diff --git a/CONTRIBUTORS b/CONTRIBUTORS index 52daeb8..8b49d5d 100644 --- a/CONTRIBUTORS +++ b/CONTRIBUTORS @@ -1 +1,2 @@ Thibault Allançon +Franck Bret diff --git a/PKG-INFO b/PKG-INFO index 7c119de..51b7b02 100644 --- a/PKG-INFO +++ b/PKG-INFO @@ -1,39 +1,39 @@ Metadata-Version: 2.1 Name: swh.core -Version: 2.11 +Version: 2.12 Summary: Software Heritage core utilities Home-page: https://forge.softwareheritage.org/diffusion/DCORE/ Author: Software Heritage developers Author-email: swh-devel@inria.fr Project-URL: Bug Reports, https://forge.softwareheritage.org/maniphest Project-URL: Funding, https://www.softwareheritage.org/donate Project-URL: Source, https://forge.softwareheritage.org/source/swh-core Project-URL: Documentation, https://docs.softwareheritage.org/devel/swh-core/ Classifier: Programming Language :: Python :: 3 Classifier: Intended Audience :: Developers Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3) Classifier: Operating System :: OS Independent Classifier: Development Status :: 5 - Production/Stable Requires-Python: >=3.7 Description-Content-Type: text/x-rst Provides-Extra: testing-core Provides-Extra: logging Provides-Extra: db Provides-Extra: http Provides-Extra: github Provides-Extra: testing License-File: LICENSE License-File: AUTHORS Software Heritage - Core foundations ==================================== Low-level utilities and helpers used by almost all other modules in the stack. core library for swh's modules: - config parser - serialization - logging mechanism - database connection - http-based RPC client/server diff --git a/swh.core.egg-info/PKG-INFO b/swh.core.egg-info/PKG-INFO index 7c119de..51b7b02 100644 --- a/swh.core.egg-info/PKG-INFO +++ b/swh.core.egg-info/PKG-INFO @@ -1,39 +1,39 @@ Metadata-Version: 2.1 Name: swh.core -Version: 2.11 +Version: 2.12 Summary: Software Heritage core utilities Home-page: https://forge.softwareheritage.org/diffusion/DCORE/ Author: Software Heritage developers Author-email: swh-devel@inria.fr Project-URL: Bug Reports, https://forge.softwareheritage.org/maniphest Project-URL: Funding, https://www.softwareheritage.org/donate Project-URL: Source, https://forge.softwareheritage.org/source/swh-core Project-URL: Documentation, https://docs.softwareheritage.org/devel/swh-core/ Classifier: Programming Language :: Python :: 3 Classifier: Intended Audience :: Developers Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3) Classifier: Operating System :: OS Independent Classifier: Development Status :: 5 - Production/Stable Requires-Python: >=3.7 Description-Content-Type: text/x-rst Provides-Extra: testing-core Provides-Extra: logging Provides-Extra: db Provides-Extra: http Provides-Extra: github Provides-Extra: testing License-File: LICENSE License-File: AUTHORS Software Heritage - Core foundations ==================================== Low-level utilities and helpers used by almost all other modules in the stack. core library for swh's modules: - config parser - serialization - logging mechanism - database connection - http-based RPC client/server diff --git a/swh.core.egg-info/SOURCES.txt b/swh.core.egg-info/SOURCES.txt index ac416f8..a2de7a2 100644 --- a/swh.core.egg-info/SOURCES.txt +++ b/swh.core.egg-info/SOURCES.txt @@ -1,142 +1,143 @@ .git-blame-ignore-revs .gitignore .pre-commit-config.yaml AUTHORS CODE_OF_CONDUCT.md