2021-04-12 09:17:40 -03:00
|
|
|
from decimal import Decimal
|
2022-05-16 11:02:37 -03:00
|
|
|
from test.support import verbose, is_android, is_emscripten, is_wasi
|
2020-07-09 10:25:10 -03:00
|
|
|
from test.support.warnings_helper import check_warnings
|
2023-06-06 11:55:21 -03:00
|
|
|
from test.support.import_helper import import_fresh_module
|
|
|
|
from unittest import mock
|
2018-06-25 21:11:06 -03:00
|
|
|
import unittest
|
2001-04-13 05:09:50 -03:00
|
|
|
import locale
|
2001-04-15 10:15:56 -03:00
|
|
|
import sys
|
2018-06-25 21:11:06 -03:00
|
|
|
import codecs
|
2019-07-01 13:28:25 -03:00
|
|
|
|
2001-04-13 05:09:50 -03:00
|
|
|
|
2008-07-26 18:02:53 -03:00
|
|
|
class BaseLocalizedTest(unittest.TestCase):
|
|
|
|
#
|
|
|
|
# Base class for tests using a real locale
|
|
|
|
#
|
2008-07-25 18:45:08 -03:00
|
|
|
|
2013-07-17 07:23:45 -03:00
|
|
|
@classmethod
|
|
|
|
def setUpClass(cls):
|
2018-06-25 21:11:06 -03:00
|
|
|
if sys.platform == 'darwin':
|
2013-07-17 07:23:45 -03:00
|
|
|
import os
|
|
|
|
tlocs = ("en_US.UTF-8", "en_US.ISO8859-1", "en_US")
|
|
|
|
if int(os.uname().release.split('.')[0]) < 10:
|
|
|
|
# The locale test work fine on OSX 10.6, I (ronaldoussoren)
|
|
|
|
# haven't had time yet to verify if tests work on OSX 10.5
|
|
|
|
# (10.4 is known to be bad)
|
|
|
|
raise unittest.SkipTest("Locale support on MacOSX is minimal")
|
|
|
|
elif sys.platform.startswith("win"):
|
|
|
|
tlocs = ("En", "English")
|
|
|
|
else:
|
|
|
|
tlocs = ("en_US.UTF-8", "en_US.ISO8859-1",
|
|
|
|
"en_US.US-ASCII", "en_US")
|
|
|
|
try:
|
|
|
|
oldlocale = locale.setlocale(locale.LC_NUMERIC)
|
|
|
|
for tloc in tlocs:
|
|
|
|
try:
|
|
|
|
locale.setlocale(locale.LC_NUMERIC, tloc)
|
|
|
|
except locale.Error:
|
|
|
|
continue
|
|
|
|
break
|
|
|
|
else:
|
|
|
|
raise unittest.SkipTest("Test locale not supported "
|
|
|
|
"(tried %s)" % (', '.join(tlocs)))
|
|
|
|
cls.enUS_locale = tloc
|
|
|
|
finally:
|
|
|
|
locale.setlocale(locale.LC_NUMERIC, oldlocale)
|
|
|
|
|
2008-07-25 18:45:08 -03:00
|
|
|
def setUp(self):
|
2013-07-17 07:23:45 -03:00
|
|
|
oldlocale = locale.setlocale(self.locale_type)
|
|
|
|
self.addCleanup(locale.setlocale, self.locale_type, oldlocale)
|
|
|
|
locale.setlocale(self.locale_type, self.enUS_locale)
|
2018-06-25 21:11:06 -03:00
|
|
|
if verbose:
|
2013-07-17 07:23:45 -03:00
|
|
|
print("testing with %r..." % self.enUS_locale, end=' ', flush=True)
|
2008-07-25 18:45:08 -03:00
|
|
|
|
|
|
|
|
|
|
|
class BaseCookedTest(unittest.TestCase):
|
|
|
|
#
|
|
|
|
# Base class for tests using cooked localeconv() values
|
|
|
|
#
|
|
|
|
|
|
|
|
def setUp(self):
|
|
|
|
locale._override_localeconv = self.cooked_values
|
|
|
|
|
|
|
|
def tearDown(self):
|
|
|
|
locale._override_localeconv = {}
|
|
|
|
|
|
|
|
class CCookedTest(BaseCookedTest):
|
|
|
|
# A cooked "C" locale
|
|
|
|
|
|
|
|
cooked_values = {
|
|
|
|
'currency_symbol': '',
|
|
|
|
'decimal_point': '.',
|
|
|
|
'frac_digits': 127,
|
|
|
|
'grouping': [],
|
|
|
|
'int_curr_symbol': '',
|
|
|
|
'int_frac_digits': 127,
|
|
|
|
'mon_decimal_point': '',
|
|
|
|
'mon_grouping': [],
|
|
|
|
'mon_thousands_sep': '',
|
|
|
|
'n_cs_precedes': 127,
|
|
|
|
'n_sep_by_space': 127,
|
|
|
|
'n_sign_posn': 127,
|
|
|
|
'negative_sign': '',
|
|
|
|
'p_cs_precedes': 127,
|
|
|
|
'p_sep_by_space': 127,
|
|
|
|
'p_sign_posn': 127,
|
|
|
|
'positive_sign': '',
|
|
|
|
'thousands_sep': ''
|
|
|
|
}
|
|
|
|
|
|
|
|
class EnUSCookedTest(BaseCookedTest):
|
|
|
|
# A cooked "en_US" locale
|
|
|
|
|
|
|
|
cooked_values = {
|
|
|
|
'currency_symbol': '$',
|
|
|
|
'decimal_point': '.',
|
|
|
|
'frac_digits': 2,
|
|
|
|
'grouping': [3, 3, 0],
|
|
|
|
'int_curr_symbol': 'USD ',
|
|
|
|
'int_frac_digits': 2,
|
|
|
|
'mon_decimal_point': '.',
|
|
|
|
'mon_grouping': [3, 3, 0],
|
|
|
|
'mon_thousands_sep': ',',
|
|
|
|
'n_cs_precedes': 1,
|
|
|
|
'n_sep_by_space': 0,
|
|
|
|
'n_sign_posn': 1,
|
|
|
|
'negative_sign': '-',
|
|
|
|
'p_cs_precedes': 1,
|
|
|
|
'p_sep_by_space': 0,
|
|
|
|
'p_sign_posn': 1,
|
|
|
|
'positive_sign': '',
|
|
|
|
'thousands_sep': ','
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2009-03-13 21:13:13 -03:00
|
|
|
class FrFRCookedTest(BaseCookedTest):
