test_pep277.py 7.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204
  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. filenames = [
  7. '1_abc',
  8. u'2_ascii',
  9. u'3_Gr\xfc\xdf-Gott',
  10. u'4_\u0393\u03b5\u03b9\u03ac-\u03c3\u03b1\u03c2',
  11. u'5_\u0417\u0434\u0440\u0430\u0432\u0441\u0442\u0432\u0443\u0439\u0442\u0435',
  12. u'6_\u306b\u307d\u3093',
  13. u'7_\u05d4\u05e9\u05e7\u05e6\u05e5\u05e1',
  14. u'8_\u66e8\u66e9\u66eb',
  15. u'9_\u66e8\u05e9\u3093\u0434\u0393\xdf',
  16. # Specific code points: fn, NFC(fn) and NFKC(fn) all differents
  17. u'10_\u1fee\u1ffd',
  18. ]
  19. # Mac OS X decomposes Unicode names, using Normal Form D.
  20. # http://developer.apple.com/mac/library/qa/qa2001/qa1173.html
  21. # "However, most volume formats do not follow the exact specification for
  22. # these normal forms. For example, HFS Plus uses a variant of Normal Form D
  23. # in which U+2000 through U+2FFF, U+F900 through U+FAFF, and U+2F800 through
  24. # U+2FAFF are not decomposed."
  25. if sys.platform != 'darwin':
  26. filenames.extend([
  27. # Specific code points: NFC(fn), NFD(fn), NFKC(fn) and NFKD(fn) all differents
  28. u'11_\u0385\u03d3\u03d4',
  29. u'12_\u00a8\u0301\u03d2\u0301\u03d2\u0308', # == NFD(u'\u0385\u03d3\u03d4')
  30. u'13_\u0020\u0308\u0301\u038e\u03ab', # == NFKC(u'\u0385\u03d3\u03d4')
  31. u'14_\u1e9b\u1fc1\u1fcd\u1fce\u1fcf\u1fdd\u1fde\u1fdf\u1fed',
  32. # Specific code points: fn, NFC(fn) and NFKC(fn) all differents
  33. u'15_\u1fee\u1ffd\ufad1',
  34. u'16_\u2000\u2000\u2000A',
  35. u'17_\u2001\u2001\u2001A',
  36. u'18_\u2003\u2003\u2003A', # == NFC(u'\u2001\u2001\u2001A')
  37. u'19_\u0020\u0020\u0020A', # u'\u0020' == u' ' == NFKC(u'\u2000') ==
  38. # NFKC(u'\u2001') == NFKC(u'\u2003')
  39. ])
  40. # Is it Unicode-friendly?
  41. if not os.path.supports_unicode_filenames:
  42. fsencoding = sys.getfilesystemencoding() or sys.getdefaultencoding()
  43. try:
  44. for name in filenames:
  45. name.encode(fsencoding)
  46. except UnicodeEncodeError:
  47. raise unittest.SkipTest("only NT+ and systems with "
  48. "Unicode-friendly filesystem encoding")
  49. # Destroy directory dirname and all files under it, to one level.
  50. def deltree(dirname):
  51. # Don't hide legitimate errors: if one of these suckers exists, it's
  52. # an error if we can't remove it.
  53. if os.path.exists(dirname):
  54. # must pass unicode to os.listdir() so we get back unicode results.
  55. for fname in os.listdir(unicode(dirname)):
  56. os.unlink(os.path.join(dirname, fname))
  57. os.rmdir(dirname)
  58. class UnicodeFileTests(unittest.TestCase):
  59. files = set(filenames)
  60. normal_form = None
  61. def setUp(self):
  62. try:
  63. os.mkdir(test_support.TESTFN)
  64. except OSError:
  65. pass
  66. files = set()
  67. for name in self.files:
  68. name = os.path.join(test_support.TESTFN, self.norm(name))
  69. with open(name, 'w') as f:
  70. f.write((name+'\n').encode("utf-8"))
  71. os.stat(name)
  72. files.add(name)
  73. self.files = files
  74. def tearDown(self):
  75. deltree(test_support.TESTFN)
  76. def norm(self, s):
  77. if self.normal_form and isinstance(s, unicode):
  78. return normalize(self.normal_form, s)
  79. return s
  80. def _apply_failure(self, fn, filename, expected_exception,
  81. check_fn_in_exception = True):
  82. with self.assertRaises(expected_exception) as c:
  83. fn(filename)
  84. exc_filename = c.exception.filename
  85. # the "filename" exception attribute may be encoded
  86. if isinstance(exc_filename, str):
  87. filename = filename.encode(sys.getfilesystemencoding())
  88. if check_fn_in_exception:
  89. self.assertEqual(exc_filename, filename, "Function '%s(%r) failed "
  90. "with bad filename in the exception: %r" %
  91. (fn.__name__, filename, exc_filename))
  92. def test_failures(self):
  93. # Pass non-existing Unicode filenames all over the place.
