dev-python/pymongo: Backport upstream changes to fix crash

Signed-off-by: Michał Górny <mgorny@gentoo.org>
2026-09-24 04:59:14 -07:00 · 2025-10-06 20:11:03 +02:00
parent 3fed013aab
commit 0c322fe28a
2 changed files with 419 additions and 0 deletions
--- a/dev-python/pymongo/files/pymongo-4.15.2-backports.patch
+++ b/dev-python/pymongo/files/pymongo-4.15.2-backports.patch
@@ -0,0 +1,212 @@
+diff --git a/bson/__init__.py b/bson/__init__.py
+index b655e30c2c..6b2ba293a6 100644
+--- a/bson/__init__.py
+++ b/bson/__init__.py
+@@ -1009,7 +1009,7 @@ def _dict_to_bson(
+                 try:
+                     elements.append(_element_to_bson(key, value, check_keys, opts))
+                 except InvalidDocument as err:
+-                    raise InvalidDocument(f"Invalid document {doc} | {err}") from err
+                    raise InvalidDocument(f"Invalid document: {err}", doc) from err
+     except AttributeError:
+         raise TypeError(f"encoder expected a mapping type but got: {doc!r}") from None
+ 
+diff --git a/bson/_cbsonmodule.c b/bson/_cbsonmodule.c
+index be91e41734..bee7198567 100644
+--- a/bson/_cbsonmodule.c
+++ b/bson/_cbsonmodule.c
+@@ -1645,11 +1645,11 @@ static int write_raw_doc(buffer_t buffer, PyObject* raw, PyObject* _raw_str) {
+ }
+ 
+ 
+-/* Update Invalid Document error message to include doc.
+/* Update Invalid Document error to include doc as a property.
+  */
+ void handle_invalid_doc_error(PyObject* dict) {
+     PyObject *etype = NULL, *evalue = NULL, *etrace = NULL;
+-    PyObject *msg = NULL, *dict_str = NULL, *new_msg = NULL;
+    PyObject *msg = NULL, *new_msg = NULL, *new_evalue = NULL;
+     PyErr_Fetch(&etype, &evalue, &etrace);
+     PyObject *InvalidDocument = _error("InvalidDocument");
+     if (InvalidDocument == NULL) {
+@@ -1659,26 +1659,22 @@ void handle_invalid_doc_error(PyObject* dict) {
+     if (evalue && PyErr_GivenExceptionMatches(etype, InvalidDocument)) {
+         PyObject *msg = PyObject_Str(evalue);
+         if (msg) {
+-            // Prepend doc to the existing message
+-            PyObject *dict_str = PyObject_Str(dict);
+-            if (dict_str == NULL) {
+-                goto cleanup;
+-            }
+-            const char * dict_str_utf8 = PyUnicode_AsUTF8(dict_str);
+-            if (dict_str_utf8 == NULL) {
+-                goto cleanup;
+-            }
+             const char * msg_utf8 = PyUnicode_AsUTF8(msg);
+             if (msg_utf8 == NULL) {
+                 goto cleanup;
+             }
+-            PyObject *new_msg = PyUnicode_FromFormat("Invalid document %s | %s", dict_str_utf8, msg_utf8);
+            PyObject *new_msg = PyUnicode_FromFormat("Invalid document: %s", msg_utf8);
+            if (new_msg == NULL) {
+                goto cleanup;
+            }
+            // Add doc to the error instance as a property.
+            PyObject *new_evalue = PyObject_CallFunctionObjArgs(InvalidDocument, new_msg, dict, NULL);
+             Py_DECREF(evalue);
+             Py_DECREF(etype);
+             etype = InvalidDocument;
+             InvalidDocument = NULL;
+-            if (new_msg) {
+-                evalue = new_msg;
+            if (new_evalue) {
+                evalue = new_evalue;
+             } else {
+                 evalue = msg;
+             }
+@@ -1689,7 +1685,7 @@ void handle_invalid_doc_error(PyObject* dict) {
+     PyErr_Restore(etype, evalue, etrace);
+     Py_XDECREF(msg);
+     Py_XDECREF(InvalidDocument);
+-    Py_XDECREF(dict_str);
+    Py_XDECREF(new_evalue);
+     Py_XDECREF(new_msg);
+ }
+ 
+diff --git a/bson/errors.py b/bson/errors.py
+index a3699e704c..ffc117f7ac 100644
+--- a/bson/errors.py
+++ b/bson/errors.py
+@@ -15,6 +15,8 @@
+ """Exceptions raised by the BSON package."""
+ from __future__ import annotations
+ 
+from typing import Any, Optional
+
+ 
+ class BSONError(Exception):
+     """Base class for all BSON exceptions."""
