Lib/test/test_str.py

import unittest
import struct
import sys
from test import test_support, string_tests


class StrSubclass(str):
    pass

class StrTest(
    string_tests.CommonTest,
    string_tests.MixinStrUnicodeUserStringTest,
    string_tests.MixinStrUserStringTest,
    string_tests.MixinStrUnicodeTest,
    ):

    type2test = str

    # We don't need to propagate to str
    def fixtype(self, obj):
        return obj

    def test_basic_creation(self):
        self.assertEqual(str(''), '')
        self.assertEqual(str(0), '0')
        self.assertEqual(str(0L), '0')
        self.assertEqual(str(()), '()')
        self.assertEqual(str([]), '[]')
        self.assertEqual(str({}), '{}')
        a = []
        a.append(a)
        self.assertEqual(str(a), '[[...]]')
        a = {}
        a[0] = a
        self.assertEqual(str(a), '{0: {...}}')

    def test_formatting(self):
        string_tests.MixinStrUnicodeUserStringTest.test_formatting(self)
        self.assertRaises(OverflowError, '%c'.__mod__, 0x1234)

    @test_support.cpython_only
    def test_formatting_huge_precision(self):
        from _testcapi import INT_MAX
        format_string = "%.{}f".format(INT_MAX + 1)
        with self.assertRaises(ValueError):
            result = format_string % 2.34

    def test_formatting_huge_width(self):
        format_string = "%{}f".format(sys.maxsize + 1)
        with self.assertRaises(ValueError):
            result = format_string % 2.34

    def test_conversion(self):
        # Make sure __str__() behaves properly
        class Foo0:
            def __unicode__(self):
                return u"foo"

        class Foo1:
            def __str__(self):
                return "foo"

        class Foo2(object):
            def __str__(self):
                return "foo"

        class Foo3(object):
            def __str__(self):
                return u"foo"

        class Foo4(unicode):
            def __str__(self):
                return u"foo"

        class Foo5(str):
            def __str__(self):
                return u"foo"

        class Foo6(str):
            def __str__(self):
                return "foos"

            def __unicode__(self):
                return u"foou"

        class Foo7(unicode):
            def __str__(self):
                return "foos"
            def __unicode__(self):
                return u"foou"

        class Foo8(str):
            def __new__(cls, content=""):
                return str.__new__(cls, 2*content)
            def __str__(self):
                return self

        class Foo9(str):
            def __str__(self):
                return "string"
            def __unicode__(self):
                return "not unicode"

        self.assertTrue(str(Foo0()).startswith("<")) # this is different from __unicode__
        self.assertEqual(str(Foo1()), "foo")
        self.assertEqual(str(Foo2()), "foo")
        self.assertEqual(str(Foo3()), "foo")
        self.assertEqual(str(Foo4("bar")), "foo")
        self.assertEqual(str(Foo5("bar")), "foo")
        self.assertEqual(str(Foo6("bar")), "foos")
        self.assertEqual(str(Foo7("bar")), "foos")
        self.assertEqual(str(Foo8("foo")), "foofoo")
        self.assertIs(type(str(Foo8("foo"))), Foo8)
        self.assertEqual(StrSubclass(Foo8("foo")), "foofoo")
        self.assertIs(type(StrSubclass(Foo8("foo"))), StrSubclass)
        self.assertEqual(str(Foo9("foo")), "string")
        self.assertEqual(unicode(Foo9("foo")), u"not unicode")

    # This test only affects 32-bit platforms because expandtabs can only take
    # an int as the max value, not a 64-bit C long.  If expandtabs is changed
    # to take a 64-bit long, this test should apply to all platforms.
    @unittest.skipIf(sys.maxint > (1 << 32) or struct.calcsize('P') != 4,
                     'only applies to 32-bit platforms')
    def test_expandtabs_overflows_gracefully(self):
        self.assertRaises(OverflowError, 't\tt\t'.expandtabs, sys.maxint)

    def test__format__(self):
        def test(value, format, expected):
            # test both with and without the trailing 's'
            self.assertEqual(value.__format__(format), expected)
            self.assertEqual(value.__format__(format + 's'), expected)

