#5057: the peepholer no longer optimizes subscription on unicode literals (e.g. u"foo"[0]) in order to produce compatible pyc files between narrow and wide builds.

1 files changed, 7 insertions, 20 deletions

diff --git a/Python/peephole.c b/Python/peephole.c
index ae84efa..fb6cd03 100644
--- a/Python/peephole.c
+++ b/Python/peephole.c
@@ -128,27 +128,14 @@ fold_binops_on_constants(unsigned char *codestr, PyObject *consts)
             newconst = PyNumber_Subtract(v, w);
             break;
         case BINARY_SUBSCR:
-            newconst = PyObject_GetItem(v, w);
             /* #5057: if v is unicode, there might be differences between
-               wide and narrow builds in cases like u'\U00012345'[0].
-               Wide builds will return a non-BMP char, whereas narrow builds
-               will return a surrogate.  In both the cases skip the
-               optimization in order to produce compatible pycs.
-             */
-#ifdef Py_USING_UNICODE
-            if (newconst != NULL &&
-                PyUnicode_Check(v) && PyUnicode_Check(newconst)) {
-                Py_UNICODE ch = PyUnicode_AS_UNICODE(newconst)[0];
-#ifdef Py_UNICODE_WIDE
-                if (ch > 0xFFFF) {
-#else
-                if (ch >= 0xD800 && ch <= 0xDFFF) {
-#endif
-                    Py_DECREF(newconst);
-                    return 0;
-                }
-            }
-#endif
+               wide and narrow builds in cases like '\U00012345'[0] or
+               '\U00012345abcdef'[3], so it's better to skip the optimization
+               in order to produce compatible pycs.
+            */
+            if (PyUnicode_Check(v))
+                return 0;
+            newconst = PyObject_GetItem(v, w);
             break;
         case BINARY_LSHIFT:
             newconst = PyNumber_Lshift(v, w);