From: Bénédikt Tran <10796600+picnixz@users.noreply.github.com>
Date: Wed, 8 Oct 2025 15:22:44 +0000 (+0200)
Subject: gh-139748: fix leaks in AC error paths when using unicode FS-based converters (#139765)
X-Git-Tag: v3.15.0a1~82
X-Git-Url: http://git.ipfire.org/gitweb.cgi?a=commitdiff_plain;h=b04a57deef66ce08233be57d1ab5873388df2cea;p=thirdparty%2FPython%2Fcpython.git

gh-139748: fix leaks in AC error paths when using unicode FS-based converters (#139765)
---

diff --git a/Lib/test/test_clinic.py b/Lib/test/test_clinic.py
index d54dd546ea36..e0dbb062eb03 100644
--- a/Lib/test/test_clinic.py
+++ b/Lib/test/test_clinic.py
@@ -2980,6 +2980,8 @@ class ClinicExternalTest(TestCase):
             "uint64",
             "uint8",
             "unicode",
+            "unicode_fs_decoded",
+            "unicode_fs_encoded",
             "unsigned_char",
             "unsigned_int",
             "unsigned_long",
diff --git a/Lib/test/test_compile.py b/Lib/test/test_compile.py
index 277a2a187543..1660dabe6815 100644
--- a/Lib/test/test_compile.py
+++ b/Lib/test/test_compile.py
@@ -651,6 +651,21 @@ class TestSpecifics(unittest.TestCase):
                 compile('pass', filename, 'exec')
         self.assertRaises(TypeError, compile, 'pass', list(b'file.py'), 'exec')
 
+    def test_compile_filename_refleak(self):
+        # Regression tests for reference leak in PyUnicode_FSDecoder.
+        # See https://github.com/python/cpython/issues/139748.
+        mortal_str = 'this is a mortal string'
+        # check error path when 'mode' AC conversion failed
+        self.assertRaises(TypeError, compile, b'', mortal_str, mode=1234)
+        # check error path when 'optimize' AC conversion failed
+        self.assertRaises(OverflowError, compile, b'', mortal_str,
+                          'exec', optimize=1 << 1000)
+        # check error path when 'dont_inherit' AC conversion failed
+        class EvilBool:
+            def __bool__(self): raise ValueError
+        self.assertRaises(ValueError, compile, b'', mortal_str,
+                          'exec', dont_inherit=EvilBool())
+
     @support.cpython_only
     def test_same_filename_used(self):
         s = """def f(): pass\ndef g(): pass"""
diff --git a/Lib/test/test_symtable.py b/Lib/test/test_symtable.py
index 26c75fcbc2bd..943e63fc13c9 100644
--- a/Lib/test/test_symtable.py
+++ b/Lib/test/test_symtable.py
@@ -579,6 +579,13 @@ class SymtableTest(unittest.TestCase):
         self.assertEqual(sorted(st.get_identifiers()), [".0", "y"])
         self.assertEqual(st.get_children(), [])
 
+    def test__symtable_refleak(self):
+        # Regression test for reference leak in PyUnicode_FSDecoder.
+        # See https://github.com/python/cpython/issues/139748.
+        mortal_str = 'this is a mortal string'
+        # check error path when 'compile_type' AC conversion failed
+        self.assertRaises(TypeError, symtable.symtable, '', mortal_str, 1)
+
 
 class ComprehensionTests(unittest.TestCase):
     def get_identifiers_recursive(self, st, res):
diff --git a/Misc/NEWS.d/next/Core_and_Builtins/2025-10-08-13-52-00.gh-issue-139748.jq0yFJ.rst b/Misc/NEWS.d/next/Core_and_Builtins/2025-10-08-13-52-00.gh-issue-139748.jq0yFJ.rst
new file mode 100644
index 000000000000..3b190f3f6df6
--- /dev/null
+++ b/Misc/NEWS.d/next/Core_and_Builtins/2025-10-08-13-52-00.gh-issue-139748.jq0yFJ.rst
@@ -0,0 +1,2 @@
+Fix reference leaks in error branches of functions accepting path strings or
+bytes such as :func:`compile` and :func:`os.system`. Patch by Bénédikt Tran.
diff --git a/Modules/_ssl.c b/Modules/_ssl.c
index e1bdc4033ba1..1fa44ef1de4e 100644
--- a/Modules/_ssl.c
+++ b/Modules/_ssl.c
@@ -1858,14 +1858,14 @@ _certificate_to_der(_sslmodulestate *state, X509 *certificate)
 
