Google Git Google Git
Sign in
chromium/external/github.com/python/cpython/9512cff6a0c8ec05be498dc97104697346bc4b45/./Lib/test/test_bytes.py
blob: 96190a7f58217098a636e3b585785abdfcdf8fc6 [file]
"""Unit tests for the bytes and bytearray types.
XXX This is a mess. Common tests should be unified with string_tests.py (and
the latter should be modernized).
"""
import array
import codecs
import contextlib
import copy
import functools
import operator
import os
import pickle
import re
import sys
import tempfile
import textwrap
import threading
import unittest
from _codecs import _unregister_error as _codecs_unregister_error
import test.support
from test import support
from test.support import import_helper
from test.support import threading_helper
from test.support import warnings_helper
import test.string_tests
import test.list_tests
from test.support import bigaddrspacetest, MAX_Py_ssize_t
from test.support.script_helper import assert_python_failure
if sys.flags.bytes_warning:
def check_bytes_warnings(func):
@functools.wraps(func)
def wrapper(*args, **kw):
with warnings_helper.check_warnings(('', BytesWarning)):
return func(*args, **kw)
return wrapper
else:
# no-op
def check_bytes_warnings(func):
return func
class Indexable:
def __init__(self, value=0):
self.value = value
def __index__(self):
return self.value
@contextlib.contextmanager
def inject_memory_error(testcase, start=0):
with testcase.assertRaises(MemoryError):
with support.inject_memory_error_cm(start):
yield
class BaseBytesTest:
def assertTypedEqual(self, actual, expected):
self.assertIs(type(actual), type(expected))
self.assertEqual(actual, expected)
def test_basics(self):
b = self.type2test()
self.assertEqual(type(b), self.type2test)
self.assertEqual(b.__class__, self.type2test)
def test_copy(self):
a = self.type2test(b"abcd")
for copy_method in (copy.copy, copy.deepcopy):
b = copy_method(a)
self.assertEqual(a, b)
self.assertEqual(type(a), type(b))
def test_empty_sequence(self):
b = self.type2test()
self.assertEqual(len(b), 0)
self.assertRaises(IndexError, lambda: b[0])
self.assertRaises(IndexError, lambda: b[1])
self.assertRaises(IndexError, lambda: b[sys.maxsize])
self.assertRaises(IndexError, lambda: b[sys.maxsize+1])
self.assertRaises(IndexError, lambda: b[10**100])
self.assertRaises(IndexError, lambda: b[-1])
self.assertRaises(IndexError, lambda: b[-2])
self.assertRaises(IndexError, lambda: b[-sys.maxsize])
self.assertRaises(IndexError, lambda: b[-sys.maxsize-1])
self.assertRaises(IndexError, lambda: b[-sys.maxsize-2])
self.assertRaises(IndexError, lambda: b[-10**100])
def test_from_iterable(self):
b = self.type2test(range(256))
self.assertEqual(len(b), 256)
self.assertEqual(list(b), list(range(256)))
# Non-sequence iterable.
b = self.type2test({42})
self.assertEqual(b, b"*")
b = self.type2test({43, 45})
self.assertIn(tuple(b), {(43, 45), (45, 43)})
# Iterator that has a __length_hint__.
b = self.type2test(iter(range(256)))
self.assertEqual(len(b), 256)
self.assertEqual(list(b), list(range(256)))
# Iterator that doesn't have a __length_hint__.
b = self.type2test(i for i in range(256) if i % 2)
self.assertEqual(len(b), 128)
self.assertEqual(list(b), list(range(256))[1::2])
# Sequence without __iter__.
class S:
def __getitem__(self, i):
return (1, 2, 3)[i]
b = self.type2test(S())
self.assertEqual(b, b"\x01\x02\x03")
def test_from_tuple(self):
# There is a special case for tuples.
b = self.type2test(tuple(range(256)))
self.assertEqual(len(b), 256)
self.assertEqual(list(b), list(range(256)))
b = self.type2test((1, 2, 3))
self.assertEqual(b, b"\x01\x02\x03")
def test_from_list(self):
# There is a special case for lists.
b = self.type2test(list(range(256)))
self.assertEqual(len(b), 256)
self.assertEqual(list(b), list(range(256)))
b = self.type2test([1, 2, 3])
self.assertEqual(b, b"\x01\x02\x03")
def test_from_mutating_list(self):
# Issue #34973: Crash in bytes constructor with mutating list.
class X:
def __index__(self):
a.clear()
return 42
a = [X(), X()]
self.assertEqual(bytes(a), b'*')
class Y:
def __index__(self):
if len(a) < 1000:
a.append(self)
return 42
a = [Y()]
self.assertEqual(bytes(a), b'*' * 1000) # should not crash
def test_from_index(self):
b = self.type2test([Indexable(), Indexable(1), Indexable(254),
Indexable(255)])
self.assertEqual(list(b), [0, 1, 254, 255])
self.assertRaises(ValueError, self.type2test, [Indexable(-1)])
self.assertRaises(ValueError, self.type2test, [Indexable(256)])
def test_from_buffer(self):
a = self.type2test(array.array('B', [1, 2, 3]))
self.assertEqual(a, b"\x01\x02\x03")
a = self.type2test(b"\x01\x02\x03")
self.assertEqual(a, b"\x01\x02\x03")
# Issues #29159 and #34974.
# Fallback when __index__ raises a TypeError
class B(bytes):
def __index__(self):
raise TypeError
self.assertEqual(self.type2test(B(b"foobar")), b"foobar")
def test_from_ssize(self):
self.assertEqual(self.type2test(0), b'')
self.assertEqual(self.type2test(1), b'\x00')
self.assertEqual(self.type2test(5), b'\x00\x00\x00\x00\x00')
self.assertRaises(ValueError, self.type2test, -1)
self.assertEqual(self.type2test('0', 'ascii'), b'0')
self.assertEqual(self.type2test(b'0'), b'0')
self.assertRaises(OverflowError, self.type2test, sys.maxsize + 1)
def test_constructor_type_errors(self):
self.assertRaises(TypeError, self.type2test, 0.0)
class C:
pass
self.assertRaises(TypeError, self.type2test, ["0"])
self.assertRaises(TypeError, self.type2test, [0.0])
self.assertRaises(TypeError, self.type2test, [None])
self.assertRaises(TypeError, self.type2test, [C()])
self.assertRaises(TypeError, self.type2test, encoding='ascii')
self.assertRaises(TypeError, self.type2test, errors='ignore')
self.assertRaises(TypeError, self.type2test, 0, 'ascii')
self.assertRaises(TypeError, self.type2test, b'', 'ascii')
self.assertRaises(TypeError, self.type2test, 0, errors='ignore')
self.assertRaises(TypeError, self.type2test, b'', errors='ignore')
self.assertRaises(TypeError, self.type2test, '')
self.assertRaises(TypeError, self.type2test, '', errors='ignore')
self.assertRaises(TypeError, self.type2test, '', b'ascii')
self.assertRaises(TypeError, self.type2test, '', 'ascii', b'ignore')
def test_constructor_value_errors(self):
self.assertRaises(ValueError, self.type2test, [-1])
self.assertRaises(ValueError, self.type2test, [-sys.maxsize])
self.assertRaises(ValueError, self.type2test, [-sys.maxsize-1])
self.assertRaises(ValueError, self.type2test, [-sys.maxsize-2])
self.assertRaises(ValueError, self.type2test, [-10**100])
self.assertRaises(ValueError, self.type2test, [256])
self.assertRaises(ValueError, self.type2test, [257])
self.assertRaises(ValueError, self.type2test, [sys.maxsize])
self.assertRaises(ValueError, self.type2test, [sys.maxsize+1])
self.assertRaises(ValueError, self.type2test, [10**100])
@bigaddrspacetest
def test_constructor_overflow(self):
size = MAX_Py_ssize_t
self.assertRaises((OverflowError, MemoryError), self.type2test, size)
try:
# Should either pass or raise an error (e.g. on debug builds with
# additional malloc() overhead), but shouldn't crash.
bytearray(size - 4)
except (OverflowError, MemoryError):
pass
def test_constructor_exceptions(self):
# Issue #34974: bytes and bytearray constructors replace unexpected
# exceptions.
class BadInt:
def __index__(self):
1/0
self.assertRaises(ZeroDivisionError, self.type2test, BadInt())
self.assertRaises(ZeroDivisionError, self.type2test, [BadInt()])
class BadIterable:
def __iter__(self):
1/0
self.assertRaises(ZeroDivisionError, self.type2test, BadIterable())
def test_compare(self):
b1 = self.type2test([1, 2, 3])
b2 = self.type2test([1, 2, 3])
b3 = self.type2test([1, 3])
self.assertEqual(b1, b2)
self.assertTrue(b2 != b3)
self.assertTrue(b1 <= b2)
self.assertTrue(b1 <= b3)
self.assertTrue(b1 < b3)
self.assertTrue(b1 >= b2)
self.assertTrue(b3 >= b2)
self.assertTrue(b3 > b2)
self.assertFalse(b1 != b2)
self.assertFalse(b2 == b3)
self.assertFalse(b1 > b2)
self.assertFalse(b1 > b3)
self.assertFalse(b1 >= b3)
self.assertFalse(b1 < b2)
self.assertFalse(b3 < b2)
self.assertFalse(b3 <= b2)
@check_bytes_warnings
def test_compare_to_str(self):
# Byte comparisons with unicode should always fail!
# Test this for all expected byte orders and Unicode character
# sizes.
self.assertEqual(self.type2test(b"\0a\0b\0c") == "abc", False)
self.assertEqual(self.type2test(b"\0\0\0a\0\0\0b\0\0\0c") == "abc",
False)
self.assertEqual(self.type2test(b"a\0b\0c\0") == "abc", False)
self.assertEqual(self.type2test(b"a\0\0\0b\0\0\0c\0\0\0") == "abc",
False)
self.assertEqual(self.type2test() == str(), False)
self.assertEqual(self.type2test() != str(), True)
def test_reversed(self):
input = list(map(ord, "Hello"))
b = self.type2test(input)
output = list(reversed(b))
input.reverse()
self.assertEqual(output, input)
def test_getslice(self):
def by(s):
return self.type2test(map(ord, s))
b = by("Hello, world")
self.assertEqual(b[:5], by("Hello"))
self.assertEqual(b[1:5], by("ello"))
self.assertEqual(b[5:7], by(", "))
self.assertEqual(b[7:], by("world"))
self.assertEqual(b[7:12], by("world"))
self.assertEqual(b[7:100], by("world"))
self.assertEqual(b[:-7], by("Hello"))
self.assertEqual(b[-11:-7], by("ello"))
self.assertEqual(b[-7:-5], by(", "))
self.assertEqual(b[-5:], by("world"))
self.assertEqual(b[-5:12], by("world"))
self.assertEqual(b[-5:100], by("world"))
self.assertEqual(b[-100:5], by("Hello"))
def test_extended_getslice(self):
# Test extended slicing by comparing with list slicing.
L = list(range(255))
b = self.type2test(L)
indices = (0, None, 1, 3, 19, 100, sys.maxsize, -1, -2, -31, -100)
for start in indices:
for stop in indices:
# Skip step 0 (invalid)
for step in indices[1:]:
self.assertEqual(b[start:stop:step], self.type2test(L[start:stop:step]))
def test_encoding(self):
sample = "Hello world\n\u1234\u5678\u9abc"
for enc in ("utf-8", "utf-16"):
b = self.type2test(sample, enc)
self.assertEqual(b, self.type2test(sample.encode(enc)))
self.assertRaises(UnicodeEncodeError, self.type2test, sample, "latin-1")
b = self.type2test(sample, "latin-1", "ignore")
self.assertEqual(b, self.type2test(sample[:-3], "utf-8"))
def test_decode(self):
sample = "Hello world\n\u1234\u5678\u9abc"
for enc in ("utf-8", "utf-16"):
b = self.type2test(sample, enc)
self.assertEqual(b.decode(enc), sample)
sample = "Hello world\n\x80\x81\xfe\xff"
b = self.type2test(sample, "latin-1")
self.assertRaises(UnicodeDecodeError, b.decode, "utf-8")
self.assertEqual(b.decode("utf-8", "ignore"), "Hello world\n")
self.assertEqual(b.decode(errors="ignore", encoding="utf-8"),
"Hello world\n")
# Default encoding is utf-8
self.assertEqual(self.type2test(b'\xe2\x98\x83').decode(), '\u2603')
def test_check_encoding_errors(self):
# bpo-37388: bytes(str) and bytes.encode() must check encoding
# and errors arguments in dev mode
invalid = 'Boom, Shaka Laka, Boom!'
encodings = ('ascii', 'utf8', 'latin1')
code = textwrap.dedent(f'''
import sys
type2test = {self.type2test.__name__}
encodings = {encodings!r}
for data in ('', 'short string'):
try:
type2test(data, encoding={invalid!r})
except LookupError:
pass
else:
sys.exit(21)
for encoding in encodings:
try:
type2test(data, encoding=encoding, errors={invalid!r})
except LookupError:
pass
else:
sys.exit(22)
for data in (b'', b'short string'):
data = type2test(data)
print(repr(data))
try:
data.decode(encoding={invalid!r})
except LookupError:
sys.exit(10)
else:
sys.exit(23)
try:
data.decode(errors={invalid!r})
except LookupError:
pass
else:
sys.exit(24)
for encoding in encodings:
try:
data.decode(encoding=encoding, errors={invalid!r})
except LookupError:
pass
else:
sys.exit(25)
sys.exit(10)
''')
proc = assert_python_failure('-X', 'dev', '-c', code)
self.assertEqual(proc.rc, 10, proc)
def test_from_int(self):
b = self.type2test(0)
self.assertEqual(b, self.type2test())
b = self.type2test(10)
self.assertEqual(b, self.type2test([0]*10))
b = self.type2test(10000)
self.assertEqual(b, self.type2test([0]*10000))
def test_concat(self):
b1 = self.type2test(b"abc")
b2 = self.type2test(b"def")
self.assertEqual(b1 + b2, b"abcdef")
self.assertEqual(b1 + bytes(b"def"), b"abcdef")
self.assertEqual(bytes(b"def") + b1, b"defabc")
self.assertRaises(TypeError, lambda: b1 + "def")
self.assertRaises(TypeError, lambda: "abc" + b2)
self.assertEqual(self.type2test() + memoryview(b''), b'')
def test_repeat(self):
for b in b"abc", self.type2test(b"abc"):
self.assertEqual(b * 3, b"abcabcabc")
self.assertEqual(b * 0, b"")
self.assertEqual(b * -1, b"")
self.assertRaises(TypeError, lambda: b * 3.14)
self.assertRaises(TypeError, lambda: 3.14 * b)
