diff --git a/swh/indexer/tests/test_metadata.py b/swh/indexer/tests/test_metadata.py
--- a/swh/indexer/tests/test_metadata.py
+++ b/swh/indexer/tests/test_metadata.py
@@ -7,7 +7,6 @@
 import unittest
 
 from hypothesis import given, strategies
-import xmltodict
 
 from swh.model.hashutil import hash_to_bytes
 
@@ -23,7 +22,8 @@
 
 from .utils import (
     BASE_TEST_CONFIG, fill_obj_storage, fill_storage,
-    YARN_PARSER_METADATA, json_document_strategy
+    YARN_PARSER_METADATA, json_document_strategy,
+    xml_document_strategy,
 )
 
 
@@ -1069,11 +1069,12 @@
         raw = json.dumps(doc).encode()
         self.codemeta_mapping.translate(raw)
 
-    @given(json_document_strategy(
-        keys=list(MAPPINGS['MavenMapping'].mapping)))
+    @given(xml_document_strategy(
+        keys=list(MAPPINGS['MavenMapping'].mapping),
+        root='project',
+        xmlns='http://maven.apache.org/POM/4.0.0'))
     def test_maven_adversarial(self, doc):
-        raw = xmltodict.unparse({'project': doc}, pretty=True)
-        self.maven_mapping.translate(raw)
+        self.maven_mapping.translate(doc)
 
     @given(strategies.dictionaries(
         # keys
diff --git a/swh/indexer/tests/utils.py b/swh/indexer/tests/utils.py
--- a/swh/indexer/tests/utils.py
+++ b/swh/indexer/tests/utils.py
@@ -5,8 +5,10 @@
 
 import abc
 import datetime
+import functools
 import hashlib
 import random
+import unittest
 
 from hypothesis import strategies
 
@@ -401,6 +403,7 @@
     strategies.characters(),
     *map(strategies.just, ['type', 'url', 'name', 'email', '@id',
                            '@context', 'repository', 'license',
+                           'repositories', 'licenses'
                            ]),
 )
 """Hypothesis strategy that generates strings, with an emphasis on those
@@ -421,7 +424,7 @@
 
 def json_document_strategy(keys=None):
     """Generates an hypothesis strategy that generates metadata files
-    for a format that uses the given keys."""
+    for a JSON-based format that uses the given keys."""
     if keys is None:
         keys = strategies.characters()
     else:
@@ -430,6 +433,98 @@
     return strategies.dictionaries(keys, generic_json_document, min_size=1)
 
 
+def _tree_to_xml(root, xmlns, data):
+    def encode(s):
+        "Skips unpaired surrogates generated by json_document_strategy"
+        return s.encode('utf8', 'replace')
+
+    def to_xml(data, indent=b' '):
+        if data is None:
+            return b''
+        elif isinstance(data, (bool, str, int, float)):
+            return indent + encode(str(data))
+        elif isinstance(data, list):
+            return b'\n'.join(to_xml(v, indent=indent) for v in data)
+        elif isinstance(data, dict):
+            lines = []
+            for (key, value) in data.items():
+                lines.append(indent + encode('<{}>'.format(key)))
+                lines.append(to_xml(value, indent=indent+b' '))
+                lines.append(indent + encode('</{}>'.format(key)))
+            return b'\n'.join(lines)
+        else:
+            raise TypeError(data)
+
+    return b'\n'.join([
+        '<{} xmlns="{}">'.format(root, xmlns).encode(),
+        to_xml(data),
+        '</{}>'.format(root).encode(),
+    ])
+
+
+class TreeToXmlTest(unittest.TestCase):
+    def test_leaves(self):
+        self.assertEqual(
+            _tree_to_xml('root', 'http://example.com', None),
+            b'<root xmlns="http://example.com">\n\n</root>'
+        )
+        self.assertEqual(
+            _tree_to_xml('root', 'http://example.com', True),
+            b'<root xmlns="http://example.com">\n True\n</root>'
+        )
+        self.assertEqual(
+            _tree_to_xml('root', 'http://example.com', 'abc'),
+            b'<root xmlns="http://example.com">\n abc\n</root>'
+        )
+        self.assertEqual(
+            _tree_to_xml('root', 'http://example.com', 42),
+            b'<root xmlns="http://example.com">\n 42\n</root>'
+        )
+        self.assertEqual(
+            _tree_to_xml('root', 'http://example.com', 3.14),
+            b'<root xmlns="http://example.com">\n 3.14\n</root>'
+        )
+
+    def test_dict(self):
+        self.assertIn(
+            _tree_to_xml('root', 'http://example.com', {
+                'foo': 'bar',
+                'baz': 'qux'
+            }),
+            [
+                b'<root xmlns="http://example.com">\n'
+                b' <foo>\n  bar\n </foo>\n'
+                b' <baz>\n  qux\n </baz>\n'
+                b'</root>',
+                b'<root xmlns="http://example.com">\n'
+                b' <baz>\n  qux\n </baz>\n'
+                b' <foo>\n  bar\n </foo>\n'
+                b'</root>'
+            ]
+        )
+
+    def test_list(self):
+        self.assertEqual(
+            _tree_to_xml('root', 'http://example.com', [
+                {'foo': 'bar'},
+                {'foo': 'baz'},
+            ]),
+            b'<root xmlns="http://example.com">\n'
+            b' <foo>\n  bar\n </foo>\n'
+            b' <foo>\n  baz\n </foo>\n'
+            b'</root>'
+        )
+
+
+def xml_document_strategy(keys, root, xmlns):
+    """Generates an hypothesis strategy that generates metadata files
+    for an XML format that uses the given keys."""
+
+    return strategies.builds(
+        functools.partial(_tree_to_xml, root, xmlns),
+        json_document_strategy(keys))
+
+
 def filter_dict(d, keys):
     'return a copy of the dict with keys deleted'
     if not isinstance(keys, (list, tuple)):