 /*[clinic input]
 _ssl._test_decode_cert
-    path: object(converter="PyUnicode_FSConverter")
+    path: unicode_fs_encoded
     /
 
 [clinic start generated code]*/
 
 static PyObject *
 _ssl__test_decode_cert_impl(PyObject *module, PyObject *path)
-/*[clinic end generated code: output=96becb9abb23c091 input=cdeaaf02d4346628]*/
+/*[clinic end generated code: output=96becb9abb23c091 input=cb4988d5e651a4f8]*/
 {
     PyObject *retval = NULL;
     X509 *x=NULL;
@@ -1895,7 +1895,6 @@ _ssl__test_decode_cert_impl(PyObject *module, PyObject *path)
     X509_free(x);
 
   fail0:
-    Py_DECREF(path);
     if (cert != NULL) BIO_free(cert);
     return retval;
 }
diff --git a/Modules/clinic/_ssl.c.h b/Modules/clinic/_ssl.c.h
index 0e8e4a7c8a73..d1fb024903e1 100644
--- a/Modules/clinic/_ssl.c.h
+++ b/Modules/clinic/_ssl.c.h
@@ -48,7 +48,7 @@ static PyObject *
 _ssl__test_decode_cert(PyObject *module, PyObject *arg)
 {
     PyObject *return_value = NULL;
-    PyObject *path;
+    PyObject *path = NULL;
 
     if (!PyUnicode_FSConverter(arg, &path)) {
         goto exit;
@@ -56,6 +56,9 @@ _ssl__test_decode_cert(PyObject *module, PyObject *arg)
     return_value = _ssl__test_decode_cert_impl(module, path);
 
 exit:
+    /* Cleanup for path */
+    Py_XDECREF(path);
+
     return return_value;
 }
 
@@ -3322,4 +3325,4 @@ exit:
 #ifndef _SSL_ENUM_CRLS_METHODDEF
     #define _SSL_ENUM_CRLS_METHODDEF
 #endif /* !defined(_SSL_ENUM_CRLS_METHODDEF) */
-/*[clinic end generated code: output=5a630a1e83927d47 input=a9049054013a1b77]*/
+/*[clinic end generated code: output=3b6c9cbfc4660ecb input=a9049054013a1b77]*/
diff --git a/Modules/clinic/socketmodule.c.h b/Modules/clinic/socketmodule.c.h
index 0cedab597db7..e0cc1c50dcba 100644
--- a/Modules/clinic/socketmodule.c.h
+++ b/Modules/clinic/socketmodule.c.h
@@ -479,7 +479,7 @@ static PyObject *
 _socket_if_nametoindex(PyObject *module, PyObject *arg)
 {
     PyObject *return_value = NULL;
-    PyObject *oname;
+    PyObject *oname = NULL;
 
     if (!PyUnicode_FSConverter(arg, &oname)) {
         goto exit;
@@ -487,6 +487,9 @@ _socket_if_nametoindex(PyObject *module, PyObject *arg)
     return_value = _socket_if_nametoindex_impl(module, oname);
 
 exit:
+    /* Cleanup for oname */
+    Py_XDECREF(oname);
+
     return return_value;
 }
 
@@ -538,4 +541,4 @@ exit:
 #ifndef _SOCKET_IF_INDEXTONAME_METHODDEF
     #define _SOCKET_IF_INDEXTONAME_METHODDEF
 #endif /* !defined(_SOCKET_IF_INDEXTONAME_METHODDEF) */
-/*[clinic end generated code: output=0376c46b76ae2bce input=a9049054013a1b77]*/
+/*[clinic end generated code: output=36051ebf6ad1e6f8 input=a9049054013a1b77]*/
diff --git a/Modules/clinic/symtablemodule.c.h b/Modules/clinic/symtablemodule.c.h
index 2ecd3afc00d2..bd55d77c5409 100644
--- a/Modules/clinic/symtablemodule.c.h
+++ b/Modules/clinic/symtablemodule.c.h
@@ -22,7 +22,7 @@ _symtable_symtable(PyObject *module, PyObject *const *args, Py_ssize_t nargs)
 {
     PyObject *return_value = NULL;
     PyObject *source;
-    PyObject *filename;
+    PyObject *filename = NULL;
     const char *startstr;
 
