gh-105156: Deprecate the old Py_UNICODE type in C API (#105157)

vstinner · methane · web-flow · commit 8ed705c083e8 · 2023-06-01T08:56:35.000+02:00
Deprecate the old Py_UNICODE and PY_UNICODE_TYPE types in the C API:
use wchar_t instead.

Replace Py_UNICODE with wchar_t in multiple C files.

Co-authored-by: Inada Naoki &lt;songofacandy@gmail.com&gt;
diff --git a/Doc/c-api/unicode.rst b/Doc/c-api/unicode.rst
@@ -52,6 +52,8 @@ Python:
       whether you selected a "narrow" or "wide" Unicode version of Python at
       build time.
 
+   .. deprecated-removed:: 3.13 3.15
+
 
 .. c:type:: PyASCIIObject
             PyCompactUnicodeObject
diff --git a/Doc/whatsnew/3.13.rst b/Doc/whatsnew/3.13.rst
@@ -352,6 +352,11 @@ Porting to Python 3.13
 Deprecated
 ----------
 
+* Deprecate the old ``Py_UNICODE`` and ``PY_UNICODE_TYPE`` types: use directly
+  the ``wchar_t`` type instead. Since Python 3.3, ``Py_UNICODE`` and
+  ``PY_UNICODE_TYPE`` are just aliases to ``wchar_t``.
+  (Contributed by Victor Stinner in :gh:`105156`.)
+
 Removed
 -------
 
diff --git a/Include/cpython/unicodeobject.h b/Include/cpython/unicodeobject.h
@@ -6,8 +6,8 @@
    Python and represents a single Unicode element in the Unicode type.
    With PEP 393, Py_UNICODE is deprecated and replaced with a
    typedef to wchar_t. */
-#define PY_UNICODE_TYPE wchar_t
-/* Py_DEPRECATED(3.3) */ typedef wchar_t Py_UNICODE;
+Py_DEPRECATED(3.13) typedef wchar_t PY_UNICODE_TYPE;
+Py_DEPRECATED(3.13) typedef wchar_t Py_UNICODE;
 
 /* --- Internal Unicode Operations ---------------------------------------- */
 
diff --git a/Misc/NEWS.d/next/C API/2023-05-31-18-37-57.gh-issue-105156.R4El5V.rst b/Misc/NEWS.d/next/C API/2023-05-31-18-37-57.gh-issue-105156.R4El5V.rst
@@ -0,0 +1,4 @@
+Deprecate the old ``Py_UNICODE`` and ``PY_UNICODE_TYPE`` types: use directly
+the ``wchar_t`` type instead. Since Python 3.3, ``Py_UNICODE`` and
+``PY_UNICODE_TYPE`` are just aliases to ``wchar_t``. Patch by Victor
+Stinner.
diff --git a/Modules/_io/fileio.c b/Modules/_io/fileio.c
@@ -231,7 +231,7 @@ _io_FileIO___init___impl(fileio *self, PyObject *nameobj, const char *mode,
 /*[clinic end generated code: output=23413f68e6484bbd input=588aac967e0ba74b]*/
 {
 #ifdef MS_WINDOWS
-    Py_UNICODE *widename = NULL;
+    wchar_t *widename = NULL;
 #else
     const char *name = NULL;
 #endif
diff --git a/Modules/_testcapi/getargs.c b/Modules/_testcapi/getargs.c
@@ -594,7 +594,7 @@ getargs_y_hash(PyObject *self, PyObject *args)
 static PyObject *
 getargs_u(PyObject *self, PyObject *args)
 {
-    Py_UNICODE *str;
+    wchar_t *str;
     if (!PyArg_ParseTuple(args, "u", &str)) {
         return NULL;
     }
@@ -604,7 +604,7 @@ getargs_u(PyObject *self, PyObject *args)
 static PyObject *
 getargs_u_hash(PyObject *self, PyObject *args)
 {
-    Py_UNICODE *str;
+    wchar_t *str;
     Py_ssize_t size;
     if (!PyArg_ParseTuple(args, "u#", &str, &size)) {
         return NULL;
@@ -615,7 +615,7 @@ getargs_u_hash(PyObject *self, PyObject *args)
 static PyObject *
 getargs_Z(PyObject *self, PyObject *args)
 {
-    Py_UNICODE *str;
+    wchar_t *str;
     if (!PyArg_ParseTuple(args, "Z", &str)) {
         return NULL;
     }
@@ -628,7 +628,7 @@ getargs_Z(PyObject *self, PyObject *args)
 static PyObject *
 getargs_Z_hash(PyObject *self, PyObject *args)
 {
-    Py_UNICODE *str;
+    wchar_t *str;
     Py_ssize_t size;
     if (!PyArg_ParseTuple(args, "Z#", &str, &size)) {
         return NULL;
diff --git a/Modules/arraymodule.c b/Modules/arraymodule.c
@@ -1830,10 +1830,10 @@ typecode_to_mformat_code(char typecode)
         return UNSIGNED_INT8;
 
