GH-110109: Churn pathlib.PurePath methods (#112012)

Re-arrange `pathlib.PurePath` methods in source code. No other changes.

The `PurePath` implementations of certain special methods, such as
`__eq__()` and `__hash__()`, are not usually applicable to user subclasses
of `_PathBase`. To facilitate their removal, another patch will split the
`PurePath` class into `_PurePathBase` and `PurePath`, with the latter
providing these special methods.

This patch prepares the ground for splitting `PurePath`. It's similar to
e8d77b0, which preceded splitting `Path`. By churning the methods here,
subsequent patches will be easier to review and less likely to break
things.
This commit is contained in:
Barney Gale 2023-11-17 15:32:50 +00:00 committed by GitHub
parent 7c50800241
commit 2dbb2e035b
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 204 additions and 204 deletions

View file

@ -246,44 +246,6 @@ class PurePath:
)
pathmod = os.path
def __new__(cls, *args, **kwargs):
"""Construct a PurePath from one or several strings and or existing
PurePath objects. The strings and path objects are combined so as
to yield a canonicalized path, which is incorporated into the
new PurePath object.
"""
if cls is PurePath:
cls = PureWindowsPath if os.name == 'nt' else PurePosixPath
return object.__new__(cls)
def __reduce__(self):
# Using the parts tuple helps share interned path parts
# when pickling related paths.
return (self.__class__, self.parts)
def __init__(self, *args):
paths = []
for arg in args:
if isinstance(arg, PurePath):
if arg.pathmod is ntpath and self.pathmod is posixpath:
# GH-103631: Convert separators for backwards compatibility.
paths.extend(path.replace('\\', '/') for path in arg._raw_paths)
else:
paths.extend(arg._raw_paths)
else:
try:
path = os.fspath(arg)
except TypeError:
path = arg
if not isinstance(path, str):
raise TypeError(
"argument should be a str or an os.PathLike "
"object where __fspath__ returns a str, "
f"not {type(path).__name__!r}")
paths.append(path)
self._raw_paths = paths
self._resolving = False
def with_segments(self, *pathsegments):
"""Construct a new path object from any number of path-like objects.
Subclasses may override this method to customize how new path objects
@ -351,96 +313,14 @@ def __str__(self):
self._tail) or '.'
return self._str
def __fspath__(self):
return str(self)
def as_posix(self):
"""Return the string representation of the path with forward (/)
slashes."""
return str(self).replace(self.pathmod.sep, '/')
def __bytes__(self):
"""Return the bytes representation of the path. This is only
recommended to use under Unix."""
return os.fsencode(self)
def __repr__(self):
return "{}({!r})".format(self.__class__.__name__, self.as_posix())
def as_uri(self):
"""Return the path as a URI."""
if not self.is_absolute():
raise ValueError("relative path can't be expressed as a file URI")
drive = self.drive
if len(drive) == 2 and drive[1] == ':':
# It's a path on a local drive => 'file:///c:/a/b'
prefix = 'file:///' + drive
path = self.as_posix()[2:]
elif drive:
# It's a path on a network drive => 'file://host/share/a/b'
prefix = 'file:'
path = self.as_posix()
else:
# It's a posix path => 'file:///etc/hosts'
prefix = 'file://'
path = str(self)
from urllib.parse import quote_from_bytes
return prefix + quote_from_bytes(os.fsencode(path))
@property
def _str_normcase(self):
# String with normalized case, for hashing and equality checks
try:
return self._str_normcase_cached
except AttributeError:
if _is_case_sensitive(self.pathmod):
self._str_normcase_cached = str(self)
else:
self._str_normcase_cached = str(self).lower()
return self._str_normcase_cached
@property
def _parts_normcase(self):
# Cached parts with normalized case, for comparisons.
try:
return self._parts_normcase_cached
except AttributeError:
self._parts_normcase_cached = self._str_normcase.split(self.pathmod.sep)
return self._parts_normcase_cached
def __eq__(self, other):
if not isinstance(other, PurePath):
return NotImplemented
return self._str_normcase == other._str_normcase and self.pathmod is other.pathmod
def __hash__(self):
try:
return self._hash
except AttributeError:
self._hash = hash(self._str_normcase)
return self._hash
def __lt__(self, other):
if not isinstance(other, PurePath) or self.pathmod is not other.pathmod:
return NotImplemented
return self._parts_normcase < other._parts_normcase
