Add support for Path.absolute()

[DimitrisJim]

BPO	29688
Nosy	@brettcannon, @pitrou, @methane, @serhiy-storchaka, @eryksun, @DimitrisJim, @barneygale
PRs	bpo-29688: Document Path.absolute #384 bpo-29688: document and test pathlib.Path.absolute(). #26153

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = None
created_at = <Date 2017-03-01.17:37:28.703>
labels = ['3.8', '3.9', '3.10', 'type-feature', 'library', 'docs']
title = 'Add support for Path.absolute()'
updated_at = <Date 2022-01-31.23:20:11.549>
user = 'https://github.com/DimitrisJim'

bugs.python.org fields:

activity = <Date 2022-01-31.23:20:11.549>
actor = 'eryksun'
assignee = 'docs@python'
closed = False
closed_date = None
closer = None
components = ['Documentation', 'Library (Lib)']
creation = <Date 2017-03-01.17:37:28.703>
creator = 'Jim Fasarakis-Hilliard'
dependencies = []
files = []
hgrepos = []
issue_num = 29688
keywords = ['patch']
message_count = 20.0
messages = ['288767', '288923', '288931', '289413', '289436', '289452', '289460', '289507', '289509', '289526', '358642', '393731', '395689', '412044', '412099', '412102', '412219', '412221', '412222', '412223']
nosy_count = 9.0
nosy_names = ['brett.cannon', 'pitrou', 'methane', 'docs@python', 'serhiy.storchaka', 'eryksun', 'Jim Fasarakis-Hilliard', '4-launchpad-kalvdans-no-ip-org', 'barneygale']
pr_nums = ['384', '26153']
priority = 'normal'
resolution = None
stage = 'patch review'
status = 'open'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue29688'
versions = ['Python 3.8', 'Python 3.9', 'Python 3.10']

[DimitrisJim]

Method absolute of Path objects lacked documentation, proposed PR adds relevant method to docs.

[brettcannon]

As brought up on the PR, it turns out Path.absolute() is extremely under-tested.

Perhaps we should deprecate Path.absolute() instead of document it and properly test it (and the testing will be necessary to move forward with the documentation)? Path.resolve() handles absolute paths already while also resolving '.' and '..': https://docs.python.org/3/library/pathlib.html#pathlib.Path.resolve. It also works with non-existent paths so unless there's some performance issue I'm not aware of for resolving '.' and '..', then I say we deprecate Path.absolute().

[brettcannon]

I've closed the PR on GitHub until we decide whether we just want to deprecate Path.absolute() in favour of Path.resolve().

[brettcannon]

I'm still thinking about this but I'm still leaning towards deprecating pathlib.absolute().

[eryksun]

resolve() can't replace absolute(). They serve different purposes. Sometimes one wants an absolute path, but without resolving symbolic links.

absolute() processes a path as a string, which will continue to be true if it's updated to call nt._getfullpathname (GetFullPathName) on Windows. OTOH, resolve() can outright fail on Windows. I can write up a list of examples (I can think of 5 or 6 unhandled error codes), but it's not directly relevant to this issue.

[brettcannon]

I know it has it's uses (avoiding stat calls is one of them), but it is still undocumented, untested, and has two comments in it saying it needs work. Because of all that it might as well not exist since it doesn't meet our standards of quality.

If someone wants to fix all those issues then we can properly document it as supported, but if no one is willing to then I don't think we should leave unsupported code lying around that people might discover through dir().

And it doesn't serve a _different_ purpose compared to resolve(), it serves a _subset_ of resolve()'s purpose since resolve() calls absolute() unconditionally.

[eryksun]

What's the rationale for not calling self._flavour.pathmod.abspath() to implement absolute()? For example:

    >>> p = pathlib.Path('C:/con')
    >>> p._flavour.pathmod.abspath(p)
    '\\\\.\\con'
    >>> p._from_parts((p._flavour.pathmod.abspath(p),), init=False)
    WindowsPath('//./con/')

That's almost right except for an unrelated problem that pathlib shouldn't append a trailing slash for \\.\ local device paths. Doing so creates a different path, which may be invalid. \\.\con is a symbolic link to \Device\ConDrv\Console, and adding a trailing backslash after the "Console" filename is invalid. An example where the resulting path is valid but wrong is the volume device \\.\C:, which is a link to something like \Device\HarddiskVolume2. Appending a backslash refers to the root directory of the file system on the volume.