|
|
|
|
# A cooked "fr_FR" locale with a space character as decimal separator
|
|
|
|
# and a non-ASCII currency symbol.
|
|
|
|
|
|
|
|
cooked_values = {
|
|
|
|
'currency_symbol': '\u20ac',
|
|
|
|
'decimal_point': ',',
|
|
|
|
'frac_digits': 2,
|
|
|
|
'grouping': [3, 3, 0],
|
|
|
|
'int_curr_symbol': 'EUR ',
|
|
|
|
'int_frac_digits': 2,
|
|
|
|
'mon_decimal_point': ',',
|
|
|
|
'mon_grouping': [3, 3, 0],
|
|
|
|
'mon_thousands_sep': ' ',
|
|
|
|
'n_cs_precedes': 0,
|
|
|
|
'n_sep_by_space': 1,
|
|
|
|
'n_sign_posn': 1,
|
|
|
|
'negative_sign': '-',
|
|
|
|
'p_cs_precedes': 0,
|
|
|
|
'p_sep_by_space': 1,
|
|
|
|
'p_sign_posn': 1,
|
|
|
|
'positive_sign': '',
|
|
|
|
'thousands_sep': ' '
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2008-07-25 18:45:08 -03:00
|
|
|
class BaseFormattingTest(object):
|
|
|
|
#
|
|
|
|
# Utility functions for formatting tests
|
|
|
|
#
|
|
|
|
|
|
|
|
def _test_format_string(self, format, value, out, **format_opts):
|
2022-06-26 07:41:19 -03:00
|
|
|
self.assertEqual(
|
|
|
|
locale.format_string(format, value, **format_opts), out)
|
2008-07-25 18:45:08 -03:00
|
|
|
|
|
|
|
def _test_currency(self, value, out, **format_opts):
|
|
|
|
self.assertEqual(locale.currency(value, **format_opts), out)
|
|
|
|
|
|
|
|
|
|
|
|
class EnUSNumberFormatting(BaseFormattingTest):
|
2008-07-26 18:02:53 -03:00
|
|
|
# XXX there is a grouping + padding bug when the thousands separator
|
|
|
|
# is empty but the grouping array contains values (e.g. Solaris 10)
|
2008-07-25 18:45:08 -03:00
|
|
|
|
|
|
|
def setUp(self):
|
|
|
|
self.sep = locale.localeconv()['thousands_sep']
|
|
|
|
|
|
|
|
def test_grouping(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%f", 1024, grouping=1, out='1%s024.000000' % self.sep)
|
|
|
|
self._test_format_string("%f", 102, grouping=1, out='102.000000')
|
|
|
|
self._test_format_string("%f", -42, grouping=1, out='-42.000000')
|
|
|
|
self._test_format_string("%+f", -42, grouping=1, out='-42.000000')
|
2008-07-25 18:45:08 -03:00
|
|
|
|
|
|
|
def test_grouping_and_padding(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%20.f", -42, grouping=1, out='-42'.rjust(20))
|
2008-07-26 18:02:53 -03:00
|
|
|
if self.sep:
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%+10.f", -4200, grouping=1,
|
2008-07-26 18:02:53 -03:00
|
|
|
out=('-4%s200' % self.sep).rjust(10))
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%-10.f", -4200, grouping=1,
|
2008-07-26 18:02:53 -03:00
|
|
|
out=('-4%s200' % self.sep).ljust(10))
|
2008-07-25 18:45:08 -03:00
|
|
|
|
|
|
|
def test_integer_grouping(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%d", 4200, grouping=True, out='4%s200' % self.sep)
|
|
|
|
self._test_format_string("%+d", 4200, grouping=True, out='+4%s200' % self.sep)
|
|
|
|
self._test_format_string("%+d", -4200, grouping=True, out='-4%s200' % self.sep)
|
2008-07-25 18:45:08 -03:00
|
|
|
|
2009-03-13 21:13:13 -03:00
|
|
|
def test_integer_grouping_and_padding(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%10d", 4200, grouping=True,
|
2009-03-13 21:13:13 -03:00
|
|
|
out=('4%s200' % self.sep).rjust(10))
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%-10d", -4200, grouping=True,
|
2009-03-13 21:13:13 -03:00
|
|
|
out=('-4%s200' % self.sep).ljust(10))
|
|
|
|
|
2008-07-25 18:45:08 -03:00
|
|
|
def test_simple(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%f", 1024, grouping=0, out='1024.000000')
|
|
|
|
self._test_format_string("%f", 102, grouping=0, out='102.000000')
|
|
|
|
self._test_format_string("%f", -42, grouping=0, out='-42.000000')
|
|
|
|
self._test_format_string("%+f", -42, grouping=0, out='-42.000000')
|
2008-07-25 18:45:08 -03:00
|
|
|
|
|
|
|
def test_padding(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%20.f", -42, grouping=0, out='-42'.rjust(20))
|
|
|
|
self._test_format_string("%+10.f", -4200, grouping=0, out='-4200'.rjust(10))
|
|
|
|
self._test_format_string("%-10.f", 4200, grouping=0, out='4200'.ljust(10))
|
2017-03-28 12:43:38 -03:00
|
|
|
|
2008-07-25 18:45:08 -03:00
|
|
|
def test_complex_formatting(self):
|
|
|
|
# Spaces in formatting string
|
|
|
|
self._test_format_string("One million is %i", 1000000, grouping=1,
|
|
|
|
out='One million is 1%s000%s000' % (self.sep, self.sep))
|
|
|
|
self._test_format_string("One million is %i", 1000000, grouping=1,
|
|
|
|
out='One million is 1%s000%s000' % (self.sep, self.sep))
|
|
|
|
# Dots in formatting string
|
|
|
|
self._test_format_string(".%f.", 1000.0, out='.1000.000000.')