        test('', '', '')
        test('abc', '', 'abc')
        test('abc', '.3', 'abc')
        test('ab', '.3', 'ab')
        test('abcdef', '.3', 'abc')
        test('abcdef', '.0', '')
        test('abc', '3.3', 'abc')
        test('abc', '2.3', 'abc')
        test('abc', '2.2', 'ab')
        test('abc', '3.2', 'ab ')
        test('result', 'x<0', 'result')
        test('result', 'x<5', 'result')
        test('result', 'x<6', 'result')
        test('result', 'x<7', 'resultx')
        test('result', 'x<8', 'resultxx')
        test('result', ' <7', 'result ')
        test('result', '<7', 'result ')
        test('result', '>7', ' result')
        test('result', '>8', '  result')
        test('result', '^8', ' result ')
        test('result', '^9', ' result  ')
        test('result', '^10', '  result  ')
        test('a', '10000', 'a' + ' ' * 9999)
        test('', '10000', ' ' * 10000)
        test('', '10000000', ' ' * 10000000)

    def test_format(self):
        self.assertEqual(''.format(), '')
        self.assertEqual('a'.format(), 'a')
        self.assertEqual('ab'.format(), 'ab')
        self.assertEqual('a{{'.format(), 'a{')
        self.assertEqual('a}}'.format(), 'a}')
        self.assertEqual('{{b'.format(), '{b')
        self.assertEqual('}}b'.format(), '}b')
        self.assertEqual('a{{b'.format(), 'a{b')

        # examples from the PEP:
        import datetime
        self.assertEqual("My name is {0}".format('Fred'), "My name is Fred")
        self.assertEqual("My name is {0[name]}".format(dict(name='Fred')),
                         "My name is Fred")
        self.assertEqual("My name is {0} :-{{}}".format('Fred'),
                         "My name is Fred :-{}")

        d = datetime.date(2007, 8, 18)
        self.assertEqual("The year is {0.year}".format(d),
                         "The year is 2007")

        # classes we'll use for testing
        class C:
            def __init__(self, x=100):
                self._x = x
            def __format__(self, spec):
                return spec

        class D:
            def __init__(self, x):
                self.x = x
            def __format__(self, spec):
                return str(self.x)

        # class with __str__, but no __format__
        class E:
            def __init__(self, x):
                self.x = x
            def __str__(self):
                return 'E(' + self.x + ')'

        # class with __repr__, but no __format__ or __str__
        class F:
            def __init__(self, x):
                self.x = x
            def __repr__(self):
                return 'F(' + self.x + ')'

        # class with __format__ that forwards to string, for some format_spec's
        class G:
            def __init__(self, x):
                self.x = x
            def __str__(self):
                return "string is " + self.x
            def __format__(self, format_spec):
                if format_spec == 'd':
                    return 'G(' + self.x + ')'
                return object.__format__(self, format_spec)

        # class that returns a bad type from __format__
        class H:
            def __format__(self, format_spec):
                return 1.0

        class I(datetime.date):
            def __format__(self, format_spec):
                return self.strftime(format_spec)

        class J(int):
            def __format__(self, format_spec):
                return int.__format__(self * 2, format_spec)


        self.assertEqual(''.format(), '')
        self.assertEqual('abc'.format(), 'abc')
        self.assertEqual('{0}'.format('abc'), 'abc')
        self.assertEqual('{0:}'.format('abc'), 'abc')
        self.assertEqual('X{0}'.format('abc'), 'Xabc')
        self.assertEqual('{0}X'.format('abc'), 'abcX')
        self.assertEqual('X{0}Y'.format('abc'), 'XabcY')
        self.assertEqual('{1}'.format(1, 'abc'), 'abc')
        self.assertEqual('X{1}'.format(1, 'abc'), 'Xabc')
        self.assertEqual('{1}X'.format(1, 'abc'), 'abcX')
        self.assertEqual('X{1}Y'.format(1, 'abc'), 'XabcY')
        self.assertEqual('{0}'.format(-15), '-15')
        self.assertEqual('{0}{1}'.format(-15, 'abc'), '-15abc')
        self.assertEqual('{0}X{1}'.format(-15, 'abc'), '-15Xabc')
        self.assertEqual('{{'.format(), '{')
        self.assertEqual('}}'.format(), '}')
        self.assertEqual('{{}}'.format(), '{}')
        self.assertEqual('{{x}}'.format(), '{x}')
        self.assertEqual('{{{0}}}'.format(123), '{123}')
        self.assertEqual('{{{{0}}}}'.format(), '{{0}}')
        self.assertEqual('}}{{'.format(), '}{')
        self.assertEqual('}}x{{'.format(), '}x{')