# XXX Shouldn't bytes and bytearray agree on what to raise?
with self.assertRaises((OverflowError, MemoryError)):
c = b * sys.maxsize
with self.assertRaises((OverflowError, MemoryError)):
b *= sys.maxsize
def test_repeat_1char(self):
self.assertEqual(self.type2test(b'x')*100, self.type2test([ord('x')]*100))
def test_contains(self):
b = self.type2test(b"abc")
self.assertIn(ord('a'), b)
self.assertIn(int(ord('a')), b)
self.assertNotIn(200, b)
self.assertRaises(ValueError, lambda: 300 in b)
self.assertRaises(ValueError, lambda: -1 in b)
self.assertRaises(ValueError, lambda: sys.maxsize+1 in b)
self.assertRaises(TypeError, lambda: None in b)
self.assertRaises(TypeError, lambda: float(ord('a')) in b)
self.assertRaises(TypeError, lambda: "a" in b)
for f in bytes, bytearray:
self.assertIn(f(b""), b)
self.assertIn(f(b"a"), b)
self.assertIn(f(b"b"), b)
self.assertIn(f(b"c"), b)
self.assertIn(f(b"ab"), b)
self.assertIn(f(b"bc"), b)
self.assertIn(f(b"abc"), b)
self.assertNotIn(f(b"ac"), b)
self.assertNotIn(f(b"d"), b)
self.assertNotIn(f(b"dab"), b)
self.assertNotIn(f(b"abd"), b)
def test_fromhex(self):
self.assertRaises(TypeError, self.type2test.fromhex)
self.assertRaises(TypeError, self.type2test.fromhex, 1)
self.assertEqual(self.type2test.fromhex(''), self.type2test())
b = bytearray([0x1a, 0x2b, 0x30])
self.assertEqual(self.type2test.fromhex('1a2B30'), b)
self.assertEqual(self.type2test.fromhex(' 1A 2B 30 '), b)
# check that ASCII whitespace is ignored
self.assertEqual(self.type2test.fromhex(' 1A\n2B\t30\v'), b)
self.assertEqual(self.type2test.fromhex(b' 1A\n2B\t30\v'), b)
for c in "\x09\x0A\x0B\x0C\x0D\x20":
self.assertEqual(self.type2test.fromhex(c), self.type2test())
for c in "\x1C\x1D\x1E\x1F\x85\xa0\u2000\u2002\u2028":
self.assertRaises(ValueError, self.type2test.fromhex, c)
# Check that we can parse bytes and bytearray
tests = [
("bytes", bytes),
("bytearray", bytearray),
("memoryview", memoryview),
("array.array", lambda bs: array.array('B', bs)),
]
for name, factory in tests:
with self.subTest(name=name):
self.assertEqual(self.type2test.fromhex(factory(b' 1A 2B 30 ')), b)
# Invalid bytes are rejected
for u8 in b"\0\x1C\x1D\x1E\x1F\x85\xa0":
b = bytes([30, 31, u8])
self.assertRaises(ValueError, self.type2test.fromhex, b)
self.assertEqual(self.type2test.fromhex('0000'), b'\0\0')
with self.assertRaisesRegex(
TypeError,
r'fromhex\(\) argument must be str or bytes-like, not tuple',
):
self.type2test.fromhex(())
self.assertRaises(ValueError, self.type2test.fromhex, 'a')
self.assertRaises(ValueError, self.type2test.fromhex, 'rt')
self.assertRaises(ValueError, self.type2test.fromhex, '1a b cd')
self.assertRaises(ValueError, self.type2test.fromhex, '\x00')
self.assertRaises(ValueError, self.type2test.fromhex, '12 \x00 34')
# For odd number of character(s)
for value in ("a", "aaa", "deadbee"):
with self.assertRaises(ValueError) as cm:
self.type2test.fromhex(value)
self.assertIn("fromhex() arg must contain an even number of hexadecimal digits", str(cm.exception))
for value, position in (("a ", 1), (" aa a ", 5), (" aa a a ", 5)):
with self.assertRaises(ValueError) as cm:
self.type2test.fromhex(value)
self.assertIn(f"non-hexadecimal number found in fromhex() arg at position {position}", str(cm.exception))
for data, pos in (
# invalid first hexadecimal character
('12 x4 56', 3),
# invalid second hexadecimal character
('12 3x 56', 4),
# two invalid hexadecimal characters
('12 xy 56', 3),
# test non-ASCII string
('12 3\xff 56', 4),
):
with self.assertRaises(ValueError) as cm:
self.type2test.fromhex(data)
self.assertIn('at position %s' % pos, str(cm.exception))
# gh-158583: Check for out of bounds reads (uninitialized bytes).
# Create an array from a list to not overallocate.
a = array.array('B', list(b'1234 ')) # Py_ISSPACE() loop
self.assertEqual(self.type2test.fromhex(a), b'\x12\x34')
a = array.array('B', list(b'12345')) # Missing second digit
with self.assertRaises(ValueError):
self.type2test.fromhex(a)
def test_hex(self):
self.assertRaises(TypeError, self.type2test.hex)
self.assertRaises(TypeError, self.type2test.hex, 1)
self.assertEqual(self.type2test(b"").hex(), "")
self.assertEqual(bytearray([0x1a, 0x2b, 0x30]).hex(), '1a2b30')
self.assertEqual(self.type2test(b"\x1a\x2b\x30").hex(), '1a2b30')
self.assertEqual(memoryview(b"\x1a\x2b\x30").hex(), '1a2b30')
def test_hex_separator_basics(self):
three_bytes = self.type2test(b'\xb9\x01\xef')
self.assertEqual(three_bytes.hex(), 'b901ef')
with self.assertRaises(ValueError):
three_bytes.hex('')
with self.assertRaises(ValueError):
three_bytes.hex('xx')
self.assertEqual(three_bytes.hex(':', 0), 'b901ef')
with self.assertRaises(TypeError):
three_bytes.hex(None, 0)
with self.assertRaises(ValueError):
three_bytes.hex('\xff')
with self.assertRaises(ValueError):
three_bytes.hex(b'\xff')
with self.assertRaises(ValueError):
three_bytes.hex(b'\x80')
with self.assertRaises(ValueError):
three_bytes.hex(chr(0x100))
self.assertEqual(three_bytes.hex(':', 0), 'b901ef')
self.assertEqual(three_bytes.hex(b'\x00'), 'b9\x0001\x00ef')
self.assertEqual(three_bytes.hex('\x00'), 'b9\x0001\x00ef')
self.assertEqual(three_bytes.hex(b'\x7f'), 'b9\x7f01\x7fef')
self.assertEqual(three_bytes.hex('\x7f'), 'b9\x7f01\x7fef')
self.assertEqual(three_bytes.hex(':', 3), 'b901ef')
self.assertEqual(three_bytes.hex(':', 4), 'b901ef')
self.assertEqual(three_bytes.hex(':', -4), 'b901ef')
self.assertEqual(three_bytes.hex(':'), 'b9:01:ef')
self.assertEqual(three_bytes.hex(b'$'), 'b9$01$ef')
self.assertEqual(three_bytes.hex(':', 1), 'b9:01:ef')
self.assertEqual(three_bytes.hex(':', -1), 'b9:01:ef')
self.assertEqual(three_bytes.hex(':', 2), 'b9:01ef')
self.assertEqual(three_bytes.hex(':', 1), 'b9:01:ef')
self.assertEqual(three_bytes.hex('*', -2), 'b901*ef')
self.assertEqual(three_bytes.hex(sep=':', bytes_per_sep=2), 'b9:01ef')
self.assertEqual(three_bytes.hex(sep='*', bytes_per_sep=-2), 'b901*ef')
for bytes_per_sep in 3, -3, sys.maxsize, -sys.maxsize:
with self.subTest(bytes_per_sep=bytes_per_sep):
self.assertEqual(three_bytes.hex(':', bytes_per_sep), 'b901ef')
for bytes_per_sep in sys.maxsize+1, -sys.maxsize-1, 2**1000, -2**1000:
with self.subTest(bytes_per_sep=bytes_per_sep):
try:
self.assertEqual(three_bytes.hex(':', bytes_per_sep), 'b901ef')
except OverflowError:
pass
value = b'{s\005\000\000\000worldi\002\000\000\000s\005\000\000\000helloi\001\000\000\0000'
self.assertEqual(value.hex('.', 8), '7b7305000000776f.726c646902000000.730500000068656c.6c6f690100000030')
def test_hex_separator_five_bytes(self):
five_bytes = self.type2test(range(90,95))
self.assertEqual(five_bytes.hex(), '5a5b5c5d5e')
def test_hex_separator_six_bytes(self):
six_bytes = self.type2test(x*3 for x in range(1, 7))
self.assertEqual(six_bytes.hex(), '0306090c0f12')
self.assertEqual(six_bytes.hex('.', 1), '03.06.09.0c.0f.12')
self.assertEqual(six_bytes.hex(' ', 2), '0306 090c 0f12')
self.assertEqual(six_bytes.hex('-', 3), '030609-0c0f12')
self.assertEqual(six_bytes.hex(':', 4), '0306:090c0f12')
self.assertEqual(six_bytes.hex(':', 5), '03:06090c0f12')
self.assertEqual(six_bytes.hex(':', 6), '0306090c0f12')
self.assertEqual(six_bytes.hex(':', 95), '0306090c0f12')
self.assertEqual(six_bytes.hex('_', -3), '030609_0c0f12')
self.assertEqual(six_bytes.hex(':', -4), '0306090c:0f12')
self.assertEqual(six_bytes.hex(b'@', -5), '0306090c0f@12')
self.assertEqual(six_bytes.hex(':', -6), '0306090c0f12')
self.assertEqual(six_bytes.hex(' ', -95), '0306090c0f12')
def test_hex_simd_boundaries(self):
# Test lengths around the SIMD threshold (16 bytes).
# SIMD processes 16 bytes at a time; smaller inputs use scalar code.
for length in (14, 15, 16, 17, 31, 32, 33, 64, 65):
data = self.type2test(bytes(range(length)))
expected = ''.join(f'{b:02x}' for b in range(length))
with self.subTest(length=length):
self.assertEqual(data.hex(), expected)
def test_hex_nibble_boundaries(self):
# Test the nibble value boundary at 9/10 (where '9' becomes 'a').
# SIMD uses signed comparison for efficiency; verify correctness
# at this boundary for various nibble combinations.
boundary_bytes = self.type2test(bytes([
0x09, # both nibbles: 0, 9
0x0a, # both nibbles: 0, 10
0x90, # both nibbles: 9, 0
0x99, # both nibbles: 9, 9 (max all-digit)
0x9a, # both nibbles: 9, 10
0xa0, # both nibbles: 10, 0
0xa9, # both nibbles: 10, 9
0xaa, # both nibbles: 10, 10 (min all-letter)
0x00, # min value
0xff, # max value
]))
self.assertEqual(boundary_bytes.hex(), '090a90999aa0a9aa00ff')
# Repeat with 16+ bytes to exercise SIMD path
simd_boundary = self.type2test(boundary_bytes * 2)
self.assertEqual(simd_boundary.hex(), '090a90999aa0a9aa00ff' * 2)
def test_join(self):
self.assertEqual(self.type2test(b"").join([]), b"")
self.assertEqual(self.type2test(b"").join([b""]), b"")
for lst in [[b"abc"], [b"a", b"bc"], [b"ab", b"c"], [b"a", b"b", b"c"]]:
lst = list(map(self.type2test, lst))
self.assertEqual(self.type2test(b"").join(lst), b"abc")
self.assertEqual(self.type2test(b"").join(tuple(lst)), b"abc")
self.assertEqual(self.type2test(b"").join(iter(lst)), b"abc")
dot_join = self.type2test(b".:").join
self.assertEqual(dot_join([b"ab", b"cd"]), b"ab.:cd")
self.assertEqual(dot_join([memoryview(b"ab"), b"cd"]), b"ab.:cd")
self.assertEqual(dot_join([b"ab", memoryview(b"cd")]), b"ab.:cd")
self.assertEqual(dot_join([bytearray(b"ab"), b"cd"]), b"ab.:cd")
self.assertEqual(dot_join([b"ab", bytearray(b"cd")]), b"ab.:cd")
# Stress it with many items
seq = [b"abc"] * 100000
expected = b"abc" + b".:abc" * 99999
self.assertEqual(dot_join(seq), expected)
# Stress test with empty separator
seq = [b"abc"] * 100000
expected = b"abc" * 100000
self.assertEqual(self.type2test(b"").join(seq), expected)
self.assertRaises(TypeError, self.type2test(b" ").join, None)
# Error handling and cleanup when some item in the middle of the
# sequence has the wrong type.
with self.assertRaises(TypeError):
dot_join([bytearray(b"ab"), "cd", b"ef"])
with self.assertRaises(TypeError):
dot_join([memoryview(b"ab"), "cd", b"ef"])
def test_join_concurrent_buffer_mutation(self):
# __buffer__() can release the GIL, letting another thread concurrently
# mutate the joined sequence (simulated here by mutating in __buffer__).
# See: https://github.com/python/cpython/issues/151295
def make_seq(mutate):
# Item is only referenced from the list slot, so mutate() frees it.
class Item:
def __buffer__(self, flags):
mutate(seq)
return memoryview(b'x')
seq = [b'a', Item(), b'c']
return seq
for sep in (self.type2test(b''), self.type2test(b'::')):
with self.subTest(sep=sep):
# Changing the list length is reported as a RuntimeError.
seq = make_seq(lambda seq: seq.clear())
self.assertRaises(RuntimeError, sep.join, seq)