+@@ -31,6 +33,17 @@ class InvalidStringData(BSONError):
+ class InvalidDocument(BSONError):
+     """Raised when trying to create a BSON object from an invalid document."""
+ 
+    def __init__(self, message: str, document: Optional[Any] = None) -> None:
+        super().__init__(message)
+        self._document = document
+
+    @property
+    def document(self) -> Any:
+        """The invalid document that caused the error.
+
+        ..versionadded:: 4.16"""
+        return self._document
+
+ 
+ class InvalidId(BSONError):
+     """Raised when trying to create an ObjectId from invalid data."""
+diff --git a/doc/changelog.rst b/doc/changelog.rst
+index 082c22fafc..7270043d41 100644
+--- a/doc/changelog.rst
+++ b/doc/changelog.rst
+@@ -1,6 +1,15 @@
+ Changelog
+ =========
+ 
+Changes in Version 4.16.0 (XXXX/XX/XX)
+--------------------------------------
+
+PyMongo 4.16 brings a number of changes including:
+
+- Removed invalid documents from :class:`bson.errors.InvalidDocument` error messages as
+  doing so may leak sensitive user data.
+  Instead, invalid documents are stored in :attr:`bson.errors.InvalidDocument.document`.
+
+ Changes in Version 4.15.1 (2025/09/16)
+ --------------------------------------
+ 
+diff --git a/test/test_bson.py b/test/test_bson.py
+index e4cf85c46c..f792db1e89 100644
+--- a/test/test_bson.py
+++ b/test/test_bson.py
+@@ -1163,7 +1163,7 @@ def __repr__(self):
+         ):
+             encode({"t": Wrapper(1)})
+ 
+-    def test_doc_in_invalid_document_error_message(self):
+    def test_doc_in_invalid_document_error_as_property(self):
+         class Wrapper:
+             def __init__(self, val):
+                 self.val = val
+@@ -1173,10 +1173,11 @@ def __repr__(self):
+ 
+         self.assertEqual("1", repr(Wrapper(1)))
+         doc = {"t": Wrapper(1)}
+-        with self.assertRaisesRegex(InvalidDocument, f"Invalid document {doc}"):
+        with self.assertRaisesRegex(InvalidDocument, "Invalid document:") as cm:
+             encode(doc)
+        self.assertEqual(cm.exception.document, doc)
+ 
+-    def test_doc_in_invalid_document_error_message_mapping(self):
+    def test_doc_in_invalid_document_error_as_property_mapping(self):
+         class MyMapping(abc.Mapping):
+             def keys(self):
+                 return ["t"]
+@@ -1192,6 +1193,11 @@ def __len__(self):
+             def __iter__(self):
+                 return iter(["t"])
+ 
+            def __eq__(self, other):
+                if isinstance(other, MyMapping):
+                    return True
+                return False
+
+         class Wrapper:
+             def __init__(self, val):
+                 self.val = val
+@@ -1201,8 +1207,9 @@ def __repr__(self):
+ 
+         self.assertEqual("1", repr(Wrapper(1)))
+         doc = MyMapping()
+-        with self.assertRaisesRegex(InvalidDocument, f"Invalid document {doc}"):
+        with self.assertRaisesRegex(InvalidDocument, "Invalid document:") as cm:
+             encode(doc)
+        self.assertEqual(cm.exception.document, doc)
+ 
+ 
+ class TestCodecOptions(unittest.TestCase):
+diff --git a/bson/_cbsonmodule.c b/bson/_cbsonmodule.c
+index bee7198567..7d184641c5 100644
+--- a/bson/_cbsonmodule.c
+++ b/bson/_cbsonmodule.c
+@@ -1657,26 +1657,28 @@ void handle_invalid_doc_error(PyObject* dict) {
+     }
+ 
+     if (evalue && PyErr_GivenExceptionMatches(etype, InvalidDocument)) {
+-        PyObject *msg = PyObject_Str(evalue);
+        msg = PyObject_Str(evalue);
+         if (msg) {
+             const char * msg_utf8 = PyUnicode_AsUTF8(msg);
+             if (msg_utf8 == NULL) {
+                 goto cleanup;
+             }
+-            PyObject *new_msg = PyUnicode_FromFormat("Invalid document: %s", msg_utf8);
+            new_msg = PyUnicode_FromFormat("Invalid document: %s", msg_utf8);
+             if (new_msg == NULL) {
+                 goto cleanup;
+             }
+             // Add doc to the error instance as a property.
