Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode. Fixes #931848.

Lib/sre_compile.py

@@ -308,14 +308,16 @@ def _optimize_unicode(charset, fixup):
             block = block + 1
             data = data + _mk_bitmap(chunk)
     header = [block]
-    if MAXCODE == 65535:
+    if _sre.CODESIZE == 2:
         code = 'H'
     else:
-        code = 'L'
+        code = 'I'
     # Convert block indices to byte array of 256 bytes
     mapping = array.array('b', mapping).tostring()
     # Convert byte array to word array
-    header = header + array.array(code, mapping).tolist()
+    mapping = array.array(code, mapping)
+    assert mapping.itemsize == _sre.CODESIZE
+    header = header + mapping.tolist()
     data[0:0] = header
     return [(BIGCHARSET, data)]
 

Misc/NEWS

@@ -42,6 +42,8 @@ Core and builtins
 Library
 -------
 
+- Bug #931848: Fix _sre.CODESIZE on 64-bit machines in UCS-4 mode.
+
 - Bug #938076: Consider output encoding in XMLGenerator.
 
 - Bug #936637: Properly delegate startElementNS in saxutils.XMLFilterBase.

Modules/sre.h

@@ -16,7 +16,7 @@
 /* size of a code word (must be unsigned short or larger, and
    large enough to hold a Py_UNICODE character) */
 #ifdef Py_UNICODE_WIDE
-#define SRE_CODE unsigned long
+#define SRE_CODE Py_UCS4
 #else
 #define SRE_CODE unsigned short
 #endif