# The list length is unchanged, so the size-change recheck
# cannot fire: only keeping the item alive avoids the crash.
def replace(seq):
seq[1] = b'z'
seq = make_seq(replace)
self.assertEqual(sep.join(seq), sep.join([b'a', b'x', b'c']))
def test_count(self):
b = self.type2test(b'mississippi')
i = 105
p = 112
w = 119
self.assertEqual(b.count(b'i'), 4)
self.assertEqual(b.count(b'ss'), 2)
self.assertEqual(b.count(b'w'), 0)
self.assertEqual(b.count(i), 4)
self.assertEqual(b.count(w), 0)
self.assertEqual(b.count(b'i', 6), 2)
self.assertEqual(b.count(b'p', 6), 2)
self.assertEqual(b.count(b'i', 1, 3), 1)
self.assertEqual(b.count(b'p', 7, 9), 1)
self.assertEqual(b.count(i, 6), 2)
self.assertEqual(b.count(p, 6), 2)
self.assertEqual(b.count(i, 1, 3), 1)
self.assertEqual(b.count(p, 7, 9), 1)
def test_startswith(self):
b = self.type2test(b'hello')
self.assertFalse(self.type2test().startswith(b"anything"))
self.assertTrue(b.startswith(b"hello"))
self.assertTrue(b.startswith(b"hel"))
self.assertTrue(b.startswith(b"h"))
self.assertFalse(b.startswith(b"hellow"))
self.assertFalse(b.startswith(b"ha"))
with self.assertRaises(TypeError) as cm:
b.startswith([b'h'])
exc = str(cm.exception)
self.assertIn('bytes', exc)
self.assertIn('tuple', exc)
def test_endswith(self):
b = self.type2test(b'hello')
self.assertFalse(bytearray().endswith(b"anything"))
self.assertTrue(b.endswith(b"hello"))
self.assertTrue(b.endswith(b"llo"))
self.assertTrue(b.endswith(b"o"))
self.assertFalse(b.endswith(b"whello"))
self.assertFalse(b.endswith(b"no"))
with self.assertRaises(TypeError) as cm:
b.endswith([b'o'])
exc = str(cm.exception)
self.assertIn('bytes', exc)
self.assertIn('tuple', exc)
def test_find(self):
b = self.type2test(b'mississippi')
i = 105
w = 119
self.assertEqual(b.find(b'ss'), 2)
self.assertEqual(b.find(b'w'), -1)
self.assertEqual(b.find(b'mississippian'), -1)
self.assertEqual(b.find(i), 1)
self.assertEqual(b.find(w), -1)
self.assertEqual(b.find(b'ss', 3), 5)
self.assertEqual(b.find(b'ss', 1, 7), 2)
self.assertEqual(b.find(b'ss', 1, 3), -1)
self.assertEqual(b.find(i, 6), 7)
self.assertEqual(b.find(i, 1, 3), 1)
self.assertEqual(b.find(w, 1, 3), -1)
for index in (-1, 256, sys.maxsize + 1):
self.assertRaisesRegex(
ValueError, r'byte must be in range\(0, 256\)',
b.find, index)
def test_rfind(self):
b = self.type2test(b'mississippi')
i = 105
w = 119
self.assertEqual(b.rfind(b'ss'), 5)
self.assertEqual(b.rfind(b'w'), -1)
self.assertEqual(b.rfind(b'mississippian'), -1)
self.assertEqual(b.rfind(i), 10)
self.assertEqual(b.rfind(w), -1)
self.assertEqual(b.rfind(b'ss', 3), 5)
self.assertEqual(b.rfind(b'ss', 0, 6), 2)
self.assertEqual(b.rfind(i, 1, 3), 1)
self.assertEqual(b.rfind(i, 3, 9), 7)
self.assertEqual(b.rfind(w, 1, 3), -1)
def test_index(self):
b = self.type2test(b'mississippi')
i = 105
w = 119
self.assertEqual(b.index(b'ss'), 2)
self.assertRaises(ValueError, b.index, b'w')
self.assertRaises(ValueError, b.index, b'mississippian')
self.assertEqual(b.index(i), 1)
self.assertRaises(ValueError, b.index, w)
self.assertEqual(b.index(b'ss', 3), 5)
self.assertEqual(b.index(b'ss', 1, 7), 2)
self.assertRaises(ValueError, b.index, b'ss', 1, 3)
self.assertEqual(b.index(i, 6), 7)
self.assertEqual(b.index(i, 1, 3), 1)
self.assertRaises(ValueError, b.index, w, 1, 3)
def test_rindex(self):
b = self.type2test(b'mississippi')
i = 105
w = 119
self.assertEqual(b.rindex(b'ss'), 5)
self.assertRaises(ValueError, b.rindex, b'w')
self.assertRaises(ValueError, b.rindex, b'mississippian')
self.assertEqual(b.rindex(i), 10)
self.assertRaises(ValueError, b.rindex, w)
self.assertEqual(b.rindex(b'ss', 3), 5)
self.assertEqual(b.rindex(b'ss', 0, 6), 2)
self.assertEqual(b.rindex(i, 1, 3), 1)
self.assertEqual(b.rindex(i, 3, 9), 7)
self.assertRaises(ValueError, b.rindex, w, 1, 3)
def test_mod(self):
b = self.type2test(b'hello, %b!')
orig = b
b = b % b'world'
self.assertEqual(b, b'hello, world!')
self.assertEqual(orig, b'hello, %b!')
self.assertFalse(b is orig)
b = self.type2test(b'%s / 100 = %d%%')
a = b % (b'seventy-nine', 79)
self.assertEqual(a, b'seventy-nine / 100 = 79%')
self.assertIs(type(a), self.type2test)
# issue 29714
b = self.type2test(b'hello,\x00%b!')
b = b % b'world'
self.assertEqual(b, b'hello,\x00world!')
self.assertIs(type(b), self.type2test)
def check(fmt, vals, result):
b = self.type2test(fmt)
b = b % vals
self.assertEqual(b, result)
self.assertIs(type(b), self.type2test)
# A set of tests adapted from test_unicode:UnicodeTest.test_formatting
check(b'...%(foo)b...', {b'foo':b"abc"}, b'...abc...')
check(b'...%(f(o)o)b...', {b'f(o)o':b"abc", b'foo':b'bar'}, b'...abc...')
check(b'...%(foo)b...', {b'foo':b"abc",b'def':123}, b'...abc...')
check(b'%*b', (5, b'abc',), b' abc')
check(b'%*b', (-5, b'abc',), b'abc ')
check(b'%*.*b', (5, 2, b'abc',), b' ab')
check(b'%*.*b', (5, 3, b'abc',), b' abc')
check(b'%i %*.*b', (10, 5, 3, b'abc',), b'10 abc')
check(b'%i%b %*.*b', (10, b'3', 5, 3, b'abc',), b'103 abc')
check(b'%c', b'a', b'a')
class PseudoFloat:
def __init__(self, value):
self.value = float(value)
def __int__(self):
return int(self.value)
pi = PseudoFloat(3.1415)
exceptions_params = [
('%x requires an integer, not float', b'%x', 3.14),
('%X requires an integer, not float', b'%X', 2.11),
('%o requires an integer, not float', b'%o', 1.79),
(r'%x requires an integer, not .*\.PseudoFloat', b'%x', pi),
('%x requires an integer, not complex', b'%x', 3j),
('%X requires an integer, not complex', b'%X', 2j),
('%o requires an integer, not complex', b'%o', 1j),
('%u requires a real number, not complex', b'%u', 3j),
('%i requires a real number, not complex', b'%i', 2j),
('%d requires a real number, not complex', b'%d', 2j),
(
r'%c requires an integer in range\(256\)'
r' or a single byte, not .*\.PseudoFloat',
b'%c', pi
),
]
for msg, format_bytes, value in exceptions_params:
with self.assertRaisesRegex(TypeError, 'format argument: ' + msg):
operator.mod(format_bytes, value)
def test_memory_leak_gh_140939(self):
# gh-140939: MemoryError is raised without leaking
_testcapi = import_helper.import_module('_testcapi')
with self.assertRaises((MemoryError, OverflowError)):
b = self.type2test(b'%*b')
b % (_testcapi.PY_SSIZE_T_MAX, b'abc')
def test_imod(self):
b = self.type2test(b'hello, %b!')
orig = b
b %= b'world'
self.assertEqual(b, b'hello, world!')
self.assertEqual(orig, b'hello, %b!')
self.assertFalse(b is orig)
b = self.type2test(b'%s / 100 = %d%%')
b %= (b'seventy-nine', 79)
self.assertEqual(b, b'seventy-nine / 100 = 79%')
self.assertIs(type(b), self.type2test)
# issue 29714
b = self.type2test(b'hello,\x00%b!')
b %= b'world'
self.assertEqual(b, b'hello,\x00world!')
self.assertIs(type(b), self.type2test)
def test_rmod(self):
with self.assertRaises(TypeError):
object() % self.type2test(b'abc')
self.assertIs(self.type2test(b'abc').__rmod__('%r'), NotImplemented)
def test_replace(self):
b = self.type2test(b'mississippi')
self.assertEqual(b.replace(b'i', b'a'), b'massassappa')
self.assertEqual(b.replace(b'ss', b'x'), b'mixixippi')
def test_replace_count_keyword(self):
b = self.type2test(b'aa')
self.assertEqual(b.replace(b'a', b'b', count=0), b'aa')
self.assertEqual(b.replace(b'a', b'b', count=1), b'ba')
self.assertEqual(b.replace(b'a', b'b', count=2), b'bb')
self.assertEqual(b.replace(b'a', b'b', count=3), b'bb')
def test_replace_int_error(self):
self.assertRaises(TypeError, self.type2test(b'a b').replace, 32, b'')
def test_split_string_error(self):
self.assertRaises(TypeError, self.type2test(b'a b').split, ' ')
self.assertRaises(TypeError, self.type2test(b'a b').rsplit, ' ')
def test_split_int_error(self):
self.assertRaises(TypeError, self.type2test(b'a b').split, 32)
self.assertRaises(TypeError, self.type2test(b'a b').rsplit, 32)
def test_split_unicodewhitespace(self):
for b in (b'a\x1Cb', b'a\x1Db', b'a\x1Eb', b'a\x1Fb'):
b = self.type2test(b)
self.assertEqual(b.split(), [b])
b = self.type2test(b"\x09\x0A\x0B\x0C\x0D\x1C\x1D\x1E\x1F")
self.assertEqual(b.split(), [b'\x1c\x1d\x1e\x1f'])
def test_rsplit_unicodewhitespace(self):
b = self.type2test(b"\x09\x0A\x0B\x0C\x0D\x1C\x1D\x1E\x1F")
self.assertEqual(b.rsplit(), [b'\x1c\x1d\x1e\x1f'])
def test_partition(self):
b = self.type2test(b'mississippi')
self.assertEqual(b.partition(b'ss'), (b'mi', b'ss', b'issippi'))
self.assertEqual(b.partition(b'w'), (b'mississippi', b'', b''))
def test_rpartition(self):
b = self.type2test(b'mississippi')
self.assertEqual(b.rpartition(b'ss'), (b'missi', b'ss', b'ippi'))
self.assertEqual(b.rpartition(b'i'), (b'mississipp', b'i', b''))
self.assertEqual(b.rpartition(b'w'), (b'', b'', b'mississippi'))
def test_partition_string_error(self):
self.assertRaises(TypeError, self.type2test(b'a b').partition, ' ')
self.assertRaises(TypeError, self.type2test(b'a b').rpartition, ' ')
def test_partition_int_error(self):
self.assertRaises(TypeError, self.type2test(b'a b').partition, 32)
self.assertRaises(TypeError, self.type2test(b'a b').rpartition, 32)
def test_pickling(self):
for proto in range(pickle.HIGHEST_PROTOCOL + 1):
for b in b"", b"a", b"abc", b"\xffab\x80", b"\0\0\377\0\0":
b = self.type2test(b)
ps = pickle.dumps(b, proto)
q = pickle.loads(ps)
self.assertEqual(b, q)
def test_iterator_pickling(self):
for proto in range(pickle.HIGHEST_PROTOCOL + 1):
for b in b"", b"a", b"abc", b"\xffab\x80", b"\0\0\377\0\0":
it = itorg = iter(self.type2test(b))
data = list(self.type2test(b))
d = pickle.dumps(it, proto)
it = pickle.loads(d)
self.assertEqual(type(itorg), type(it))
self.assertEqual(list(it), data)
it = pickle.loads(d)
if not b:
continue
next(it)
d = pickle.dumps(it, proto)
it = pickle.loads(d)
self.assertEqual(list(it), data[1:])
def test_strip_bytearray(self):
self.assertEqual(self.type2test(b'abc').strip(memoryview(b'ac')), b'b')
self.assertEqual(self.type2test(b'abc').lstrip(memoryview(b'ac')), b'bc')
self.assertEqual(self.type2test(b'abc').rstrip(memoryview(b'ac')), b'ab')
def test_strip_string_error(self):
self.assertRaises(TypeError, self.type2test(b'abc').strip, 'ac')
self.assertRaises(TypeError, self.type2test(b'abc').lstrip, 'ac')
self.assertRaises(TypeError, self.type2test(b'abc').rstrip, 'ac')
def test_strip_int_error(self):
self.assertRaises(TypeError, self.type2test(b' abc ').strip, 32)
self.assertRaises(TypeError, self.type2test(b' abc ').lstrip, 32)
self.assertRaises(TypeError, self.type2test(b' abc ').rstrip, 32)
def test_center(self):
# Fill character can be either bytes or bytearray (issue 12380)
b = self.type2test(b'abc')
for fill_type in (bytes, bytearray):
self.assertEqual(b.center(7, fill_type(b'-')),
self.type2test(b'--abc--'))
def test_ljust(self):
# Fill character can be either bytes or bytearray (issue 12380)
b = self.type2test(b'abc')
for fill_type in (bytes, bytearray):
self.assertEqual(b.ljust(7, fill_type(b'-')),
self.type2test(b'abc----'))
def test_rjust(self):
# Fill character can be either bytes or bytearray (issue 12380)
b = self.type2test(b'abc')
for fill_type in (bytes, bytearray):
self.assertEqual(b.rjust(7, fill_type(b'-')),
self.type2test(b'----abc'))
def test_xjust_int_error(self):
self.assertRaises(TypeError, self.type2test(b'abc').center, 7, 32)
self.assertRaises(TypeError, self.type2test(b'abc').ljust, 7, 32)
self.assertRaises(TypeError, self.type2test(b'abc').rjust, 7, 32)
def test_ord(self):
b = self.type2test(b'\0A\x7f\x80\xff')
self.assertEqual([ord(b[i:i+1]) for i in range(len(b))],
[0, 65, 127, 128, 255])
def test_maketrans(self):
transtable = b'\000\001\002\003\004\005\006\007\010\011\012\013\014\015\016\017\020\021\022\023\024\025\026\027\030\031\032\033\034\035\036\037 !"#$%&\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`xyzdefghijklmnopqrstuvwxyz{|}~\177\200\201\202\203\204\205\206\207\210\211\212\213\214\215\216\217\220\221\222\223\224\225\226\227\230\231\232\233\234\235\236\237\240\241\242\243\244\245\246\247\250\251\252\253\254\255\256\257\260\261\262\263\264\265\266\267\270\271\272\273\274\275\276\277\300\301\302\303\304\305\306\307\310\311\312\313\314\315\316\317\320\321\322\323\324\325\326\327\330\331\332\333\334\335\336\337\340\341\342\343\344\345\346\347\350\351\352\353\354\355\356\357\360\361\362\363\364\365\366\367\370\371\372\373\374\375\376\377'
self.assertEqual(self.type2test.maketrans(b'abc', b'xyz'), transtable)
transtable = b'\000\001\002\003\004\005\006\007\010\011\012\013\014\015\016\017\020\021\022\023\024\025\026\027\030\031\032\033\034\035\036\037 !"#$%&\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~\177\200\201\202\203\204\205\206\207\210\211\212\213\214\215\216\217\220\221\222\223\224\225\226\227\230\231\232\233\234\235\236\237\240\241\242\243\244\245\246\247\250\251\252\253\254\255\256\257\260\261\262\263\264\265\266\267\270\271\272\273\274\275\276\277\300\301\302\303\304\305\306\307\310\311\312\313\314\315\316\317\320\321\322\323\324\325\326\327\330\331\332\333\334\335\336\337\340\341\342\343\344\345\346\347\350\351\352\353\354\355\356\357\360\361\362\363\364\365\366\367\370\371\372\373\374xyz'
self.assertEqual(self.type2test.maketrans(b'\375\376\377', b'xyz'), transtable)
self.assertRaises(ValueError, self.type2test.maketrans, b'abc', b'xyzq')
self.assertRaises(TypeError, self.type2test.maketrans, 'abc', 'def')
def test_none_arguments(self):
# issue 11828
b = self.type2test(b'hello')
l = self.type2test(b'l')
h = self.type2test(b'h')
x = self.type2test(b'x')
o = self.type2test(b'o')
self.assertEqual(2, b.find(l, None))
self.assertEqual(3, b.find(l, -2, None))
self.assertEqual(2, b.find(l, None, -2))
self.assertEqual(0, b.find(h, None, None))
self.assertEqual(3, b.rfind(l, None))
self.assertEqual(3, b.rfind(l, -2, None))
self.assertEqual(2, b.rfind(l, None, -2))
self.assertEqual(0, b.rfind(h, None, None))
self.assertEqual(2, b.index(l, None))
self.assertEqual(3, b.index(l, -2, None))
self.assertEqual(2, b.index(l, None, -2))
self.assertEqual(0, b.index(h, None, None))
self.assertEqual(3, b.rindex(l, None))
self.assertEqual(3, b.rindex(l, -2, None))
self.assertEqual(2, b.rindex(l, None, -2))
self.assertEqual(0, b.rindex(h, None, None))
self.assertEqual(2, b.count(l, None))
self.assertEqual(1, b.count(l, -2, None))
self.assertEqual(1, b.count(l, None, -2))
self.assertEqual(0, b.count(x, None, None))
self.assertEqual(True, b.endswith(o, None))
self.assertEqual(True, b.endswith(o, -2, None))
self.assertEqual(True, b.endswith(l, None, -2))
self.assertEqual(False, b.endswith(x, None, None))
self.assertEqual(True, b.startswith(h, None))
self.assertEqual(True, b.startswith(l, -2, None))
self.assertEqual(True, b.startswith(h, None, -2))
self.assertEqual(False, b.startswith(x, None, None))
def test_integer_arguments_out_of_byte_range(self):
b = self.type2test(b'hello')
for method in (b.count, b.find, b.index, b.rfind, b.rindex):
self.assertRaises(ValueError, method, -1)
self.assertRaises(ValueError, method, 256)
self.assertRaises(ValueError, method, 9999)
def test_find_etc_raise_correct_error_messages(self):
# issue 11828
b = self.type2test(b'hello')
x = self.type2test(b'x')
self.assertRaisesRegex(TypeError, r'\bfind\b', b.find,
x, None, None, None)
self.assertRaisesRegex(TypeError, r'\brfind\b', b.rfind,
x, None, None, None)
self.assertRaisesRegex(TypeError, r'\bindex\b', b.index,
x, None, None, None)
self.assertRaisesRegex(TypeError, r'\brindex\b', b.rindex,
x, None, None, None)
self.assertRaisesRegex(TypeError, r'\bcount\b', b.count,
x, None, None, None)
self.assertRaisesRegex(TypeError, r'\bstartswith\b', b.startswith,
x, None, None, None)
self.assertRaisesRegex(TypeError, r'\bendswith\b', b.endswith,
x, None, None, None)
def test_free_after_iterating(self):
test.support.check_free_after_iterating(self, iter, self.type2test)
test.support.check_free_after_iterating(self, reversed, self.type2test)
def test_translate(self):
b = self.type2test(b'hello')
rosetta = bytearray(range(256))
rosetta[ord('o')] = ord('e')
self.assertRaises(TypeError, b.translate)
self.assertRaises(TypeError, b.translate, None, None)
self.assertRaises(ValueError, b.translate, bytes(range(255)))
c = b.translate(rosetta, b'hello')
self.assertEqual(c, b'')
self.assertEqual(type(c), self.type2test)
c = b.translate(rosetta)
d = b.translate(rosetta, b'')
self.assertEqual(c, b'helle')
self.assertEqual(type(c), self.type2test)
self.assertEqual(d, b'helle')
c = b.translate(rosetta, b'l')
self.assertEqual(c, b'hee')
c = b.translate(None, b'e')
self.assertEqual(c, b'hllo')
# test delete as a keyword argument
c = b.translate(rosetta, delete=b'')
self.assertEqual(c, b'helle')
c = b.translate(rosetta, delete=b'l')
self.assertEqual(c, b'hee')
c = b.translate(None, delete=b'e')
self.assertEqual(c, b'hllo')
# short inputs starting with NUL bytes
table = bytes.maketrans(b'\x00', b'Z')
for data in b'\x00', b'\x00' * 8, b'\x00' * 8 + b'a' * 247:
c = self.type2test(data).translate(table)
self.assertEqual(c, data.replace(b'\x00', b'Z'))
@support.cpython_only
def test_translate_unchanged(self):
if self.type2test != bytes:
self.skipTest("test specific bytes.translate()")
# bytes.translate() returns the input string unchanged
# if no byte is modified
size = 1024
b = b'hell' + b'o' * size
rosetta = bytearray(range(256))
rosetta[ord('#')] = ord('?')