        # weird field names
        self.assertEqual("{0[foo-bar]}".format({'foo-bar':'baz'}), 'baz')
        self.assertEqual("{0[foo bar]}".format({'foo bar':'baz'}), 'baz')
        self.assertEqual("{0[ ]}".format({' ':3}), '3')

        self.assertEqual('{foo._x}'.format(foo=C(20)), '20')
        self.assertEqual('{1}{0}'.format(D(10), D(20)), '2010')
        self.assertEqual('{0._x.x}'.format(C(D('abc'))), 'abc')
        self.assertEqual('{0[0]}'.format(['abc', 'def']), 'abc')
        self.assertEqual('{0[1]}'.format(['abc', 'def']), 'def')
        self.assertEqual('{0[1][0]}'.format(['abc', ['def']]), 'def')
        self.assertEqual('{0[1][0].x}'.format(['abc', [D('def')]]), 'def')

        # strings
        self.assertEqual('{0:.3s}'.format('abc'), 'abc')
        self.assertEqual('{0:.3s}'.format('ab'), 'ab')
        self.assertEqual('{0:.3s}'.format('abcdef'), 'abc')
        self.assertEqual('{0:.0s}'.format('abcdef'), '')
        self.assertEqual('{0:3.3s}'.format('abc'), 'abc')
        self.assertEqual('{0:2.3s}'.format('abc'), 'abc')
        self.assertEqual('{0:2.2s}'.format('abc'), 'ab')
        self.assertEqual('{0:3.2s}'.format('abc'), 'ab ')
        self.assertEqual('{0:x<0s}'.format('result'), 'result')
        self.assertEqual('{0:x<5s}'.format('result'), 'result')
        self.assertEqual('{0:x<6s}'.format('result'), 'result')
        self.assertEqual('{0:x<7s}'.format('result'), 'resultx')
        self.assertEqual('{0:x<8s}'.format('result'), 'resultxx')
        self.assertEqual('{0: <7s}'.format('result'), 'result ')
        self.assertEqual('{0:<7s}'.format('result'), 'result ')
        self.assertEqual('{0:>7s}'.format('result'), ' result')
        self.assertEqual('{0:>8s}'.format('result'), '  result')
        self.assertEqual('{0:^8s}'.format('result'), ' result ')
        self.assertEqual('{0:^9s}'.format('result'), ' result  ')
        self.assertEqual('{0:^10s}'.format('result'), '  result  ')
        self.assertEqual('{0:10000}'.format('a'), 'a' + ' ' * 9999)
        self.assertEqual('{0:10000}'.format(''), ' ' * 10000)
        self.assertEqual('{0:10000000}'.format(''), ' ' * 10000000)

        # format specifiers for user defined type
        self.assertEqual('{0:abc}'.format(C()), 'abc')

        # !r and !s coercions
        self.assertEqual('{0!s}'.format('Hello'), 'Hello')
        self.assertEqual('{0!s:}'.format('Hello'), 'Hello')
        self.assertEqual('{0!s:15}'.format('Hello'), 'Hello          ')
        self.assertEqual('{0!s:15s}'.format('Hello'), 'Hello          ')
        self.assertEqual('{0!r}'.format('Hello'), "'Hello'")
        self.assertEqual('{0!r:}'.format('Hello'), "'Hello'")
        self.assertEqual('{0!r}'.format(F('Hello')), 'F(Hello)')

