docutils/test/test_error_reporting.py

   1 #! /usr/bin/env python
   2 # .. coding: utf-8
   3 # $Id$
   4 # Author: Günter Milde <milde@users.sourceforge.net>
   5 # Copyright: This module has been placed in the public domain.
   6
   7 """
   8 Test `EnvironmentError` reporting.
   9
  10 In some locales, the `errstr` argument of IOError and OSError contains
  11 non-ASCII chars.
  12
  13 In Python 2, converting an exception instance to `str` or `unicode`
  14 might fail, with non-ASCII chars in arguments and the default encoding
  15 and errors ('ascii', 'strict').
  16
  17 Therefore, Docutils must not use string interpolation with exception
  18 instances like, e.g., ::
  19
  20   try:
  21     something
  22   except IOError, error:
  23     print 'Found %s' % error
  24
  25 unless the minimal required Python version has this problem fixed.
  26 """
  27
  28 import unittest
  29 import sys, os
  30 import codecs
  31 try: # from standard library module `io`
  32     from io import StringIO, BytesIO
  33 except ImportError: # new in Python 2.6
  34     from StringIO import StringIO
  35     BytesIO = StringIO
  36
  37 import DocutilsTestSupport              # must be imported before docutils
  38 from docutils import core, parsers, frontend, utils
  39 from docutils.utils.error_reporting import SafeString, ErrorString, ErrorOutput
  40 from docutils._compat import b, bytes
  41
  42 oldlocale = None
  43 if sys.version_info < (3,0): # problems solved in py3k
  44     try:
  45         import locale # module missing in Jython
  46         oldlocale = locale.getlocale()
  47         # Why does getlocale return the defaultlocale in Python 3.2 ????
  48         # oldlocale = (None, None) # test suite runs without locale
  49     except ImportError:
  50         print ('cannot test error reporting with problematic locales,\n'
  51             '`import locale` failed.')
  52
  53
  54 # locales confirmed to use non-ASCII chars in the IOError message
  55 # for a missing file (https://bugs.gentoo.org/show_bug.cgi?id=349101)
  56 # TODO: add more confirmed problematic locales
  57 problematic_locales = ['cs_CZ', 'cs_CZ.UTF8',
  58                        'el_GR', 'el_GR.UTF-8',
  59                        # 'fr_FR.UTF-8', # only OSError
  60                        'ja_JP.UTF-8',
  61                        'ru_RU', 'ru_RU.KOI8-R',
  62                        'ru_RU.UTF-8',
  63                        '',  # default locale: might be non-problematic
  64                        ]
  65
  66 if oldlocale is not None:
  67     # find a supported problematic locale:
  68     for testlocale in problematic_locales:
  69         try:
  70             locale.setlocale(locale.LC_ALL, testlocale)
  71         except locale.Error:
  72             testlocale = None
  73         else:
  74             break
  75     locale.setlocale(locale.LC_ALL, oldlocale) # reset
  76 else:
  77     testlocale = None
  78
  79 class SafeStringTests(unittest.TestCase):
  80     # the error message in EnvironmentError instances comes from the OS
  81     # and in some locales (e.g. ru_RU), contains high bit chars.
  82     # -> see the test in test_error_reporting.py
  83
  84     # test data:
  85     bs = b('\xfc')     # unicode(bs) fails, str(bs) in Python 3 return repr()
  86     us = u'\xfc'       # bytes(us) fails; str(us) fails in Python 2
  87     be = Exception(bs) # unicode(be) fails
  88     ue = Exception(us) # bytes(ue) fails, str(ue) fails in Python 2;
  89                        # unicode(ue) fails in Python < 2.6 (issue2517_)
  90                        # .. _issue2517: http://bugs.python.org/issue2517
  91     # wrapped test data:
  92     wbs = SafeString(bs)
  93     wus = SafeString(us)
  94     wbe = SafeString(be)
  95     wue = SafeString(ue)
  96
  97     def test_7bit(self):
  98         # wrapping (not required with 7-bit chars) must not change the
  99         # result of conversions:
 100         bs7 = b('foo')
 101         us7 = u'foo'
 102         be7 = Exception(bs7)
 103         ue7 = Exception(us7)
 104         self.assertEqual(str(42), str(SafeString(42)))
 105         self.assertEqual(str(bs7), str(SafeString(bs7)))
 106         self.assertEqual(str(us7), str(SafeString(us7)))
 107         self.assertEqual(str(be7), str(SafeString(be7)))
 108         self.assertEqual(str(ue7), str(SafeString(ue7)))
 109         self.assertEqual(unicode(7), unicode(SafeString(7)))
 110         self.assertEqual(unicode(bs7), unicode(SafeString(bs7)))
 111         self.assertEqual(unicode(us7), unicode(SafeString(us7)))
 112         self.assertEqual(unicode(be7), unicode(SafeString(be7)))
 113         self.assertEqual(unicode(ue7), unicode(SafeString(ue7)))
 114
 115     def test_ustr(self):
 116         """Test conversion to a unicode-string."""