self.assertIs(b.translate(rosetta), b)
# bytes.translate() always create a new object
# if the input string is a bytes subclass
class bytes_subclass(bytes):
pass
b = bytes_subclass(b)
result = b.translate(rosetta)
self.assertIsNot(result, b)
self.assertEqual(result, b)
def test_sq_item(self):
_testlimitedcapi = import_helper.import_module('_testlimitedcapi')
obj = self.type2test((42,))
with self.assertRaises(IndexError):
_testlimitedcapi.sequence_getitem(obj, -2)
with self.assertRaises(IndexError):
_testlimitedcapi.sequence_getitem(obj, 1)
self.assertEqual(_testlimitedcapi.sequence_getitem(obj, 0), 42)
class BytesTest(BaseBytesTest, unittest.TestCase):
type2test = bytes
def test__bytes__(self):
foo = b'foo\x00bar'
self.assertEqual(foo.__bytes__(), foo)
self.assertEqual(type(foo.__bytes__()), self.type2test)
class bytes_subclass(bytes):
pass
bar = bytes_subclass(b'bar\x00foo')
self.assertEqual(bar.__bytes__(), bar)
self.assertEqual(type(bar.__bytes__()), self.type2test)
def test_getitem_error(self):
b = b'python'
msg = "byte indices must be integers or slices"
with self.assertRaisesRegex(TypeError, msg):
b['a']
def test_buffer_is_readonly(self):
fd = os.open(__file__, os.O_RDONLY)
with open(fd, "rb", buffering=0) as f:
self.assertRaises(TypeError, f.readinto, b"")
def test_custom(self):
self.assertEqual(bytes(BytesSubclass(b'abc')), b'abc')
self.assertEqual(BytesSubclass(OtherBytesSubclass(b'abc')),
BytesSubclass(b'abc'))
self.assertEqual(bytes(WithBytes(b'abc')), b'abc')
self.assertEqual(BytesSubclass(WithBytes(b'abc')), BytesSubclass(b'abc'))
class NoBytes: pass
self.assertRaises(TypeError, bytes, NoBytes())
self.assertRaises(TypeError, bytes, WithBytes('abc'))
self.assertRaises(TypeError, bytes, WithBytes(None))
class IndexWithBytes:
def __bytes__(self):
return b'a'
def __index__(self):
return 42
self.assertEqual(bytes(IndexWithBytes()), b'a')
# Issue #25766
class StrWithBytes(str):
def __new__(cls, value):
self = str.__new__(cls, '\u20ac')
self.value = value
return self
def __bytes__(self):
return self.value
self.assertEqual(bytes(StrWithBytes(b'abc')), b'abc')
self.assertEqual(bytes(StrWithBytes(b'abc'), 'iso8859-15'), b'\xa4')
self.assertEqual(bytes(StrWithBytes(BytesSubclass(b'abc'))), b'abc')
self.assertEqual(BytesSubclass(StrWithBytes(b'abc')), BytesSubclass(b'abc'))
self.assertEqual(BytesSubclass(StrWithBytes(b'abc'), 'iso8859-15'),
BytesSubclass(b'\xa4'))
self.assertEqual(BytesSubclass(StrWithBytes(BytesSubclass(b'abc'))),
BytesSubclass(b'abc'))
self.assertEqual(BytesSubclass(StrWithBytes(OtherBytesSubclass(b'abc'))),
BytesSubclass(b'abc'))
# Issue #24731
self.assertTypedEqual(bytes(WithBytes(BytesSubclass(b'abc'))), BytesSubclass(b'abc'))
self.assertTypedEqual(BytesSubclass(WithBytes(BytesSubclass(b'abc'))),
BytesSubclass(b'abc'))
self.assertTypedEqual(BytesSubclass(WithBytes(OtherBytesSubclass(b'abc'))),
BytesSubclass(b'abc'))
class BytesWithBytes(bytes):
def __new__(cls, value):
self = bytes.__new__(cls, b'\xa4')
self.value = value
return self
def __bytes__(self):
return self.value
self.assertTypedEqual(bytes(BytesWithBytes(b'abc')), b'abc')
self.assertTypedEqual(BytesSubclass(BytesWithBytes(b'abc')),
BytesSubclass(b'abc'))
self.assertTypedEqual(bytes(BytesWithBytes(BytesSubclass(b'abc'))),
BytesSubclass(b'abc'))
self.assertTypedEqual(BytesSubclass(BytesWithBytes(BytesSubclass(b'abc'))),
BytesSubclass(b'abc'))
self.assertTypedEqual(BytesSubclass(BytesWithBytes(OtherBytesSubclass(b'abc'))),
BytesSubclass(b'abc'))
# Test PyBytes_FromFormat()
def test_from_format(self):
ctypes = import_helper.import_module('ctypes')
_testcapi = import_helper.import_module('_testcapi')
from ctypes import pythonapi, py_object
from ctypes import (
c_int, c_uint,
c_long, c_ulong,
c_size_t, c_ssize_t,
c_char_p)
PyBytes_FromFormat = pythonapi.PyBytes_FromFormat
PyBytes_FromFormat.argtypes = (c_char_p,)
PyBytes_FromFormat.restype = py_object
# basic tests
self.assertEqual(PyBytes_FromFormat(b'format'),
b'format')
self.assertEqual(PyBytes_FromFormat(b'Hello %s !', b'world'),
b'Hello world !')
# test formatters
self.assertEqual(PyBytes_FromFormat(b'c=%c', c_int(0)),
b'c=\0')
self.assertEqual(PyBytes_FromFormat(b'c=%c', c_int(ord('@'))),
b'c=@')
self.assertEqual(PyBytes_FromFormat(b'c=%c', c_int(255)),
b'c=\xff')
self.assertEqual(PyBytes_FromFormat(b'd=%d ld=%ld zd=%zd',
c_int(1), c_long(2),
c_size_t(3)),
b'd=1 ld=2 zd=3')
self.assertEqual(PyBytes_FromFormat(b'd=%d ld=%ld zd=%zd',
c_int(-1), c_long(-2),
c_size_t(-3)),
b'd=-1 ld=-2 zd=-3')
self.assertEqual(PyBytes_FromFormat(b'u=%u lu=%lu zu=%zu',
c_uint(123), c_ulong(456),
c_size_t(789)),
b'u=123 lu=456 zu=789')
self.assertEqual(PyBytes_FromFormat(b'i=%i', c_int(123)),
b'i=123')
self.assertEqual(PyBytes_FromFormat(b'i=%i', c_int(-123)),
b'i=-123')
self.assertEqual(PyBytes_FromFormat(b'x=%x', c_int(0xabc)),
b'x=abc')
sizeof_ptr = ctypes.sizeof(c_char_p)
if os.name == 'nt':
# Windows (MSCRT)
ptr_format = '0x%0{}X'.format(2 * sizeof_ptr)
def ptr_formatter(ptr):
return (ptr_format % ptr)
else:
# UNIX (glibc)
def ptr_formatter(ptr):
return '%#x' % ptr
ptr = 0xabcdef
self.assertEqual(PyBytes_FromFormat(b'ptr=%p', c_char_p(ptr)),
('ptr=' + ptr_formatter(ptr)).encode('ascii'))
self.assertEqual(PyBytes_FromFormat(b's=%s', c_char_p(b'cstr')),
b's=cstr')
# test minimum and maximum integer values
size_max = c_size_t(-1).value
for formatstr, ctypes_type, value, py_formatter in (
(b'%d', c_int, _testcapi.INT_MIN, str),
(b'%d', c_int, _testcapi.INT_MAX, str),
(b'%ld', c_long, _testcapi.LONG_MIN, str),
(b'%ld', c_long, _testcapi.LONG_MAX, str),
(b'%lu', c_ulong, _testcapi.ULONG_MAX, str),
(b'%zd', c_ssize_t, _testcapi.PY_SSIZE_T_MIN, str),
(b'%zd', c_ssize_t, _testcapi.PY_SSIZE_T_MAX, str),
(b'%zu', c_size_t, size_max, str),
(b'%p', c_char_p, size_max, ptr_formatter),
):
self.assertEqual(PyBytes_FromFormat(formatstr, ctypes_type(value)),
py_formatter(value).encode('ascii')),
# width and precision (width is currently ignored)
self.assertEqual(PyBytes_FromFormat(b'%5s', b'a'),
b'a')
self.assertEqual(PyBytes_FromFormat(b'%.3s', b'abcdef'),
b'abc')
# '%%' formatter
self.assertEqual(PyBytes_FromFormat(b'%%'),
b'%')
self.assertEqual(PyBytes_FromFormat(b'[%%]'),
b'[%]')
self.assertEqual(PyBytes_FromFormat(b'%%%c', c_int(ord('_'))),
b'%_')
self.assertEqual(PyBytes_FromFormat(b'%%s'),
b'%s')
# Invalid formats and partial formatting
self.assertEqual(PyBytes_FromFormat(b'%'), b'%')
self.assertEqual(PyBytes_FromFormat(b'x=%i y=%', c_int(2), c_int(3)),
b'x=2 y=%')
# Issue #19969: %c must raise OverflowError for values
# not in the range [0; 255]
self.assertRaises(OverflowError,
PyBytes_FromFormat, b'%c', c_int(-1))
self.assertRaises(OverflowError,
PyBytes_FromFormat, b'%c', c_int(256))
# Issue #33817: empty strings
self.assertEqual(PyBytes_FromFormat(b''),
b'')
self.assertEqual(PyBytes_FromFormat(b'%s', b''),
b'')
def test_bytes_blocking(self):
class IterationBlocked(list):
__bytes__ = None
i = [0, 1, 2, 3]
self.assertEqual(bytes(i), b'\x00\x01\x02\x03')
self.assertRaises(TypeError, bytes, IterationBlocked(i))
# At least in CPython, because bytes.__new__ and the C API
# PyBytes_FromObject have different fallback rules, integer
# fallback is handled specially, so test separately.
class IntBlocked(int):
__bytes__ = None
self.assertEqual(bytes(3), b'\0\0\0')
self.assertRaises(TypeError, bytes, IntBlocked(3))
