Home | History | Annotate | Download | only in test
      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
      3 import sys, os, unittest
      4 from unicodedata import normalize
      5 from test import test_support
      6 
      7 filenames = [
      8     '1_abc',
      9     u'2_ascii',
     10     u'3_Gr\xfc\xdf-Gott',
     11     u'4_\u0393\u03b5\u03b9\u03ac-\u03c3\u03b1\u03c2',
     12     u'5_\u0417\u0434\u0440\u0430\u0432\u0441\u0442\u0432\u0443\u0439\u0442\u0435',
     13     u'6_\u306b\u307d\u3093',
     14     u'7_\u05d4\u05e9\u05e7\u05e6\u05e5\u05e1',
     15     u'8_\u66e8\u66e9\u66eb',
     16     u'9_\u66e8\u05e9\u3093\u0434\u0393\xdf',
     17     # Specific code points: fn, NFC(fn) and NFKC(fn) all differents
     18     u'10_\u1fee\u1ffd',
     19     ]
     20 
     21 # Mac OS X decomposes Unicode names, using Normal Form D.
     22 # http://developer.apple.com/mac/library/qa/qa2001/qa1173.html
     23 # "However, most volume formats do not follow the exact specification for
     24 # these normal forms.  For example, HFS Plus uses a variant of Normal Form D
     25 # in which U+2000 through U+2FFF, U+F900 through U+FAFF, and U+2F800 through
     26 # U+2FAFF are not decomposed."
     27 if sys.platform != 'darwin':
     28     filenames.extend([
     29         # Specific code points: NFC(fn), NFD(fn), NFKC(fn) and NFKD(fn) all differents
     30         u'11_\u0385\u03d3\u03d4',
     31         u'12_\u00a8\u0301\u03d2\u0301\u03d2\u0308',    # == NFD(u'\u0385\u03d3\u03d4')
     32         u'13_\u0020\u0308\u0301\u038e\u03ab',          # == NFKC(u'\u0385\u03d3\u03d4')
     33         u'14_\u1e9b\u1fc1\u1fcd\u1fce\u1fcf\u1fdd\u1fde\u1fdf\u1fed',
     34 
     35         # Specific code points: fn, NFC(fn) and NFKC(fn) all differents
     36         u'15_\u1fee\u1ffd\ufad1',
     37         u'16_\u2000\u2000\u2000A',
     38         u'17_\u2001\u2001\u2001A',
     39         u'18_\u2003\u2003\u2003A', # == NFC(u'\u2001\u2001\u2001A')
     40         u'19_\u0020\u0020\u0020A', # u'\u0020' == u' ' == NFKC(u'\u2000') ==
     41                                    #   NFKC(u'\u2001') == NFKC(u'\u2003')
     42 ])
     43 
     44 
     45 # Is it Unicode-friendly?
     46 if not os.path.supports_unicode_filenames:
     47     fsencoding = sys.getfilesystemencoding() or sys.getdefaultencoding()
     48     try:
     49         for name in filenames:
     50             name.encode(fsencoding)
     51     except UnicodeEncodeError:
     52         raise unittest.SkipTest("only NT+ and systems with "
     53                                 "Unicode-friendly filesystem encoding")
     54 
     55 
     56 # Destroy directory dirname and all files under it, to one level.
     57 def deltree(dirname):
     58     # Don't hide legitimate errors:  if one of these suckers exists, it's
     59     # an error if we can't remove it.
     60     if os.path.exists(dirname):
     61         # must pass unicode to os.listdir() so we get back unicode results.
     62         for fname in os.listdir(unicode(dirname)):
     63             os.unlink(os.path.join(dirname, fname))
     64         os.rmdir(dirname)
     65 
     66 
     67 class UnicodeFileTests(unittest.TestCase):
     68     files = set(filenames)
     69     normal_form = None
     70 
     71     def setUp(self):
     72         try:
     73             os.mkdir(test_support.TESTFN)
     74         except OSError:
     75             pass
     76         files = set()
     77         for name in self.files:
     78             name = os.path.join(test_support.TESTFN, self.norm(name))
     79             with open(name, 'w') as f:
     80                 f.write((name+'\n').encode("utf-8"))
     81             os.stat(name)
     82             files.add(name)
     83         self.files = files
     84 
     85     def tearDown(self):
     86         deltree(test_support.TESTFN)
     87 
     88     def norm(self, s):
     89         if self.normal_form and isinstance(s, unicode):
     90             return normalize(self.normal_form, s)
     91         return s
     92 
     93     def _apply_failure(self, fn, filename, expected_exception,
     94                        check_fn_in_exception = True):
     95         with self.assertRaises(expected_exception) as c:
     96             fn(filename)
     97         exc_filename = c.exception.filename
     98         # the "filename" exception attribute may be encoded
     99         if isinstance(exc_filename, str):
    100             filename = filename.encode(sys.getfilesystemencoding())
    101         if check_fn_in_exception:
    102             self.assertEqual(exc_filename, filename, "Function '%s(%r) failed "
    103                              "with bad filename in the exception: %r" %
    104                              (fn.__name__, filename, exc_filename))
    105 
    106     def test_failures(self):
    107         # Pass non-existing Unicode filenames all over the place.