|
|
|
|
# Padding
|
2008-07-26 18:02:53 -03:00
|
|
|
if self.sep:
|
|
|
|
self._test_format_string("--> %10.2f", 4200, grouping=1,
|
|
|
|
out='--> ' + ('4%s200.00' % self.sep).rjust(10))
|
2008-07-25 18:45:08 -03:00
|
|
|
# Asterisk formats
|
|
|
|
self._test_format_string("%10.*f", (2, 1000), grouping=0,
|
|
|
|
out='1000.00'.rjust(10))
|
2008-07-26 18:02:53 -03:00
|
|
|
if self.sep:
|
|
|
|
self._test_format_string("%*.*f", (10, 2, 1000), grouping=1,
|
|
|
|
out=('1%s000.00' % self.sep).rjust(10))
|
2008-07-25 18:45:08 -03:00
|
|
|
# Test more-in-one
|
2008-07-26 18:02:53 -03:00
|
|
|
if self.sep:
|
|
|
|
self._test_format_string("int %i float %.2f str %s",
|
|
|
|
(1000, 1000.0, 'str'), grouping=1,
|
|
|
|
out='int 1%s000 float 1%s000.00 str str' %
|
|
|
|
(self.sep, self.sep))
|
2008-07-25 18:45:08 -03:00
|
|
|
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("total=%i%%", 100, out='total=100%')
|
|
|
|
self._test_format_string("newline: %i\n", 3, out='newline: 3\n')
|
|
|
|
self._test_format_string("extra: %ii", 3, out='extra: 3i')
|
2010-04-26 23:45:53 -03:00
|
|
|
|
|
|
|
|
|
|
|
class TestLocaleFormatString(unittest.TestCase):
|
|
|
|
"""General tests on locale.format_string"""
|
|
|
|
|
|
|
|
def test_percent_escape(self):
|
|
|
|
self.assertEqual(locale.format_string('%f%%', 1.0), '%f%%' % 1.0)
|
|
|
|
self.assertEqual(locale.format_string('%d %f%%d', (1, 1.0)),
|
|
|
|
'%d %f%%d' % (1, 1.0))
|
|
|
|
self.assertEqual(locale.format_string('%(foo)s %%d', {'foo': 'bar'}),
|
|
|
|
('%(foo)s %%d' % {'foo': 'bar'}))
|
|
|
|
|
|
|
|
def test_mapping(self):
|
|
|
|
self.assertEqual(locale.format_string('%(foo)s bing.', {'foo': 'bar'}),
|
|
|
|
('%(foo)s bing.' % {'foo': 'bar'}))
|
|
|
|
self.assertEqual(locale.format_string('%(foo)s', {'foo': 'bar'}),
|
|
|
|
('%(foo)s' % {'foo': 'bar'}))
|
|
|
|
|
2009-04-01 00:42:00 -03:00
|
|
|
|
|
|
|
|
2008-07-25 18:45:08 -03:00
|
|
|
class TestNumberFormatting(BaseLocalizedTest, EnUSNumberFormatting):
|
|
|
|
# Test number formatting with a real English locale.
|
|
|
|
|
|
|
|
locale_type = locale.LC_NUMERIC
|
|
|
|
|
|
|
|
def setUp(self):
|
|
|
|
BaseLocalizedTest.setUp(self)
|
|
|
|
EnUSNumberFormatting.setUp(self)
|
|
|
|
|
|
|
|
|
|
|
|
class TestEnUSNumberFormatting(EnUSCookedTest, EnUSNumberFormatting):
|
|
|
|
# Test number formatting with a cooked "en_US" locale.
|
|
|
|
|
|
|
|
def setUp(self):
|
|
|
|
EnUSCookedTest.setUp(self)
|
|
|
|
EnUSNumberFormatting.setUp(self)
|
|
|
|
|
|
|
|
def test_currency(self):
|
|
|
|
self._test_currency(50000, "$50000.00")
|
|
|
|
self._test_currency(50000, "$50,000.00", grouping=True)
|
|
|
|
self._test_currency(50000, "USD 50,000.00",
|
|
|
|
grouping=True, international=True)
|
|
|
|
|
|
|
|
|
|
|
|
class TestCNumberFormatting(CCookedTest, BaseFormattingTest):
|
|
|
|
# Test number formatting with a cooked "C" locale.
|
|
|
|
|
|
|
|
def test_grouping(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%.2f", 12345.67, grouping=True, out='12345.67')
|
2008-07-25 18:45:08 -03:00
|
|
|
|
|
|
|
def test_grouping_and_padding(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%9.2f", 12345.67, grouping=True, out=' 12345.67')
|
2008-07-25 18:45:08 -03:00
|
|
|
|
|
|
|
|
2009-03-13 21:13:13 -03:00
|
|
|
class TestFrFRNumberFormatting(FrFRCookedTest, BaseFormattingTest):