        # test fallback to object.__format__
        self.assertEqual('{0}'.format({}), '{}')
        self.assertEqual('{0}'.format([]), '[]')
        self.assertEqual('{0}'.format([1]), '[1]')
        self.assertEqual('{0}'.format(E('data')), 'E(data)')
        self.assertEqual('{0:d}'.format(G('data')), 'G(data)')
        self.assertEqual('{0!s}'.format(G('data')), 'string is data')

        msg = 'object.__format__ with a non-empty format string is deprecated'
        with test_support.check_warnings((msg, PendingDeprecationWarning)):
            self.assertEqual('{0:^10}'.format(E('data')), ' E(data)  ')
            self.assertEqual('{0:^10s}'.format(E('data')), ' E(data)  ')
            self.assertEqual('{0:>15s}'.format(G('data')), ' string is data')

        self.assertEqual("{0:date: %Y-%m-%d}".format(I(year=2007,
                                                       month=8,
                                                       day=27)),
                         "date: 2007-08-27")

        # test deriving from a builtin type and overriding __format__
        self.assertEqual("{0}".format(J(10)), "20")


        # string format specifiers
        self.assertEqual('{0:}'.format('a'), 'a')

        # computed format specifiers
        self.assertEqual("{0:.{1}}".format('hello world', 5), 'hello')
        self.assertEqual("{0:.{1}s}".format('hello world', 5), 'hello')
        self.assertEqual("{0:.{precision}s}".format('hello world', precision=5), 'hello')
        self.assertEqual("{0:{width}.{precision}s}".format('hello world', width=10, precision=5), 'hello     ')
        self.assertEqual("{0:{width}.{precision}s}".format('hello world', width='10', precision='5'), 'hello     ')

        # test various errors
        self.assertRaises(ValueError, '{'.format)
        self.assertRaises(ValueError, '}'.format)
        self.assertRaises(ValueError, 'a{'.format)
        self.assertRaises(ValueError, 'a}'.format)
        self.assertRaises(ValueError, '{a'.format)
        self.assertRaises(ValueError, '}a'.format)
        self.assertRaises(IndexError, '{0}'.format)
        self.assertRaises(IndexError, '{1}'.format, 'abc')
        self.assertRaises(KeyError,   '{x}'.format)
        self.assertRaises(ValueError, "}{".format)
        self.assertRaises(ValueError, "{".format)
        self.assertRaises(ValueError, "}".format)
        self.assertRaises(ValueError, "abc{0:{}".format)
        self.assertRaises(ValueError, "{0".format)
        self.assertRaises(IndexError, "{0.}".format)
        self.assertRaises(ValueError, "{0.}".format, 0)
        self.assertRaises(IndexError, "{0[}".format)
        self.assertRaises(ValueError, "{0[}".format, [])
        self.assertRaises(KeyError,   "{0]}".format)
        self.assertRaises(ValueError, "{0.[]}".format, 0)
        self.assertRaises(ValueError, "{0..foo}".format, 0)
        self.assertRaises(ValueError, "{0[0}".format, 0)
        self.assertRaises(ValueError, "{0[0:foo}".format, 0)
        self.assertRaises(KeyError,   "{c]}".format)
        self.assertRaises(ValueError, "{{ {{{0}}".format, 0)
        self.assertRaises(ValueError, "{0}}".format, 0)
        self.assertRaises(KeyError,   "{foo}".format, bar=3)
        self.assertRaises(ValueError, "{0!x}".format, 3)
        self.assertRaises(ValueError, "{0!}".format, 0)
        self.assertRaises(ValueError, "{0!rs}".format, 0)
        self.assertRaises(ValueError, "{!}".format)
        self.assertRaises(IndexError, "{:}".format)
        self.assertRaises(IndexError, "{:s}".format)
        self.assertRaises(IndexError, "{}".format)

        # issue 6089
        self.assertRaises(ValueError, "{0[0]x}".format, [None])
        self.assertRaises(ValueError, "{0[0](10)}".format, [None])

        # can't have a replacement on the field name portion
        self.assertRaises(TypeError, '{0[{1}]}'.format, 'abcdefg', 4)

