gh-89653: PEP 670: unicodeobject.h uses _Py_CAST() (GH-92696) (GH-92703)

Use _Py_CAST() and _Py_STATIC_CAST() in macros wrapping static inline
functions of unicodeobject.h.

Change also the kind type from unsigned int to int: same parameter
type than PyUnicode_FromKindAndData().

The limited API version 3.11 no longer casts arguments to expected
types.
(cherry picked from commit d0c9353a79c2003385c83892db5dfd4e443474c9)

Co-authored-by: Victor Stinner <vstinner@python.org>
Co-authored-by: Victor Stinner <vstinner@python.org>

diff --git a/Doc/c-api/unicode.rst b/Doc/c-api/unicode.rst
index d139112578ca99652e35d0a60b0c00d99fb0eb98..ae24477fa45c84092705b0405ea5ed7ed95fceac 100644 (file)
--- a/Doc/c-api/unicode.rst
+++ b/Doc/c-api/unicode.rst
@@ -149,7 +149,7 @@ access to internal read-only data of Unicode objects:
       ``PyUnicode_WCHAR_KIND`` is deprecated.
 
 
-.. c:function:: unsigned int PyUnicode_KIND(PyObject *o)
+.. c:function:: int PyUnicode_KIND(PyObject *o)
 
    Return one of the PyUnicode kind constants (see above) that indicate how many
    bytes per character this Unicode object uses to store its data.  *o* has to
@@ -168,7 +168,7 @@ access to internal read-only data of Unicode objects:
    .. versionadded:: 3.3
 
 
-.. c:function:: void PyUnicode_WRITE(unsigned int kind, void *data, \
+.. c:function:: void PyUnicode_WRITE(int kind, void *data, \
                                      Py_ssize_t index, Py_UCS4 value)
 
    Write into a canonical representation *data* (as obtained with
@@ -181,7 +181,7 @@ access to internal read-only data of Unicode objects:
    .. versionadded:: 3.3
 
 
-.. c:function:: Py_UCS4 PyUnicode_READ(unsigned int kind, void *data, \
+.. c:function:: Py_UCS4 PyUnicode_READ(int kind, void *data, \
                                        Py_ssize_t index)
 
    Read a code point from a canonical representation *data* (as obtained with

diff --git a/Include/cpython/unicodeobject.h b/Include/cpython/unicodeobject.h
index 8e182d0fbf799950f605ca7d11b09c59bb58feed..fdbcb1cf403ae6eba4d8f52b931a8c5828128b53 100644 (file)
--- a/Include/cpython/unicodeobject.h
+++ b/Include/cpython/unicodeobject.h
@@ -367,7 +367,7 @@ static inline Py_ssize_t PyUnicode_GET_LENGTH(PyObject *op) {
    kind and data pointers obtained from other function calls.
    index is the index in the string (starts at 0) and value is the new
    code point value which should be written to that location. */
-static inline void PyUnicode_WRITE(unsigned int kind, void *data,
+static inline void PyUnicode_WRITE(int kind, void *data,
                                    Py_ssize_t index, Py_UCS4 value)
 {
     if (kind == PyUnicode_1BYTE_KIND) {
@@ -384,12 +384,15 @@ static inline void PyUnicode_WRITE(unsigned int kind, void *data,
         _Py_STATIC_CAST(Py_UCS4*, data)[index] = value;
     }
 }
+#if !defined(Py_LIMITED_API) || Py_LIMITED_API+0 < 0x030b0000
 #define PyUnicode_WRITE(kind, data, index, value) \
-    PyUnicode_WRITE((unsigned int)(kind), (void*)(data), (index), (Py_UCS4)(value))
+    PyUnicode_WRITE(_Py_STATIC_CAST(int, kind), _Py_CAST(void*, data), \
+                    (index), _Py_STATIC_CAST(Py_UCS4, value))
+#endif
 
 /* Read a code point from the string's canonical representation.  No checks
    or ready calls are performed. */
-static inline Py_UCS4 PyUnicode_READ(unsigned int kind,
+static inline Py_UCS4 PyUnicode_READ(int kind,
                                      const void *data, Py_ssize_t index)
 {
     if (kind == PyUnicode_1BYTE_KIND) {
@@ -401,8 +404,11 @@ static inline Py_UCS4 PyUnicode_READ(unsigned int kind,
     assert(kind == PyUnicode_4BYTE_KIND);
     return _Py_STATIC_CAST(const Py_UCS4*, data)[index];
 }
+#if !defined(Py_LIMITED_API) || Py_LIMITED_API+0 < 0x030b0000
 #define PyUnicode_READ(kind, data, index) \
-    PyUnicode_READ((unsigned int)(kind), (const void*)(data), (index))
+    PyUnicode_READ(_Py_STATIC_CAST(int, kind), _Py_CAST(const void*, data), \
+                   (index))
+#endif
 
 /* PyUnicode_READ_CHAR() is less efficient than PyUnicode_READ() because it
    calls PyUnicode_KIND() and might call it twice.  For single reads, use
@@ -411,7 +417,7 @@ static inline Py_UCS4 PyUnicode_READ(unsigned int kind,
 static inline Py_UCS4 PyUnicode_READ_CHAR(PyObject *unicode, Py_ssize_t index)
 {
     assert(PyUnicode_IS_READY(unicode));
-    unsigned int kind = PyUnicode_KIND(unicode);
+    int kind = PyUnicode_KIND(unicode);
     if (kind == PyUnicode_1BYTE_KIND) {
         return PyUnicode_1BYTE_DATA(unicode)[index];
     }
@@ -436,7 +442,7 @@ static inline Py_UCS4 PyUnicode_MAX_CHAR_VALUE(PyObject *op)
         return 0x7fU;
     }
 
-    unsigned int kind = PyUnicode_KIND(op);
+    int kind = PyUnicode_KIND(op);
     if (kind == PyUnicode_1BYTE_KIND) {
        return 0xffU;
     }