# While there is no separately-defined rule for handling bytes
# subclasses differently from other buffer-interface classes,
# an implementation may well special-case them (as CPython 2.x
# str did), so test them separately.
class BytesSubclassBlocked(bytes):
__bytes__ = None
self.assertEqual(bytes(b'ab'), b'ab')
self.assertRaises(TypeError, bytes, BytesSubclassBlocked(b'ab'))
class BufferBlocked(bytearray):
__bytes__ = None
ba, bb = bytearray(b'ab'), BufferBlocked(b'ab')
self.assertEqual(bytes(ba), b'ab')
self.assertRaises(TypeError, bytes, bb)
def test_repeat_id_preserving(self):
a = b'123abc1@'
b = b'456zyx-+'
self.assertEqual(id(a), id(a))
self.assertNotEqual(id(a), id(b))
self.assertNotEqual(id(a), id(a * -4))
self.assertNotEqual(id(a), id(a * 0))
self.assertEqual(id(a), id(a * 1))
self.assertEqual(id(a), id(1 * a))
self.assertNotEqual(id(a), id(a * 2))
class SubBytes(bytes):
pass
s = SubBytes(b'qwerty()')
self.assertEqual(id(s), id(s))
self.assertNotEqual(id(s), id(s * -4))
self.assertNotEqual(id(s), id(s * 0))
self.assertNotEqual(id(s), id(s * 1))
self.assertNotEqual(id(s), id(1 * s))
self.assertNotEqual(id(s), id(s * 2))
@support.cpython_only
def test_concat_cpython(self):
# Test optimizations
empty = b''
abc = b"abc"
self.assertIs(abc + empty, abc)
self.assertIs(empty + abc, abc)
self.assertIs(empty + bytearray(b''), empty)
self.assertIs(empty + memoryview(b''), empty)
class Subclass(bytes):
pass
empty_subclass = Subclass(b"")
self.assertIs(empty + empty_subclass, empty)
self.assertIs(empty_subclass + empty, empty)
self.assertIs(abc + empty_subclass, abc)
self.assertIs(empty_subclass + abc, abc)
def assert_copy(result, expected):
self.assertIsNot(result, expected)
self.assertEqual(type(result), bytes)
self.assertEqual(result, expected)
# Copy the string if it's a subclass
def_subclass = Subclass(b"def")
assert_copy(def_subclass + empty, b'def')
assert_copy(empty + def_subclass, b'def')
assert_copy(abc + def_subclass, b'abcdef')
assert_copy(def_subclass + abc, b'defabc')
@support.cpython_only
def test_repeat_cpython(self):
# Test optimizations
for b in b'', b'x', b"abc":
self.assertIs(b * 1, b)
self.assertIs(b * 0, b'')
self.assertIs(b * -23, b'')
class ByteArrayTest(BaseBytesTest, unittest.TestCase):
type2test = bytearray
_testlimitedcapi = import_helper.import_module('_testlimitedcapi')
def test_getitem_error(self):
b = bytearray(b'python')
msg = "bytearray indices must be integers or slices"
with self.assertRaisesRegex(TypeError, msg):
b['a']
def test_setitem_error(self):
b = bytearray(b'python')
msg = "bytearray indices must be integers or slices"
with self.assertRaisesRegex(TypeError, msg):
b['a'] = "python"
def test_nohash(self):
self.assertRaises(TypeError, hash, bytearray())
def test_bytearray_api(self):
short_sample = b"Hello world\n"
sample = short_sample + b"\0"*(20 - len(short_sample))
tfn = tempfile.mktemp()
try:
# Prepare
with open(tfn, "wb") as f:
f.write(short_sample)
# Test readinto
with open(tfn, "rb") as f:
b = bytearray(20)
n = f.readinto(b)
self.assertEqual(n, len(short_sample))
self.assertEqual(list(b), list(sample))
# Test writing in binary mode
with open(tfn, "wb") as f:
f.write(b)
with open(tfn, "rb") as f:
self.assertEqual(f.read(), sample)
# Text mode is ambiguous; don't test
finally:
try:
os.remove(tfn)
except OSError:
pass
def test_mod_concurrent_mutation(self):
# Prevent crash in __mod__ when formatting mutates the bytearray.
# Regression test for https://github.com/python/cpython/issues/142557.
fmt = bytearray(b"%a end")
class S:
def __repr__(self):
fmt.clear()
return "E"
self.assertRaises(BufferError, fmt.__mod__, S())
def test_reverse(self):
b = bytearray(b'hello')
self.assertEqual(b.reverse(), None)
self.assertEqual(b, b'olleh')
b = bytearray(b'hello1') # test even number of items
b.reverse()
self.assertEqual(b, b'1olleh')
b = bytearray()
b.reverse()
self.assertFalse(b)
def test_clear(self):
b = bytearray(b'python')
b.clear()
self.assertEqual(b, b'')
b = bytearray(b'')
b.clear()
self.assertEqual(b, b'')
b = bytearray(b'')
b.append(ord('r'))
b.clear()
b.append(ord('p'))
self.assertEqual(b, b'p')
# Cleared object should be empty.
b = bytearray(b'abc')
b.clear()
self.assertEqual(b.__alloc__(), 0)
base_size = sys.getsizeof(bytearray())
self.assertEqual(sys.getsizeof(b), base_size)
c = b.copy()
self.assertEqual(c.__alloc__(), 0)
self.assertEqual(sys.getsizeof(c), base_size)
def test_copy(self):
b = bytearray(b'abc')
bb = b.copy()
self.assertEqual(bb, b'abc')
b = bytearray(b'')
bb = b.copy()
self.assertEqual(bb, b'')
# test that it's indeed a copy and not a reference
b = bytearray(b'abc')
bb = b.copy()
self.assertEqual(b, bb)
self.assertIsNot(b, bb)
bb.append(ord('d'))
self.assertEqual(bb, b'abcd')
self.assertEqual(b, b'abc')
def test_regexps(self):
def by(s):
return bytearray(map(ord, s))
b = by("Hello, world")
self.assertEqual(re.findall(br"\w+", b), [by("Hello"), by("world")])
def test_resize(self):
ba = bytearray(b'abcdef')
self.assertIsNone(ba.resize(3))
self.assertEqual(ba, bytearray(b'abc'))
self.assertIsNone(ba.resize(10))
self.assertEqual(len(ba), 10)
# Bytes beyond set values must be cleared.
self.assertEqual(ba, bytearray(b'abc\0\0\0\0\0\0\0'))
ba[3:10] = b'defghij'
self.assertEqual(ba, bytearray(b'abcdefghij'))
self.assertIsNone(ba.resize(2 ** 20))
self.assertEqual(len(ba), 2**20)
self.assertEqual(ba, bytearray(b'abcdefghij' + b'\0' * (2 ** 20 - 10)))
self.assertIsNone(ba.resize(0))
self.assertEqual(ba, bytearray())
self.assertIsNone(ba.resize(10))
self.assertEqual(ba, bytearray(b'\0' * 10))
# Subclass
ba = ByteArraySubclass(b'abcdef')
self.assertIsNone(ba.resize(3))
self.assertEqual(ba, bytearray(b'abc'))
# Check arguments
self.assertRaises(TypeError, bytearray().resize)
self.assertRaises(TypeError, bytearray().resize, (10, 10))
self.assertRaises(ValueError, bytearray().resize, -1)
self.assertRaises(ValueError, bytearray().resize, -200)
self.assertRaises(MemoryError, bytearray().resize, sys.maxsize)
self.assertRaises(MemoryError, bytearray(1000).resize, sys.maxsize)
@support.nomemtest
def test_resize_error(self):
# gh-157242: If bytearray.resize() fails (MemoryError),
# the bytearray must be left unchanged.
offset = 3
for logical_offset in (False, True):
with self.subTest(logical_offset=logical_offset):
# grow bytearray
ba = bytearray(b'0123456789')
if logical_offset:
expected = ba[offset:]
del ba[:offset]
else:
expected = ba.copy()
with inject_memory_error(self):
ba.resize(1024)
self.assertEqual(ba, expected)
# shrink bytearray
ba = bytearray(b'0123456789')
if logical_offset:
expected = ba[offset:]
del ba[:offset]
else:
expected = ba.copy()
with inject_memory_error(self):
ba.resize(1)
self.assertEqual(ba, expected)
def test_take_bytes(self):
ba = bytearray(b'ab')
self.assertEqual(ba.take_bytes(), b'ab')
self.assertEqual(len(ba), 0)
self.assertEqual(ba, bytearray(b''))
self.assertEqual(ba.__alloc__(), 0)
base_size = sys.getsizeof(bytearray())
self.assertEqual(sys.getsizeof(ba), base_size)
# Positive and negative slicing.
ba = bytearray(b'abcdef')
self.assertEqual(ba.take_bytes(1), b'a')
self.assertEqual(ba, bytearray(b'bcdef'))
self.assertEqual(len(ba), 5)
self.assertEqual(ba.take_bytes(-5), b'')
self.assertEqual(ba, bytearray(b'bcdef'))
self.assertEqual(len(ba), 5)
self.assertEqual(ba.take_bytes(-3), b'bc')
self.assertEqual(ba, bytearray(b'def'))
self.assertEqual(len(ba), 3)
self.assertEqual(ba.take_bytes(3), b'def')
self.assertEqual(ba, bytearray(b''))
self.assertEqual(len(ba), 0)
# Take nothing from emptiness.
self.assertEqual(ba.take_bytes(0), b'')
self.assertEqual(ba.take_bytes(), b'')
self.assertEqual(ba.take_bytes(None), b'')
# Out of bounds, bad take value.
self.assertRaises(IndexError, ba.take_bytes, -1)
self.assertRaises(TypeError, ba.take_bytes, 3.14)
ba = bytearray(b'abcdef')
self.assertRaises(IndexError, ba.take_bytes, 7)
# Offset between physical and logical start (ob_bytes != ob_start).
ba = bytearray(b'abcde')
del ba[:2]
self.assertEqual(ba, bytearray(b'cde'))
self.assertEqual(ba.take_bytes(), b'cde')
# Overallocation at end.
ba = bytearray(b'abcde')
del ba[-2:]
self.assertEqual(ba, bytearray(b'abc'))
self.assertEqual(ba.take_bytes(), b'abc')
ba = bytearray(b'abcde')
ba.resize(4)
self.assertEqual(ba.take_bytes(), b'abcd')
# Take of a bytearray with references should fail.
ba = bytearray(b'abc')
with memoryview(ba) as mv:
self.assertRaises(BufferError, ba.take_bytes)
self.assertEqual(ba.take_bytes(), b'abc')
# Leaving one byte must not adopt the shared single-byte bytes object
# as the buffer.
ba = bytearray(b'abc')
self.assertEqual(ba.take_bytes(2), b'ab')
ba[0] = ord('A')
self.assertEqual(ba, bytearray(b'A'))
self.assertEqual(ord(b'c'), ord('c'))
@support.nomemtest
def test_take_bytes_error(self):
# gh-157242: If bytearray.take_bytes() fails (MemoryError),
# the bytearray must be left unchanged.
for logical_offset, to_take, start_list in (
(True, 5, (0, 1)),
(False, 5, (0, 1)),
(True, None, (0,)),
):
for start in start_list:
with self.subTest(logical_offset=logical_offset, start=start):
ba = bytearray(b'0123456789')
if logical_offset:
expected = ba[3:]
del ba[:3]
else:
expected = ba.copy()
with inject_memory_error(self, start):
ba.take_bytes(to_take)
self.assertEqual(ba, expected)
@support.cpython_only # tests an implementation detail
def test_take_bytes_optimization(self):
# Validate optimization around taking lots of little chunks out of a
# much bigger buffer. Save work by only copying a little rather than
# moving a lot.
ba = bytearray(b'abcdef' + b'0' * 1000)
start_alloc = ba.__alloc__()
# Take two bytes at a time, checking alloc doesn't change.
self.assertEqual(ba.take_bytes(2), b'ab')
self.assertEqual(ba.__alloc__(), start_alloc)
self.assertEqual(len(ba), 4 + 1000)
self.assertEqual(ba.take_bytes(2), b'cd')
self.assertEqual(ba.__alloc__(), start_alloc)
self.assertEqual(len(ba), 2 + 1000)
self.assertEqual(ba.take_bytes(2), b'ef')
self.assertEqual(ba.__alloc__(), start_alloc)
self.assertEqual(len(ba), 0 + 1000)
self.assertEqual(ba.__alloc__(), start_alloc)
# Take over half, alloc shrinks to exact size.
self.assertEqual(ba.take_bytes(501), b'0' * 501)
self.assertEqual(len(ba), 499)
bytes_header_size = sys.getsizeof(b'')
self.assertEqual(ba.__alloc__(), 499 + bytes_header_size)
def test_take_bytes_hash(self):
# gh-158219: bytearray constructor must not use a bytes object
# if its hash value is already cached.
def encode(string, errors='strict'):
encoded = string.encode('utf-8')
hash(encoded) # a codec may hash its own output
return encoded, len(string)
def hashing_codec(name):
if name != 'test_take_bytes_hash':
return None
return codecs.CodecInfo(encode, None, name=name)
codecs.register(hashing_codec)
self.addCleanup(codecs.unregister, hashing_codec)
ba = bytearray('hello', 'test_take_bytes_hash')
ba[0] = ord('H')
taken = ba.take_bytes()
self.assertEqual(taken, b'Hello')
self.assertEqual(hash(taken), hash(b'Hello'))
def test_take_bytes_reentrant_resize(self):
# gh-153570: n.__index__() can resize the bytearray, so take_bytes()
# must re-read the size afterwards. It cached the size before the
# call and used it for the bounds check and the buffer reads, so a
# reentrant clear() returned freed memory (a use-after-free read).
def take(target, resize, n):
class Evil:
def __index__(self):
resize(target)
return n
return target.take_bytes(Evil())
# clear() during __index__: nothing is left to take.
ba = bytearray(b'abcdefgh')
with self.assertRaises(IndexError):
take(ba, lambda b: b.clear(), 8)
self.assertEqual(ba, b'')
# shrink during __index__: n past the new size is out of range.
ba = bytearray(b'abcdefgh')
with self.assertRaises(IndexError):
take(ba, lambda b: b.__delitem__(slice(4, None)), 8)
self.assertEqual(ba, b'abcd')