|
|
|
|
# Test number formatting with a cooked "fr_FR" locale.
|
|
|
|
|
|
|
|
def test_decimal_point(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%.2f", 12345.67, out='12345,67')
|
2009-03-13 21:13:13 -03:00
|
|
|
|
|
|
|
def test_grouping(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%.2f", 345.67, grouping=True, out='345,67')
|
|
|
|
self._test_format_string("%.2f", 12345.67, grouping=True, out='12 345,67')
|
2009-03-13 21:13:13 -03:00
|
|
|
|
|
|
|
def test_grouping_and_padding(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%6.2f", 345.67, grouping=True, out='345,67')
|
|
|
|
self._test_format_string("%7.2f", 345.67, grouping=True, out=' 345,67')
|
|
|
|
self._test_format_string("%8.2f", 12345.67, grouping=True, out='12 345,67')
|
|
|
|
self._test_format_string("%9.2f", 12345.67, grouping=True, out='12 345,67')
|
|
|
|
self._test_format_string("%10.2f", 12345.67, grouping=True, out=' 12 345,67')
|
|
|
|
self._test_format_string("%-6.2f", 345.67, grouping=True, out='345,67')
|
|
|
|
self._test_format_string("%-7.2f", 345.67, grouping=True, out='345,67 ')
|
|
|
|
self._test_format_string("%-8.2f", 12345.67, grouping=True, out='12 345,67')
|
|
|
|
self._test_format_string("%-9.2f", 12345.67, grouping=True, out='12 345,67')
|
|
|
|
self._test_format_string("%-10.2f", 12345.67, grouping=True, out='12 345,67 ')
|
2009-03-13 21:13:13 -03:00
|
|
|
|
|
|
|
def test_integer_grouping(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%d", 200, grouping=True, out='200')
|
|
|
|
self._test_format_string("%d", 4200, grouping=True, out='4 200')
|
2009-03-13 21:13:13 -03:00
|
|
|
|
|
|
|
def test_integer_grouping_and_padding(self):
|
2022-06-26 07:41:19 -03:00
|
|
|
self._test_format_string("%4d", 4200, grouping=True, out='4 200')
|
|
|
|
self._test_format_string("%5d", 4200, grouping=True, out='4 200')
|
|
|
|
self._test_format_string("%10d", 4200, grouping=True, out='4 200'.rjust(10))
|
|
|
|
self._test_format_string("%-4d", 4200, grouping=True, out='4 200')
|
|
|
|
self._test_format_string("%-5d", 4200, grouping=True, out='4 200')
|
|
|
|
self._test_format_string("%-10d", 4200, grouping=True, out='4 200'.ljust(10))
|
2009-03-13 21:13:13 -03:00
|
|
|
|
|
|
|
def test_currency(self):
|
|
|
|
euro = '\u20ac'
|
|
|
|
self._test_currency(50000, "50000,00 " + euro)
|
|
|
|
self._test_currency(50000, "50 000,00 " + euro, grouping=True)
|
2020-01-19 23:45:50 -04:00
|
|
|
self._test_currency(50000, "50 000,00 EUR",
|
2009-03-13 21:13:13 -03:00
|
|
|
grouping=True, international=True)
|
|
|
|
|
|
|
|
|
2009-10-19 16:43:09 -03:00
|
|
|
class TestCollation(unittest.TestCase):
|
|
|
|
# Test string collation functions
|
|
|
|
|
|
|
|
def test_strcoll(self):
|
|
|
|
self.assertLess(locale.strcoll('a', 'b'), 0)
|
|
|
|
self.assertEqual(locale.strcoll('a', 'a'), 0)
|
|
|
|
self.assertGreater(locale.strcoll('b', 'a'), 0)
|
2017-06-28 02:30:06 -03:00
|
|
|
# embedded null character
|
|
|
|
self.assertRaises(ValueError, locale.strcoll, 'a\0', 'a')
|
|
|
|
self.assertRaises(ValueError, locale.strcoll, 'a', 'a\0')
|
2009-10-19 16:43:09 -03:00
|
|
|
|
|
|
|
def test_strxfrm(self):
|
|
|
|
self.assertLess(locale.strxfrm('a'), locale.strxfrm('b'))
|
2017-06-28 02:30:06 -03:00
|
|
|
# embedded null character
|
|
|
|
self.assertRaises(ValueError, locale.strxfrm, 'a\0')
|
2009-10-19 16:43:09 -03:00
|
|
|
|
|
|
|
|
|
|
|
class TestEnUSCollation(BaseLocalizedTest, TestCollation):
|
|
|
|
# Test string collation functions with a real English locale
|
|
|
|
|
|
|
|
locale_type = locale.LC_ALL
|
|
|
|
|
|
|
|
def setUp(self):
|
2022-04-21 22:39:24 -03:00
|
|
|
enc = codecs.lookup(locale.getencoding() or 'ascii').name