     if (!_PyArg_CheckPositional("symtable", nargs, 3, 3)) {
@@ -48,6 +48,9 @@ _symtable_symtable(PyObject *module, PyObject *const *args, Py_ssize_t nargs)
     return_value = _symtable_symtable_impl(module, source, filename, startstr);
 
 exit:
+    /* Cleanup for filename */
+    Py_XDECREF(filename);
+
     return return_value;
 }
-/*[clinic end generated code: output=931964a76a72f850 input=a9049054013a1b77]*/
+/*[clinic end generated code: output=7a8545d9a1efe837 input=a9049054013a1b77]*/
diff --git a/Modules/posixmodule.c b/Modules/posixmodule.c
index f50167c223e2..4189d3008560 100644
--- a/Modules/posixmodule.c
+++ b/Modules/posixmodule.c
@@ -3137,17 +3137,6 @@ class dev_t_return_converter(unsigned_long_return_converter):
     conversion_fn = '_PyLong_FromDev'
     unsigned_cast = '(dev_t)'
 
-class FSConverter_converter(CConverter):
-    type = 'PyObject *'
-    converter = 'PyUnicode_FSConverter'
-    def converter_init(self):
-        if self.default is not unspecified:
-            fail("FSConverter_converter does not support default values")
-        self.c_default = 'NULL'
-
-    def cleanup(self):
-        return "Py_XDECREF(" + self.name + ");\n"
-
 class pid_t_converter(CConverter):
     type = 'pid_t'
     format_unit = '" _Py_PARSE_PID "'
@@ -3211,7 +3200,7 @@ class confname_converter(CConverter):
         """, argname=argname, converter=self.converter, table=self.table)
 
 [python start generated code]*/
-/*[python end generated code: output=da39a3ee5e6b4b0d input=8189d5ae78244626]*/
+/*[python end generated code: output=da39a3ee5e6b4b0d input=d2759f2332cd39b3]*/
 
 /*[clinic input]
 
@@ -6135,14 +6124,14 @@ os_system_impl(PyObject *module, const wchar_t *command)
 /*[clinic input]
 os.system -> long
 
-    command: FSConverter
+    command: unicode_fs_encoded
 
 Execute the command in a subshell.
 [clinic start generated code]*/
 
 static long
 os_system_impl(PyObject *module, PyObject *command)
-/*[clinic end generated code: output=290fc437dd4f33a0 input=86a58554ba6094af]*/
+/*[clinic end generated code: output=290fc437dd4f33a0 input=47c6f24b6dc92881]*/
 {
     long result;
     const char *bytes = PyBytes_AsString(command);
@@ -9328,7 +9317,7 @@ error:
 /*[clinic input]
 os.initgroups
 
-    username as oname: FSConverter
+    username as oname: unicode_fs_encoded
     gid: int
     /
 
@@ -9341,12 +9330,12 @@ group id.
 
 static PyObject *
 os_initgroups_impl(PyObject *module, PyObject *oname, int gid)
-/*[clinic end generated code: output=7f074d30a425fd3a input=df3d54331b0af204]*/
+/*[clinic end generated code: output=7f074d30a425fd3a input=984e60c7fed88cb4]*/
 #else
 /*[clinic input]
 os.initgroups
 
-    username as oname: FSConverter
+    username as oname: unicode_fs_encoded
     gid: gid_t
     /
 
@@ -9359,7 +9348,7 @@ group id.
 
 static PyObject *
 os_initgroups_impl(PyObject *module, PyObject *oname, gid_t gid)
-/*[clinic end generated code: output=59341244521a9e3f input=0cb91bdc59a4c564]*/
+/*[clinic end generated code: output=59341244521a9e3f input=17d8fbe2dea42ca4]*/
 #endif
 {
     const char *username = PyBytes_AS_STRING(oname);
@@ -13115,8 +13104,8 @@ os_putenv_impl(PyObject *module, PyObject *name, PyObject *value)
 /*[clinic input]
 os.putenv
 
-    name: FSConverter
-    value: FSConverter
+    name: unicode_fs_encoded
+    value: unicode_fs_encoded
     /
 
 Change or add an environment variable.
@@ -13124,7 +13113,7 @@ Change or add an environment variable.
 