 117         # unicode(self.bs) fails
 118         self.assertEqual(unicode, type(unicode(self.wbs)))
 119         self.assertEqual(unicode(self.us), unicode(self.wus))
 120         # unicode(self.be) fails
 121         self.assertEqual(unicode, type(unicode(self.wbe)))
 122         # unicode(ue) fails in Python < 2.6 (issue2517_)
 123         self.assertEqual(unicode, type(unicode(self.wue)))
 124         self.assertEqual(self.us, unicode(self.wue))
 125
 126     def test_str(self):
 127         """Test conversion to a string (bytes in Python 2, unicode in Python 3)."""
 128         self.assertEqual(str(self.bs), str(self.wbs))
 129         self.assertEqual(str(self.be), str(self.be))
 130         # str(us) fails in Python 2
 131         self.assertEqual(str, type(str(self.wus)))
 132         # str(ue) fails in Python 2
 133         self.assertEqual(str, type(str(self.wue)))
 134
 135
 136 class ErrorStringTests(unittest.TestCase):
 137     bs = b('\xfc')     # unicode(bs) fails, str(bs) in Python 3 return repr()
 138     us = u'\xfc'       # bytes(us) fails; str(us) fails in Python 2
 139
 140     def test_str(self):
 141         self.assertEqual('Exception: spam',
 142                          str(ErrorString(Exception('spam'))))
 143         self.assertEqual('IndexError: '+str(self.bs),
 144                          str(ErrorString(IndexError(self.bs))))
 145         self.assertEqual('ImportError: %s' % SafeString(self.us),
 146                          str(ErrorString(ImportError(self.us))))
 147
 148     def test_unicode(self):
 149         self.assertEqual(u'Exception: spam',
 150                          unicode(ErrorString(Exception(u'spam'))))
 151         self.assertEqual(u'IndexError: '+self.us,
 152                          unicode(ErrorString(IndexError(self.us))))
 153         self.assertEqual(u'ImportError: %s' % SafeString(self.bs),
 154                          unicode(ErrorString(ImportError(self.bs))))
 155
 156
 157 # ErrorOutput tests
 158 # -----------------
 159
 160 # Stub: Buffer with 'strict' auto-conversion of input to byte string:
 161 class BBuf(BytesIO, object): # super class object required by Python <= 2.5
 162     def write(self, data):
 163         if isinstance(data, unicode):
 164             data.encode('ascii', 'strict')
 165         super(BBuf, self).write(data)
 166
 167 # Stub: Buffer expecting unicode string:
 168 class UBuf(StringIO, object): # super class object required by Python <= 2.5
 169     def write(self, data):
 170         # emulate Python 3 handling of stdout, stderr
 171         if isinstance(data, bytes):
 172             raise TypeError('must be unicode, not bytes')
 173         super(UBuf, self).write(data)
 174
 175 class ErrorOutputTests(unittest.TestCase):
 176     def test_defaults(self):
 177         e = ErrorOutput()
 178         self.assertEqual(e.stream, sys.stderr)
 179
 180     def test_bbuf(self):
 181         buf = BBuf() # buffer storing byte string
 182         e = ErrorOutput(buf, encoding='ascii')
 183         # write byte-string as-is
 184         e.write(b('b\xfc'))
 185         self.assertEqual(buf.getvalue(), b('b\xfc'))
 186         # encode unicode data with backslashescape fallback replacement:
 187         e.write(u' u\xfc')
 188         self.assertEqual(buf.getvalue(), b('b\xfc u\\xfc'))
 189         # handle Exceptions with Unicode string args
 190         # unicode(Exception(u'e\xfc')) # fails in Python < 2.6
 191         e.write(AttributeError(u' e\xfc'))
 192         self.assertEqual(buf.getvalue(), b('b\xfc u\\xfc e\\xfc'))
 193         # encode with `encoding` attribute
 194         e.encoding = 'utf8'
 195         e.write(u' u\xfc')
 196         self.assertEqual(buf.getvalue(), b('b\xfc u\\xfc e\\xfc u\xc3\xbc'))
 197
 198     def test_ubuf(self):
 199         buf = UBuf() # buffer only accepting unicode string
 200         # decode of binary strings
 201         e = ErrorOutput(buf, encoding='ascii')
 202         e.write(b('b\xfc'))
 203         self.assertEqual(buf.getvalue(), u'b\ufffd') # use REPLACEMENT CHARACTER
 204         # write Unicode string and Exceptions with Unicode args
 205         e.write(u' u\xfc')
 206         self.assertEqual(buf.getvalue(), u'b\ufffd u\xfc')
 207         e.write(AttributeError(u' e\xfc'))
 208         self.assertEqual(buf.getvalue(), u'b\ufffd u\xfc e\xfc')
 209         # decode with `encoding` attribute
 210         e.encoding = 'latin1'
 211         e.write(b(' b\xfc'))
 212         self.assertEqual(buf.getvalue(), u'b\ufffd u\xfc e\xfc b\xfc')
 213
 214
 215
 216 class SafeStringTests_locale(unittest.TestCase):
 217     """