[brettcannon]

"What's the rationale for not calling self._flavour.pathmod.abspath() to implement absolute()?" Beats me. :) It's just how Antoine wrote it and that's all I know.

[serhiy-storchaka]

posixpath.abspath() collapses "<path>/<symlink>/.." to "<path>", this is not correct. Not sure about ntpath.abspath().

[methane]

posixpath.abspath() collapses "<path>/<symlink>/.." to "<path>", this is not correct. Not sure about ntpath.abspath().

I feel this is reasonable limitation, and expected behavior for some cases.
So I think adding note about this in document is enough.

[brettcannon]

I have opened https://bugs.python.org/issue39090 to track updating the pathlib docs to have a section on getting the absolute path in various ways along with what the trade-offs are for each approach.

[barneygale]

New PR up here: #26153

The correspondence between pathlib and os.path is as follows:

• Path.resolve() is roughly os.path.realpath()
• Path.absolute() is roughly os.path.abspath()

Differences:

• resolve() always raises RuntimeError on symlink loops, whereas realpath() either raises OSError or nothing depending on strict.
• absolute() doesn't normalize the path, whereas abspath() does. Normalizing without resolving symlinks can change the meaning of the path, so pathlib does the better thing here.

[barneygale]

Hi - please could a core dev review PR 26153? It adds documentation and tests for Path.absolute(). Thank you!

[brettcannon]

New changeset 18cb2ef by Barney Gale in branch 'main':
bpo-29688: document and test pathlib.Path.absolute() (GH-26153)
18cb2ef

[barneygale]

Now that GH 26153 is merged, I think this bug can be resolved.

[eryksun]

In Windows, paths that are relative to the current directory on a drive aren't resolved. The following should be resolved by the current code:

    >>> os.chdir('C:/Temp')
    >>> pathlib.Path('C:').absolute()
    WindowsPath('C:')

But _from_parts() has bugs with drive-relative paths.

Assuming the bugs are fixed, when a path has a drive, Path.absolute() should resolve against abspath(self.drive) instead of getcwd().

[brettcannon]

@eryksun I'm not seeing what's wrong with your example. Would you mind pointing out what you expect the result to be?

And are you saying on Windows you have to resolve the drive separately from the working directory and then concatenate them?

[eryksun]

I'm not seeing what's wrong with your example.

"C:" or "C:spam\\eggs" are not absolute paths. They depend on the effective working directory on the drive. An absolute path should never depend on a working directory, which can change at random.

WinAPI SetEnvironmentVariableW() allows applications to set environment variables with names that begin with "=". These names are effectively reserved for special use by the OS, at least as documented. In particular, names of the form "=X:", where "X" is a drive letter, are used to store the working directory on a drive. The C runtime _[w]chdir() function sets these per-drive environment variables, as does Python's os.chdir(). As environment variables, they can be inherited by child processes.

When then Windows API resolves a file path to access a file, or in GetFullPathNameW(), a drive-relative path such as "X:" or "X:spam\\eggs" is resolved against either the current working directory (if it's on the drive) or the value of the "=X:" environment variable for the drive. If the latter isn't defined, it defaults to the root directory, e.g. "X:\\". If the current working directory is on the drive, the system updates the value of the "=X:" environment variable, if it exists.

on Windows you have to resolve the drive separately from the
working directory and then concatenate them?

No, if self.drive is defined, then abspath(self.drive) should be called instead of getcwd(). In Windows, ntpath.abspath() calls WinAPI GetFullPathNameW(), which resolves the working directory on the drive.

[barneygale]

@eryksun thanks for flagging, a couple thoughts:

I'd imagine that bug is reproducible with Path('C:\\Temp', 'C:') already, right? If that's the case, should it logged as a separate issue?

I'm planning to /experimentally/ throw away pathlib's internal
path parsing logic and defer to posixpath / ntpath instead. I suspect this bug and others will be fixed by that change, but I need to see what the performance impact will be.

[eryksun]

I'd imagine that bug is reproducible with Path('C:\\Temp', 'C:')
already, right? If that's the case, should it logged as a
separate issue?

Yes, it's a separate issue that affects the _from_parts() call in absolute().

How about designing absolute() to create a new instance from an absolute path that's created by os.path? For example:

join(abspath(self.drive) if self.drive else getcwd(), self)

Of course use accessor functions.

[barneygale]

This feature request was implemented in #26153. The bug Eryk mentioned is logged as #94909 (fixed) and #100809 (PR available). Resolving this issue!