 static PyObject *
 os_putenv_impl(PyObject *module, PyObject *name, PyObject *value)
-/*[clinic end generated code: output=d29a567d6b2327d2 input=a97bc6152f688d31]*/
+/*[clinic end generated code: output=d29a567d6b2327d2 input=84fcd30f873c8c45]*/
 {
     const char *name_string = PyBytes_AS_STRING(name);
     const char *value_string = PyBytes_AS_STRING(value);
@@ -13167,7 +13156,7 @@ os_unsetenv_impl(PyObject *module, PyObject *name)
 #else
 /*[clinic input]
 os.unsetenv
-    name: FSConverter
+    name: unicode_fs_encoded
     /
 
 Delete an environment variable.
@@ -13175,7 +13164,7 @@ Delete an environment variable.
 
 static PyObject *
 os_unsetenv_impl(PyObject *module, PyObject *name)
-/*[clinic end generated code: output=54c4137ab1834f02 input=2bb5288a599c7107]*/
+/*[clinic end generated code: output=54c4137ab1834f02 input=78ff12e505ade80a]*/
 {
     if (PySys_Audit("os.unsetenv", "(O)", name) < 0) {
         return NULL;
@@ -15229,14 +15218,14 @@ os_urandom_impl(PyObject *module, Py_ssize_t size)
 /*[clinic input]
 os.memfd_create
 
-    name: FSConverter
+    name: unicode_fs_encoded
     flags: unsigned_int(bitwise=True, c_default="MFD_CLOEXEC") = MFD_CLOEXEC
 
 [clinic start generated code]*/
 
 static PyObject *
 os_memfd_create_impl(PyObject *module, PyObject *name, unsigned int flags)
-/*[clinic end generated code: output=6681ede983bdb9a6 input=a42cfc199bcd56e9]*/
+/*[clinic end generated code: output=6681ede983bdb9a6 input=cd0eb092cfac474b]*/
 {
     int fd;
     const char *bytes = PyBytes_AS_STRING(name);
diff --git a/Modules/socketmodule.c b/Modules/socketmodule.c
index fa153efd8b72..dd4b68929774 100644
--- a/Modules/socketmodule.c
+++ b/Modules/socketmodule.c
@@ -7278,7 +7278,7 @@ Returns a list of network interface information (index, name) tuples.");
 
 /*[clinic input]
 _socket.if_nametoindex
-    oname: object(converter="PyUnicode_FSConverter")
+    oname: unicode_fs_encoded
     /
 
 Returns the interface index corresponding to the interface name if_name.
@@ -7286,7 +7286,7 @@ Returns the interface index corresponding to the interface name if_name.
 
 static PyObject *
 _socket_if_nametoindex_impl(PyObject *module, PyObject *oname)
-/*[clinic end generated code: output=289a411614f30244 input=01e0f1205307fb77]*/
+/*[clinic end generated code: output=289a411614f30244 input=6125dc20683560cf]*/
 {
 #ifdef MS_WINDOWS
     NET_IFINDEX index;
@@ -7295,7 +7295,6 @@ _socket_if_nametoindex_impl(PyObject *module, PyObject *oname)
 #endif
 
     index = if_nametoindex(PyBytes_AS_STRING(oname));
-    Py_DECREF(oname);
     if (index == 0) {
         /* if_nametoindex() doesn't set errno */
         PyErr_SetString(PyExc_OSError, "no interface with this name");
diff --git a/Modules/symtablemodule.c b/Modules/symtablemodule.c
index d0d5223e5ace..d353f406831e 100644
--- a/Modules/symtablemodule.c
+++ b/Modules/symtablemodule.c
@@ -13,7 +13,7 @@ module _symtable
 _symtable.symtable
 
     source:    object
-    filename:  object(converter='PyUnicode_FSDecoder')
+    filename:  unicode_fs_decoded
     startstr:  str
     /
 
