gh-76535: Add C API functions for changing case of a single codepoint

Author: lysnikolaou

Doc/c-api/unicode.rst

@@ -223,6 +223,14 @@ the Python configuration.
 
    Return ``1`` or ``0`` depending on whether *ch* is a titlecase character.
 
+.. c:function:: int Py_UNICODE_ISCASED(Py_UCS4 ch)
+
+   Return ``1`` or ``0`` depending on whether *ch* is a cased character.
+
+.. c:function:: int Py_UNICODE_ISCASEIGNORABLE(Py_UCS4 ch)
+
+   Return ``1`` or ``0`` depending on whether *ch* is a case-ignorable character.
+
 
 .. c:function:: int Py_UNICODE_ISLINEBREAK(Py_UCS4 ch)
 
@@ -322,6 +330,26 @@ These APIs can be used to work with surrogates:
    surrogate pair. *high* must be in the range [0xD800; 0xDBFF] and *low* must
    be in the range [0xDC00; 0xDFFF].
 
+.. c:function:: Py_ssize_t PyUnicode_ToLower(Py_UCS4 ch, Py_UCS4 *buffer, \
+                                             Py_ssize_t size)
+
+   Convert *ch* to lower case, store result in *buffer*, which should be
+   able to hold *size* characters, and return the number of characters stored.
+   If the buffer is not big enough, return -1.
+
+.. c:function:: Py_ssize_t PyUnicode_ToUpper(Py_UCS4 ch, Py_UCS4 *buffer, \
+                                             Py_ssize_t size)
+
+   Convert *ch* to upper case, store result in *buffer*, which should be
+   able to hold *size* characters, and return the number of characters stored.
+   If the buffer is not big enough, return -1.
+
+.. c:function:: Py_ssize_t PyUnicode_ToTitle(Py_UCS4 ch, Py_UCS4 *buffer, \
+                                             Py_ssize_t size)
+
+   Convert *ch* to title case, store result in *buffer*, which should be
+   able to hold *size* characters, and return the number of characters stored.
+   If the buffer is not big enough, return -1.
 
 Creating and accessing Unicode strings
 """"""""""""""""""""""""""""""""""""""

Include/cpython/unicodeobject.h

@@ -602,6 +602,14 @@ PyAPI_FUNC(int) _PyUnicode_IsTitlecase(
     Py_UCS4 ch       /* Unicode character */
     );
 
+PyAPI_FUNC(int) _PyUnicode_IsCased(
+    Py_UCS4 ch       /* Unicode character */
+    );
+
+PyAPI_FUNC(int) _PyUnicode_IsCaseIgnorable(
+    Py_UCS4 ch       /* Unicode character */
+    );
+
 PyAPI_FUNC(int) _PyUnicode_IsWhitespace(
     const Py_UCS4 ch         /* Unicode character */
     );
@@ -671,6 +679,8 @@ static inline int Py_UNICODE_ISSPACE(Py_UCS4 ch) {
 #define Py_UNICODE_ISLOWER(ch) _PyUnicode_IsLowercase(ch)
 #define Py_UNICODE_ISUPPER(ch) _PyUnicode_IsUppercase(ch)
 #define Py_UNICODE_ISTITLE(ch) _PyUnicode_IsTitlecase(ch)
+#define Py_UNICODE_ISCASED(ch) _PyUnicode_IsCased(ch)
+#define Py_UNICODE_ISCASEIGNORABLE(ch) _PyUnicode_IsCaseIgnorable(ch)
 #define Py_UNICODE_ISLINEBREAK(ch) _PyUnicode_IsLinebreak(ch)
 
 #define Py_UNICODE_TOLOWER(ch) _PyUnicode_ToLowercase(ch)

Include/internal/pycore_unicodeobject.h

@@ -21,8 +21,6 @@ extern int _PyUnicode_ToLowerFull(Py_UCS4 ch, Py_UCS4 *res);
 extern int _PyUnicode_ToTitleFull(Py_UCS4 ch, Py_UCS4 *res);
 extern int _PyUnicode_ToUpperFull(Py_UCS4 ch, Py_UCS4 *res);
 extern int _PyUnicode_ToFoldedFull(Py_UCS4 ch, Py_UCS4 *res);
-extern int _PyUnicode_IsCaseIgnorable(Py_UCS4 ch);
-extern int _PyUnicode_IsCased(Py_UCS4 ch);
 
 /* --- Unicode API -------------------------------------------------------- */
 

Include/unicodeobject.h

@@ -1007,6 +1007,30 @@ PyAPI_FUNC(int) PyUnicode_Contains(
 
 PyAPI_FUNC(int) PyUnicode_IsIdentifier(PyObject *s);
 
+/* Lowercases character and adds result to buffer */
+
+PyAPI_FUNC(Py_ssize_t) PyUnicode_ToLower(
+    Py_UCS4 ch,
+    Py_UCS4 *buffer,
+    Py_ssize_t size
+    );
+
+/* Uppercases character and adds result to buffer */
+
+PyAPI_FUNC(Py_ssize_t) PyUnicode_ToUpper(
+    Py_UCS4 ch,
+    Py_UCS4 *buffer,
+    Py_ssize_t size
+    );
+
+/* Titlecases character and adds result to buffer */
+
+PyAPI_FUNC(Py_ssize_t) PyUnicode_ToTitle(
+    Py_UCS4 ch,
+    Py_UCS4 *buffer,
+    Py_ssize_t size
+    );
+
 /* === Characters Type APIs =============================================== */
 
 #ifndef Py_LIMITED_API

Objects/unicodeobject.c

@@ -9616,6 +9616,36 @@ case_operation(PyObject *self,
     return res;
 }
 
+Py_ssize_t
+PyUnicode_ToLower(Py_UCS4 ch, Py_UCS4 *buffer, Py_ssize_t size)
+{
+    Py_ssize_t n = _PyUnicode_ToLowerFull(ch, buffer);
+    if (n > size) {
+        return -1;
+    }
+    return n;
+}
+
+Py_ssize_t
+PyUnicode_ToUpper(Py_UCS4 ch, Py_UCS4 *buffer, Py_ssize_t size)
+{
+    Py_ssize_t n = _PyUnicode_ToUpperFull(ch, buffer);
+    if (n > size) {
+        return -1;
+    }
+    return n;
+}
+
+Py_ssize_t
+PyUnicode_ToTitle(Py_UCS4 ch, Py_UCS4 *buffer, Py_ssize_t size)
+{
+    Py_ssize_t n = _PyUnicode_ToTitleFull(ch, buffer);
+    if (n > size) {
+        return -1;
+    }
+    return n;
+}
+
 PyObject *
 PyUnicode_Join(PyObject *separator, PyObject *seq)
 {