+-            PyObject *new_evalue = PyObject_CallFunctionObjArgs(InvalidDocument, new_msg, dict, NULL);
+            new_evalue = PyObject_CallFunctionObjArgs(InvalidDocument, new_msg, dict, NULL);
+             Py_DECREF(evalue);
+             Py_DECREF(etype);
+             etype = InvalidDocument;
+             InvalidDocument = NULL;
+             if (new_evalue) {
+                 evalue = new_evalue;
+                new_evalue = NULL;
+             } else {
+                 evalue = msg;
+                msg = NULL;
+             }
+         }
+         PyErr_NormalizeException(&etype, &evalue, &etrace);
--- a/dev-python/pymongo/pymongo-4.15.2-r1.ebuild
+++ b/dev-python/pymongo/pymongo-4.15.2-r1.ebuild
@@ -0,0 +1,207 @@
+# Copyright 1999-2025 Gentoo Authors
+# Distributed under the terms of the GNU General Public License v2
+
+EAPI=8
+
+DISTUTILS_EXT=1
+DISTUTILS_USE_PEP517=hatchling
+PYTHON_COMPAT=( pypy3_11 python3_{11..14} )
+
+inherit check-reqs distutils-r1
+
+MY_P=mongo-python-driver-${PV}
+DESCRIPTION="Python driver for MongoDB"
+HOMEPAGE="
+	https://github.com/mongodb/mongo-python-driver/
+	https://pypi.org/project/pymongo/
+"
+SRC_URI="
+	https://github.com/mongodb/mongo-python-driver/archive/${PV}.tar.gz
+		-> ${MY_P}.gh.tar.gz
+"
+S=${WORKDIR}/${MY_P}
+
+LICENSE="Apache-2.0"
+SLOT="0"
+KEYWORDS="~alpha ~amd64 ~arm ~arm64 ~hppa ~loong ~mips ~ppc ~ppc64 ~riscv ~s390 ~sparc ~x86"
+IUSE="doc kerberos +native-extensions +test-full"
+
+RDEPEND="
+	<dev-python/dnspython-3.0.0[${PYTHON_USEDEP}]
+	kerberos? ( dev-python/kerberos[${PYTHON_USEDEP}] )
+"
+BDEPEND="
+	dev-python/setuptools[${PYTHON_USEDEP}]
+	test? (
+		test-full? (
+			>=dev-db/mongodb-2.6.0
+		)
+	)
+"
+
+distutils_enable_sphinx doc
+
+EPYTEST_PLUGINS=( pytest-asyncio )
+EPYTEST_RERUNS=5
+distutils_enable_tests pytest
+
+reqcheck() {
+	if use test && use test-full; then
+		# During the tests, database size reaches 1.5G.
+		local CHECKREQS_DISK_BUILD=1536M
+
+		check-reqs_${1}
+	fi
+}
+
+pkg_pretend() {
+	reqcheck pkg_pretend
+}
+
+pkg_setup() {
+	reqcheck pkg_setup
+}
+
+src_prepare() {
+	local PATCHES=(
+		# https://github.com/mongodb/mongo-python-driver/pull/2539
+		# https://github.com/mongodb/mongo-python-driver/pull/2573
+		"${FILESDIR}/${P}-backports.patch"
+	)
+
+	distutils-r1_src_prepare
+	# we do not want hatch-requirements-txt and its ton of NIH deps
+	sed -i -e '/requirements/d' pyproject.toml || die
+}
+
+python_compile() {
+	# causes build errors to be fatal
+	local -x TOX_ENV_NAME=whatever
+	local DISTUTILS_ARGS=()
+	# unconditionally implicitly disabled on pypy3
+	if ! use native-extensions; then
+		export NO_EXT=1
+	else
+		export PYMONGO_C_EXT_MUST_BUILD=1
+		unset NO_EXT
+	fi
+
+	distutils-r1_python_compile
+
+	# upstream forces setup.py build_ext -i in their setuptools hack
+	find -name '*.so' -delete || die
+}
+
+python_test() {
+	rm -rf bson pymongo || die
+
+	local EPYTEST_DESELECT=(
+		# network-sandbox
+		test/asynchronous/test_async_loop_unblocked.py::TestClientLoopUnblocked::test_client_does_not_block_loop
+		test/asynchronous/test_client.py::AsyncClientUnitTest::test_connection_timeout_ms_propagates_to_DNS_resolver
+		test/asynchronous/test_client.py::AsyncClientUnitTest::test_detected_environment_logging
+		test/asynchronous/test_client.py::AsyncClientUnitTest::test_detected_environment_warning
+		test/asynchronous/test_client.py::TestClient::test_service_name_from_kwargs
+		test/asynchronous/test_client.py::TestClient::test_srv_max_hosts_kwarg
+		test/test_client.py::ClientUnitTest::test_connection_timeout_ms_propagates_to_DNS_resolver
+		test/test_client.py::ClientUnitTest::test_detected_environment_logging
+		test/test_client.py::ClientUnitTest::test_detected_environment_warning
