gh-116622: Don't expose FICLONE ioctl on Android

[mhsmith]

Although this ioctl exists in the system headers, it's blocked by SELinux with a message like this:

type=1400 audit(0.0:18729): avc: denied{ioctl } for path=2F646174612F646174612F6F72672E707974686F6E2E746573746265642F63616368652F746573745F707974686F6E5F776F726B65725F36373532C3A62F40746573745F363735325F746D70C3A62F636F707941 dev="dm-39" ino=369303 ioctlcmd=0x9409 scontext=u:r:untrusted_app:s0:c225,c256,c512,c768 tcontext=u:object_r:app_data_file:s0:c225,c256,c512,c768 tclass=file permissive=0 app=org.python.testbed

On Python 3.14 this breaks the test for Path.copy, which was added in #119058:

====================================================================== ERROR: test_copytree_to_existing_directory_dirs_exist_ok (test.test_pathlib.test_pathlib.PosixPathTest.test_copytree_to_existing_directory_dirs_exist_ok) ---------------------------------------------------------------------- Traceback (most recent call last): File "/data/user/0/org.python.testbed/files/python/lib/python3.14/test/test_pathlib/test_pathlib_abc.py", line 1899, in test_copytree_to_existing_directory_dirs_exist_ok source.copytree(target, dirs_exist_ok=True) ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/data/user/0/org.python.testbed/files/python/lib/python3.14/pathlib/_abc.py", line 870, in copytree on_error(err) ~~~~~~~~^^^^^ File "/data/user/0/org.python.testbed/files/python/lib/python3.14/pathlib/_abc.py", line 848, in on_error raise err File "/data/user/0/org.python.testbed/files/python/lib/python3.14/pathlib/_abc.py", line 868, in copytree on_error(err) ~~~~~~~~^^^^^ File "/data/user/0/org.python.testbed/files/python/lib/python3.14/pathlib/_abc.py", line 848, in on_error raise err File "/data/user/0/org.python.testbed/files/python/lib/python3.14/pathlib/_abc.py", line 864, in copytree source.copy(target_dir.joinpath(source.name), ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ follow_symlinks=follow_symlinks, ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ preserve_metadata=preserve_metadata) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/data/user/0/org.python.testbed/files/python/lib/python3.14/pathlib/_abc.py", line 827, in copy copyfileobj(source_f, target_f) ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^ File "/data/user/0/org.python.testbed/files/python/lib/python3.14/pathlib/_os.py", line 174, in copyfileobj raise err File "/data/user/0/org.python.testbed/files/python/lib/python3.14/pathlib/_os.py", line 164, in copyfileobj raise err File "/data/user/0/org.python.testbed/files/python/lib/python3.14/pathlib/_os.py", line 160, in copyfileobj clonefd(source_fd, target_fd) ~~~~~~~^^^^^^^^^^^^^^^^^^^^^^ File "/data/user/0/org.python.testbed/files/python/lib/python3.14/pathlib/_os.py", line 58, in clonefd fcntl.ioctl(target_fd, fcntl.FICLONE, source_fd) ~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ PermissionError: [Errno 13] Permission denied: '/data/data/org.python.testbed/cache/test_python_worker_7506æ/@test_7506_tmpæ/dirC/fileC' -> '/data/data/org.python.testbed/cache/test_python_worker_7506æ/@test_7506_tmpæ/copyC/fileC' 

On Python 3.13 I don't think Python ever uses this ioctl itself, but it's still worth backporting this PR for the benefit of user code.

• Issue: Add support for Android as a target platform #116622

[freakboy3742]

A question and a comment:

Q: Where is the FICLONE and FICLONERANGE constant coming from? Could this be managed as an autoconf check, rather than an explicit Android #ifdef?

Comment: This needs docs noting the feature exclusion on Android.

[bedevere-app]

A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

[mhsmith]

Where is the FICLONE and FICLONERANGE constant coming from? Could this be managed as an autoconf check, rather than an explicit Android #ifdef?

They're coming from the operating system headers. fcntlmodule.c has #ifdefs for each one, so they don't need autoconf checks.

This needs docs noting the feature exclusion on Android.

I think this is already covered by the statement in the fcntl docs that "The values used for cmd are operating system dependent". There is no list of the available values in the Python documentation.

[freakboy3742]

Ok - those both make sense.

[miss-islington-app]

Thanks @mhsmith for the PR, and @freakboy3742 for merging it 🌮🎉.. I'm working now to backport this PR to: 3.13.
🐍🍒⛏🤖

[bedevere-app]

GH-122539 is a backport of this pull request to the 3.13 branch.