bpo-36346: Prepare for removing the legacy Unicode C API (AC only). (p…

…ythonGH-21223)
elprans · Jun 30, 2020 · 349f76c · 349f76c
1 parent b333266
commit 349f76c
Show file tree

Hide file tree

Showing 7 changed files with 523 additions and 45 deletions.
diff --git a/Include/cpython/unicodeobject.h b/Include/cpython/unicodeobject.h
@@ -11,6 +11,8 @@
 
 /* --- Internal Unicode Operations ---------------------------------------- */
 
+#define USE_UNICODE_WCHAR_CACHE 1
+
 /* Since splitting on whitespace is an important use case, and
    whitespace in most situations is solely ASCII whitespace, we
    optimize for the common case by using a quick look-up table
@@ -1169,4 +1171,7 @@ PyAPI_FUNC(PyObject*) _PyUnicode_FromId(_Py_Identifier*);
    and where the hash values are equal (i.e. a very probable match) */
 PyAPI_FUNC(int) _PyUnicode_EQ(PyObject *, PyObject *);
 
+PyAPI_FUNC(int) _PyUnicode_WideCharString_Converter(PyObject *, void *);
+PyAPI_FUNC(int) _PyUnicode_WideCharString_Opt_Converter(PyObject *, void *);
+
 PyAPI_FUNC(Py_ssize_t) _PyUnicode_ScanIdentifier(PyObject *);
diff --git a/Lib/test/clinic.test b/Lib/test/clinic.test
@@ -1813,13 +1813,26 @@ test_Py_UNICODE_converter(PyObject *module, PyObject *const *args, Py_ssize_t na
     const Py_UNICODE *e;
     Py_ssize_clean_t e_length;
 
-    if (!_PyArg_ParseStack(args, nargs, "uuZu#Z#:test_Py_UNICODE_converter",
-        &a, &b, &c, &d, &d_length, &e, &e_length)) {
+    if (!_PyArg_ParseStack(args, nargs, "O&O&O&u#Z#:test_Py_UNICODE_converter",
+        _PyUnicode_WideCharString_Converter, &a, _PyUnicode_WideCharString_Converter, &b, _PyUnicode_WideCharString_Opt_Converter, &c, &d, &d_length, &e, &e_length)) {
         goto exit;
     }
     return_value = test_Py_UNICODE_converter_impl(module, a, b, c, d, d_length, e, e_length);
 
 exit:
+    /* Cleanup for a */
+    #if !USE_UNICODE_WCHAR_CACHE
+    PyMem_Free((void *)a);
+    #endif /* USE_UNICODE_WCHAR_CACHE */
+    /* Cleanup for b */
+    #if !USE_UNICODE_WCHAR_CACHE
+    PyMem_Free((void *)b);
+    #endif /* USE_UNICODE_WCHAR_CACHE */
+    /* Cleanup for c */
+    #if !USE_UNICODE_WCHAR_CACHE
+    PyMem_Free((void *)c);
+    #endif /* USE_UNICODE_WCHAR_CACHE */
+
     return return_value;
 }
 
@@ -1830,7 +1843,7 @@ test_Py_UNICODE_converter_impl(PyObject *module, const Py_UNICODE *a,
                                Py_ssize_clean_t d_length,
                                const Py_UNICODE *e,
                                Py_ssize_clean_t e_length)
-/*[clinic end generated code: output=dd0a09a1b772e57b input=064a3b68ad7f04b0]*/
+/*[clinic end generated code: output=ef45e982fedf0b3d input=064a3b68ad7f04b0]*/
 
 
 /*[clinic input]

diff --git a/Modules/clinic/_winapi.c.h b/Modules/clinic/_winapi.c.h
diff --git a/Modules/clinic/posixmodule.c.h b/Modules/clinic/posixmodule.c.h
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
@@ -3273,6 +3273,80 @@ PyUnicode_AsWideCharString(PyObject *unicode,
 
 #endif /* HAVE_WCHAR_H */
 
+int
+_PyUnicode_WideCharString_Converter(PyObject *obj, void *ptr)
+{
+    wchar_t **p = (wchar_t **)ptr;
+    if (obj == NULL) {
+#if !USE_UNICODE_WCHAR_CACHE
+        PyMem_Free(*p);
+#endif /* USE_UNICODE_WCHAR_CACHE */
+        *p = NULL;
+        return 1;
+    }
+    if (PyUnicode_Check(obj)) {
+#if USE_UNICODE_WCHAR_CACHE
+_Py_COMP_DIAG_PUSH
+_Py_COMP_DIAG_IGNORE_DEPR_DECLS
+        *p = (wchar_t *)_PyUnicode_AsUnicode(obj);
+        if (*p == NULL) {
+            return 0;
+        }
+        return 1;
+_Py_COMP_DIAG_POP
+#else /* USE_UNICODE_WCHAR_CACHE */
+        *p = PyUnicode_AsWideCharString(obj, NULL);
+        if (*p == NULL) {
+            return 0;
+        }
+        return Py_CLEANUP_SUPPORTED;
+#endif /* USE_UNICODE_WCHAR_CACHE */
+    }
+    PyErr_Format(PyExc_TypeError,
+                 "argument must be str, not %.50s",
+                 obj->ob_type->tp_name);
+    return 0;
+}
+
+int
+_PyUnicode_WideCharString_Opt_Converter(PyObject *obj, void *ptr)
+{
+    wchar_t **p = (wchar_t **)ptr;
+    if (obj == NULL) {
+#if !USE_UNICODE_WCHAR_CACHE
+        PyMem_Free(*p);
+#endif /* USE_UNICODE_WCHAR_CACHE */
+        *p = NULL;
+        return 1;
+    }
+    if (obj == Py_None) {
+        *p = NULL;
+        return 1;
+    }
+    if (PyUnicode_Check(obj)) {
+#if USE_UNICODE_WCHAR_CACHE
+_Py_COMP_DIAG_PUSH
+_Py_COMP_DIAG_IGNORE_DEPR_DECLS
+        *p = (wchar_t *)_PyUnicode_AsUnicode(obj);
+        if (*p == NULL) {
+            return 0;
+        }
+        return 1;
+_Py_COMP_DIAG_POP
+#else /* USE_UNICODE_WCHAR_CACHE */
+        *p = PyUnicode_AsWideCharString(obj, NULL);
+        if (*p == NULL) {
+            return 0;
+        }
+        return Py_CLEANUP_SUPPORTED;
+#endif /* USE_UNICODE_WCHAR_CACHE */
+    }
+    PyErr_Format(PyExc_TypeError,
+                 "argument must be str or None, not %.50s",
+                 obj->ob_type->tp_name);
+    return 0;
+}
+
 PyObject *
 PyUnicode_FromOrdinal(int ordinal)
 {