# grow during __index__: the take runs against the new, larger size.
ba = bytearray(b'abcd')
self.assertEqual(take(ba, lambda b: b.extend(b'efgh'), 8), b'abcdefgh')
self.assertEqual(ba, b'')
def test_setitem(self):
def setitem_as_mapping(b, i, val):
b[i] = val
def setitem_as_sequence(b, i, val):
self._testlimitedcapi.sequence_setitem(b, i, val)
def do_tests(setitem):
b = bytearray([1, 2, 3])
setitem(b, 1, 100)
self.assertEqual(b, bytearray([1, 100, 3]))
setitem(b, -1, 200)
self.assertEqual(b, bytearray([1, 100, 200]))
setitem(b, 0, Indexable(10))
self.assertEqual(b, bytearray([10, 100, 200]))
try:
setitem(b, 3, 0)
self.fail("Didn't raise IndexError")
except IndexError:
pass
try:
setitem(b, -10, 0)
self.fail("Didn't raise IndexError")
except IndexError:
pass
try:
setitem(b, 0, 256)
self.fail("Didn't raise ValueError")
except ValueError:
pass
try:
setitem(b, 0, Indexable(-1))
self.fail("Didn't raise ValueError")
except ValueError:
pass
try:
setitem(b, 0, object())
self.fail("Didn't raise TypeError")
except TypeError:
pass
with self.subTest("tp_as_mapping"):
do_tests(setitem_as_mapping)
with self.subTest("tp_as_sequence"):
do_tests(setitem_as_sequence)
def test_delitem(self):
def del_as_mapping(b, i):
del b[i]
def del_as_sequence(b, i):
self._testlimitedcapi.sequence_delitem(b, i)
def do_tests(delete):
b = bytearray(range(10))
delete(b, 0)
self.assertEqual(b, bytearray(range(1, 10)))
delete(b, -1)
self.assertEqual(b, bytearray(range(1, 9)))
delete(b, 4)
self.assertEqual(b, bytearray([1, 2, 3, 4, 6, 7, 8]))
with self.subTest("tp_as_mapping"):
do_tests(del_as_mapping)
with self.subTest("tp_as_sequence"):
do_tests(del_as_sequence)
def test_setslice(self):
b = bytearray(range(10))
self.assertEqual(list(b), list(range(10)))
b[0:5] = bytearray([1, 1, 1, 1, 1])
self.assertEqual(b, bytearray([1, 1, 1, 1, 1, 5, 6, 7, 8, 9]))
del b[0:-5]
self.assertEqual(b, bytearray([5, 6, 7, 8, 9]))
b[0:0] = bytearray([0, 1, 2, 3, 4])
self.assertEqual(b, bytearray(range(10)))
b[-7:-3] = bytearray([100, 101])
self.assertEqual(b, bytearray([0, 1, 2, 100, 101, 7, 8, 9]))
b[3:5] = [3, 4, 5, 6]
self.assertEqual(b, bytearray(range(10)))
b[3:0] = [42, 42, 42]
self.assertEqual(b, bytearray([0, 1, 2, 42, 42, 42, 3, 4, 5, 6, 7, 8, 9]))
b[3:] = b'foo'
self.assertEqual(b, bytearray([0, 1, 2, 102, 111, 111]))
b[:3] = memoryview(b'foo')
self.assertEqual(b, bytearray([102, 111, 111, 102, 111, 111]))
b[3:4] = []
self.assertEqual(b, bytearray([102, 111, 111, 111, 111]))
for elem in [5, -5, 0, int(10e20), 'str', 2.3,
['a', 'b'], [b'a', b'b'], [[]]]:
with self.assertRaises(TypeError):
b[3:4] = elem
for elem in [[254, 255, 256], [-256, 9000]]:
with self.assertRaises(ValueError):
b[3:4] = elem
def test_setslice_extend(self):
# Exercise the resizing logic (see issue #19087)
b = bytearray(range(100))
self.assertEqual(list(b), list(range(100)))
del b[:10]
self.assertEqual(list(b), list(range(10, 100)))
b.extend(range(100, 110))
self.assertEqual(list(b), list(range(10, 110)))
def test_fifo_overrun(self):
# Test for issue #23985, a buffer overrun when implementing a FIFO
# Build Python in pydebug mode for best results.
b = bytearray(10)
b.pop() # Defeat expanding buffer off-by-one quirk
del b[:1] # Advance start pointer without reallocating
b += bytes(2) # Append exactly the number of deleted bytes
del b # Free memory buffer, allowing pydebug verification
def test_del_expand(self):
# Reducing the size should not expand the buffer (issue #23985)
b = bytearray(10)
size = sys.getsizeof(b)
del b[:1]
self.assertLessEqual(sys.getsizeof(b), size)
def test_extended_set_del_slice(self):
indices = (0, None, 1, 3, 19, 300, 1<<333, sys.maxsize,
-1, -2, -31, -300)
for start in indices:
for stop in indices:
# Skip invalid step 0
for step in indices[1:]:
L = list(range(255))
b = bytearray(L)
# Make sure we have a slice of exactly the right length,
# but with different data.
data = L[start:stop:step]
data.reverse()
L[start:stop:step] = data
b[start:stop:step] = data
self.assertEqual(b, bytearray(L))
del L[start:stop:step]
del b[start:stop:step]
self.assertEqual(b, bytearray(L))
def test_setslice_trap(self):
# This test verifies that we correctly handle assigning self
# to a slice of self (the old Lambert Meertens trap).
b = bytearray(range(256))
b[8:] = b
self.assertEqual(b, bytearray(list(range(8)) + list(range(256))))
def test_setslice_reentrant_resize(self):
# gh-153578: a buffer argument whose __buffer__ resizes the bytearray
# while the buffer is being acquired must not leave the slice bounds
# with lo > hi, which drove a negative-size memmove (an out-of-bounds
# write) in the setslice path reached through extend().
class Evil:
def __init__(self, resize):
self.resize = resize
def __buffer__(self, flags):
self.resize()
return memoryview(b'ABCDEFGH')
# clear() during __buffer__: extend appends to the emptied bytearray.
b = bytearray(b'x' * 100)
b.extend(Evil(b.clear))
self.assertEqual(b, b'ABCDEFGH')
# partial shrink during __buffer__.
b = bytearray(b'x' * 100)
b.extend(Evil(lambda: b.__delitem__(slice(30, None))))
self.assertEqual(b, b'x' * 30 + b'ABCDEFGH')
# grow during __buffer__: the data lands at the original end.
b = bytearray(b'x' * 10)
b.extend(Evil(lambda: b.extend(b'y' * 100)))
self.assertEqual(b, b'x' * 10 + b'ABCDEFGH' + b'y' * 100)
def test_iconcat(self):
b = bytearray(b"abc")
b1 = b
b += b"def"
self.assertEqual(b, b"abcdef")
self.assertEqual(b, b1)
self.assertIs(b, b1)
b += b"xyz"
self.assertEqual(b, b"abcdefxyz")
try:
b += ""
except TypeError:
pass
else:
self.fail("bytes += unicode didn't raise TypeError")
def test_irepeat(self):
b = bytearray(b"abc")
b1 = b
b *= 3
self.assertEqual(b, b"abcabcabc")
self.assertEqual(b, b1)
self.assertIs(b, b1)
def test_irepeat_1char(self):
b = bytearray(b"x")
b1 = b
b *= 100
self.assertEqual(b, b"x"*100)
self.assertEqual(b, b1)
self.assertIs(b, b1)
def test_alloc(self):
b = bytearray()
alloc = b.__alloc__()
self.assertGreaterEqual(alloc, 0)
seq = [alloc]
for i in range(100):
b += b"x"
alloc = b.__alloc__()
self.assertGreater(alloc, len(b)) # including trailing null byte
if alloc not in seq:
seq.append(alloc)
def test_init_alloc(self):
b = bytearray()
def g():
for i in range(1, 100):
yield i
a = list(b)
self.assertEqual(a, list(range(1, len(a)+1)))
self.assertEqual(len(b), len(a))
self.assertLessEqual(len(b), i)
alloc = b.__alloc__()
self.assertGreater(alloc, len(b)) # including trailing null byte
b.__init__(g())
self.assertEqual(list(b), list(range(1, 100)))
self.assertEqual(len(b), 99)
alloc = b.__alloc__()
self.assertGreater(alloc, len(b))
def test_extend(self):
orig = b'hello'
a = bytearray(orig)
a.extend(a)
self.assertEqual(a, orig + orig)
self.assertEqual(a[5:], orig)
a = bytearray(b'')
# Test iterators that don't have a __length_hint__
a.extend(map(int, orig * 25))
a.extend(int(x) for x in orig * 25)
self.assertEqual(a, orig * 50)
self.assertEqual(a[-5:], orig)
a = bytearray(b'')
a.extend(iter(map(int, orig * 50)))
self.assertEqual(a, orig * 50)
self.assertEqual(a[-5:], orig)
a = bytearray(b'')
a.extend(list(map(int, orig * 50)))
self.assertEqual(a, orig * 50)
self.assertEqual(a[-5:], orig)
a = bytearray(b'')
self.assertRaises(ValueError, a.extend, [0, 1, 2, 256])
self.assertRaises(ValueError, a.extend, [0, 1, 2, -1])
self.assertEqual(len(a), 0)
a = bytearray(b'')
a.extend([Indexable(ord('a'))])
self.assertEqual(a, b'a')
a = bytearray(b'abc')
self.assertRaisesRegex(TypeError, # Override for string.
"expected iterable of integers; got: 'str'",
a.extend, 'def')
self.assertRaisesRegex(TypeError, # But not for others.
"can't extend bytearray with float",
a.extend, 1.0)
def test_remove(self):
b = bytearray(b'hello')
b.remove(ord('l'))
self.assertEqual(b, b'helo')
b.remove(ord('l'))
self.assertEqual(b, b'heo')
self.assertRaises(ValueError, lambda: b.remove(ord('l')))
self.assertRaises(ValueError, lambda: b.remove(400))
self.assertRaises(TypeError, lambda: b.remove('e'))
# remove first and last
b.remove(ord('o'))
b.remove(ord('h'))
self.assertEqual(b, b'e')
self.assertRaises(TypeError, lambda: b.remove(b'e'))
b.remove(Indexable(ord('e')))
self.assertEqual(b, b'')
# test values outside of the ascii range: (0, 127)
c = bytearray([126, 127, 128, 129])
c.remove(127)
self.assertEqual(c, bytes([126, 128, 129]))
c.remove(129)
self.assertEqual(c, bytes([126, 128]))
def test_pop(self):
b = bytearray(b'world')
self.assertEqual(b.pop(), ord('d'))
self.assertEqual(b.pop(0), ord('w'))
self.assertEqual(b.pop(-2), ord('r'))
self.assertRaises(IndexError, lambda: b.pop(10))
self.assertRaises(IndexError, lambda: bytearray().pop())
# test for issue #6846
self.assertEqual(bytearray(b'\xff').pop(), 0xff)
def test_nosort(self):
self.assertRaises(AttributeError, lambda: bytearray().sort())
def test_append(self):
b = bytearray(b'hell')
b.append(ord('o'))
self.assertEqual(b, b'hello')
self.assertEqual(b.append(100), None)
b = bytearray()
b.append(ord('A'))
self.assertEqual(len(b), 1)
self.assertRaises(TypeError, lambda: b.append(b'o'))
b = bytearray()
b.append(Indexable(ord('A')))
self.assertEqual(b, b'A')
def test_insert(self):
b = bytearray(b'msssspp')
b.insert(1, ord('i'))
b.insert(4, ord('i'))
b.insert(-2, ord('i'))
b.insert(1000, ord('i'))
self.assertEqual(b, b'mississippi')
self.assertRaises(TypeError, lambda: b.insert(0, b'1'))
b = bytearray()
b.insert(0, Indexable(ord('A')))
self.assertEqual(b, b'A')
def test_copied(self):
# Issue 4348. Make sure that operations that don't mutate the array
# copy the bytes.
b = bytearray(b'abc')
self.assertIsNot(b, b.replace(b'abc', b'cde', 0))
t = bytearray([i for i in range(256)])
x = bytearray(b'')
self.assertIsNot(x, x.translate(t))
def test_partition_bytearray_doesnt_share_nullstring(self):
a, b, c = bytearray(b"x").partition(b"y")
self.assertEqual(b, b"")
self.assertEqual(c, b"")
self.assertIsNot(b, c)
b += b"!"
self.assertEqual(c, b"")
a, b, c = bytearray(b"x").partition(b"y")
self.assertEqual(b, b"")
self.assertEqual(c, b"")
# Same for rpartition
b, c, a = bytearray(b"x").rpartition(b"y")
self.assertEqual(b, b"")
self.assertEqual(c, b"")
self.assertIsNot(b, c)
b += b"!"
self.assertEqual(c, b"")
c, b, a = bytearray(b"x").rpartition(b"y")
self.assertEqual(b, b"")
self.assertEqual(c, b"")
def test_resize_forbidden(self):
# #4509: can't resize a bytearray when there are buffer exports, even
# if it wouldn't reallocate the underlying buffer.
# Furthermore, no destructive changes to the buffer may be applied
# before raising the error.
b = bytearray(10)
v = memoryview(b)
def manual_resize(n):
b[1:-1] = range(n + 1, 2*n - 1)
b.resize(10)
orig = b[:]
self.assertRaises(BufferError, b.resize, 11)
self.assertRaises(BufferError, manual_resize, 11)
self.assertEqual(b, orig)
self.assertRaises(BufferError, b.resize, 9)
self.assertEqual(b, orig)
self.assertRaises(BufferError, b.resize, 0)
self.assertEqual(b, orig)
# Other operations implying resize
self.assertRaises(BufferError, b.pop, 0)
self.assertEqual(b, orig)
self.assertRaises(BufferError, b.remove, b[1])
self.assertEqual(b, orig)
def delitem():
del b[1]
self.assertRaises(BufferError, delitem)
self.assertEqual(b, orig)
# deleting a non-contiguous slice
def delslice():
b[1:-1:2] = b""
self.assertRaises(BufferError, delslice)
self.assertEqual(b, orig)
def test_decode_resize_forbidden(self):
# The storage is pinned while it is decoded, so an error handler
# cannot resize the bytearray.
b = bytearray(b'ab\xffcd')
errors = 'test.bytearray_decode_resize'
def handler(exc):
self.assertRaises(BufferError, b.clear)
self.assertRaises(BufferError, b.append, 0)
return ('?', exc.end)
self.addCleanup(_codecs_unregister_error, errors)
codecs.register_error(errors, handler)
for encoding in 'utf-8', 'utf-8-sig':
with self.subTest(encoding=encoding):
self.assertEqual(b.decode(encoding, errors), 'ab?cd')
self.assertEqual(b, b'ab\xffcd')
def test_decode_subclass_buffer(self):
# decode() decodes the buffer that the object exports.
class B(bytearray):
def __buffer__(self, flags):
return memoryview(b'other')
self.assertEqual(B(b'mine').decode(), 'other')
self.assertEqual(B(b'mine').decode('latin-1'), 'other')
@test.support.cpython_only
def test_obsolete_write_lock(self):
_testcapi = import_helper.import_module('_testcapi')
self.assertRaises(BufferError, _testcapi.getbuffer_with_null_view, bytearray())
def test_iterator_pickling2(self):
orig = bytearray(b'abc')
data = list(b'qwerty')
for proto in range(pickle.HIGHEST_PROTOCOL + 1):
# initial iterator
itorig = iter(orig)
d = pickle.dumps((itorig, orig), proto)
it, b = pickle.loads(d)
b[:] = data
self.assertEqual(type(it), type(itorig))
self.assertEqual(list(it), data)
# running iterator
next(itorig)
d = pickle.dumps((itorig, orig), proto)
it, b = pickle.loads(d)
b[:] = data
self.assertEqual(type(it), type(itorig))
self.assertEqual(list(it), data[1:])
# empty iterator
for i in range(1, len(orig)):
next(itorig)
d = pickle.dumps((itorig, orig), proto)
it, b = pickle.loads(d)
b[:] = data
self.assertEqual(type(it), type(itorig))
self.assertEqual(list(it), data[len(orig):])
# exhausted iterator
self.assertRaises(StopIteration, next, itorig)
d = pickle.dumps((itorig, orig), proto)
it, b = pickle.loads(d)
b[:] = data
self.assertEqual(list(it), [])
test_exhausted_iterator = test.list_tests.CommonTest.test_exhausted_iterator
def test_iterator_length_hint(self):
# Issue 27443: __length_hint__ can return negative integer
ba = bytearray(b'ab')
it = iter(ba)
next(it)
ba.clear()
# Shouldn't raise an error
self.assertEqual(list(it), [])
def test_repeat_after_setslice(self):
# bpo-42924: * used to copy from the wrong memory location
b = bytearray(b'abc')
b[:2] = b'x'
b1 = b * 1
b3 = b * 3
self.assertEqual(b1, b'xc')
self.assertEqual(b1, b)
self.assertEqual(b3, b'xcxcxc')
def test_mutating_index(self):
# bytearray slice assignment can call into python code
# that reallocates the internal buffer
# See gh-91153
class Boom:
def __index__(self):
b.clear()
return 0
with self.subTest("tp_as_mapping"):
b = bytearray(b'Now you see me...')