|
2009-10-19 16:43:09 -03:00
|
|
|
if enc not in ('utf-8', 'iso8859-1', 'cp1252'):
|
|
|
|
raise unittest.SkipTest('encoding not suitable')
|
2018-06-25 21:11:06 -03:00
|
|
|
if enc != 'iso8859-1' and (sys.platform == 'darwin' or is_android or
|
2009-10-19 16:43:09 -03:00
|
|
|
sys.platform.startswith('freebsd')):
|
|
|
|
raise unittest.SkipTest('wcscoll/wcsxfrm have known bugs')
|
2010-09-23 12:20:15 -03:00
|
|
|
BaseLocalizedTest.setUp(self)
|
2009-10-19 16:43:09 -03:00
|
|
|
|
2017-04-04 05:35:15 -03:00
|
|
|
@unittest.skipIf(sys.platform.startswith('aix'),
|
|
|
|
'bpo-29972: broken test on AIX')
|
2022-05-16 11:02:37 -03:00
|
|
|
@unittest.skipIf(
|
|
|
|
is_emscripten or is_wasi,
|
|
|
|
"musl libc issue on Emscripten/WASI, bpo-46390"
|
|
|
|
)
|
2024-09-16 20:00:36 -03:00
|
|
|
@unittest.skipIf(sys.platform.startswith("netbsd"),
|
|
|
|
"gh-124108: NetBSD doesn't support UTF-8 for LC_COLLATE")
|
2009-10-19 16:43:09 -03:00
|
|
|
def test_strcoll_with_diacritic(self):
|
|
|
|
self.assertLess(locale.strcoll('à', 'b'), 0)
|
|
|
|
|
2017-04-04 05:35:15 -03:00
|
|
|
@unittest.skipIf(sys.platform.startswith('aix'),
|
|
|
|
'bpo-29972: broken test on AIX')
|
2022-05-16 11:02:37 -03:00
|
|
|
@unittest.skipIf(
|
|
|
|
is_emscripten or is_wasi,
|
|
|
|
"musl libc issue on Emscripten/WASI, bpo-46390"
|
|
|
|
)
|
2024-09-16 20:00:36 -03:00
|
|
|
@unittest.skipIf(sys.platform.startswith("netbsd"),
|
|
|
|
"gh-124108: NetBSD doesn't support UTF-8 for LC_COLLATE")
|
2009-10-19 16:43:09 -03:00
|
|
|
def test_strxfrm_with_diacritic(self):
|
|
|
|
self.assertLess(locale.strxfrm('à'), locale.strxfrm('b'))
|
|
|
|
|
|
|
|
|
2013-12-19 15:21:40 -04:00
|
|
|
class NormalizeTest(unittest.TestCase):
|
|
|
|
def check(self, localename, expected):
|
|
|
|
self.assertEqual(locale.normalize(localename), expected, msg=localename)
|
|
|
|
|
|
|
|
def test_locale_alias(self):
|
|
|
|
for localename, alias in locale.locale_alias.items():
|
|
|
|
with self.subTest(locale=(localename, alias)):
|
|
|
|
self.check(localename, alias)
|
|
|
|
|
|
|
|
def test_empty(self):
|
|
|
|
self.check('', '')
|
|
|
|
|
|
|
|
def test_c(self):
|
|
|
|
self.check('c', 'C')
|
|
|
|
self.check('posix', 'C')
|
|
|
|
|
|
|
|
def test_english(self):
|
|
|
|
self.check('en', 'en_US.ISO8859-1')
|
|
|
|
self.check('EN', 'en_US.ISO8859-1')
|
2013-12-26 18:56:53 -04:00
|
|
|
self.check('en.iso88591', 'en_US.ISO8859-1')
|
2013-12-19 15:21:40 -04:00
|
|
|
self.check('en_US', 'en_US.ISO8859-1')
|
|
|
|
self.check('en_us', 'en_US.ISO8859-1')
|
|
|
|
self.check('en_GB', 'en_GB.ISO8859-1')
|
|
|
|
self.check('en_US.UTF-8', 'en_US.UTF-8')
|
|
|
|
self.check('en_US.utf8', 'en_US.UTF-8')
|
|
|
|
self.check('en_US:UTF-8', 'en_US.UTF-8')
|
|
|
|
self.check('en_US.ISO8859-1', 'en_US.ISO8859-1')
|
|
|
|
self.check('en_US.US-ASCII', 'en_US.ISO8859-1')
|
2013-12-26 18:56:53 -04:00
|
|
|
self.check('en_US.88591', 'en_US.ISO8859-1')
|
|
|
|
self.check('en_US.885915', 'en_US.ISO8859-15')
|
2013-12-19 15:21:40 -04:00
|
|
|
self.check('english', 'en_EN.ISO8859-1')
|
2013-12-26 18:56:53 -04:00
|
|
|
self.check('english_uk.ascii', 'en_GB.ISO8859-1')
|
2013-12-19 15:21:40 -04:00
|
|
|
|
|
|
|
def test_hyphenated_encoding(self):
|
|
|
|
self.check('az_AZ.iso88599e', 'az_AZ.ISO8859-9E')
|
|
|
|
self.check('az_AZ.ISO8859-9E', 'az_AZ.ISO8859-9E')
|
|
|
|
self.check('tt_RU.koi8c', 'tt_RU.KOI8-C')
|
|
|
|
self.check('tt_RU.KOI8-C', 'tt_RU.KOI8-C')
|
|
|
|
self.check('lo_LA.cp1133', 'lo_LA.IBM-CP1133')
|
|
|
|
self.check('lo_LA.ibmcp1133', 'lo_LA.IBM-CP1133')
|
|
|
|
self.check('lo_LA.IBM-CP1133', 'lo_LA.IBM-CP1133')
|
|
|
|
