1
# Test the Unicode versions of normal file functions
2
# open, os.open, os.stat. os.listdir, os.rename, os.remove, os.mkdir, os.chdir, os.rmdir
7
from unicodedata import normalize
8
from test import support
14
'4_\u0393\u03b5\u03b9\u03ac-\u03c3\u03b1\u03c2',
15
'5_\u0417\u0434\u0440\u0430\u0432\u0441\u0442\u0432\u0443\u0439\u0442\u0435',
16
'6_\u306b\u307d\u3093',
17
'7_\u05d4\u05e9\u05e7\u05e6\u05e5\u05e1',
18
'8_\u66e8\u66e9\u66eb',
19
'9_\u66e8\u05e9\u3093\u0434\u0393\xdf',
20
# Specific code points: fn, NFC(fn) and NFKC(fn) all differents
24
# Mac OS X decomposes Unicode names, using Normal Form D.
25
# http://developer.apple.com/mac/library/qa/qa2001/qa1173.html
26
# "However, most volume formats do not follow the exact specification for
27
# these normal forms. For example, HFS Plus uses a variant of Normal Form D
28
# in which U+2000 through U+2FFF, U+F900 through U+FAFF, and U+2F800 through
29
# U+2FAFF are not decomposed."
30
if sys.platform != 'darwin':
32
# Specific code points: NFC(fn), NFD(fn), NFKC(fn) and NFKD(fn) all differents
33
'11_\u0385\u03d3\u03d4',
34
'12_\u00a8\u0301\u03d2\u0301\u03d2\u0308', # == NFD('\u0385\u03d3\u03d4')
35
'13_\u0020\u0308\u0301\u038e\u03ab', # == NFKC('\u0385\u03d3\u03d4')
36
'14_\u1e9b\u1fc1\u1fcd\u1fce\u1fcf\u1fdd\u1fde\u1fdf\u1fed',
38
# Specific code points: fn, NFC(fn) and NFKC(fn) all differents
39
'15_\u1fee\u1ffd\ufad1',
40
'16_\u2000\u2000\u2000A',
41
'17_\u2001\u2001\u2001A',
42
'18_\u2003\u2003\u2003A', # == NFC('\u2001\u2001\u2001A')
43
'19_\u0020\u0020\u0020A', # '\u0020' == ' ' == NFKC('\u2000') ==
44
# NFKC('\u2001') == NFKC('\u2003')
48
# Is it Unicode-friendly?
49
if not os.path.supports_unicode_filenames:
50
fsencoding = sys.getfilesystemencoding()
52
for name in filenames:
53
name.encode(fsencoding)
54
except UnicodeEncodeError:
55
raise unittest.SkipTest("only NT+ and systems with "
56
"Unicode-friendly filesystem encoding")
59
# Destroy directory dirname and all files under it, to one level.
61
# Don't hide legitimate errors: if one of these suckers exists, it's
62
# an error if we can't remove it.
63
if os.path.exists(dirname):
64
# must pass unicode to os.listdir() so we get back unicode results.
65
for fname in os.listdir(str(dirname)):
66
os.unlink(os.path.join(dirname, fname))
70
class UnicodeFileTests(unittest.TestCase):
71
files = set(filenames)
76
os.mkdir(support.TESTFN)
77
except FileExistsError:
80
for name in self.files:
81
name = os.path.join(support.TESTFN, self.norm(name))
82
with open(name, 'wb') as f:
83
f.write((name+'\n').encode("utf-8"))
89
deltree(support.TESTFN)
93
return normalize(self.normal_form, s)
96
def _apply_failure(self, fn, filename,
97
expected_exception=FileNotFoundError,
99
with self.assertRaises(expected_exception) as c:
101
exc_filename = c.exception.filename
103
self.assertEqual(exc_filename, filename, "Function '%s(%a) failed "
104
"with bad filename in the exception: %a" %
105
(fn.__name__, filename, exc_filename))
107
def test_failures(self):
108
# Pass non-existing Unicode filenames all over the place.
109
for name in self.files:
111
self._apply_failure(open, name)
112
self._apply_failure(os.stat, name)
113
self._apply_failure(os.chdir, name)
114
self._apply_failure(os.rmdir, name)
115
self._apply_failure(os.remove, name)
116
self._apply_failure(os.listdir, name)
118
if sys.platform == 'win32':
119
# Windows is lunatic. Issue #13366.
120
_listdir_failure = NotADirectoryError, FileNotFoundError
122
_listdir_failure = NotADirectoryError
125
for name in self.files:
127
f.write((name+'\n').encode("utf-8"))
130
self._apply_failure(os.listdir, name, self._listdir_failure)
132
# Skip the test on darwin, because darwin does normalize the filename to
133
# NFD (a variant of Unicode NFD form). Normalize the filename to NFC, NFKC,
134
# NFKD in Python is useless, because darwin will normalize it later and so
135
# open(), os.stat(), etc. don't raise any exception.
136
@unittest.skipIf(sys.platform == 'darwin', 'irrelevant test on Mac OS X')
137
def test_normalize(self):
138
files = set(self.files)
140
for nf in set(['NFC', 'NFD', 'NFKC', 'NFKD']):
141
others |= set(normalize(nf, file) for file in files)
144
self._apply_failure(open, name)
145
self._apply_failure(os.stat, name)
146
self._apply_failure(os.chdir, name)
147
self._apply_failure(os.rmdir, name)
148
self._apply_failure(os.remove, name)
149
self._apply_failure(os.listdir, name)
151
# Skip the test on darwin, because darwin uses a normalization different
152
# than Python NFD normalization: filenames are different even if we use
153
# Python NFD normalization.
154
@unittest.skipIf(sys.platform == 'darwin', 'irrelevant test on Mac OS X')
155
def test_listdir(self):
156
sf0 = set(self.files)
157
with warnings.catch_warnings():
158
warnings.simplefilter("ignore", DeprecationWarning)
159
f1 = os.listdir(support.TESTFN.encode(sys.getfilesystemencoding()))
160
f2 = os.listdir(support.TESTFN)
161
sf2 = set(os.path.join(support.TESTFN, f) for f in f2)
162
self.assertEqual(sf0, sf2, "%a != %a" % (sf0, sf2))
163
self.assertEqual(len(f1), len(f2))
165
def test_rename(self):
166
for name in self.files:
167
os.rename(name, "tmp")
168
os.rename("tmp", name)
170
def test_directory(self):
171
dirname = os.path.join(support.TESTFN, 'Gr\xfc\xdf-\u66e8\u66e9\u66eb')
172
filename = '\xdf-\u66e8\u66e9\u66eb'
177
with open(filename, 'wb') as f:
178
f.write((filename + '\n').encode("utf-8"))
179
os.access(filename,os.R_OK)
186
class UnicodeNFCFileTests(UnicodeFileTests):
190
class UnicodeNFDFileTests(UnicodeFileTests):
194
class UnicodeNFKCFileTests(UnicodeFileTests):
198
class UnicodeNFKDFileTests(UnicodeFileTests):
204
support.run_unittest(
208
UnicodeNFKCFileTests,
209
UnicodeNFKDFileTests,
212
deltree(support.TESTFN)
215
if __name__ == "__main__":