From 6f80f5d4446f06d15274ad519cae6929a3565cc0 Mon Sep 17 00:00:00 2001 From: Antoine Pitrou Date: Sun, 23 Sep 2012 19:55:21 +0200 Subject: [PATCH] Issue #15379: Fix passing of non-BMP characters as integers for the charmap decoder (already working as unicode strings). Patch by Serhiy Storchaka. --- Lib/test/test_codeccallbacks.py | 2 +- Lib/test/test_codecs.py | 105 ++++++++++++++++++++++++++++++++ Misc/NEWS | 3 + Objects/unicodeobject.c | 28 ++++++++- 4 files changed, 135 insertions(+), 3 deletions(-) diff --git a/Lib/test/test_codeccallbacks.py b/Lib/test/test_codeccallbacks.py index c5b1e25c208..e656d2fe8f7 100644 --- a/Lib/test/test_codeccallbacks.py +++ b/Lib/test/test_codeccallbacks.py @@ -744,7 +744,7 @@ class CodecCallbackTest(unittest.TestCase): raise ValueError self.assertRaises(UnicodeError, codecs.charmap_decode, b"\xff", "strict", {0xff: None}) self.assertRaises(ValueError, codecs.charmap_decode, b"\xff", "strict", D()) - self.assertRaises(TypeError, codecs.charmap_decode, b"\xff", "strict", {0xff: sys.maxunicode+1}) + self.assertRaises(TypeError, codecs.charmap_decode, b"\xff", "strict", {0xff: 0x110000}) def test_encodehelper(self): # enhance coverage of: diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py index 3426a4dde9a..f342d88b9f6 100644 --- a/Lib/test/test_codecs.py +++ b/Lib/test/test_codecs.py @@ -1546,6 +1546,10 @@ class CharmapTest(unittest.TestCase): ("abc", 3) ) + self.assertRaises(UnicodeDecodeError, + codecs.charmap_decode, b"\x00\x01\x02", "strict", "ab" + ) + self.assertEqual( codecs.charmap_decode(b"\x00\x01\x02", "replace", "ab"), ("ab\ufffd", 3) @@ -1572,6 +1576,107 @@ class CharmapTest(unittest.TestCase): ("", len(allbytes)) ) + def test_decode_with_int2str_map(self): + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "strict", + {0: 'a', 1: 'b', 2: 'c'}), + ("abc", 3) + ) + + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "strict", + {0: 'Aa', 1: 'Bb', 2: 'Cc'}), + ("AaBbCc", 3) + ) + + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "strict", + {0: '\U0010FFFF', 1: 'b', 2: 'c'}), + ("\U0010FFFFbc", 3) + ) + + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "strict", + {0: 'a', 1: 'b', 2: ''}), + ("ab", 3) + ) + + self.assertRaises(UnicodeDecodeError, + codecs.charmap_decode, b"\x00\x01\x02", "strict", + {0: 'a', 1: 'b'} + ) + + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "replace", + {0: 'a', 1: 'b'}), + ("ab\ufffd", 3) + ) + + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "replace", + {0: 'a', 1: 'b', 2: None}), + ("ab\ufffd", 3) + ) + + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "ignore", + {0: 'a', 1: 'b'}), + ("ab", 3) + ) + + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "ignore", + {0: 'a', 1: 'b', 2: None}), + ("ab", 3) + ) + + allbytes = bytes(range(256)) + self.assertEqual( + codecs.charmap_decode(allbytes, "ignore", {}), + ("", len(allbytes)) + ) + + def test_decode_with_int2int_map(self): + a = ord('a') + b = ord('b') + c = ord('c') + + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "strict", + {0: a, 1: b, 2: c}), + ("abc", 3) + ) + + # Issue #15379 + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "strict", + {0: 0x10FFFF, 1: b, 2: c}), + ("\U0010FFFFbc", 3) + ) + + self.assertRaises(TypeError, + codecs.charmap_decode, b"\x00\x01\x02", "strict", + {0: 0x110000, 1: b, 2: c} + ) + + self.assertRaises(UnicodeDecodeError, + codecs.charmap_decode, b"\x00\x01\x02", "strict", + {0: a, 1: b}, + ) + + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "replace", + {0: a, 1: b}), + ("ab\ufffd", 3) + ) + + self.assertEqual( + codecs.charmap_decode(b"\x00\x01\x02", "ignore", + {0: a, 1: b}), + ("ab", 3) + ) + + class WithStmtTest(unittest.TestCase): def test_encodedfile(self): f = io.BytesIO(b"\xc3\xbc") diff --git a/Misc/NEWS b/Misc/NEWS index 74782aac68c..57ed006fa76 100644 --- a/Misc/NEWS +++ b/Misc/NEWS @@ -10,6 +10,9 @@ What's New in Python 3.2.4 Core and Builtins ----------------- +- Issue #15379: Fix passing of non-BMP characters as integers for the charmap + decoder (already working as unicode strings). Patch by Serhiy Storchaka. + - Issue #13992: The trashcan mechanism is now thread-safe. This eliminates sporadic crashes in multi-thread programs when several long deallocator chains ran concurrently and involved subclasses of built-in container diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c index 8b782b4065e..f59db36dd3d 100644 --- a/Objects/unicodeobject.c +++ b/Objects/unicodeobject.c @@ -5250,12 +5250,36 @@ PyObject *PyUnicode_DecodeCharmap(const char *s, /* Apply mapping */ if (PyLong_Check(x)) { long value = PyLong_AS_LONG(x); - if (value < 0 || value > 65535) { + if (value < 0 || value > 0x10FFFF) { PyErr_SetString(PyExc_TypeError, - "character mapping must be in range(65536)"); + "character mapping must be in range(0x110000)"); Py_DECREF(x); goto onError; } + +#ifndef Py_UNICODE_WIDE + if (value > 0xFFFF) { + /* see the code for 1-n mapping below */ + if (extrachars < 2) { + /* resize first */ + Py_ssize_t oldpos = p - PyUnicode_AS_UNICODE(v); + Py_ssize_t needed = 10 - extrachars; + extrachars += needed; + /* XXX overflow detection missing */ + if (_PyUnicode_Resize(&v, + PyUnicode_GET_SIZE(v) + needed) < 0) { + Py_DECREF(x); + goto onError; + } + p = PyUnicode_AS_UNICODE(v) + oldpos; + } + value -= 0x10000; + *p++ = 0xD800 | (value >> 10); + *p++ = 0xDC00 | (value & 0x3FF); + extrachars -= 2; + } + else +#endif *p++ = (Py_UNICODE)value; } else if (x == Py_None) {