A small Python module for determining appropriate platform-specific dirs, e.g. a "user data dir".

Running into issues with appdirs 1.4.1, which appears to have been released today. Am getting the issue below. More details can be found in this log. Issues exist for other Python versions as well.

FileNotFoundError: [Errno 2] No such file or directory: '/home/travis/virtualenv/python3.6.0/lib/python3.6/site-packages/appdirs-1.4.1.dist-info/METADATA'

cc @alimanfoo

Fixes Issue #16 Bumps version to 1.3.0 Removes gratuitous case mangling on the case-sensitive Linux, since that's not wise Fixes the uterly wrong behaviour in site_data_dir, return result based on XDG_DATA_DIRS and make room for respecting the standard which specifies XDG_DATA_DIRS is a multiple-value variable Add *_config_dir which are distinct on nix-es, according to XDG specs; on Windows and Mac return the corresponding *_data_dir (Fixes Issue #6)

test_api.py only tests the type of the return values. we need tests to cover the behaviour of appdirs functions (i.e., return values) on each platform.

Another attempt at https://github.com/ActiveState/appdirs/pull/118 cc @pradyunsg

Closes https://github.com/ActiveState/appdirs/pull/107 (I ran into something similar) Closes https://github.com/ActiveState/appdirs/issues/108

It's unclear who's maintaining this so cc @zoofood @daved @rawktron @MDrakos @Naatan @autarch @jdufresne

Could we also get a release? I saw https://github.com/ActiveState/appdirs/issues/79, indicating this library may not be getting the love it deserves. That is quite unfortunate since many projects rely on appdirs:

[email protected] ~\Desktop $ pypinfo --days 365 appdirs
Served from cache: False
Data processed: 638.08 GiB
Data billed: 638.08 GiB
Estimated cost: $3.12

| download_count |
| -------------- |
|     39,354,264 |

• Python version: 3.9
• appdirs version: 1.4.4

Linux:

>>> appdirs.user_data_dir('foo')
'/home/user/.local/share/foo'

Windows:

>>> appdirs.user_data_dir('foo')
'C:\\Users\\user\\AppData\\Local\\foo\\foo'

Specifying an empty string as second argument works to avoid this issue.

The e-mail address [email protected] as indicated in maintainers points to a non-existent domain.

$ dig srid.name                                                                                                                                           

; <<>> DiG 9.10.6 <<>> srid.name
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 46493
;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;srid.name.                     IN      A

;; Query time: 71 msec
;; SERVER: 192.168.162.1#53(192.168.162.1)
;; WHEN: Sun Mar 29 09:49:05 EDT 2020
;; MSG SIZE  rcvd: 38

wycliff:appdirs-1.2.0 $ python setup.py nosetests
running nosetests
running egg_info
writing lib/appdirs.egg-info/PKG-INFO
writing top-level names to lib/appdirs.egg-info/top_level.txt
writing dependency_links to lib/appdirs.egg-info/dependency_links.txt
reading manifest file 'lib/appdirs.egg-info/SOURCES.txt'
reading manifest template 'MANIFEST.in'
writing manifest file 'lib/appdirs.egg-info/SOURCES.txt'
running build_ext
E.EEEEEEE
======================================================================
ERROR: test_api (test_appdirs.DocTestsTestCase)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/matej/build/BUILD/appdirs-1.2.0/test/test_appdirs.py", line 18, in test_api
    test = doctest.DocFileTest("api.doctests")
  File "/usr/lib64/python2.7/doctest.py", line 2424, in DocFileTest
    doc, path = _load_testfile(path, package, module_relative)
  File "/usr/lib64/python2.7/doctest.py", line 219, in _load_testfile
    with open(filename) as f:
IOError: [Errno 2] No such file or directory: '/home/matej/build/BUILD/appdirs-1.2.0/test/api.doctests'

======================================================================
ERROR: Failure: TypeError (testmod_paths_from_testdir() takes exactly 1 argument (0 given))
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/loader.py", line 231, in generate
    for test in g():
TypeError: testmod_paths_from_testdir() takes exactly 1 argument (0 given)

======================================================================
ERROR: Failure: TypeError (testmods_from_testdir() takes exactly 1 argument (0 given))
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/loader.py", line 231, in generate
    for test in g():
TypeError: testmods_from_testdir() takes exactly 1 argument (0 given)