  94. for name in self.files:
  95. name = "not_" + name
  96. self._apply_failure(open, name, IOError)
  97. self._apply_failure(os.stat, name, OSError)
  98. self._apply_failure(os.chdir, name, OSError)
  99. self._apply_failure(os.rmdir, name, OSError)
  100. self._apply_failure(os.remove, name, OSError)
  101. # listdir may append a wildcard to the filename, so dont check
  102. self._apply_failure(os.listdir, name, OSError, False)
  103. def test_open(self):
  104. for name in self.files:
  105. f = open(name, 'w')
  106. f.write((name+'\n').encode("utf-8"))
  107. f.close()
  108. os.stat(name)
  109. # Skip the test on darwin, because darwin does normalize the filename to
  110. # NFD (a variant of Unicode NFD form). Normalize the filename to NFC, NFKC,
  111. # NFKD in Python is useless, because darwin will normalize it later and so
  112. # open(), os.stat(), etc. don't raise any exception.
  113. @unittest.skipIf(sys.platform == 'darwin', 'irrelevant test on Mac OS X')
  114. def test_normalize(self):
  115. files = set(f for f in self.files if isinstance(f, unicode))
  116. others = set()
  117. for nf in set(['NFC', 'NFD', 'NFKC', 'NFKD']):
  118. others |= set(normalize(nf, file) for file in files)
  119. others -= files
  120. for name in others:
  121. self._apply_failure(open, name, IOError)
  122. self._apply_failure(os.stat, name, OSError)
  123. self._apply_failure(os.chdir, name, OSError)
  124. self._apply_failure(os.rmdir, name, OSError)
  125. self._apply_failure(os.remove, name, OSError)
  126. # listdir may append a wildcard to the filename, so dont check
  127. self._apply_failure(os.listdir, name, OSError, False)
  128. # Skip the test on darwin, because darwin uses a normalization different
  129. # than Python NFD normalization: filenames are different even if we use
  130. # Python NFD normalization.
  131. @unittest.skipIf(sys.platform == 'darwin', 'irrelevant test on Mac OS X')
  132. def test_listdir(self):
  133. sf0 = set(self.files)
  134. f1 = os.listdir(test_support.TESTFN)
  135. f2 = os.listdir(unicode(test_support.TESTFN,
  136. sys.getfilesystemencoding()))
  137. sf2 = set(os.path.join(unicode(test_support.TESTFN), f) for f in f2)
  138. self.assertEqual(sf0, sf2)
  139. self.assertEqual(len(f1), len(f2))
  140. def test_rename(self):
  141. for name in self.files:
  142. os.rename(name, "tmp")
  143. os.rename("tmp", name)
  144. def test_directory(self):
  145. dirname = os.path.join(test_support.TESTFN,
  146. u'Gr\xfc\xdf-\u66e8\u66e9\u66eb')
  147. filename = u'\xdf-\u66e8\u66e9\u66eb'
  148. with test_support.temp_cwd(dirname):
  149. with open(filename, 'w') as f:
  150. f.write((filename + '\n').encode("utf-8"))
  151. os.access(filename,os.R_OK)
  152. os.remove(filename)
  153. class UnicodeNFCFileTests(UnicodeFileTests):
  154. normal_form = 'NFC'
  155. class UnicodeNFDFileTests(UnicodeFileTests):
  156. normal_form = 'NFD'
  157. class UnicodeNFKCFileTests(UnicodeFileTests):
  158. normal_form = 'NFKC'
  159. class UnicodeNFKDFileTests(UnicodeFileTests):
  160. normal_form = 'NFKD'
  161. def test_main():
  162. try:
  163. test_support.run_unittest(
  164. UnicodeFileTests,
  165. UnicodeNFCFileTests,
  166. UnicodeNFDFileTests,
  167. UnicodeNFKCFileTests,
  168. UnicodeNFKDFileTests,
  169. )
  170. finally:
  171. deltree(test_support.TESTFN)
  172. if __name__ == "__main__":
  173. test_main()