with self.assertRaises(IndexError):
b[0] = Boom()
with self.subTest("tp_as_sequence"):
b = bytearray(b'Now you see me...')
with self.assertRaises(IndexError):
self._testlimitedcapi.sequence_setitem(b, 0, Boom())
def test_mutating_index_inbounds(self):
# gh-91153 continued
# Ensure buffer is not broken even if length is correct
class MutatesOnIndex:
def __init__(self):
self.ba = bytearray(0x180)
def __index__(self):
self.ba.clear()
self.new_ba = bytearray(0x180) # to catch out-of-bounds writes
self.ba.extend([0] * 0x180) # to check bounds checks
return 0
with self.subTest("skip_bounds_safety"):
instance = MutatesOnIndex()
instance.ba[instance] = ord("?")
self.assertEqual(instance.ba[0], ord("?"), "Assigned bytearray not altered")
self.assertEqual(instance.new_ba, bytearray(0x180), "Wrong object altered")
with self.subTest("skip_bounds_safety_capi"):
instance = MutatesOnIndex()
instance.ba[instance] = ord("?")
self._testlimitedcapi.sequence_setitem(instance.ba, instance, ord("?"))
self.assertEqual(instance.ba[0], ord("?"), "Assigned bytearray not altered")
self.assertEqual(instance.new_ba, bytearray(0x180), "Wrong object altered")
with self.subTest("skip_bounds_safety_slice"):
instance = MutatesOnIndex()
instance.ba[instance:1] = [ord("?")]
self.assertEqual(instance.ba[0], ord("?"), "Assigned bytearray not altered")
self.assertEqual(instance.new_ba, bytearray(0x180), "Wrong object altered")
def test_search_methods_reentrancy_raises_buffererror(self):
# gh-142560: Raise BufferError if buffer mutates during search arg conversion.
class Evil:
def __init__(self, ba):
self.ba = ba
def __buffer__(self, flags):
self.ba.clear()
return memoryview(self.ba)
def __release_buffer__(self, view: memoryview) -> None:
view.release()
def __index__(self):
self.ba.clear()
return ord("A")
def make_case():
ba = bytearray(b"A")
return ba, Evil(ba)
for name in ("find", "count", "index", "rindex", "rfind"):
ba, evil = make_case()
with self.subTest(name):
with self.assertRaises(BufferError):
getattr(ba, name)(evil)
ba, evil = make_case()
with self.assertRaises(BufferError):
evil in ba
with self.assertRaises(BufferError):
ba.split(evil)
with self.assertRaises(BufferError):
ba.rsplit(evil)
def test_extend_empty_buffer_overflow(self):
# gh-143003
class EvilIter:
def __iter__(self):
return self
def __next__(self):
return next(source)
def __length_hint__(self):
return 0
# Use ASCII digits so float() takes the fast path that expects a NUL terminator.
source = iter(b'42')
ba = bytearray()
ba.extend(EvilIter())
self.assertRaises(ValueError, float, bytearray())
def test_hex_use_after_free(self):
# Prevent UAF in bytearray.hex(sep) with re-entrant sep.__len__.
# Regression test for https://github.com/python/cpython/issues/143195.
ba = bytearray(b'\xAA')
class S(bytes):
def __len__(self):
ba.clear()
return 1
self.assertRaises(BufferError, ba.hex, S(b':'))
def test_no_init_called(self):
# A bytearray created without calling bytearray.__init__
# should not crash the interpreter (see gh-153419).
def bytearray_new():
return bytearray.__new__(bytearray)
bytearray_new().insert(0, 1)
bytearray_new().extend(b"x")
bytearray_new().extend([1, 2, 3])
bytearray_new().resize(4)
bytearray_new().__init__(5)
bytearray_new().__init__(b"xyz")
bytearray_new().take_bytes()
bytearray_new().take_bytes(0)
a = bytearray_new()
a.append(1)
a = bytearray_new()
a += b"x"
a = bytearray_new()
a[:] = b"xyz"
def test_reinit_length(self):
# There is a shortcut taken when resizing, where alloc/2 < newsize.
# In this case, the existing buffer is reused, rather than reset.
# If this happens when newsize == 0 and alloc == 1, then various
# code assumptions can be violated. This test should catch those
# in debug builds. (see gh-153419)
a = bytearray(1)
a.__init__()
self.assertEqual(a, b"")
def test_reinit_with_view(self):
a = bytearray()
with memoryview(a):
self.assertRaises(BufferError, a.__init__, "x", "ascii")
self.assertEqual(a, b"")
class AssortedBytesTest(unittest.TestCase):
#
# Test various combinations of bytes and bytearray
#
def test_bytes_repr(self, f=repr):
self.assertEqual(f(b''), "b''")
self.assertEqual(f(b"abc"), "b'abc'")
self.assertEqual(f(bytes([92])), r"b'\\'")
self.assertEqual(f(bytes([0, 1, 254, 255])), r"b'\x00\x01\xfe\xff'")
self.assertEqual(f(b'\a\b\t\n\v\f\r'), r"b'\x07\x08\t\n\x0b\x0c\r'")
self.assertEqual(f(b'"'), """b'"'""") # '"'
self.assertEqual(f(b"'"), '''b"'"''') # "'"
self.assertEqual(f(b"'\""), r"""b'\'"'""") # '\'"'
self.assertEqual(f(b"\"'\""), r"""b'"\'"'""") # '"\'"'
self.assertEqual(f(b"'\"'"), r"""b'\'"\''""") # '\'"\''
self.assertEqual(f(BytesSubclass(b"abc")), "b'abc'")
def test_bytearray_repr(self, f=repr):
self.assertEqual(f(bytearray()), "bytearray(b'')")
self.assertEqual(f(bytearray(b'abc')), "bytearray(b'abc')")
self.assertEqual(f(bytearray([92])), r"bytearray(b'\\')")
self.assertEqual(f(bytearray([0, 1, 254, 255])),
r"bytearray(b'\x00\x01\xfe\xff')")
self.assertEqual(f(bytearray([7, 8, 9, 10, 11, 12, 13])),
r"bytearray(b'\x07\x08\t\n\x0b\x0c\r')")
self.assertEqual(f(bytearray(b'"')), """bytearray(b'"')""") # '"'
self.assertEqual(f(bytearray(b"'")), '''bytearray(b"'")''') # "'"
self.assertEqual(f(bytearray(b"'\"")), r"""bytearray(b'\'"')""") # '\'"'
self.assertEqual(f(bytearray(b"\"'\"")), r"""bytearray(b'"\'"')""") # '"\'"'
self.assertEqual(f(bytearray(b'\'"\'')), r"""bytearray(b'\'"\'')""") # '\'"\''
self.assertEqual(f(ByteArraySubclass(b"abc")), "ByteArraySubclass(b'abc')")
self.assertEqual(f(ByteArraySubclass.Nested(b"abc")), "Nested(b'abc')")
self.assertEqual(f(ByteArraySubclass.Ŭñıçöđë(b"abc")), "Ŭñıçöđë(b'abc')")
@check_bytes_warnings
def test_bytes_str(self):
self.test_bytes_repr(str)
@check_bytes_warnings
def test_bytearray_str(self):
self.test_bytearray_repr(str)
@check_bytes_warnings
def test_format(self):
for b in b'abc', bytearray(b'abc'):
self.assertEqual(format(b), str(b))
self.assertEqual(format(b, ''), str(b))
with self.assertRaisesRegex(TypeError,
r'\b%s\b' % re.escape(type(b).__name__)):
format(b, 's')
def test_compare_bytes_to_bytearray(self):
self.assertEqual(b"abc" == bytes(b"abc"), True)
self.assertEqual(b"ab" != bytes(b"abc"), True)
self.assertEqual(b"ab" <= bytes(b"abc"), True)
self.assertEqual(b"ab" < bytes(b"abc"), True)
self.assertEqual(b"abc" >= bytes(b"ab"), True)
self.assertEqual(b"abc" > bytes(b"ab"), True)
self.assertEqual(b"abc" != bytes(b"abc"), False)
self.assertEqual(b"ab" == bytes(b"abc"), False)
self.assertEqual(b"ab" > bytes(b"abc"), False)
self.assertEqual(b"ab" >= bytes(b"abc"), False)
self.assertEqual(b"abc" < bytes(b"ab"), False)
self.assertEqual(b"abc" <= bytes(b"ab"), False)
self.assertEqual(bytes(b"abc") == b"abc", True)
self.assertEqual(bytes(b"ab") != b"abc", True)
self.assertEqual(bytes(b"ab") <= b"abc", True)
self.assertEqual(bytes(b"ab") < b"abc", True)
self.assertEqual(bytes(b"abc") >= b"ab", True)
self.assertEqual(bytes(b"abc") > b"ab", True)
self.assertEqual(bytes(b"abc") != b"abc", False)
self.assertEqual(bytes(b"ab") == b"abc", False)
self.assertEqual(bytes(b"ab") > b"abc", False)
self.assertEqual(bytes(b"ab") >= b"abc", False)
self.assertEqual(bytes(b"abc") < b"ab", False)
self.assertEqual(bytes(b"abc") <= b"ab", False)
@test.support.requires_docstrings
def test_doc(self):
self.assertIsNotNone(bytearray.__doc__)
self.assertStartsWith(bytearray.__doc__, "bytearray(")
self.assertIsNotNone(bytes.__doc__)
self.assertStartsWith(bytes.__doc__, "bytes(")
def test_from_bytearray(self):
sample = bytes(b"Hello world\n\x80\x81\xfe\xff")
buf = memoryview(sample)
b = bytearray(buf)
self.assertEqual(b, bytearray(sample))
def test_literal(self):
tests = [
(b"Wonderful spam", "Wonderful spam"),
(br"Wonderful spam too", "Wonderful spam too"),
(b"\xaa\x00\000\200", "\xaa\x00\000\200"),
(br"\xaa\x00\000\200", r"\xaa\x00\000\200"),
]
for b, s in tests:
self.assertEqual(b, bytearray(s, 'latin-1'))
for c in range(128, 256):
self.assertRaises(SyntaxError, eval,
'b"%s"' % chr(c))
def test_split_bytearray(self):
self.assertEqual(b'a b'.split(memoryview(b' ')), [b'a', b'b'])
def test_rsplit_bytearray(self):
self.assertEqual(b'a b'.rsplit(memoryview(b' ')), [b'a', b'b'])
def test_return_self(self):
# bytearray.replace must always return a new bytearray
b = bytearray()
self.assertIsNot(b.replace(b'', b''), b)
@unittest.skipUnless(sys.flags.bytes_warning,
"BytesWarning is needed for this test: use -bb option")
def test_compare(self):
def bytes_warning():
return warnings_helper.check_warnings(('', BytesWarning))
with bytes_warning():
b'' == ''
with bytes_warning():
'' == b''
with bytes_warning():
b'' != ''
with bytes_warning():
'' != b''
with bytes_warning():
bytearray(b'') == ''
with bytes_warning():
'' == bytearray(b'')
with bytes_warning():
bytearray(b'') != ''
with bytes_warning():
'' != bytearray(b'')
with bytes_warning():
b'\0' == 0
with bytes_warning():
0 == b'\0'
with bytes_warning():
b'\0' != 0
with bytes_warning():
0 != b'\0'
# Optimizations:
# __iter__? (optimization)
# __reversed__? (optimization)
# XXX More string methods? (Those that don't use character properties)
# There are tests in string_tests.py that are more
# comprehensive for things like partition, etc.
# Unfortunately they are all bundled with tests that
# are not appropriate for bytes
# I've started porting some of those into bytearray_tests.py, we should port
# the rest that make sense (the code can be cleaned up to use modern
# unittest methods at the same time).
class BytearrayPEP3137Test(unittest.TestCase):
def marshal(self, x):
return bytearray(x)
def test_returns_new_copy(self):
val = self.marshal(b'1234')
# On immutable types these MAY return a reference to themselves
# but on mutable types like bytearray they MUST return a new copy.
for methname in ('zfill', 'rjust', 'ljust', 'center'):
method = getattr(val, methname)
newval = method(3)
self.assertEqual(val, newval)
self.assertIsNot(val, newval,
methname+' returned self on a mutable object')
for expr in ('val.split()[0]', 'val.rsplit()[0]',
'val.partition(b".")[0]', 'val.rpartition(b".")[2]',
'val.splitlines()[0]', 'val.replace(b"", b"")'):
newval = eval(expr)
self.assertEqual(val, newval)
self.assertIsNot(val, newval,
expr+' returned val on a mutable object')
sep = self.marshal(b'')
newval = sep.join([val])
self.assertEqual(val, newval)
self.assertIsNot(val, newval)
class FixedStringTest(test.string_tests.BaseTest):
def fixtype(self, obj):
if isinstance(obj, str):
return self.type2test(obj.encode("utf-8"))
return super().fixtype(obj)
contains_bytes = True
def test_mixed_cmp(self):
a = self.type2test(b'ab')
for t in bytes, bytearray, BytesSubclass, ByteArraySubclass:
with self.subTest(t.__name__):
self._assert_cmp(a, t(b'ab'), 0)
self._assert_cmp(a, t(b'a'), 1)
self._assert_cmp(a, t(b'ac'), -1)
class ByteArrayAsStringTest(FixedStringTest, unittest.TestCase):
type2test = bytearray
class BytesAsStringTest(FixedStringTest, unittest.TestCase):
type2test = bytes
class SubclassTest:
def test_basic(self):
self.assertIsSubclass(self.type2test, self.basetype)
self.assertIsInstance(self.type2test(), self.basetype)
a, b = b"abcd", b"efgh"
_a, _b = self.type2test(a), self.type2test(b)
# test comparison operators with subclass instances
self.assertTrue(_a == _a)
self.assertTrue(_a != _b)
self.assertTrue(_a < _b)
self.assertTrue(_a <= _b)
self.assertTrue(_b >= _a)
self.assertTrue(_b > _a)
self.assertIsNot(_a, a)
# test concat of subclass instances
self.assertEqual(a + b, _a + _b)
self.assertEqual(a + b, a + _b)
self.assertEqual(a + b, _a + b)
# test repeat
self.assertTrue(a*5 == _a*5)
def test_join(self):
