Merge pull request #9 from jaraco/feature/native-tarfile

Provide native tarfile functionality

conftest.py

@@ -0,0 +1,32 @@
+import http.server
+import io
+import functools
+import tarfile
+import threading
+
+import portend
+import pytest
+
+
+@pytest.fixture
+def tarfile_served(tmp_path_factory):
+    """
+    Start an HTTP server serving a tarfile.
+    """
+    tmp_path = tmp_path_factory.mktemp('www')
+    fn = tmp_path / 'served.tgz'
+    tf = tarfile.open(fn, mode='w:gz')
+    info = tarfile.TarInfo('served/contents.txt')
+    tf.addfile(info, io.BytesIO('hello, contents'.encode()))
+    tf.close()
+    httpd, url = start_server(tmp_path)
+    with httpd:
+        yield url + '/served.tgz'
+
+
+def start_server(path):
+    _host, port = addr = ('', portend.find_available_local_port())
+    Handler = functools.partial(http.server.SimpleHTTPRequestHandler, directory=path)
+    httpd = http.server.HTTPServer(addr, Handler)
+    threading.Thread(target=httpd.serve_forever, daemon=True).start()
+    return httpd, f'http://localhost:{port}'

jaraco/context.py

@@ -6,11 +6,19 @@
 import os
 import shutil
 import subprocess
+import sys
 import tempfile
+import urllib.request
 import warnings
 from typing import Iterator
 
 
+if sys.version_info < (3, 12):
+    from backports import tarfile
+else:
+    import tarfile
+
+
 @contextlib.contextmanager
 def pushd(dir: str | os.PathLike) -> Iterator[str | os.PathLike]:
     """
@@ -34,23 +42,38 @@ def tarball(
 ) -> Iterator[str | os.PathLike]:
     """
     Get a tarball, extract it, yield, then clean up.
+
+    >>> import urllib.request
+    >>> url = getfixture('tarfile_served')
+    >>> target = getfixture('tmp_path') / 'out'
+    >>> tb = tarball(url, target_dir=target)
+    >>> import pathlib
+    >>> with tb as extracted:
+    ...     contents = pathlib.Path(extracted, 'contents.txt').read_text()
+    >>> assert not os.path.exists(extracted)
     """
     if target_dir is None:
         target_dir = os.path.basename(url).replace('.tar.gz', '').replace('.tgz', '')
-    runner = functools.partial(subprocess.check_call, shell=True)
     # In the tar command, use --strip-components=1 to strip the first path and
     #  then
     #  use -C to cause the files to be extracted to {target_dir}. This ensures
     #  that we always know where the files were extracted.
-    runner('mkdir {target_dir}'.format(**vars()))
+    os.mkdir(target_dir)
     try:
-        getter = 'wget {url} -O -'
-        extract = 'tar x{compression} --strip-components=1 -C {target_dir}'
-        cmd = ' | '.join((getter, extract))
-        runner(cmd.format(compression=infer_compression(url), **vars()))
+        req = urllib.request.urlopen(url)
+        with tarfile.open(fileobj=req, mode='r|gz') as tf:
+            tf.extractall(path=target_dir, filter=strip_first_component)
         yield target_dir
     finally:
-        runner('rm -Rf {target_dir}'.format(**vars()))
+        shutil.rmtree(target_dir)
+
+
+def strip_first_component(
+    member: tarfile.TarInfo,
+    path,
+) -> tarfile.TarInfo:
+    _, member.name = member.name.split('/', 1)
+    return member
 
 
 def _compose(*cmgrs):

newsfragments/5.feature.rst

@@ -0,0 +1 @@
+Implemented tarfile using native functionality and avoiding subprocessing, making it portable.

setup.cfg

@@ -16,6 +16,7 @@ classifiers =
 include_package_data = true
 python_requires = >=3.8
 install_requires =
+	backports.tarfile; python_version < "3.12"
 
 [options.extras_require]
 testing =
@@ -28,6 +29,7 @@ testing =
 	pytest-ruff >= 0.2.1
 
 	# local
+	portend
 
 docs =
 	# upstream