#9200: merge with 3.2.

diff --cc Lib/test/test_unicode.py
Simple merge

diff --cc Misc/NEWS
Simple merge

diff --cc Objects/unicodeobject.c
index 9d8fcb54e947eaa6d71e2bd258e3a4fb49dc3487,8567a9f2db14a2d0c78413df9eea88dd19703952..94b7b61802fb7d69d7d470eb66cf978eac2b21f4
--- 1/Objects/unicodeobject.c
--- 2/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@@ -6514,11 -6442,41 +6514,33 @@@ PyUnicode_EncodeDecimal(Py_UNICODE *s
              start = 0;                          \
      }
  
 -
 -/* helper macros used by _Py_UNICODE_NEXT */
 -#define _Py_UNICODE_IS_HIGH_SURROGATE(ch) (0xD800 <= ch && ch <= 0xDBFF)
 -#define _Py_UNICODE_IS_LOW_SURROGATE(ch) (0xDC00 <= ch && ch <= 0xDFFF)
 -/* Join two surrogate characters and return a single Py_UCS4 value. */
 -#define _Py_UNICODE_JOIN_SURROGATES(high, low)  \
 -    (((((Py_UCS4)(high) & 0x03FF) << 10) |      \
 -      ((Py_UCS4)(low) & 0x03FF)) + 0x10000)
 -
+ /* _Py_UNICODE_NEXT is a private macro used to retrieve the character pointed
+  * by 'ptr', possibly combining surrogate pairs on narrow builds.
+  * 'ptr' and 'end' must be Py_UNICODE*, with 'ptr' pointing at the character
+  * that should be returned and 'end' pointing to the end of the buffer.
+  * ('end' is used on narrow builds to detect a lone surrogate at the
+  * end of the buffer that should be returned unchanged.)
+  * The ptr and end arguments should be side-effect free and ptr must an lvalue.
+  * The type of the returned char is always Py_UCS4.
+  *
+  * Note: the macro advances ptr to next char, so it might have side-effects
+  *       (especially if used with other macros).
+  */
 -#define _Py_UNICODE_NEXT(ptr, end)                                      \
 -     (((_Py_UNICODE_IS_HIGH_SURROGATE(*(ptr)) && (ptr) < (end)) &&      \
 -        _Py_UNICODE_IS_LOW_SURROGATE((ptr)[1])) ?                       \
 -       ((ptr) += 2,_Py_UNICODE_JOIN_SURROGATES((ptr)[-2], (ptr)[-1])) : \
+ #ifdef Py_UNICODE_WIDE
+ #define _Py_UNICODE_NEXT(ptr, end) *(ptr)++
+ #else
 -Py_ssize_t PyUnicode_Count(PyObject *str,
 -                           PyObject *substr,
 -                           Py_ssize_t start,
 -                           Py_ssize_t end)
++#define _Py_UNICODE_NEXT(ptr, end)                                     \
++     (((Py_UNICODE_IS_HIGH_SURROGATE(*(ptr)) && (ptr) < (end)) &&      \
++        Py_UNICODE_IS_LOW_SURROGATE((ptr)[1])) ?                       \
++       ((ptr) += 2,Py_UNICODE_JOIN_SURROGATES((ptr)[-2], (ptr)[-1])) : \
+        (Py_UCS4)*(ptr)++)
+ #endif
+ 
 +Py_ssize_t
 +PyUnicode_Count(PyObject *str,
 +                PyObject *substr,
 +                Py_ssize_t start,
 +                Py_ssize_t end)
  {
      Py_ssize_t result;
      PyUnicodeObject* str_obj;