GH-125413: Revert addition of pathlib.Path.scandir() method (#127377)

Remove documentation for `pathlib.Path.scandir()`, and rename the method to
`_scandir()`. In the private pathlib ABCs, make `iterdir()` abstract and
call it from `_scandir()`.

It's not worthwhile to add this method at the moment - see discussion:
https://discuss.python.org/t/ergonomics-of-new-pathlib-path-scandir/71721

Co-authored-by: Steve Dower <steve.dower@microsoft.com>
This commit is contained in:
Barney Gale 2024-12-05 21:39:43 +00:00 committed by GitHub
parent f4f530804b
commit 8b3cccf3f9
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
7 changed files with 22 additions and 85 deletions

View file

@ -1289,35 +1289,6 @@ Reading directories
raised.
.. method:: Path.scandir()
When the path points to a directory, return an iterator of
:class:`os.DirEntry` objects corresponding to entries in the directory. The
returned iterator supports the :term:`context manager` protocol. It is
implemented using :func:`os.scandir` and gives the same guarantees.
Using :meth:`~Path.scandir` instead of :meth:`~Path.iterdir` can
significantly increase the performance of code that also needs file type or
file attribute information, because :class:`os.DirEntry` objects expose
this information if the operating system provides it when scanning a
directory.
The following example displays the names of subdirectories. The
``entry.is_dir()`` check will generally not make an additional system call::
>>> p = Path('docs')
>>> with p.scandir() as entries:
... for entry in entries:
... if entry.is_dir():
... entry.name
...
'_templates'
'_build'
'_static'
.. versionadded:: 3.14
.. method:: Path.glob(pattern, *, case_sensitive=None, recurse_symlinks=False)
Glob the given relative *pattern* in the directory represented by this path,

View file

@ -532,12 +532,6 @@ pathlib
(Contributed by Barney Gale in :gh:`73991`.)
* Add :meth:`pathlib.Path.scandir` to scan a directory and return an iterator
of :class:`os.DirEntry` objects. This is exactly equivalent to calling
:func:`os.scandir` on a path object.
(Contributed by Barney Gale in :gh:`125413`.)
pdb
---

View file

@ -94,7 +94,7 @@ class PathGlobber(_GlobberBase):
lexists = operator.methodcaller('exists', follow_symlinks=False)
add_slash = operator.methodcaller('joinpath', '')
scandir = operator.methodcaller('scandir')
scandir = operator.methodcaller('_scandir')
@staticmethod
def concat_path(path, text):
@ -632,13 +632,14 @@ class PathBase(PurePathBase):
with self.open(mode='w', encoding=encoding, errors=errors, newline=newline) as f:
return f.write(data)
def scandir(self):
"""Yield os.DirEntry objects of the directory contents.
def _scandir(self):
"""Yield os.DirEntry-like objects of the directory contents.
The children are yielded in arbitrary order, and the
special entries '.' and '..' are not included.
"""
raise UnsupportedOperation(self._unsupported_msg('scandir()'))
import contextlib
return contextlib.nullcontext(self.iterdir())
def iterdir(self):
"""Yield path objects of the directory contents.
@ -646,9 +647,7 @@ class PathBase(PurePathBase):
The children are yielded in arbitrary order, and the
special entries '.' and '..' are not included.
"""
with self.scandir() as entries:
names = [entry.name for entry in entries]
return map(self.joinpath, names)
raise UnsupportedOperation(self._unsupported_msg('iterdir()'))
def _glob_selector(self, parts, case_sensitive, recurse_symlinks):
if case_sensitive is None:
@ -698,7 +697,7 @@ class PathBase(PurePathBase):
if not top_down:
paths.append((path, dirnames, filenames))
try:
with path.scandir() as entries:
with path._scandir() as entries:
for entry in entries:
name = entry.name
try:

View file

@ -634,8 +634,8 @@ class Path(PathBase, PurePath):
path_str = path_str[:-1]
yield path_str
def scandir(self):
"""Yield os.DirEntry objects of the directory contents.
def _scandir(self):
"""Yield os.DirEntry-like objects of the directory contents.
The children are yielded in arbitrary order, and the
special entries '.' and '..' are not included.

View file

@ -1,5 +1,4 @@
import collections
import contextlib
import io
import os
import errno
@ -1418,24 +1417,6 @@ DummyPathStatResult = collections.namedtuple(
'st_mode st_ino st_dev st_nlink st_uid st_gid st_size st_atime st_mtime st_ctime')
class DummyDirEntry:
"""
Minimal os.DirEntry-like object. Returned from DummyPath.scandir().
"""
__slots__ = ('name', '_is_symlink', '_is_dir')
def __init__(self, name, is_symlink, is_dir):
self.name = name
self._is_symlink = is_symlink
self._is_dir = is_dir
def is_symlink(self):
return self._is_symlink
def is_dir(self, *, follow_symlinks=True):
return self._is_dir and (follow_symlinks or not self._is_symlink)
class DummyPath(PathBase):
"""
Simple implementation of PathBase that keeps files and directories in
@ -1503,25 +1484,14 @@ class DummyPath(PathBase):
stream = io.TextIOWrapper(stream, encoding=encoding, errors=errors, newline=newline)
return stream
@contextlib.contextmanager
def scandir(self):
path = self.resolve()
path_str = str(path)
if path_str in self._files:
raise NotADirectoryError(errno.ENOTDIR, "Not a directory", path_str)
elif path_str in self._directories:
yield iter([path.joinpath(name)._dir_entry for name in self._directories[path_str]])
def iterdir(self):
path = str(self.resolve())
if path in self._files:
raise NotADirectoryError(errno.ENOTDIR, "Not a directory", path)
elif path in self._directories:
return iter([self / name for name in self._directories[path]])
else:
raise FileNotFoundError(errno.ENOENT, "File not found", path_str)
@property
def _dir_entry(self):
path_str = str(self)
is_symlink = path_str in self._symlinks
is_directory = (path_str in self._directories
if not is_symlink
else self._symlinks[path_str][1])
return DummyDirEntry(self.name, is_symlink, is_directory)
raise FileNotFoundError(errno.ENOENT, "File not found", path)
def mkdir(self, mode=0o777, parents=False, exist_ok=False):
path = str(self.parent.resolve() / self.name)
@ -2214,9 +2184,9 @@ class DummyPathTest(DummyPurePathTest):
def test_scandir(self):
p = self.cls(self.base)
with p.scandir() as entries:
with p._scandir() as entries:
self.assertTrue(list(entries))
with p.scandir() as entries:
with p._scandir() as entries:
for entry in entries:
child = p / entry.name
self.assertIsNotNone(entry)

View file

@ -597,7 +597,7 @@ TypeError is now raised instead of ValueError for some logical errors.
.. nonce: Jat5kq
.. section: Library
Add :meth:`pathlib.Path.scandir` method to efficiently fetch directory
Add :meth:`!pathlib.Path.scandir` method to efficiently fetch directory
children and their file attributes. This is a trivial wrapper of
:func:`os.scandir`.

View file

@ -0,0 +1,3 @@
Revert addition of :meth:`!pathlib.Path.scandir`. This method was added in
3.14.0a2. The optimizations remain for file system paths, but other
subclasses should only have to implement :meth:`pathlib.Path.iterdir`.