalid XML: %s" % exc) from exc
 
 
 def qn(tag: str) -> str:

src/docx/opc/part.py

@@ -6,6 +6,9 @@
 
 from typing import TYPE_CHECKING, Callable, Type, cast
 
+from lxml.etree import XMLSyntaxError
+
+from docx.opc.exceptions import OpcError
 from docx.opc.oxml import serialize_part_xml
 from docx.opc.packuri import PackURI
 from docx.opc.rel import Relationships
@@ -228,7 +231,14 @@ def element(self):
 
     @classmethod
     def load(cls, partname: PackURI, content_type: str, blob: bytes, package: Package):
-        element = parse_xml(blob)
+        try:
+            element = parse_xml(blob)
+        except XMLSyntaxError as exc:
+            # A part with malformed XML would otherwise raise a bare lxml
+            # XMLSyntaxError out of Document()/open; report it as a package error.
+            raise OpcError(
+                "part '%s' contains invalid XML: %s" % (partname, exc)
+            ) from exc
         return cls(partname, content_type, element, package)
 
     @property

tests/opc/test_part.py

@@ -345,6 +345,14 @@ def it_can_be_constructed_by_PartFactory(
         __init_.assert_called_once_with(ANY, partname_, content_type_, element_, package_)
         assert isinstance(part, XmlPart)
 
+    def it_raises_OpcError_on_a_part_with_invalid_xml(
+        self, partname_, content_type_, package_
+    ):
+        from docx.opc.exceptions import OpcError
+
+        with pytest.raises(OpcError):
+            XmlPart.load(partname_, content_type_, b"<not well formed", package_)
+
     def it_can_serialize_to_xml(self, blob_fixture):
         xml_part, element_, serialize_part_xml_ = blob_fixture
         blob = xml_part.blob