gh-87868: Sort and remove duplicates in getenvironment()

Lib/test/test_subprocess.py

@@ -785,6 +785,19 @@ def test_env(self):
             stdout, stderr = p.communicate()
             self.assertEqual(stdout, b"orange")
 
+    @unittest.skipUnless(sys.platform == "win32", "Windows only issue")
+    def test_win32_duplicate_envs(self):
+        newenv = os.environ.copy()
+        newenv["fRUit"] = "cherry"
+        newenv["fruit"] = "lemon"
+        newenv["FRUIT"] = "orange"
+        newenv["frUit"] = "banana"
+        with subprocess.Popen(["CMD", "/c", "SET", "fruit"],
+                              stdout=subprocess.PIPE,
+                              env=newenv) as p:
+            stdout, stderr = p.communicate()
+            self.assertEqual(stdout, b"fRUit=cherry\r\n")
+
     # Windows requires at least the SYSTEMROOT environment variable to start
     # Python
     @unittest.skipIf(sys.platform == 'win32',

Misc/NEWS.d/next/Windows/2023-03-15-23-53-45.gh-issue-87868.4C36oQ.rst

@@ -0,0 +1,2 @@
+Correctly sort and remove duplicate environment variables in
+``_winapi.CreateProcess``.

Modules/_winapi.c

@@ -774,23 +774,111 @@ gethandle(PyObject* obj, const char* name)
     return ret;
 }
 
+static PyObject* sortenvironmentkey(PyObject *module, PyObject *item) {
+    return _winapi_LCMapStringEx_impl(NULL, LOCALE_NAME_INVARIANT, LCMAP_UPPERCASE, item);
+}
+
+static PyMethodDef sortenvironmentkey_def = {
+    "sortenvironmentkey",
+    _PyCFunction_CAST(sortenvironmentkey),
+    METH_O,
+    ""
+};
+
+static PyObject *
+normalize_environment(PyObject* environment) {
+    PyObject *result, *keys, *keyfunc, *sort, *args, *kwargs;
+
+    keys = PyMapping_Keys(environment);
+    if (keys == NULL) {
+        return NULL;
+    }
+
+    keyfunc = PyCFunction_New(&sortenvironmentkey_def, NULL);
+    sort = PyObject_GetAttrString(keys, "sort");
+    args = PyTuple_New(0);
+    kwargs = PyDict_New();
+    PyDict_SetItemString(kwargs, "key", keyfunc);
+    if (PyObject_Call(sort, args, kwargs) == NULL) {
+        goto error;
+    }
+
+    result = PyDict_New();
+
+    for (int i=0; i<PyList_GET_SIZE(keys); i++) {
+        PyObject *key = PyList_GET_ITEM(keys, i);
+        PyObject* value = PyObject_GetItem(environment, key);
+
+        if (! PyUnicode_Check(key) || ! PyUnicode_Check(value)) {
+            PyErr_SetString(PyExc_TypeError,
+                "environment can only contain strings");
+            Py_DECREF(result);
+            result = NULL;
+            goto error;
+        }
+        if (PyUnicode_FindChar(key, '\0', 0, PyUnicode_GET_LENGTH(key), 1) != -1 ||
+            PyUnicode_FindChar(value, '\0', 0, PyUnicode_GET_LENGTH(value), 1) != -1)
+        {
+            PyErr_SetString(PyExc_ValueError, "embedded null character");
+            Py_DECREF(result);
+            result = NULL;
+            goto error;
+        }
+        /* Search from index 1 because on Windows starting '=' is allowed for
+           defining hidden environment variables. */
+        if (PyUnicode_GET_LENGTH(key) == 0 ||
+            PyUnicode_FindChar(key, '=', 1, PyUnicode_GET_LENGTH(key), 1) != -1)
+        {
+            PyErr_SetString(PyExc_ValueError, "illegal environment variable name");
+            Py_DECREF(result);
+            result = NULL;
+            goto error;
+        }
+
+        if (i == 0) {
+            continue;
+        }
+
+        wchar_t *key_string = PyUnicode_AsWideCharString(key, NULL);
+        wchar_t *prev_key_string = PyUnicode_AsWideCharString(PyList_GET_ITEM(keys, i-1), NULL);
+        if (CompareStringOrdinal(prev_key_string, -1, key_string, -1, TRUE) == CSTR_EQUAL) {
+            continue;
+        }
+        PyObject_SetItem(result, key, value);
+    }
+
+error:
+    Py_DECREF(keys);
+    Py_DECREF(keyfunc);
+    Py_DECREF(sort);
+    Py_DECREF(args);
+    Py_DECREF(kwargs);
+
+    return result;
+}
+
 static wchar_t *
-getenvironment(PyObject* environment)
+getenvironment(PyObject* env)
 {
     Py_ssize_t i, envsize, totalsize;
     wchar_t *buffer = NULL, *p, *end;
-    PyObject *keys, *values;
+    PyObject *environment = NULL, *keys = NULL, *values = NULL;
 
     /* convert environment dictionary to windows environment string */
-    if (! PyMapping_Check(environment)) {
+    if (! PyMapping_Check(env)) {
         PyErr_SetString(
             PyExc_TypeError, "environment must be dictionary or None");
         return NULL;
     }
 
+    environment = normalize_environment(env);
+    if (environment == NULL) {
+        return NULL;
+    }
+
     keys = PyMapping_Keys(environment);
     if (!keys) {
-        return NULL;
+        goto error;
     }
     values = PyMapping_Values(environment);
     if (!values) {
@@ -821,26 +909,6 @@ getenvironment(PyObject* environment)
         PyObject* value = PyList_GET_ITEM(values, i);
         Py_ssize_t size;
 
-        if (! PyUnicode_Check(key) || ! PyUnicode_Check(value)) {
-            PyErr_SetString(PyExc_TypeError,
-                "environment can only contain strings");
-            goto error;
-        }
-        if (PyUnicode_FindChar(key, '\0', 0, PyUnicode_GET_LENGTH(key), 1) != -1 ||
-            PyUnicode_FindChar(value, '\0', 0, PyUnicode_GET_LENGTH(value), 1) != -1)
-        {
-            PyErr_SetString(PyExc_ValueError, "embedded null character");
-            goto error;
-        }
-        /* Search from index 1 because on Windows starting '=' is allowed for
-           defining hidden environment variables. */
-        if (PyUnicode_GET_LENGTH(key) == 0 ||
-            PyUnicode_FindChar(key, '=', 1, PyUnicode_GET_LENGTH(key), 1) != -1)
-        {
-            PyErr_SetString(PyExc_ValueError, "illegal environment variable name");
-            goto error;
-        }
-
         size = PyUnicode_AsWideChar(key, NULL, 0);
         assert(size > 1);
         if (totalsize > PY_SSIZE_T_MAX - size) {
@@ -884,6 +952,7 @@ getenvironment(PyObject* environment)
 
 cleanup:
 error:
+    Py_XDECREF(environment);
     Py_XDECREF(keys);
     Py_XDECREF(values);
     return buffer;