Issue #14850: Now a chamap decoder treates U+FFFE as "undefined mapping"

in any mapping, not only in an unicode string.
This commit is contained in:
Serhiy Storchaka 2013-01-15 14:42:59 +02:00
parent 90b5d9288d
commit 9599745e2c
3 changed files with 79 additions and 26 deletions

View file

@ -1551,6 +1551,14 @@ class CharmapTest(unittest.TestCase):
(u"abc", 3)
)
self.assertRaises(UnicodeDecodeError,
codecs.charmap_decode, b"\x00\x01\x02", "strict", u"ab"
)
self.assertRaises(UnicodeDecodeError,
codecs.charmap_decode, "\x00\x01\x02", "strict", u"ab\ufffe"
)
self.assertEqual(
codecs.charmap_decode("\x00\x01\x02", "replace", u"ab"),
(u"ab\ufffd", 3)
@ -1566,10 +1574,6 @@ class CharmapTest(unittest.TestCase):
(u"ab", 3)
)
self.assertRaises(UnicodeDecodeError,
codecs.charmap_decode, b"\x00\x01\x02", "strict", u"ab"
)
self.assertEqual(
codecs.charmap_decode("\x00\x01\x02", "ignore", u"ab\ufffe"),
(u"ab", 3)
@ -1611,6 +1615,17 @@ class CharmapTest(unittest.TestCase):
{0: u'a', 1: u'b'}
)
self.assertRaises(UnicodeDecodeError,
codecs.charmap_decode, "\x00\x01\x02", "strict",
{0: u'a', 1: u'b', 2: None}
)
# Issue #14850
self.assertRaises(UnicodeDecodeError,
codecs.charmap_decode, "\x00\x01\x02", "strict",
{0: u'a', 1: u'b', 2: u'\ufffe'}
)
self.assertEqual(
codecs.charmap_decode("\x00\x01\x02", "replace",
{0: u'a', 1: u'b'}),
@ -1623,6 +1638,13 @@ class CharmapTest(unittest.TestCase):
(u"ab\ufffd", 3)
)
# Issue #14850
self.assertEqual(
codecs.charmap_decode("\x00\x01\x02", "replace",
{0: u'a', 1: u'b', 2: u'\ufffe'}),
(u"ab\ufffd", 3)
)
self.assertEqual(
codecs.charmap_decode("\x00\x01\x02", "ignore",
{0: u'a', 1: u'b'}),
@ -1635,7 +1657,14 @@ class CharmapTest(unittest.TestCase):
(u"ab", 3)
)
allbytes = bytes(range(256))
# Issue #14850
self.assertEqual(
codecs.charmap_decode("\x00\x01\x02", "ignore",
{0: u'a', 1: u'b', 2: u'\ufffe'}),
(u"ab", 3)
)
allbytes = "".join(chr(i) for i in xrange(256))
self.assertEqual(
codecs.charmap_decode(allbytes, "ignore", {}),
(u"", len(allbytes))
@ -1669,18 +1698,35 @@ class CharmapTest(unittest.TestCase):
{0: a, 1: b},
)
self.assertRaises(UnicodeDecodeError,
codecs.charmap_decode, "\x00\x01\x02", "strict",
{0: a, 1: b, 2: 0xFFFE},
)
self.assertEqual(
codecs.charmap_decode("\x00\x01\x02", "replace",
{0: a, 1: b}),
(u"ab\ufffd", 3)
)
self.assertEqual(
codecs.charmap_decode("\x00\x01\x02", "replace",
{0: a, 1: b, 2: 0xFFFE}),
(u"ab\ufffd", 3)
)
self.assertEqual(
codecs.charmap_decode("\x00\x01\x02", "ignore",
{0: a, 1: b}),
(u"ab", 3)
)
self.assertEqual(
codecs.charmap_decode("\x00\x01\x02", "ignore",
{0: a, 1: b, 2: 0xFFFE}),
(u"ab", 3)
)
class WithStmtTest(unittest.TestCase):
def test_encodedfile(self):