self.check('uk_ua.microsoftcp1251', 'uk_UA.CP1251')
|
|
|
|
self.check('uk_ua.microsoft-cp1251', 'uk_UA.CP1251')
|
|
|
|
self.check('ka_ge.georgianacademy', 'ka_GE.GEORGIAN-ACADEMY')
|
|
|
|
self.check('ka_GE.GEORGIAN-ACADEMY', 'ka_GE.GEORGIAN-ACADEMY')
|
|
|
|
self.check('cs_CZ.iso88592', 'cs_CZ.ISO8859-2')
|
|
|
|
self.check('cs_CZ.ISO8859-2', 'cs_CZ.ISO8859-2')
|
|
|
|
|
|
|
|
def test_euro_modifier(self):
|
|
|
|
self.check('de_DE@euro', 'de_DE.ISO8859-15')
|
|
|
|
self.check('en_US.ISO8859-15@euro', 'en_US.ISO8859-15')
|
2013-12-26 18:56:53 -04:00
|
|
|
self.check('de_DE.utf8@euro', 'de_DE.UTF-8')
|
2013-12-19 15:21:40 -04:00
|
|
|
|
|
|
|
def test_latin_modifier(self):
|
|
|
|
self.check('be_BY.UTF-8@latin', 'be_BY.UTF-8@latin')
|
|
|
|
self.check('sr_RS.UTF-8@latin', 'sr_RS.UTF-8@latin')
|
2013-12-26 18:56:53 -04:00
|
|
|
self.check('sr_RS.UTF-8@latn', 'sr_RS.UTF-8@latin')
|
2013-12-19 15:21:40 -04:00
|
|
|
|
|
|
|
def test_valencia_modifier(self):
|
|
|
|
self.check('ca_ES.UTF-8@valencia', 'ca_ES.UTF-8@valencia')
|
2018-05-06 02:46:15 -03:00
|
|
|
self.check('ca_ES@valencia', 'ca_ES.UTF-8@valencia')
|
2013-12-19 15:21:40 -04:00
|
|
|
self.check('ca@valencia', 'ca_ES.ISO8859-1@valencia')
|
|
|
|
|
2013-12-26 15:20:59 -04:00
|
|
|
def test_devanagari_modifier(self):
|
|
|
|
self.check('ks_IN.UTF-8@devanagari', 'ks_IN.UTF-8@devanagari')
|
|
|
|
self.check('ks_IN@devanagari', 'ks_IN.UTF-8@devanagari')
|
|
|
|
self.check('ks@devanagari', 'ks_IN.UTF-8@devanagari')
|
|
|
|
self.check('ks_IN.UTF-8', 'ks_IN.UTF-8')
|
|
|
|
self.check('ks_IN', 'ks_IN.UTF-8')
|
|
|
|
self.check('ks', 'ks_IN.UTF-8')
|
|
|
|
self.check('sd_IN.UTF-8@devanagari', 'sd_IN.UTF-8@devanagari')
|
|
|
|
self.check('sd_IN@devanagari', 'sd_IN.UTF-8@devanagari')
|
|
|
|
self.check('sd@devanagari', 'sd_IN.UTF-8@devanagari')
|
|
|
|
self.check('sd_IN.UTF-8', 'sd_IN.UTF-8')
|
|
|
|
self.check('sd_IN', 'sd_IN.UTF-8')
|
|
|
|
self.check('sd', 'sd_IN.UTF-8')
|
|
|
|
|
2013-12-26 18:56:53 -04:00
|
|
|
def test_euc_encoding(self):
|
|
|
|
self.check('ja_jp.euc', 'ja_JP.eucJP')
|
|
|
|
self.check('ja_jp.eucjp', 'ja_JP.eucJP')
|
|
|
|
self.check('ko_kr.euc', 'ko_KR.eucKR')
|
|
|
|
self.check('ko_kr.euckr', 'ko_KR.eucKR')
|
|
|
|
self.check('zh_cn.euc', 'zh_CN.eucCN')
|
|
|
|
self.check('zh_tw.euc', 'zh_TW.eucTW')
|
2017-03-20 03:49:43 -03:00
|
|
|
self.check('zh_tw.euctw', 'zh_TW.eucTW')
|
2013-12-26 18:56:53 -04:00
|
|
|
|
|
|
|
def test_japanese(self):
|
|
|
|
self.check('ja', 'ja_JP.eucJP')
|
|
|
|
self.check('ja.jis', 'ja_JP.JIS7')
|
|
|
|
self.check('ja.sjis', 'ja_JP.SJIS')
|
|
|
|
self.check('ja_jp', 'ja_JP.eucJP')
|
|
|
|
self.check('ja_jp.ajec', 'ja_JP.eucJP')
|
|
|
|
self.check('ja_jp.euc', 'ja_JP.eucJP')
|
|
|
|
self.check('ja_jp.eucjp', 'ja_JP.eucJP')
|
|
|
|
self.check('ja_jp.iso-2022-jp', 'ja_JP.JIS7')
|
|
|
|
self.check('ja_jp.iso2022jp', 'ja_JP.JIS7')
|
|
|
|
self.check('ja_jp.jis', 'ja_JP.JIS7')
|
|
|
|
self.check('ja_jp.jis7', 'ja_JP.JIS7')
|
|
|
|
self.check('ja_jp.mscode', 'ja_JP.SJIS')
|
|
|
|
self.check('ja_jp.pck', 'ja_JP.SJIS')
|
|
|
|
self.check('ja_jp.sjis', 'ja_JP.SJIS')
|
|
|
|
self.check('ja_jp.ujis', 'ja_JP.eucJP')
|
|
|
|
self.check('ja_jp.utf8', 'ja_JP.UTF-8')
|
|
|
|
self.check('japan', 'ja_JP.eucJP')
|
|
|
|
self.check('japanese', 'ja_JP.eucJP')
|
|
|
|
self.check('japanese-euc', 'ja_JP.eucJP')
|
|
|
|
self.check('japanese.euc', 'ja_JP.eucJP')
|
|
|
|
self.check('japanese.sjis', 'ja_JP.SJIS')
|
|
|
|
self.check('jp_jp', 'ja_JP.eucJP')
|
|
|
|
|
2013-12-19 15:21:40 -04:00
|
|
|
|
2008-07-25 18:45:08 -03:00
|
|
|
class TestMiscellaneous(unittest.TestCase):
|
2019-08-29 01:33:52 -03:00
|
|
|
def test_defaults_UTF8(self):