def __le__(self, other):
if not isinstance(other, PurePath) or self.pathmod is not other.pathmod:
return NotImplemented
return self._parts_normcase <= other._parts_normcase
def __gt__(self, other):
if not isinstance(other, PurePath) or self.pathmod is not other.pathmod:
return NotImplemented
return self._parts_normcase > other._parts_normcase
def __ge__(self, other):
if not isinstance(other, PurePath) or self.pathmod is not other.pathmod:
return NotImplemented
return self._parts_normcase >= other._parts_normcase
@property
def drive(self):
"""The drive prefix (letter or UNC path), if any."""
@ -694,6 +574,126 @@ def match(self, path_pattern, *, case_sensitive=None):
match = _compile_pattern(pattern_str, sep, case_sensitive)
return match(str(self)) is not None
def __new__(cls, *args, **kwargs):
"""Construct a PurePath from one or several strings and or existing
PurePath objects. The strings and path objects are combined so as
to yield a canonicalized path, which is incorporated into the
new PurePath object.
"""
if cls is PurePath:
cls = PureWindowsPath if os.name == 'nt' else PurePosixPath
return object.__new__(cls)
def __init__(self, *args):
paths = []
for arg in args:
if isinstance(arg, PurePath):
if arg.pathmod is ntpath and self.pathmod is posixpath:
# GH-103631: Convert separators for backwards compatibility.
paths.extend(path.replace('\\', '/') for path in arg._raw_paths)
else:
paths.extend(arg._raw_paths)
else:
try:
path = os.fspath(arg)
except TypeError:
path = arg
if not isinstance(path, str):
raise TypeError(
"argument should be a str or an os.PathLike "
"object where __fspath__ returns a str, "
f"not {type(path).__name__!r}")
paths.append(path)
self._raw_paths = paths
self._resolving = False
def __reduce__(self):
# Using the parts tuple helps share interned path parts
# when pickling related paths.
return (self.__class__, self.parts)
def __fspath__(self):
return str(self)
def __bytes__(self):
"""Return the bytes representation of the path. This is only
recommended to use under Unix."""
return os.fsencode(self)
@property
def _str_normcase(self):
# String with normalized case, for hashing and equality checks
try:
return self._str_normcase_cached
except AttributeError:
if _is_case_sensitive(self.pathmod):
self._str_normcase_cached = str(self)
else:
self._str_normcase_cached = str(self).lower()
return self._str_normcase_cached
def __hash__(self):
try:
return self._hash
except AttributeError:
self._hash = hash(self._str_normcase)
return self._hash
def __eq__(self, other):
if not isinstance(other, PurePath):
return NotImplemented
return self._str_normcase == other._str_normcase and self.pathmod is other.pathmod
@property
def _parts_normcase(self):
# Cached parts with normalized case, for comparisons.
try:
return self._parts_normcase_cached
except AttributeError:
self._parts_normcase_cached = self._str_normcase.split(self.pathmod.sep)
return self._parts_normcase_cached
def __lt__(self, other):
if not isinstance(other, PurePath) or self.pathmod is not other.pathmod:
return NotImplemented
return self._parts_normcase < other._parts_normcase
def __le__(self, other):
if not isinstance(other, PurePath) or self.pathmod is not other.pathmod:
return NotImplemented
return self._parts_normcase <= other._parts_normcase
def __gt__(self, other):
if not isinstance(other, PurePath) or self.pathmod is not other.pathmod:
return NotImplemented
return self._parts_normcase > other._parts_normcase
def __ge__(self, other):
if not isinstance(other, PurePath) or self.pathmod is not other.pathmod:
return NotImplemented
return self._parts_normcase >= other._parts_normcase
def as_uri(self):
"""Return the path as a URI."""
if not self.is_absolute():
raise ValueError("relative path can't be expressed as a file URI")
drive = self.drive
if len(drive) == 2 and drive[1] == ':':
# It's a path on a local drive => 'file:///c:/a/b'
prefix = 'file:///' + drive
path = self.as_posix()[2:]
elif drive:
# It's a path on a network drive => 'file://host/share/a/b'
prefix = 'file:'
path = self.as_posix()
else:
# It's a posix path => 'file:///etc/hosts'
prefix = 'file://'
path = str(self)
from urllib.parse import quote_from_bytes
return prefix + quote_from_bytes(os.fsencode(path))
# Subclassing os.PathLike makes isinstance() checks slower,
# which in turn makes Path construction slower. Register instead!