 218     Test docutils.SafeString with 'problematic' locales.
 219
 220     The error message in `EnvironmentError` instances comes from the OS
 221     and in some locales (e.g. ru_RU), contains high bit chars.
 222     """
 223     if testlocale:
 224         locale.setlocale(locale.LC_ALL, testlocale)
 225     # test data:
 226     bs = b('\xfc')
 227     us = u'\xfc'
 228     try:
 229         open(b('\xfc'))
 230     except IOError, e: # in Python 3 the name for the exception instance
 231         bioe = e       # is local to the except clause
 232     try:
 233         open(u'\xfc')
 234     except IOError, e:
 235         uioe = e
 236     except UnicodeEncodeError:
 237         try:
 238             open(u'\xfc'.encode(sys.getfilesystemencoding(), 'replace'))
 239         except IOError, e:
 240             uioe = e
 241     try:
 242         os.chdir(b('\xfc'))
 243     except OSError, e:
 244         bose = e
 245     try:
 246         os.chdir(u'\xfc')
 247     except OSError, e:
 248         uose = e
 249     except UnicodeEncodeError:
 250         try:
 251             os.chdir(u'\xfc'.encode(sys.getfilesystemencoding(), 'replace'))
 252         except OSError, e:
 253             uose = e
 254     # wrapped test data:
 255     wbioe = SafeString(bioe)
 256     wuioe = SafeString(uioe)
 257     wbose = SafeString(bose)
 258     wuose = SafeString(uose)
 259     # reset locale
 260     if testlocale:
 261         locale.setlocale(locale.LC_ALL, oldlocale)
 262
 263     def test_ustr(self):
 264         """Test conversion to a unicode-string."""
 265         # unicode(bioe) fails with e.g. 'ru_RU.utf8' locale
 266         self.assertEqual(unicode, type(unicode(self.wbioe)))
 267         self.assertEqual(unicode, type(unicode(self.wuioe)))
 268         self.assertEqual(unicode, type(unicode(self.wbose)))
 269         self.assertEqual(unicode, type(unicode(self.wuose)))
 270
 271     def test_str(self):
 272         """Test conversion to a string (bytes in Python 2, unicode in Python 3)."""
 273         self.assertEqual(str(self.bioe), str(self.wbioe))
 274         self.assertEqual(str(self.uioe), str(self.wuioe))
 275         self.assertEqual(str(self.bose), str(self.wbose))
 276         self.assertEqual(str(self.uose), str(self.wuose))
 277
 278
 279
 280 class ErrorReportingTests(unittest.TestCase):
 281     """
 282     Test cases where error reporting can go wrong.
 283
 284     Do not test the exact output (as this varies with the locale), just
 285     ensure that the correct exception is thrown.
 286     """
 287
 288     # These tests fail with a 'problematic locale',
 289     # Docutils revision < 7035, and Python 2:
 290
 291     parser = parsers.rst.Parser()
 292     """Parser shared by all ParserTestCases."""
 293
 294     option_parser = frontend.OptionParser(components=(parsers.rst.Parser,))
 295     settings = option_parser.get_default_values()
 296     settings.report_level = 1
 297     settings.halt_level = 1
 298     settings.warning_stream = ''
 299     document = utils.new_document('test data', settings)
 300
 301     def setUp(self):
 302         if testlocale:
 303             locale.setlocale(locale.LC_ALL, testlocale)
 304
 305     def tearDown(self):
 306         if testlocale:
 307             locale.setlocale(locale.LC_ALL, oldlocale)
 308
 309     def test_include(self):
 310         source = ('.. include:: bogus.txt')
 311         self.assertRaises(utils.SystemMessage,
 312                           self.parser.parse, source, self.document)
 313
 314     def test_raw_file(self):
 315         source = ('.. raw:: html\n'
 316                   '   :file: bogus.html\n')
 317         self.assertRaises(utils.SystemMessage,
 318                           self.parser.parse, source, self.document)
 319
 320     def test_csv_table(self):
 321         source = ('.. csv-table:: external file\n'
 322                   '   :file: bogus.csv\n')
 323         self.assertRaises(utils.SystemMessage,
 324                           self.parser.parse, source, self.document)
 325
 326 if __name__ == '__main__':
 327     unittest.main()