|
|
|
|
# Issue #18378: on (at least) macOS setting LC_CTYPE to "UTF-8" is
|
2021-10-06 20:13:48 -03:00
|
|
|
# valid. Furthermore LC_CTYPE=UTF is used by the UTF-8 locale coercing
|
2019-08-29 01:33:52 -03:00
|
|
|
# during interpreter startup (on macOS).
|
|
|
|
import _locale
|
|
|
|
import os
|
|
|
|
|
|
|
|
self.assertEqual(locale._parse_localename('UTF-8'), (None, 'UTF-8'))
|
|
|
|
|
|
|
|
if hasattr(_locale, '_getdefaultlocale'):
|
|
|
|
orig_getlocale = _locale._getdefaultlocale
|
|
|
|
del _locale._getdefaultlocale
|
|
|
|
else:
|
|
|
|
orig_getlocale = None
|
|
|
|
|
|
|
|
orig_env = {}
|
|
|
|
try:
|
|
|
|
for key in ('LC_ALL', 'LC_CTYPE', 'LANG', 'LANGUAGE'):
|
|
|
|
if key in os.environ:
|
|
|
|
orig_env[key] = os.environ[key]
|
|
|
|
del os.environ[key]
|
|
|
|
|
|
|
|
os.environ['LC_CTYPE'] = 'UTF-8'
|
|
|
|
|
2022-02-22 17:06:43 -04:00
|
|
|
with check_warnings(('', DeprecationWarning)):
|
|
|
|
self.assertEqual(locale.getdefaultlocale(), (None, 'UTF-8'))
|
2019-08-29 01:33:52 -03:00
|
|
|
|
|
|
|
finally:
|
|
|
|
for k in orig_env:
|
|
|
|
os.environ[k] = orig_env[k]
|
|
|
|
|
|
|
|
if 'LC_CTYPE' not in orig_env:
|
|
|
|
del os.environ['LC_CTYPE']
|
|
|
|
|
|
|
|
if orig_getlocale is not None:
|
|
|
|
_locale._getdefaultlocale = orig_getlocale
|
|
|
|
|
2022-04-21 22:39:24 -03:00
|
|
|
def test_getencoding(self):
|
|
|
|
# Invoke getencoding to make sure it does not cause exceptions.
|
|
|
|
enc = locale.getencoding()
|
|
|
|
self.assertIsInstance(enc, str)
|
|
|
|
self.assertNotEqual(enc, "")
|
|
|
|
# make sure it is valid
|
|
|
|
codecs.lookup(enc)
|
|
|
|
|
2023-06-06 11:55:21 -03:00
|
|
|
def test_getencoding_fallback(self):
|
|
|
|
# When _locale.getencoding() is missing, locale.getencoding() uses
|
|
|
|
# the Python filesystem
|
|
|
|
encoding = 'FALLBACK_ENCODING'
|
|
|
|
with mock.patch.object(sys, 'getfilesystemencoding',
|
|
|
|
return_value=encoding):
|
|
|
|
locale_fallback = import_fresh_module('locale', blocked=['_locale'])
|
|
|
|
self.assertEqual(locale_fallback.getencoding(), encoding)
|
|
|
|
|
2008-07-25 18:45:08 -03:00
|
|
|
def test_getpreferredencoding(self):
|
|
|
|
# Invoke getpreferredencoding to make sure it does not cause exceptions.
|
|
|
|
enc = locale.getpreferredencoding()
|
|
|
|
if enc:
|
|
|
|
# If encoding non-empty, make sure it is valid
|
|
|
|
codecs.lookup(enc)
|
|
|
|
|
2009-10-19 16:43:09 -03:00
|
|
|
def test_strcoll_3303(self):
|
|
|
|
# test crasher from bug #3303
|
|
|
|
self.assertRaises(TypeError, locale.strcoll, "a", None)
|
|
|
|
self.assertRaises(TypeError, locale.strcoll, b"a", None)
|
2008-07-25 18:45:08 -03:00
|
|
|
|
2009-12-01 17:59:18 -04:00
|
|
|
def test_setlocale_category(self):
|
|
|
|
locale.setlocale(locale.LC_ALL)
|
|
|
|
locale.setlocale(locale.LC_TIME)
|
|
|
|
locale.setlocale(locale.LC_CTYPE)
|
|
|
|
locale.setlocale(locale.LC_COLLATE)
|
|
|
|
locale.setlocale(locale.LC_MONETARY)
|
|
|
|
locale.setlocale(locale.LC_NUMERIC)
|
|
|
|
|
|
|
|
# crasher from bug #7419
|
|
|
|
self.assertRaises(locale.Error, locale.setlocale, 12345)
|
|
|
|
|
2011-07-23 21:40:25 -03:00
|
|
|
def test_getsetlocale_issue1813(self):
|
|
|
|
# Issue #1813: setting and getting the locale under a Turkish locale
|
2011-07-26 18:55:33 -03:00
|
|
|
oldlocale = locale.setlocale(locale.LC_CTYPE)
|
2011-07-23 21:40:25 -03:00
|
|
|
self.addCleanup(locale.setlocale, locale.LC_CTYPE, oldlocale)
|
|
|
|
try:
|
|
|
|
locale.setlocale(locale.LC_CTYPE, 'tr_TR')
|
|
|
|
except locale.Error:
|
|
|
|
# Unsupported locale on this system
|
|
|
|
self.skipTest('test needs Turkish locale')
|
2011-07-26 18:55:33 -03:00
|
|
|
loc = locale.getlocale(locale.LC_CTYPE)
|
2018-06-25 21:11:06 -03:00
|
|
|
if verbose:
|
2015-02-18 02:04:37 -04:00
|
|
|
print('testing with %a' % (loc,), end=' ', flush=True)
|
2021-03-31 08:01:46 -03:00
|
|
|
try:
|
|
|
|
locale.setlocale(locale.LC_CTYPE, loc)
|
|
|
|
except locale.Error as exc:
