gh-106844: Fix issues in _winapi.LCMapStringEx (GH-107832)

* Strings with length from 2**31-1 to 2**32-2 always caused MemoryError,
   it doesn't matter how much memory is available.
* Strings with length exactly 2**32-1 caused OSError.
* Strings longer than 2**32-1 characters were truncated due to integer overflow bug.
* Strings containing the null character were truncated at the first null character.

Now strings longer than 2**31-1 characters caused OverflowError and the null character is allowed.

diff --git a/Lib/test/test_ntpath.py b/Lib/test/test_ntpath.py
index 538d758624c9d65af44eadd35cb9a71b69b38b8d..78e1cb582512b0da27743c626c73b87760a8b708 100644 (file)
--- a/Lib/test/test_ntpath.py
+++ b/Lib/test/test_ntpath.py
@@ -1036,6 +1036,7 @@ class PathLikeTests(NtpathTestCase):
         self._check_function(self.path.normcase)
         if sys.platform == 'win32':
             self.assertEqual(ntpath.normcase('\u03a9\u2126'), 'ωΩ')
+            self.assertEqual(ntpath.normcase('abc\x00def'), 'abc\x00def')
 
     def test_path_isabs(self):
         self._check_function(self.path.isabs)

diff --git a/Misc/NEWS.d/next/Windows/2023-07-18-13-01-26.gh-issue-106844.mci4xO.rst b/Misc/NEWS.d/next/Windows/2023-07-18-13-01-26.gh-issue-106844.mci4xO.rst
new file mode 100644 (file)
index 0000000..1fdf162
--- /dev/null
+++ b/Misc/NEWS.d/next/Windows/2023-07-18-13-01-26.gh-issue-106844.mci4xO.rst
@@ -0,0 +1 @@
+Fix integer overflow and truncating by the null character in :func:`!_winapi.LCMapStringEx` which affects :func:`ntpath.normcase`.

diff --git a/Modules/_winapi.c b/Modules/_winapi.c
index 0edb36d18dd3ff2f6bfbe14931ca592037e7237f..eec33499b983fe718595748fe27b10ab43581c8e 100644 (file)
--- a/Modules/_winapi.c
+++ b/Modules/_winapi.c
@@ -1539,40 +1539,56 @@ _winapi.LCMapStringEx
 
     locale: LPCWSTR
     flags: DWORD
-    src: LPCWSTR
+    src: unicode
 
 [clinic start generated code]*/
 
 static PyObject *
 _winapi_LCMapStringEx_impl(PyObject *module, LPCWSTR locale, DWORD flags,
-                           LPCWSTR src)
-/*[clinic end generated code: output=cf4713d80e2b47c9 input=9fe26f95d5ab0001]*/
+                           PyObject *src)
+/*[clinic end generated code: output=b90e6b26e028ff0a input=3e3dcd9b8164012f]*/
 {
     if (flags & (LCMAP_SORTHANDLE | LCMAP_HASH | LCMAP_BYTEREV |
                  LCMAP_SORTKEY)) {
         return PyErr_Format(PyExc_ValueError, "unsupported flags");
     }
 
-    int dest_size = LCMapStringEx(locale, flags, src, -1, NULL, 0,
+    Py_ssize_t src_size;
+    wchar_t *src_ = PyUnicode_AsWideCharString(src, &src_size);
+    if (!src_) {
+        return NULL;
+    }
+    if (src_size > INT_MAX) {
+        PyMem_Free(src_);
+        PyErr_SetString(PyExc_OverflowError, "input string is too long");
+        return NULL;
+    }
+
+    int dest_size = LCMapStringEx(locale, flags, src_, (int)src_size, NULL, 0,
                                   NULL, NULL, 0);
-    if (dest_size == 0) {
-        return PyErr_SetFromWindowsErr(0);
+    if (dest_size <= 0) {
+        DWORD error = GetLastError();
+        PyMem_Free(src_);
+        return PyErr_SetFromWindowsErr(error);
     }
 
     wchar_t* dest = PyMem_NEW(wchar_t, dest_size);
     if (dest == NULL) {
+        PyMem_Free(src_);
         return PyErr_NoMemory();
     }
 
-    int nmapped = LCMapStringEx(locale, flags, src, -1, dest, dest_size,
+    int nmapped = LCMapStringEx(locale, flags, src_, (int)src_size, dest, dest_size,
                                 NULL, NULL, 0);
-    if (nmapped == 0) {
+    if (nmapped <= 0) {
         DWORD error = GetLastError();
+        PyMem_Free(src_);
         PyMem_DEL(dest);
         return PyErr_SetFromWindowsErr(error);
     }
 
-    PyObject *ret = PyUnicode_FromWideChar(dest, dest_size - 1);
+    PyMem_Free(src_);
+    PyObject *ret = PyUnicode_FromWideChar(dest, nmapped);
     PyMem_DEL(dest);
 
     return ret;

diff --git a/Modules/clinic/_winapi.c.h b/Modules/clinic/_winapi.c.h
index 8f46b8f1095e9866f0d57fe175d3eddf8a8150c3..35ac053547121c1e726c954f9f5fe8aadab66c20 100644 (file)
--- a/Modules/clinic/_winapi.c.h
+++ b/Modules/clinic/_winapi.c.h
@@ -884,7 +884,7 @@ PyDoc_STRVAR(_winapi_LCMapStringEx__doc__,
 
 static PyObject *
 _winapi_LCMapStringEx_impl(PyObject *module, LPCWSTR locale, DWORD flags,
-                           LPCWSTR src);
+                           PyObject *src);
 
 static PyObject *
 _winapi_LCMapStringEx(PyObject *module, PyObject *const *args, Py_ssize_t nargs, PyObject *kwnames)
@@ -911,16 +911,16 @@ _winapi_LCMapStringEx(PyObject *module, PyObject *const *args, Py_ssize_t nargs,
     static const char * const _keywords[] = {"locale", "flags", "src", NULL};
     static _PyArg_Parser _parser = {
         .keywords = _keywords,
-        .format = "O&kO&:LCMapStringEx",
+        .format = "O&kU:LCMapStringEx",
         .kwtuple = KWTUPLE,
     };
     #undef KWTUPLE
     LPCWSTR locale = NULL;
     DWORD flags;
-    LPCWSTR src = NULL;
+    PyObject *src;
 
     if (!_PyArg_ParseStackAndKeywords(args, nargs, kwnames, &_parser,
-        _PyUnicode_WideCharString_Converter, &locale, &flags, _PyUnicode_WideCharString_Converter, &src)) {
+        _PyUnicode_WideCharString_Converter, &locale, &flags, &src)) {
         goto exit;
     }
     return_value = _winapi_LCMapStringEx_impl(module, locale, flags, src);
@@ -928,8 +928,6 @@ _winapi_LCMapStringEx(PyObject *module, PyObject *const *args, Py_ssize_t nargs,
 exit:
     /* Cleanup for locale */
     PyMem_Free((void *)locale);
-    /* Cleanup for src */
-    PyMem_Free((void *)src);
 
     return return_value;
 }
@@ -1480,4 +1478,4 @@ exit:
 
     return return_value;
 }
-/*[clinic end generated code: output=f32fe6ecdbffd74d input=a9049054013a1b77]*/
+/*[clinic end generated code: output=ff91ab5cae8961dd input=a9049054013a1b77]*/