        # exceed maximum recursion depth
        self.assertRaises(ValueError, "{0:{1:{2}}}".format, 'abc', 's', '')
        self.assertRaises(ValueError, "{0:{1:{2:{3:{4:{5:{6}}}}}}}".format,
                          0, 1, 2, 3, 4, 5, 6, 7)

        # string format spec errors
        self.assertRaises(ValueError, "{0:-s}".format, '')
        self.assertRaises(ValueError, format, "", "-")
        self.assertRaises(ValueError, "{0:=s}".format, '')

    def test_format_huge_precision(self):
        format_string = ".{}f".format(sys.maxsize + 1)
        with self.assertRaises(ValueError):
            result = format(2.34, format_string)

    def test_format_huge_width(self):
        format_string = "{}f".format(sys.maxsize + 1)
        with self.assertRaises(ValueError):
            result = format(2.34, format_string)

    def test_format_huge_item_number(self):
        format_string = "{{{}:.6f}}".format(sys.maxsize + 1)
        with self.assertRaises(ValueError):
            result = format_string.format(2.34)

    def test_format_auto_numbering(self):
        class C:
            def __init__(self, x=100):
                self._x = x
            def __format__(self, spec):
                return spec

        self.assertEqual('{}'.format(10), '10')
        self.assertEqual('{:5}'.format('s'), 's    ')
        self.assertEqual('{!r}'.format('s'), "'s'")
        self.assertEqual('{._x}'.format(C(10)), '10')
        self.assertEqual('{[1]}'.format([1, 2]), '2')
        self.assertEqual('{[a]}'.format({'a':4, 'b':2}), '4')
        self.assertEqual('a{}b{}c'.format(0, 1), 'a0b1c')

        self.assertEqual('a{:{}}b'.format('x', '^10'), 'a    x     b')
        self.assertEqual('a{:{}x}b'.format(20, '#'), 'a0x14b')

        # can't mix and match numbering and auto-numbering
        self.assertRaises(ValueError, '{}{1}'.format, 1, 2)
        self.assertRaises(ValueError, '{1}{}'.format, 1, 2)
        self.assertRaises(ValueError, '{:{1}}'.format, 1, 2)
        self.assertRaises(ValueError, '{0:{}}'.format, 1, 2)

        # can mix and match auto-numbering and named
        self.assertEqual('{f}{}'.format(4, f='test'), 'test4')
        self.assertEqual('{}{f}'.format(4, f='test'), '4test')
        self.assertEqual('{:{f}}{g}{}'.format(1, 3, g='g', f=2), ' 1g3')
        self.assertEqual('{f:{}}{}{g}'.format(2, 4, f=1, g='g'), ' 14g')

    def test_format_c_overflow(self):
        # issue #7267
        self.assertRaises(OverflowError, '{0:c}'.format, -1)
        self.assertRaises(OverflowError, '{0:c}'.format, 256)

    def test_buffer_is_readonly(self):
        self.assertRaises(TypeError, sys.stdin.readinto, b"")

    def test_encode_and_decode_kwargs(self):
        self.assertEqual('abcde'.encode('ascii', 'replace'),
                         'abcde'.encode('ascii', errors='replace'))
        self.assertEqual('abcde'.encode('ascii', 'ignore'),
                         'abcde'.encode(encoding='ascii', errors='ignore'))
        self.assertEqual('Andr\202 x'.decode('ascii', 'ignore'),
                         'Andr\202 x'.decode('ascii', errors='ignore'))
        self.assertEqual('Andr\202 x'.decode('ascii', 'replace'),
                         'Andr\202 x'.decode(encoding='ascii', errors='replace'))

    def test_startswith_endswith_errors(self):
        with self.assertRaises(UnicodeDecodeError):
            '\xff'.startswith(u'x')
        with self.assertRaises(UnicodeDecodeError):
            '\xff'.endswith(u'x')
        for meth in ('foo'.startswith, 'foo'.endswith):
            with self.assertRaises(TypeError) as cm:
                meth(['f'])
            exc = str(cm.exception)
            self.assertIn('unicode', exc)
            self.assertIn('str', exc)
            self.assertIn('tuple', exc)

def test_main():
    test_support.run_unittest(StrTest)

if __name__ == "__main__":
    test_main()