|
|
|
|
# bpo-37945: setlocale(LC_CTYPE) fails with getlocale(LC_CTYPE)
|
|
|
|
# and the tr_TR locale on Windows. getlocale() builds a locale
|
|
|
|
# which is not recognize by setlocale().
|
|
|
|
self.skipTest(f"setlocale(LC_CTYPE, {loc!r}) failed: {exc!r}")
|
2011-07-26 18:55:33 -03:00
|
|
|
self.assertEqual(loc, locale.getlocale(locale.LC_CTYPE))
|
2011-07-23 21:40:25 -03:00
|
|
|
|
2011-11-04 16:35:07 -03:00
|
|
|
def test_invalid_locale_format_in_localetuple(self):
|
|
|
|
with self.assertRaises(TypeError):
|
|
|
|
locale.setlocale(locale.LC_ALL, b'fi_FI')
|
|
|
|
|
|
|
|
def test_invalid_iterable_in_localetuple(self):
|
|
|
|
with self.assertRaises(TypeError):
|
|
|
|
locale.setlocale(locale.LC_ALL, (b'not', b'valid'))
|
|
|
|
|
2008-07-25 18:45:08 -03:00
|
|
|
|
2014-10-23 17:52:31 -03:00
|
|
|
class BaseDelocalizeTest(BaseLocalizedTest):
|
|
|
|
|
|
|
|
def _test_delocalize(self, value, out):
|
|
|
|
self.assertEqual(locale.delocalize(value), out)
|
|
|
|
|
|
|
|
def _test_atof(self, value, out):
|
|
|
|
self.assertEqual(locale.atof(value), out)
|
|
|
|
|
|
|
|
def _test_atoi(self, value, out):
|
|
|
|
self.assertEqual(locale.atoi(value), out)
|
|
|
|
|
|
|
|
|
|
|
|
class TestEnUSDelocalize(EnUSCookedTest, BaseDelocalizeTest):
|
|
|
|
|
|
|
|
def test_delocalize(self):
|
|
|
|
self._test_delocalize('50000.00', '50000.00')
|
|
|
|
self._test_delocalize('50,000.00', '50000.00')
|
|
|
|
|
|
|
|
def test_atof(self):
|
|
|
|
self._test_atof('50000.00', 50000.)
|
|
|
|
self._test_atof('50,000.00', 50000.)
|
|
|
|
|
|
|
|
def test_atoi(self):
|
|
|
|
self._test_atoi('50000', 50000)
|
|
|
|
self._test_atoi('50,000', 50000)
|
|
|
|
|
|
|
|
|
|
|
|
class TestCDelocalizeTest(CCookedTest, BaseDelocalizeTest):
|
|
|
|
|
|
|
|
def test_delocalize(self):
|
|
|
|
self._test_delocalize('50000.00', '50000.00')
|
|
|
|
|
|
|
|
def test_atof(self):
|
|
|
|
self._test_atof('50000.00', 50000.)
|
|
|
|
|
|
|
|
def test_atoi(self):
|
|
|
|
self._test_atoi('50000', 50000)
|
|
|
|
|
|
|
|
|
|
|
|
class TestfrFRDelocalizeTest(FrFRCookedTest, BaseDelocalizeTest):
|
|
|
|
|
|
|
|
def test_delocalize(self):
|
|
|
|
self._test_delocalize('50000,00', '50000.00')
|
|
|
|
self._test_delocalize('50 000,00', '50000.00')
|
|
|
|
|
|
|
|
def test_atof(self):
|
|
|
|
self._test_atof('50000,00', 50000.)
|
|
|
|
self._test_atof('50 000,00', 50000.)
|
|
|
|
|
|
|
|
def test_atoi(self):
|
|
|
|
self._test_atoi('50000', 50000)
|
|
|
|
self._test_atoi('50 000', 50000)
|
|
|
|
|
|
|
|
|
2021-04-12 09:17:40 -03:00
|
|
|
class BaseLocalizeTest(BaseLocalizedTest):
|
|
|
|
|
|
|
|
def _test_localize(self, value, out, grouping=False):
|
|
|
|
self.assertEqual(locale.localize(value, grouping=grouping), out)
|
|
|
|
|
|
|
|
|
|
|
|
class TestEnUSLocalize(EnUSCookedTest, BaseLocalizeTest):
|
|
|
|
|
|
|
|
def test_localize(self):
|
|
|
|
self._test_localize('50000.00', '50000.00')
|
|
|
|
self._test_localize(
|
|
|
|
'{0:.16f}'.format(Decimal('1.15')), '1.1500000000000000')
|
|
|
|
|
|
|
|
|
|
|
|
class TestCLocalize(CCookedTest, BaseLocalizeTest):
|
|
|
|
|
|
|
|
def test_localize(self):
|
|
|
|
self._test_localize('50000.00', '50000.00')
|
|
|
|
|
|
|
|
|
|
|
|
class TestfrFRLocalize(FrFRCookedTest, BaseLocalizeTest):
|
|
|
|
|
|
|
|
def test_localize(self):
|
|
|
|
self._test_localize('50000.00', '50000,00')
|
|
|
|
self._test_localize('50000.00', '50 000,00', grouping=True)
|
|
|
|
|
|
|
|
|
2008-07-25 18:45:08 -03:00
|
|
|
if __name__ == '__main__':
|
2013-07-17 07:23:45 -03:00
|
|
|
unittest.main()
|