[3.14] gh-156126: Fix crash in -X importtime with unencodable module … (#156330)

[3.14] gh-156126: Fix crash in -X importtime with unencodable module names (GH-156137)

(cherry picked from commit 3a5aa68)

Author: Eclips4

Lib/test/test_cmd_line.py

@@ -1225,6 +1225,24 @@ def test_import_time(self):
         assert_python_failure('-X', 'importtime=-1', '-c', code)
         assert_python_failure('-X', 'importtime=3', '-c', code)
 
+    def test_import_time_unencodable_module_name(self):
+        code = textwrap.dedent("""
+            import sys, types
+            name = 'mod\\ud800'
+            sys.modules[name] = types.ModuleType(name)
+            __import__(name)
+            try:
+                __import__('nonexistent\\ud800')
+            except ModuleNotFoundError:
+                pass
+        """)
+        res = assert_python_ok('-X', 'importtime=2', '-c', code)
+        res_err = res.err.decode('utf-8')
+        self.assertRegex(res_err,
+                         r'import time: cached\s* \| cached\s* \| mod\\ud800')
+        self.assertRegex(res_err,
+                         r'import time: \s*\d+ \| \s*\d+ \| \s*nonexistent\\ud800')
+
     def res2int(self, res):
         out = res.out.strip().decode("utf-8")
         return tuple(int(i) for i in out.split())

Misc/NEWS.d/next/Core_and_Builtins/2026-08-20-22-13-20.gh-issue-156126.pXm4Qr.rst

@@ -0,0 +1,3 @@
+Fix a crash when importing a module whose name contains characters that
+cannot be encoded to UTF-8 (such as lone surrogates) while :option:`-X
+importtime <-X>` is enabled.

Python/import.c

@@ -241,6 +241,19 @@ import_get_module(PyThreadState *tstate, PyObject *name)
     return m;
 }
 
+static PyObject *
+get_importtime_name(PyObject *name)
+{
+    PyObject *exc = PyErr_GetRaisedException();
+    PyObject *encoded = PyUnicode_AsEncodedString(name, "utf-8",
+                                                  "backslashreplace");
+    if (encoded == NULL) {
+        PyErr_Clear();
+    }
+    PyErr_SetRaisedException(exc);
+    return encoded;
+}
+
 static int
 import_ensure_initialized(PyInterpreterState *interp, PyObject *mod, PyObject *name)
 {
@@ -278,8 +291,11 @@ import_ensure_initialized(PyInterpreterState *interp, PyObject *mod, PyObject *n
     if (_PyInterpreterState_GetConfig(interp)->import_time == 2) {
         _IMPORT_TIME_HEADER(interp);
 #define import_level FIND_AND_LOAD(interp).import_level
+        PyObject *encoded_name = get_importtime_name(name);
         fprintf(stderr, "import time: cached    | cached     | %*s\n",
-                import_level*2, PyUnicode_AsUTF8(name));
+                import_level*2,
+                encoded_name != NULL ? PyBytes_AS_STRING(encoded_name) : "?");
+        Py_XDECREF(encoded_name);
 #undef import_level
     }
 
@@ -3782,10 +3798,13 @@ import_find_and_load(PyThreadState *tstate, PyObject *abs_name)
         PyTime_t cum = t2 - t1;
 
         import_level--;
+        PyObject *encoded_name = get_importtime_name(abs_name);
         fprintf(stderr, "import time: %9ld | %10ld | %*s%s\n",
                 (long)_PyTime_AsMicroseconds(cum - accumulated, _PyTime_ROUND_CEILING),
                 (long)_PyTime_AsMicroseconds(cum, _PyTime_ROUND_CEILING),
-                import_level*2, "", PyUnicode_AsUTF8(abs_name));
+                import_level*2, "",
+                encoded_name != NULL ? PyBytes_AS_STRING(encoded_name) : "?");
+        Py_XDECREF(encoded_name);
 
         accumulated = accumulated_copy + cum;
     }