    108         for name in self.files:
    109             name = "not_" + name
    110             self._apply_failure(open, name, IOError)
    111             self._apply_failure(os.stat, name, OSError)
    112             self._apply_failure(os.chdir, name, OSError)
    113             self._apply_failure(os.rmdir, name, OSError)
    114             self._apply_failure(os.remove, name, OSError)
    115             # listdir may append a wildcard to the filename, so dont check
    116             self._apply_failure(os.listdir, name, OSError, False)
    117 
    118     def test_open(self):
    119         for name in self.files:
    120             f = open(name, 'w')
    121             f.write((name+'\n').encode("utf-8"))
    122             f.close()
    123             os.stat(name)
    124 
    125     # Skip the test on darwin, because darwin does normalize the filename to
    126     # NFD (a variant of Unicode NFD form). Normalize the filename to NFC, NFKC,
    127     # NFKD in Python is useless, because darwin will normalize it later and so
    128     # open(), os.stat(), etc. don't raise any exception.
    129     @unittest.skipIf(sys.platform == 'darwin', 'irrelevant test on Mac OS X')
    130     def test_normalize(self):
    131         files = set(f for f in self.files if isinstance(f, unicode))
    132         others = set()
    133         for nf in set(['NFC', 'NFD', 'NFKC', 'NFKD']):
    134             others |= set(normalize(nf, file) for file in files)
    135         others -= files
    136         for name in others:
    137             self._apply_failure(open, name, IOError)
    138             self._apply_failure(os.stat, name, OSError)
    139             self._apply_failure(os.chdir, name, OSError)
    140             self._apply_failure(os.rmdir, name, OSError)
    141             self._apply_failure(os.remove, name, OSError)
    142             # listdir may append a wildcard to the filename, so dont check
    143             self._apply_failure(os.listdir, name, OSError, False)
    144 
    145     # Skip the test on darwin, because darwin uses a normalization different
    146     # than Python NFD normalization: filenames are different even if we use
    147     # Python NFD normalization.
    148     @unittest.skipIf(sys.platform == 'darwin', 'irrelevant test on Mac OS X')
    149     def test_listdir(self):
    150         sf0 = set(self.files)
    151         f1 = os.listdir(test_support.TESTFN)
    152         f2 = os.listdir(unicode(test_support.TESTFN,
    153                                 sys.getfilesystemencoding()))
    154         sf2 = set(os.path.join(unicode(test_support.TESTFN), f) for f in f2)
    155         self.assertEqual(sf0, sf2)
    156         self.assertEqual(len(f1), len(f2))
    157 
    158     def test_rename(self):
    159         for name in self.files:
    160             os.rename(name, "tmp")
    161             os.rename("tmp", name)
    162 
    163     def test_directory(self):
    164         dirname = os.path.join(test_support.TESTFN,
    165                                u'Gr\xfc\xdf-\u66e8\u66e9\u66eb')
    166         filename = u'\xdf-\u66e8\u66e9\u66eb'
    167         oldwd = os.getcwd()
    168         os.mkdir(dirname)
    169         os.chdir(dirname)
    170         try:
    171             with open(filename, 'w') as f:
    172                 f.write((filename + '\n').encode("utf-8"))
    173             os.access(filename,os.R_OK)
    174             os.remove(filename)
    175         finally:
    176             os.chdir(oldwd)
    177             os.rmdir(dirname)
    178 
    179 
    180 class UnicodeNFCFileTests(UnicodeFileTests):
    181     normal_form = 'NFC'
    182 
    183 
    184 class UnicodeNFDFileTests(UnicodeFileTests):
    185     normal_form = 'NFD'
    186 
    187 
    188 class UnicodeNFKCFileTests(UnicodeFileTests):
    189     normal_form = 'NFKC'
    190 
    191 
    192 class UnicodeNFKDFileTests(UnicodeFileTests):
    193     normal_form = 'NFKD'
    194 
    195 
    196 def test_main():
    197     try:
    198         test_support.run_unittest(
    199             UnicodeFileTests,
    200             UnicodeNFCFileTests,
    201             UnicodeNFDFileTests,
    202             UnicodeNFKCFileTests,
    203             UnicodeNFKDFileTests,
    204         )
    205     finally:
    206         deltree(test_support.TESTFN)
    207 
    208 
    209 if __name__ == "__main__":
    210     test_main()
    211