     case 'u':
-        if (sizeof(Py_UNICODE) == 2) {
+        if (sizeof(wchar_t) == 2) {
             return UTF16_LE + is_big_endian;
         }
-        if (sizeof(Py_UNICODE) == 4) {
+        if (sizeof(wchar_t) == 4) {
             return UTF32_LE + is_big_endian;
         }
         return UNKNOWN_FORMAT;
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
@@ -1800,14 +1800,14 @@ PyUnicode_FromWideChar(const wchar_t *u, Py_ssize_t size)
 
     switch (PyUnicode_KIND(unicode)) {
     case PyUnicode_1BYTE_KIND:
-        _PyUnicode_CONVERT_BYTES(Py_UNICODE, unsigned char,
+        _PyUnicode_CONVERT_BYTES(wchar_t, unsigned char,
                                 u, u + size, PyUnicode_1BYTE_DATA(unicode));
         break;
     case PyUnicode_2BYTE_KIND:
 #if Py_UNICODE_SIZE == 2
         memcpy(PyUnicode_2BYTE_DATA(unicode), u, size * 2);
 #else
-        _PyUnicode_CONVERT_BYTES(Py_UNICODE, Py_UCS2,
+        _PyUnicode_CONVERT_BYTES(wchar_t, Py_UCS2,
                                 u, u + size, PyUnicode_2BYTE_DATA(unicode));
 #endif
         break;
@@ -3809,9 +3809,9 @@ PyUnicode_AsUTF8(PyObject *unicode)
 PyUnicode_GetSize() has been deprecated since Python 3.3
 because it returned length of Py_UNICODE.
 
-But this function is part of stable abi, because it don't
+But this function is part of stable abi, because it doesn't
 include Py_UNICODE in signature and it was not excluded from
-stable abi in PEP 384.
+stable ABI in PEP 384.
 */
 PyAPI_FUNC(Py_ssize_t)
 PyUnicode_GetSize(PyObject *unicode)
diff --git a/Python/modsupport.c b/Python/modsupport.c
@@ -329,7 +329,7 @@ do_mkvalue(const char **p_format, va_list *p_va)
         case 'u':
         {
             PyObject *v;
-            Py_UNICODE *u = va_arg(*p_va, Py_UNICODE *);
+            const wchar_t *u = va_arg(*p_va, wchar_t*);
             Py_ssize_t n;
             if (**p_format == '#') {
                 ++*p_format;