@@ -23,7 +23,7 @@ Return symbol and scope dictionaries used internally by compiler.
 static PyObject *
 _symtable_symtable_impl(PyObject *module, PyObject *source,
                         PyObject *filename, const char *startstr)
-/*[clinic end generated code: output=59eb0d5fc7285ac4 input=9dd8a50c0c36a4d7]*/
+/*[clinic end generated code: output=59eb0d5fc7285ac4 input=436ffff90d02e4f6]*/
 {
     struct symtable *st;
     PyObject *t;
@@ -47,12 +47,10 @@ _symtable_symtable_impl(PyObject *module, PyObject *source,
     else {
         PyErr_SetString(PyExc_ValueError,
            "symtable() arg 3 must be 'exec' or 'eval' or 'single'");
-        Py_DECREF(filename);
         Py_XDECREF(source_copy);
         return NULL;
     }
     st = _Py_SymtableStringObjectFlags(str, filename, start, &cf);
-    Py_DECREF(filename);
     Py_XDECREF(source_copy);
     if (st == NULL) {
         return NULL;
diff --git a/Python/bltinmodule.c b/Python/bltinmodule.c
index b7e08ae54da5..2551c2c961b6 100644
--- a/Python/bltinmodule.c
+++ b/Python/bltinmodule.c
@@ -745,7 +745,7 @@ builtin_chr(PyObject *module, PyObject *i)
 compile as builtin_compile
 
     source: object
-    filename: object(converter="PyUnicode_FSDecoder")
+    filename: unicode_fs_decoded
     mode: str
     flags: int = 0
     dont_inherit: bool = False
@@ -771,7 +771,7 @@ static PyObject *
 builtin_compile_impl(PyObject *module, PyObject *source, PyObject *filename,
                      const char *mode, int flags, int dont_inherit,
                      int optimize, int feature_version)
-/*[clinic end generated code: output=b0c09c84f116d3d7 input=cc78e20e7c7682ba]*/
+/*[clinic end generated code: output=b0c09c84f116d3d7 input=8f0069edbdac381b]*/
 {
     PyObject *source_copy;
     const char *str;
@@ -889,7 +889,6 @@ builtin_compile_impl(PyObject *module, PyObject *source, PyObject *filename,
 error:
     result = NULL;
 finally:
-    Py_DECREF(filename);
     return result;
 }
 
diff --git a/Python/clinic/bltinmodule.c.h b/Python/clinic/bltinmodule.c.h
index 60e20a46af5b..adb82f45c25b 100644
--- a/Python/clinic/bltinmodule.c.h
+++ b/Python/clinic/bltinmodule.c.h
@@ -296,7 +296,7 @@ builtin_compile(PyObject *module, PyObject *const *args, Py_ssize_t nargs, PyObj
     PyObject *argsbuf[7];
     Py_ssize_t noptargs = nargs + (kwnames ? PyTuple_GET_SIZE(kwnames) : 0) - 3;
     PyObject *source;
-    PyObject *filename;
+    PyObject *filename = NULL;
     const char *mode;
     int flags = 0;
     int dont_inherit = 0;
@@ -367,6 +367,9 @@ skip_optional_kwonly:
     return_value = builtin_compile_impl(module, source, filename, mode, flags, dont_inherit, optimize, feature_version);
 
 exit:
+    /* Cleanup for filename */
+    Py_XDECREF(filename);
+
     return return_value;
 }
 
@@ -1274,4 +1277,4 @@ builtin_issubclass(PyObject *module, PyObject *const *args, Py_ssize_t nargs)
 exit:
     return return_value;
 }
-/*[clinic end generated code: output=c0b72519622c849e input=a9049054013a1b77]*/
+/*[clinic end generated code: output=7eada753dc2e046f input=a9049054013a1b77]*/
diff --git a/Tools/clinic/libclinic/converters.py b/Tools/clinic/libclinic/converters.py
index 3154299e31b4..201125b9165e 100644
--- a/Tools/clinic/libclinic/converters.py
+++ b/Tools/clinic/libclinic/converters.py
@@ -916,6 +916,26 @@ class unicode_converter(CConverter):
         return super().parse_arg(argname, displayname, limited_capi=limited_capi)
 
 
+class _unicode_fs_converter_base(CConverter):
+    type = 'PyObject *'
+
+    def converter_init(self) -> None:
+        if self.default is not unspecified:
+            fail(f"{self.__class__.__name__} does not support default values")
+        self.c_default = 'NULL'
+
+    def cleanup(self) -> str:
+        return f"Py_XDECREF({self.parser_name});"
+
+
+class unicode_fs_encoded_converter(_unicode_fs_converter_base):
+    converter = 'PyUnicode_FSConverter'
+
+
+class unicode_fs_decoded_converter(_unicode_fs_converter_base):
+    converter = 'PyUnicode_FSDecoder'
+
+
 @add_legacy_c_converter('u')
 @add_legacy_c_converter('u#', zeroes=True)
 @add_legacy_c_converter('Z', accept={str, NoneType})