# Make sure join returns a NEW object for single item sequences
# involving a subclass.
# Make sure that it is of the appropriate type.
s1 = self.type2test(b"abcd")
s2 = self.basetype().join([s1])
self.assertIsNot(s1, s2)
self.assertIs(type(s2), self.basetype, type(s2))
# Test reverse, calling join on subclass
s3 = s1.join([b"abcd"])
self.assertIs(type(s3), self.basetype)
def test_pickle(self):
a = self.type2test(b"abcd")
a.x = 10
a.z = self.type2test(b"efgh")
for proto in range(pickle.HIGHEST_PROTOCOL + 1):
b = pickle.loads(pickle.dumps(a, proto))
self.assertNotEqual(id(a), id(b))
self.assertEqual(a, b)
self.assertEqual(a.x, b.x)
self.assertEqual(a.z, b.z)
self.assertEqual(type(a), type(b))
self.assertEqual(type(a.z), type(b.z))
self.assertNotHasAttr(b, 'y')
def test_copy(self):
a = self.type2test(b"abcd")
a.x = 10
a.z = self.type2test(b"efgh")
for copy_method in (copy.copy, copy.deepcopy):
b = copy_method(a)
self.assertNotEqual(id(a), id(b))
self.assertEqual(a, b)
self.assertEqual(a.x, b.x)
self.assertEqual(a.z, b.z)
self.assertEqual(type(a), type(b))
self.assertEqual(type(a.z), type(b.z))
self.assertNotHasAttr(b, 'y')
def test_fromhex(self):
b = self.type2test.fromhex('1a2B30')
self.assertEqual(b, b'\x1a\x2b\x30')
self.assertIs(type(b), self.type2test)
class B1(self.basetype):
def __new__(cls, value):
me = self.basetype.__new__(cls, value)
me.foo = 'bar'
return me
b = B1.fromhex('1a2B30')
self.assertEqual(b, b'\x1a\x2b\x30')
self.assertIs(type(b), B1)
self.assertEqual(b.foo, 'bar')
class B2(self.basetype):
def __init__(me, *args, **kwargs):
if self.basetype is not bytes:
self.basetype.__init__(me, *args, **kwargs)
me.foo = 'bar'
b = B2.fromhex('1a2B30')
self.assertEqual(b, b'\x1a\x2b\x30')
self.assertIs(type(b), B2)
self.assertEqual(b.foo, 'bar')
class ByteArraySubclass(bytearray):
class Nested(bytearray):
pass
class Ŭñıçöđë(bytearray):
pass
class ByteArraySubclassWithSlots(bytearray):
__slots__ = ('x', 'y', '__dict__')
class BytesSubclass(bytes):
pass
class OtherBytesSubclass(bytes):
pass
class WithBytes:
def __init__(self, value):
self.value = value
def __bytes__(self):
return self.value
class ByteArraySubclassTest(SubclassTest, unittest.TestCase):
basetype = bytearray
type2test = ByteArraySubclass
def test_init_override(self):
class subclass(bytearray):
def __init__(me, newarg=1, *args, **kwargs):
bytearray.__init__(me, *args, **kwargs)
x = subclass(4, b"abcd")
x = subclass(4, source=b"abcd")
self.assertEqual(x, b"abcd")
x = subclass(newarg=4, source=b"abcd")
self.assertEqual(x, b"abcd")
class ByteArraySubclassWithSlotsTest(SubclassTest, unittest.TestCase):
basetype = bytearray
type2test = ByteArraySubclassWithSlots
class BytesSubclassTest(SubclassTest, unittest.TestCase):
basetype = bytes
type2test = BytesSubclass
class FreeThreadingTest(unittest.TestCase):
@unittest.skipUnless(support.Py_GIL_DISABLED, 'this test can only possibly fail with GIL disabled')
@threading_helper.reap_threads
@threading_helper.requires_working_threading()
def test_free_threading_bytearray(self):
# Test pretty much everything that can break under free-threading.
# Non-deterministic, but at least one of these things will fail if
# bytearray module is not free-thread safe.
def clear(b, a, *args): # MODIFIES!
b.wait()
try: a.clear()
except BufferError: pass
def clear2(b, a, c): # MODIFIES c!
b.wait()
try: c.clear()
except BufferError: pass
def pop1(b, a): # MODIFIES!
b.wait()
try: a.pop()
except IndexError: pass
def append1(b, a): # MODIFIES!
b.wait()
a.append(0)
def insert1(b, a): # MODIFIES!
b.wait()
a.insert(0, 0)
def extend(b, a): # MODIFIES!
c = bytearray(b'0' * 0x400000)
b.wait()
a.extend(c)
def remove(b, a): # MODIFIES!
c = ord('0')
b.wait()
try: a.remove(c)
except ValueError: pass
def reverse(b, a): # modifies inplace
b.wait()
a.reverse()
def reduce(b, a):
b.wait()
a.__reduce__()
def reduceex2(b, a):
b.wait()
a.__reduce_ex__(2)
def reduceex3(b, a):
b.wait()
c = a.__reduce_ex__(3)
assert not c[1] or 0xdd not in c[1][0]
def count0(b, a):
b.wait()
a.count(0)
def decode(b, a):
b.wait()
a.decode()
def find(b, a):
c = bytearray(b'0' * 0x40000)
b.wait()
a.find(c)
def hex(b, a):
b.wait()
a.hex('_')
def join(b, a):
b.wait()
a.join([b'1', b'2', b'3'])
def replace(b, a):
b.wait()
a.replace(b'0', b'')
def maketrans(b, a, c):
b.wait()
try: a.maketrans(a, c)
except ValueError: pass
def translate(b, a, c):
b.wait()
a.translate(c)
def copy(b, a):
b.wait()
c = a.copy()
if c: assert c[0] == 48 # '0'
def endswith(b, a):
b.wait()
assert not a.endswith(b'\xdd')
def index(b, a):
b.wait()
try: a.index(b'\xdd')
except ValueError: return
assert False
def lstrip(b, a):
b.wait()
assert not a.lstrip(b'0')
def partition(b, a):
b.wait()
assert not a.partition(b'\xdd')[2]
def removeprefix(b, a):
b.wait()
assert not a.removeprefix(b'0')
def removesuffix(b, a):
b.wait()
assert not a.removesuffix(b'0')
def rfind(b, a):
b.wait()
assert a.rfind(b'\xdd') == -1
def rindex(b, a):
b.wait()
try: a.rindex(b'\xdd')
except ValueError: return
assert False
def rpartition(b, a):
b.wait()
assert not a.rpartition(b'\xdd')[0]
def rsplit(b, a):
b.wait()
assert len(a.rsplit(b'\xdd')) == 1
def rstrip(b, a):
b.wait()
assert not a.rstrip(b'0')
def split(b, a):
b.wait()
assert len(a.split(b'\xdd')) == 1
def splitlines(b, a):
b.wait()
l = len(a.splitlines())
assert l > 1 or l == 0
def startswith(b, a):
b.wait()
assert not a.startswith(b'\xdd')
def strip(b, a):
b.wait()
assert not a.strip(b'0')
def repeat(b, a):
b.wait()
a * 2
def contains(b, a):
b.wait()
assert 0xdd not in a
def iconcat(b, a): # MODIFIES!
c = bytearray(b'0' * 0x400000)
b.wait()
a += c
def subscript(b, a):
b.wait()
try: assert a[0] != 0xdd
except IndexError: pass
def ass_subscript(b, a): # MODIFIES!
c = bytearray(b'0' * 0x400000)
b.wait()
a[:] = c
def ass_subscript2(b, a, c): # MODIFIES!
b.wait()
a[:] = c
assert b'\xdd' not in a
def mod(b, a):
c = tuple(range(4096))
b.wait()
try: a % c
except TypeError: pass
def mod2(b, a, c):
b.wait()
d = a % c
assert b'\xdd' not in d
def repr_(b, a):
b.wait()
repr(a)
def capitalize(b, a):
b.wait()
c = a.capitalize()
assert not c or c[0] not in (0xdd, 0xcd)
def center(b, a):
b.wait()
c = a.center(0x60000)
assert not c or c[0x20000] not in (0xdd, 0xcd)
def expandtabs(b, a):
b.wait()
c = a.expandtabs()
assert not c or c[0] not in (0xdd, 0xcd)
def ljust(b, a):
b.wait()
c = a.ljust(0x600000)
assert not c or c[0] not in (0xdd, 0xcd)
def lower(b, a):
b.wait()
c = a.lower()
assert not c or c[0] not in (0xdd, 0xcd)
def rjust(b, a):
b.wait()
c = a.rjust(0x600000)
assert not c or c[-1] not in (0xdd, 0xcd)
def swapcase(b, a):
b.wait()
c = a.swapcase()
assert not c or c[-1] not in (0xdd, 0xcd)
def title(b, a):
b.wait()
c = a.title()
assert not c or c[-1] not in (0xdd, 0xcd)
def upper(b, a):
b.wait()
c = a.upper()
assert not c or c[-1] not in (0xdd, 0xcd)
def zfill(b, a):
b.wait()
c = a.zfill(0x400000)
assert not c or c[-1] not in (0xdd, 0xcd)
def resize(b, a): # MODIFIES!
b.wait()
a.resize(10)
def take_bytes(b, a): # MODIFIES!
b.wait()
c = a.take_bytes()
assert not c or c[0] == 48 # '0'
def take_bytes_n(b, a): # MODIFIES!
b.wait()
try:
c = a.take_bytes(10)
assert c == b'0123456789'
except IndexError: pass
def check(funcs, a=None, *args):
if a is None:
a = bytearray(b'0' * 0x400000)
barrier = threading.Barrier(len(funcs))
threads = []
for func in funcs:
thread = threading.Thread(target=func, args=(barrier, a, *args))
threads.append(thread)
with threading_helper.start_threads(threads):
pass
# hard errors
check([clear] + [reduce] * 10)
check([clear] + [reduceex2] * 10)
check([clear] + [append1] * 10)
check([clear] * 10)
check([clear] + [count0] * 10)
check([clear] + [decode] * 10)
check([clear] + [extend] * 10)
check([clear] + [find] * 10)
check([clear] + [hex] * 10)
check([clear] + [insert1] * 10)
check([clear] + [join] * 10)
check([clear] + [pop1] * 10)
check([clear] + [remove] * 10)
check([clear] + [replace] * 10)
check([clear] + [reverse] * 10)
check([clear, clear2] + [maketrans] * 10, bytearray(range(128)), bytearray(range(128)))
check([clear] + [translate] * 10, None, bytearray.maketrans(bytearray(range(128)), bytearray(range(128))))
check([clear] + [repeat] * 10)
check([clear] + [iconcat] * 10)
check([clear] + [ass_subscript] * 10)
check([clear] + [repr_] * 10)
# gh-148605: Do not test "a *= 2" since it allocates up to 4 GiB using
# 10 threads
# value errors
check([clear] + [reduceex3] * 10, bytearray(b'a' * 0x40000))
check([clear] + [copy] * 10)
check([clear] + [endswith] * 10)
check([clear] + [index] * 10)
check([clear] + [lstrip] * 10)
check([clear] + [partition] * 10)
check([clear] + [removeprefix] * 10, bytearray(b'0'))
check([clear] + [removesuffix] * 10, bytearray(b'0'))
check([clear] + [rfind] * 10)
check([clear] + [rindex] * 10)
check([clear] + [rpartition] * 10)
check([clear] + [rsplit] * 10, bytearray(b'0' * 0x4000))
check([clear] + [rstrip] * 10)
check([clear] + [split] * 10, bytearray(b'0' * 0x4000))
check([clear] + [splitlines] * 10, bytearray(b'\n' * 0x400))
check([clear] + [startswith] * 10)
check([clear] + [strip] * 10)
check([clear] + [resize] * 10)
check([clear] + [take_bytes] * 10)
check([take_bytes_n] * 10, bytearray(b'0123456789' * 0x400))
check([take_bytes_n] * 10, bytearray(b'0123456789' * 5))
check([clear] + [contains] * 10)
check([clear] + [subscript] * 10)
check([clear2] + [ass_subscript2] * 10, None, bytearray(b'0' * 0x400000))
check([clear] + [mod] * 10, bytearray(b'%d' * 4096))
check([clear2] + [mod2] * 10, bytearray(b'%s'), bytearray(b'0' * 0x400000))
check([clear] + [capitalize] * 10, bytearray(b'a' * 0x40000))
check([clear] + [center] * 10, bytearray(b'a' * 0x40000))
check([clear] + [expandtabs] * 10, bytearray(b'0\t' * 4096))
check([clear] + [ljust] * 10, bytearray(b'0' * 0x400000))
check([clear] + [lower] * 10, bytearray(b'A' * 0x400000))
check([clear] + [rjust] * 10, bytearray(b'0' * 0x400000))
check([clear] + [swapcase] * 10, bytearray(b'aA' * 0x200000))
check([clear] + [title] * 10, bytearray(b'aA' * 0x200000))
check([clear] + [upper] * 10, bytearray(b'a' * 0x400000))
check([clear] + [zfill] * 10, bytearray(b'1' * 0x200000))
@unittest.skipUnless(support.Py_GIL_DISABLED, 'this test can only possibly fail with GIL disabled')
@threading_helper.reap_threads
@threading_helper.requires_working_threading()
def test_free_threading_bytearrayiter(self):
# Non-deterministic but good chance to fail if bytearrayiter is not free-threading safe.
# We are fishing for a "Assertion failed: object has negative ref count" and tsan races.
def iter_next(b, it):
b.wait()
list(it)
def iter_reduce(b, it):
b.wait()
it.__reduce__()
def iter_setstate(b, it):
b.wait()
it.__setstate__(0)
def check(funcs, it):
barrier = threading.Barrier(len(funcs))
threads = []
for func in funcs:
thread = threading.Thread(target=func, args=(barrier, it))
threads.append(thread)
with threading_helper.start_threads(threads):
pass
for _ in range(10):
ba = bytearray(b'0' * 0x4000) # this is a load-bearing variable, do not remove
check([iter_next] * 10, iter(ba))
check([iter_next] + [iter_reduce] * 10, iter(ba)) # for tsan
check([iter_next] + [iter_setstate] * 10, iter(ba)) # for tsan
@unittest.skipUnless(support.Py_GIL_DISABLED, 'this test can only possibly fail with GIL disabled')
@threading_helper.reap_threads
@threading_helper.requires_working_threading()
def test_free_threading_bytearray_resize(self):
def resize_stress(ba):
for _ in range(1000):
try:
ba.resize(1000)
ba.resize(1)
except (BufferError, ValueError):
pass
ba = bytearray(100)
threads = [threading.Thread(target=resize_stress, args=(ba,)) for _ in range(4)]
with threading_helper.start_threads(threads):
pass
@threading_helper.reap_threads
@threading_helper.requires_working_threading()
def test_free_threading_bytearray_resize_other_thread(self):
# Shrinking a bytearray whose buffer another thread owns must not
# adopt the immortal single-byte bytes object a the buffer.
ba = bytearray(b'abc')
thread = threading.Thread(target=ba.resize, args=(1,))
with threading_helper.start_threads([thread]):
pass
ba[0] = ord('X')
self.assertEqual(ba, bytearray(b'X'))
self.assertEqual(ord(b'a'), ord('a'))
if __name__ == "__main__":
unittest.main()