View file

@ -117,33 +117,6 @@ def test_different_pathmods_unordered(self):
with self.assertRaises(TypeError):
p >= q
def test_bytes(self):
P = self.cls
message = (r"argument should be a str or an os\.PathLike object "
r"where __fspath__ returns a str, not 'bytes'")
with self.assertRaisesRegex(TypeError, message):
P(b'a')
with self.assertRaisesRegex(TypeError, message):
P(b'a', 'b')
with self.assertRaisesRegex(TypeError, message):
P('a', b'b')
with self.assertRaises(TypeError):
P('a').joinpath(b'b')
with self.assertRaises(TypeError):
P('a') / b'b'
with self.assertRaises(TypeError):
b'a' / P('b')
with self.assertRaises(TypeError):
P('a').match(b'b')
with self.assertRaises(TypeError):
P('a').relative_to(b'b')
with self.assertRaises(TypeError):
P('a').with_name(b'b')
with self.assertRaises(TypeError):
P('a').with_stem(b'b')
with self.assertRaises(TypeError):
P('a').with_suffix(b'b')
def _check_str_subclass(self, *args):
# Issue #21127: it should be possible to construct a PurePath object
# from a str subclass instance, and it then gets converted to
@ -273,18 +246,6 @@ def test_as_posix_common(self):
self.assertEqual(P(pathstr).as_posix(), pathstr)
# Other tests for as_posix() are in test_equivalences().
def test_as_bytes_common(self):
sep = os.fsencode(self.sep)
P = self.cls
self.assertEqual(bytes(P('a/b')), b'a' + sep + b'b')
def test_as_uri_common(self):
P = self.cls
with self.assertRaises(ValueError):
P('a').as_uri()
with self.assertRaises(ValueError):
P().as_uri()
def test_repr_common(self):
for pathstr in ('a', 'a/b', 'a/b/c', '/', '/a/b', '/a/b/c'):
with self.subTest(pathstr=pathstr):
@ -297,17 +258,6 @@ def test_repr_common(self):
inner = r[len(clsname) + 1 : -1]
self.assertEqual(eval(inner), p.as_posix())
def test_repr_roundtrips(self):
for pathstr in ('a', 'a/b', 'a/b/c', '/', '/a/b', '/a/b/c'):
with self.subTest(pathstr=pathstr):
p = self.cls(pathstr)
r = repr(p)
# The repr() roundtrips.
q = eval(r, pathlib.__dict__)
self.assertIs(q.__class__, p.__class__)
self.assertEqual(q, p)
self.assertEqual(repr(q), r)
def test_eq_common(self):
P = self.cls
self.assertEqual(P('a/b'), P('a/b'))
@ -390,34 +340,6 @@ def test_match_common(self):
self.assertTrue(P().match('**'))
self.assertFalse(P().match('**/*'))
def test_ordering_common(self):
# Ordering is tuple-alike.
def assertLess(a, b):
self.assertLess(a, b)
self.assertGreater(b, a)
P = self.cls
a = P('a')
b = P('a/b')
c = P('abc')
d = P('b')
assertLess(a, b)
assertLess(a, c)
assertLess(a, d)
assertLess(b, c)
assertLess(c, d)
P = self.cls
a = P('/a')
b = P('/a/b')
c = P('/abc')
d = P('/b')
assertLess(a, b)
assertLess(a, c)
assertLess(a, d)