+		test/test_client.py::TestClient::test_service_name_from_kwargs
+		test/test_client.py::TestClient::test_srv_max_hosts_kwarg
+		test/test_dns.py::TestCaseInsensitive::test_connect_case_insensitive
+		test/asynchronous/test_dns.py::IsolatedAsyncioTestCaseInsensitive::test_connect_case_insensitive
+		test/test_srv_polling.py
+		test/asynchronous/test_srv_polling.py
+		test/test_uri_spec.py::TestAllScenarios::test_test_uri_options_srv-options_SRV_URI_with_custom_srvServiceName
+		test/test_uri_spec.py::TestAllScenarios::test_test_uri_options_srv-options_SRV_URI_with_invalid_type_for_srvMaxHosts
+		test/test_uri_spec.py::TestAllScenarios::test_test_uri_options_srv-options_SRV_URI_with_negative_integer_for_srvMaxHosts
+		test/test_uri_spec.py::TestAllScenarios::test_test_uri_options_srv-options_SRV_URI_with_positive_srvMaxHosts_and_loadBalanced=fa
+		test/test_uri_spec.py::TestAllScenarios::test_test_uri_options_srv-options_SRV_URI_with_srvMaxHosts
+		test/test_uri_spec.py::TestAllScenarios::test_test_uri_options_srv-options_SRV_URI_with_srvMaxHosts=0_and_loadBalanced=true
+		test/test_uri_spec.py::TestAllScenarios::test_test_uri_options_srv-options_SRV_URI_with_srvMaxHosts=0_and_replicaSet
+
+		# broken regularly by changes in mypy
+		test/test_typing.py::TestMypyFails::test_mypy_failures
+
+		# fragile to timing? fails because we're getting too many logs
+		test/test_connection_logging.py::TestConnectionLoggingConnectionPoolOptions::test_maxConnecting_should_be_included_in_connection_pool_created_message_when_specified
+
+		# hangs?
+		test/asynchronous/test_grid_file.py::AsyncTestGridFile::test_small_chunks
+
+		# broken async tests?
+		test/asynchronous/test_encryption.py
+
+		# -Werror
+		test/test_read_preferences.py::TestMongosAndReadPreference::test_read_preference_hedge_deprecated
+		test/asynchronous/test_read_preferences.py::TestMongosAndReadPreference::test_read_preference_hedge_deprecated
+
+		# fragile to timing? Internet?
+		test/test_client.py::TestClient::test_repr_srv_host
+		test/asynchronous/test_client.py::TestClient::test_repr_srv_host
+		test/asynchronous/test_ssl.py::TestSSL::test_pyopenssl_ignored_in_async
+	)
+
+	if ! use test-full; then
+		# .invalid is guaranteed to return NXDOMAIN per RFC 6761
+		local -x DB_IP=mongodb.invalid
+		epytest -p asyncio
+		return
+	fi
+
+	# Yes, we need TCP/IP for that...
+	local -x DB_IP=127.0.0.1
+	local -x DB_PORT=27000
+
+	local dbpath=${TMPDIR}/mongo.db
+	local logpath=${TMPDIR}/mongod.log
+
+	local failed=
+	mkdir -p "${dbpath}" || die
+	while true; do
+		ebegin "Trying to start mongod on port ${DB_PORT}"
+
+		# mongodb is extremely inefficient
+		# https://www.mongodb.com/docs/manual/reference/ulimit/#review-and-set-resource-limits
+		ulimit -n 64000 || die
+
+		local mongod_options=(
+			--dbpath "${dbpath}"
+			--bind_ip "${DB_IP}"
+			--port "${DB_PORT}"
+			--unixSocketPrefix "${TMPDIR}"
+			--logpath "${logpath}"
+			--fork
+
+			# try to reduce resource use
+			--wiredTigerCacheSizeGB 0.25
+		)
+
+		LC_ALL=C mongod "${mongod_options[@]}" && sleep 2
+
+		# Now we need to check if the server actually started...
+		if [[ ${?} -eq 0 && -S "${TMPDIR}"/mongodb-${DB_PORT}.sock ]]; then
+			# yay!
+			eend 0
+			break
+		elif grep -q 'Address already in use' "${logpath}"; then
+			# ay, someone took our port!
+			eend 1
+			: $(( DB_PORT += 1 ))
+			continue
+		else
+			eend 1
+			eerror "Unable to start mongod for tests. See the server log:"
+			eerror "	${logpath}"
+			die "Unable to start mongod for tests."
+		fi
+	done
+
+	nonfatal epytest -m "default or default_async or encryption" || failed=1
+
+	mongod --dbpath "${dbpath}" --shutdown || die
+
+	[[ ${failed} ]] && die "Tests fail with ${EPYTHON}"
+
+	rm -rf "${dbpath}" || die
+}