add str.casefold() (closes #13752)

Objects/unicodectype.c

@@ -185,7 +185,7 @@ Py_UCS4 _PyUnicode_ToUppercase(Py_UCS4 ch)
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
 
     if (ctype->flags & EXTENDED_CASE_MASK)
-        return _PyUnicode_ExtendedCase[ctype->upper & 0xFFFFFF];
+        return _PyUnicode_ExtendedCase[ctype->upper & 0xFFFF];
     return ctype->upper ? ctype->upper : ch;
 }
 
@@ -197,7 +197,7 @@ Py_UCS4 _PyUnicode_ToLowercase(Py_UCS4 ch)
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
 
     if (ctype->flags & EXTENDED_CASE_MASK)
-        return _PyUnicode_ExtendedCase[ctype->lower & 0xFFFFFF];
+        return _PyUnicode_ExtendedCase[ctype->lower & 0xFFFF];
     return ctype->lower ? ctype->lower : ch;
 }
 
@@ -206,7 +206,7 @@ int _PyUnicode_ToLowerFull(Py_UCS4 ch, Py_UCS4 *res)
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
 
     if (ctype->flags & EXTENDED_CASE_MASK) {
-        int index = ctype->lower & 0xFFFFFF;
+        int index = ctype->lower & 0xFFFF;
         int n = ctype->lower >> 24;
         int i;
         for (i = 0; i < n; i++)
@@ -222,7 +222,7 @@ int _PyUnicode_ToTitleFull(Py_UCS4 ch, Py_UCS4 *res)
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
 
     if (ctype->flags & EXTENDED_CASE_MASK) {
-        int index = ctype->title & 0xFFFFFF;
+        int index = ctype->title & 0xFFFF;
         int n = ctype->title >> 24;
         int i;
         for (i = 0; i < n; i++)
@@ -238,7 +238,7 @@ int _PyUnicode_ToUpperFull(Py_UCS4 ch, Py_UCS4 *res)
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
 
     if (ctype->flags & EXTENDED_CASE_MASK) {
-        int index = ctype->upper & 0xFFFFFF;
+        int index = ctype->upper & 0xFFFF;
         int n = ctype->upper >> 24;
         int i;
         for (i = 0; i < n; i++)
@@ -249,6 +249,21 @@ int _PyUnicode_ToUpperFull(Py_UCS4 ch, Py_UCS4 *res)
     return 1;
 }
 
+int _PyUnicode_ToFoldedFull(Py_UCS4 ch, Py_UCS4 *res)
+{
+    const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
+
+    if (ctype->flags & EXTENDED_CASE_MASK && (ctype->lower >> 20) & 7) {
+        int index = (ctype->lower & 0xFFFF) + (ctype->lower >> 24);
+        int n = (ctype->lower >> 20) & 7;
+        int i;
+        for (i = 0; i < n; i++)
+            res[i] = _PyUnicode_ExtendedCase[index + i];
+        return n;
+    }
+    return _PyUnicode_ToLowerFull(ch, res);
+}
+
 int _PyUnicode_IsCased(Py_UCS4 ch)
 {
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);

Objects/unicodeobject.c

@@ -9576,6 +9576,24 @@ do_lower(int kind, void *data, Py_ssize_t length, Py_UCS4 *res, Py_UCS4 *maxchar
     return do_upper_or_lower(kind, data, length, res, maxchar, 1);
 }
 
+static Py_ssize_t
+do_casefold(int kind, void *data, Py_ssize_t length, Py_UCS4 *res, Py_UCS4 *maxchar)
+{
+    Py_ssize_t i, k = 0;
+
+    for (i = 0; i < length; i++) {
+        Py_UCS4 c = PyUnicode_READ(kind, data, i);
+        Py_UCS4 mapped[3];
+        int j, n_res = _PyUnicode_ToFoldedFull(c, mapped);
+        for (j = 0; j < n_res; j++) {
+            if (mapped[j] > *maxchar)
+                *maxchar = mapped[j];
+            res[k++] = mapped[j];
+        }
+    }
+    return k;
+}
+
 static Py_ssize_t
 do_title(int kind, void *data, Py_ssize_t length, Py_UCS4 *res, Py_UCS4 *maxchar)
 {
@@ -10501,6 +10519,22 @@ unicode_capitalize(PyObject *self)
     return case_operation(self, do_capitalize);
 }
 
+PyDoc_STRVAR(casefold__doc__,
+             "S.casefold() -> str\n\
+\n\
+Return a version of S suitable for caseless comparisons.");
+
+static PyObject *
+unicode_casefold(PyObject *self)
+{
+    if (PyUnicode_READY(self) == -1)
+        return NULL;
+    if (PyUnicode_IS_ASCII(self))
+        return ascii_upper_or_lower(self, 1);
+    return case_operation(self, do_casefold);
+}
+
+
 /* Argument converter.  Coerces to a single unicode character */
 
 static int
@@ -12998,6 +13032,7 @@ static PyMethodDef unicode_methods[] = {
     {"rsplit", (PyCFunction) unicode_rsplit, METH_VARARGS, rsplit__doc__},
     {"join", (PyCFunction) unicode_join, METH_O, join__doc__},
     {"capitalize", (PyCFunction) unicode_capitalize, METH_NOARGS, capitalize__doc__},
+    {"casefold", (PyCFunction) unicode_casefold, METH_NOARGS, casefold__doc__},
     {"title", (PyCFunction) unicode_title, METH_NOARGS, title__doc__},
     {"center", (PyCFunction) unicode_center, METH_VARARGS, center__doc__},
     {"count", (PyCFunction) unicode_count, METH_VARARGS, count__doc__},