Original file line number	Diff line number	Diff line change
`@@ -231,7 +231,7 @@ _io_FileIO___init___impl(fileio self, PyObject nameobj, const char *mode,`
`231`	`231`	`/[clinic end generated code: output=23413f68e6484bbd input=588aac967e0ba74b]/`
`232`	`232`	`{`
`233`	`233`	`#ifdef MS_WINDOWS`
`234`		`- Py_UNICODE *widename = NULL;`
	`234`	`+ wchar_t *widename = NULL;`
`235`	`235`	`#else`
`236`	`236`	`const char *name = NULL;`
`237`	`237`	`#endif`
Original file line number	Diff line number	Diff line change
`@@ -594,7 +594,7 @@ getargs_y_hash(PyObject self, PyObject args)`
`594`	`594`	`static PyObject *`
`595`	`595`	`getargs_u(PyObject self, PyObject args)`
`596`	`596`	`{`
`597`		`- Py_UNICODE *str;`
	`597`	`+ wchar_t *str;`
`598`	`598`	`if (!PyArg_ParseTuple(args, "u", &str)) {`
`599`	`599`	`return NULL;`
`600`	`600`	`}`
`@@ -604,7 +604,7 @@ getargs_u(PyObject self, PyObject args)`
`604`	`604`	`static PyObject *`
`605`	`605`	`getargs_u_hash(PyObject self, PyObject args)`
`606`	`606`	`{`
`607`		`- Py_UNICODE *str;`
	`607`	`+ wchar_t *str;`
`608`	`608`	`Py_ssize_t size;`
`609`	`609`	`if (!PyArg_ParseTuple(args, "u#", &str, &size)) {`
`610`	`610`	`return NULL;`
`@@ -615,7 +615,7 @@ getargs_u_hash(PyObject self, PyObject args)`
`615`	`615`	`static PyObject *`
`616`	`616`	`getargs_Z(PyObject self, PyObject args)`
`617`	`617`	`{`
`618`		`- Py_UNICODE *str;`
	`618`	`+ wchar_t *str;`
`619`	`619`	`if (!PyArg_ParseTuple(args, "Z", &str)) {`
`620`	`620`	`return NULL;`
`621`	`621`	`}`
`@@ -628,7 +628,7 @@ getargs_Z(PyObject self, PyObject args)`
`628`	`628`	`static PyObject *`
`629`	`629`	`getargs_Z_hash(PyObject self, PyObject args)`
`630`	`630`	`{`
`631`		`- Py_UNICODE *str;`
	`631`	`+ wchar_t *str;`
`632`	`632`	`Py_ssize_t size;`
`633`	`633`	`if (!PyArg_ParseTuple(args, "Z#", &str, &size)) {`
`634`	`634`	`return NULL;`
Original file line number	Diff line number	Diff line change
`@@ -1830,10 +1830,10 @@ typecode_to_mformat_code(char typecode)`
`1830`	`1830`	`return UNSIGNED_INT8;`
`1831`	`1831`
`1832`	`1832`	`case 'u':`
`1833`		`- if (sizeof(Py_UNICODE) == 2) {`
	`1833`	`+ if (sizeof(wchar_t) == 2) {`
`1834`	`1834`	`return UTF16_LE + is_big_endian;`
`1835`	`1835`	`}`
`1836`		`- if (sizeof(Py_UNICODE) == 4) {`
	`1836`	`+ if (sizeof(wchar_t) == 4) {`
`1837`	`1837`	`return UTF32_LE + is_big_endian;`
`1838`	`1838`	`}`
`1839`	`1839`	`return UNKNOWN_FORMAT;`
Original file line number	Diff line number	Diff line change
`@@ -329,7 +329,7 @@ do_mkvalue(const char *p_format, va_list p_va)`
`329`	`329`	`case 'u':`
`330`	`330`	`{`
`331`	`331`	`PyObject *v;`
`332`		`- Py_UNICODE u = va_arg(p_va, Py_UNICODE *);`
	`332`	`+ const wchar_t u = va_arg(p_va, wchar_t*);`
`333`	`333`	`Py_ssize_t n;`
`334`	`334`	`if (**p_format == '#') {`
`335`	`335`	`++*p_format;`