[pypy-commit] pypy unicode-utf8: Raise ValueError when array item is invalid unicode

[rlamy]

Author: Ronan Lamy <ronan.lamy at gmail.com>
Branch: unicode-utf8
Changeset: r93379:bdbea29862af
Date: 2017-12-12 02:44 +0000
http://bitbucket.org/pypy/pypy/changeset/bdbea29862af/

Log:	Raise ValueError when array item is invalid unicode

diff --git a/pypy/module/array/interp_array.py b/pypy/module/array/interp_array.py
--- a/pypy/module/array/interp_array.py
+++ b/pypy/module/array/interp_array.py
@@ -1030,7 +1030,11 @@
                 return space.newbytes(item)
             elif mytype.typecode == 'u':
                 code = r_uint(ord(item))
-                return space.newutf8(rutf8.unichr_as_utf8(code), 1)
+                try:
+                    return space.newutf8(rutf8.unichr_as_utf8(code), 1)
+                except ValueError:
+                    raise oefmt(space.w_ValueError,
+                        "character is not in range [U+0000; U+10ffff]")
             assert 0, "unreachable"
 
         # interface