======================================================================
ERROR: Failure: TypeError (testcases_from_testmod() takes exactly 1 argument (0 given))
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/loader.py", line 231, in generate
    for test in g():
TypeError: testcases_from_testmod() takes exactly 1 argument (0 given)

======================================================================
ERROR: Failure: TypeError (tests_from_manifest() takes exactly 1 argument (0 given))
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/loader.py", line 231, in generate
    for test in g():
TypeError: tests_from_manifest() takes exactly 1 argument (0 given)

======================================================================
ERROR: Failure: TypeError (tests_from_manifest_and_tags() takes exactly 2 arguments (0 given))
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/loader.py", line 231, in generate
    for test in g():
TypeError: tests_from_manifest_and_tags() takes exactly 2 arguments (0 given)

======================================================================
ERROR: testlib.test
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
TypeError: test() takes at least 1 argument (0 given)

======================================================================
ERROR: testlib.list_tests
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
TypeError: list_tests() takes exactly 2 arguments (0 given)

----------------------------------------------------------------------
Ran 9 tests in 0.076s

FAILED (errors=8)
wycliff:appdirs-1.2.0 $ 

This is a proposed fix for issue #154

Provides an additional fallback for Windows if ctypes, JNA, and winreg all fail.

Tested and works as expected in Windows 10 with Jython 2.7.2. Also tested to verify prior behavior is unchanged against Jython 2.7.2 with JNA and CPython 3.8.5.

This also will fix the downstream pip issue, where pip is currently broken in Jython with a standard Java install.

When setup.py reads appdirs.py without an encoding defined, it can sometimes fail having guessed the wrong encoding. For example:

UnicodeDecodeError: 'ascii' codec can't decode byte 0xc8 in position 129: ordinal not in range(128)

Whereas the builtin open does not accept the encoding argument until 3.0, io.open supports encoding under all versions supported by appdirs.

Starting with wheel 0.32.0 (2018-09-29), the "license_file" option is deprecated.

https://wheel.readthedocs.io/en/stable/news.html

The wheel will continue to include LICENSE, it is now included automatically:

https://wheel.readthedocs.io/en/stable/user_guide.html#including-license-files-in-the-generated-wheel-file

There is a fair amount of Python development on that platform that it is worth adding. Lower priority than other items in the queue, but want to make sure its here.

It looks like the config dir should be ~/Library/Application Support.

A few other projects as examples:

• https://github.com/dirs-dev/dirs-rs/blob/999efde766ed370a60c58145711db728ae9c04e8/src/mac.rs#L8
• https://github.com/adrg/xdg/blob/31e6a328d142d3b8ed43fda69c4f2b55e77d8e79/paths_darwin.go#L28
• https://github.com/dirs-dev/directories-jvm/blob/917e954cf94442399409ea384e94d658b57682b6/src/main/java/dev/dirs/BaseDirectories.java#L268

closes: #185

appdirs.user_config_dir() points to ~/Library/Preferences on macOS: https://github.com/ActiveState/appdirs/blob/193a2cbba58cce2542882fcedd0e49f6763672ed/appdirs.py#L186-L199

This goes against OS guidelines:

src: https://developer.apple.com/library/archive/documentation/FileManagement/Conceptual/FileSystemProgrammingGuide/MacOSXDirectories/MacOSXDirectories.html

Description

There are cases when win32com.shell is available but import of shellcon or shell cause crashes. e.g. when pywin is compiled for different binary. In our case we use appdirs in multiple different pythons where we can't affect it's version.

Changes

Try import all needed parts for pywin getter instead of checking just import win32com.shell .

Description

There are 2 branches but are incompatible and contain different code. PRs are into master but for pypi is used release. Also I would recommend to create main branch name instead of master.

I haven't found a package that provides that functionality and I think that it might be a good use case to handle it here. All major OSes have some form of default user directories like:

$HOME/Desktop
$HOME/Documents
$HOME/Downloads
$HOME/Music
$HOME/Pictures
$HOME/Public
$HOME/Videos

The paths may also differ based on system language preferences, so the i18n needs to be handed as well.

https://wiki.archlinux.org/title/XDG_user_directories https://support.apple.com/guide/mac-help/folders-that-come-with-your-mac-mchlp1143/mac