Avoid breaking multiprocessing with bozo_exception

docs/character-encoding.rst

@@ -99,6 +99,13 @@ sets the ``bozo`` bit to ``1`` and sets ``bozo_exception`` to
 ``feedparser.CharacterEncodingUnknown``.  In this case, parsed values will be
 strings, not Unicode strings.
 
+.. note::
+
+    When using multiprocessing, Python exceptions cannot always be safely
+    serialized between processes. In such cases, ``bozo_exception`` will
+    contain a string representation of the exception instead of the exception
+    object itself.
+
 
 Handling Incorrectly-Declared Media Types
 -----------------------------------------

feedparser/api.py

@@ -27,6 +27,7 @@
 # POSSIBILITY OF SUCH DAMAGE.
 
 import io
+import multiprocessing
 import urllib.error
 import urllib.parse
 import xml.sax
@@ -374,3 +375,10 @@ def _parse_file_inplace(
         result["namespaces"] = {}
     else:
         result["namespaces"] = feed_parser.namespaces_in_use
+
+    if "bozo_exception" in result and _is_multiprocessing():
+        result["bozo_exception"] = str(result["bozo_exception"])
+
+
+def _is_multiprocessing():
+    return multiprocessing.parent_process() is not None

tests/test_multiprocessing.py

@@ -0,0 +1,72 @@
+from concurrent.futures import ProcessPoolExecutor
+
+import pytest
+
+import feedparser
+
+base_feed_str = (
+    b"<?xml version='1.0' encoding='UTF-8'?>\n"
+    b'<rss version="2.0">\n'
+    b"<channel>\n"
+    b"<title>Foo</title>\n"
+    b"<link>https://foo.com/</link>\n"
+    b"<item>"
+    b"<title>Title 1</title>"
+    b"<link>https://foo.com/1</link>"
+    b"<pubDate>Thu, 05 Jun 2025 18:27:58 -0000</pubDate>"
+    b"</item>\n"
+    b"</channel>\n"
+    b"</rss>\n"
+)
+
+
+def _parse_and_return_full(raw_feed: bytes):
+    return feedparser.parse(raw_feed)
+
+
+@pytest.mark.parametrize(
+    "feed, expected_title, expected_bozo_exception, expected_items",
+    [
+        (
+            base_feed_str,
+            "Foo",
+            None,
+            [
+                dict(
+                    title="Title 1",
+                    link="https://foo.com/1",
+                    published="Thu, 05 Jun 2025 18:27:58 -0000",
+                )
+            ],
+        ),
+        (
+            b"\n" + base_feed_str,
+            "Foo",
+            "XML or text declaration not at start of entity",
+            [
+                dict(
+                    title="Title 1",
+                    link="https://foo.com/1",
+                    published="Thu, 05 Jun 2025 18:27:58 -0000",
+                )
+            ],
+        ),
+    ],
+    ids=["correct_feed", "leading_newline_feed"],
+)
+def test_multiprocessing_parse(
+    feed, expected_title, expected_bozo_exception, expected_items
+):
+    with ProcessPoolExecutor(1) as pool:
+        future = pool.submit(_parse_and_return_full, feed)
+        result = future.result()
+
+    assert result["feed"]["title"] == expected_title
+    if expected_bozo_exception:
+        assert expected_bozo_exception in result.get("bozo_exception")
+    else:
+        assert result.get("bozo_exception") is None
+    for observed, expected in zip(result["entries"], expected_items, strict=True):
+        assert observed["published"] == expected["published"]
+        assert observed["link"] == expected["link"]
+        assert observed["title"] == expected["title"]