assertLess(b, c)
assertLess(c, d)
with self.assertRaises(TypeError):
P() < {}
def test_parts_common(self):
# `parts` returns a tuple.
sep = self.sep
@ -430,12 +352,6 @@ def test_parts_common(self):
parts = p.parts
self.assertEqual(parts, (sep, 'a', 'b'))
def test_fspath_common(self):
P = self.cls
p = P('a/b')
self._check_str(p.__fspath__(), ('a/b',))
self._check_str(os.fspath(p), ('a/b',))
def test_equivalences(self):
for k, tuples in self.equivalences.items():
canon = k.replace('/', self.sep)
@ -787,6 +703,90 @@ def test_pickling_common(self):
self.assertEqual(hash(pp), hash(p))
self.assertEqual(str(pp), str(p))
def test_fspath_common(self):
P = self.cls
p = P('a/b')
self._check_str(p.__fspath__(), ('a/b',))
self._check_str(os.fspath(p), ('a/b',))
def test_bytes(self):
P = self.cls
message = (r"argument should be a str or an os\.PathLike object "
r"where __fspath__ returns a str, not 'bytes'")
with self.assertRaisesRegex(TypeError, message):
P(b'a')
with self.assertRaisesRegex(TypeError, message):
P(b'a', 'b')
with self.assertRaisesRegex(TypeError, message):
P('a', b'b')
with self.assertRaises(TypeError):
P('a').joinpath(b'b')
with self.assertRaises(TypeError):
P('a') / b'b'
with self.assertRaises(TypeError):
b'a' / P('b')
with self.assertRaises(TypeError):
P('a').match(b'b')
with self.assertRaises(TypeError):
P('a').relative_to(b'b')
with self.assertRaises(TypeError):
P('a').with_name(b'b')
with self.assertRaises(TypeError):
P('a').with_stem(b'b')
with self.assertRaises(TypeError):
P('a').with_suffix(b'b')
def test_as_bytes_common(self):
sep = os.fsencode(self.sep)
P = self.cls
self.assertEqual(bytes(P('a/b')), b'a' + sep + b'b')
def test_ordering_common(self):
# Ordering is tuple-alike.
def assertLess(a, b):
self.assertLess(a, b)
self.assertGreater(b, a)
P = self.cls
a = P('a')
b = P('a/b')
c = P('abc')
d = P('b')
assertLess(a, b)
assertLess(a, c)
assertLess(a, d)
assertLess(b, c)
assertLess(c, d)
P = self.cls
a = P('/a')
b = P('/a/b')
c = P('/abc')
d = P('/b')
assertLess(a, b)
assertLess(a, c)
assertLess(a, d)
assertLess(b, c)
assertLess(c, d)
with self.assertRaises(TypeError):
P() < {}
def test_as_uri_common(self):
P = self.cls
with self.assertRaises(ValueError):
P('a').as_uri()
with self.assertRaises(ValueError):
P().as_uri()
def test_repr_roundtrips(self):
for pathstr in ('a', 'a/b', 'a/b/c', '/', '/a/b', '/a/b/c'):
with self.subTest(pathstr=pathstr):
p = self.cls(pathstr)
r = repr(p)
# The repr() roundtrips.
q = eval(r, pathlib.__dict__)
self.assertIs(q.__class__, p.__class__)
self.assertEqual(q, p)
self.assertEqual(repr(q), r)
class PurePosixPathTest(PurePathTest):
cls = pathlib.PurePosixPath