CONTRIBUTORS LICENSE MANIFEST.in Makefile Makefile.local README.rst conftest.py mypy.ini pyproject.toml pytest.ini requirements-db-pytestplugin.txt requirements-db.txt requirements-github.txt requirements-http.txt requirements-logging.txt requirements-swh.txt requirements-test.txt requirements.txt setup.cfg setup.py tox.ini docs/.gitignore docs/Makefile docs/README.rst docs/cli.rst docs/conf.py docs/db.rst docs/index.rst docs/_static/.placeholder docs/_templates/.placeholder swh/__init__.py swh/__main__.py swh.core.egg-info/PKG-INFO swh.core.egg-info/SOURCES.txt swh.core.egg-info/dependency_links.txt swh.core.egg-info/entry_points.txt swh.core.egg-info/requires.txt swh.core.egg-info/top_level.txt swh/core/__init__.py swh/core/api_async.py swh/core/collections.py swh/core/config.py swh/core/logger.py swh/core/py.typed swh/core/pytest_plugin.py swh/core/sentry.py swh/core/statsd.py swh/core/tarball.py swh/core/utils.py swh/core/api/__init__.py swh/core/api/asynchronous.py swh/core/api/classes.py swh/core/api/gunicorn_config.py swh/core/api/negotiation.py swh/core/api/serializers.py swh/core/api/tests/__init__.py swh/core/api/tests/conftest.py swh/core/api/tests/server_testing.py swh/core/api/tests/test_async.py swh/core/api/tests/test_classes.py swh/core/api/tests/test_gunicorn.py swh/core/api/tests/test_init.py swh/core/api/tests/test_rpc_client.py swh/core/api/tests/test_rpc_client_server.py swh/core/api/tests/test_rpc_server.py swh/core/api/tests/test_rpc_server_asynchronous.py swh/core/api/tests/test_serializers.py swh/core/cli/__init__.py swh/core/cli/db.py swh/core/db/__init__.py swh/core/db/common.py swh/core/db/db_utils.py swh/core/db/pytest_plugin.py swh/core/db/sql/35-dbversion.sql swh/core/db/sql/36-dbmodule.sql swh/core/db/tests/__init__.py swh/core/db/tests/conftest.py swh/core/db/tests/test_cli.py swh/core/db/tests/test_db.py swh/core/db/tests/test_db_utils.py swh/core/db/tests/data/cli/sql/0-superuser-init.sql swh/core/db/tests/data/cli/sql/30-schema.sql swh/core/db/tests/data/cli/sql/40-funcs.sql swh/core/db/tests/data/cli/sql/50-data.sql swh/core/db/tests/data/cli_new/sql/0-superuser-init.sql swh/core/db/tests/data/cli_new/sql/30-schema.sql swh/core/db/tests/data/cli_new/sql/40-funcs.sql swh/core/db/tests/data/cli_new/sql/50-data.sql swh/core/db/tests/data/cli_new/sql/upgrades/001.sql swh/core/db/tests/data/cli_new/sql/upgrades/002.sql swh/core/db/tests/data/cli_new/sql/upgrades/003.sql swh/core/db/tests/data/cli_new/sql/upgrades/004.sql swh/core/db/tests/data/cli_new/sql/upgrades/005.sql swh/core/db/tests/data/cli_new/sql/upgrades/006.sql swh/core/db/tests/pytest_plugin/__init__.py swh/core/db/tests/pytest_plugin/test_pytest_plugin.py swh/core/db/tests/pytest_plugin/data/0-schema.sql swh/core/db/tests/pytest_plugin/data/1-data.sql swh/core/github/__init__.py swh/core/github/pytest_plugin.py swh/core/github/utils.py swh/core/github/tests/__init__.py swh/core/github/tests/test_github_utils.py swh/core/github/tests/test_pytest_plugin.py swh/core/tests/__init__.py swh/core/tests/test_cli.py swh/core/tests/test_collections.py swh/core/tests/test_config.py swh/core/tests/test_logger.py swh/core/tests/test_pytest_plugin.py swh/core/tests/test_sentry.py swh/core/tests/test_statsd.py swh/core/tests/test_tarball.py swh/core/tests/test_utils.py +swh/core/tests/data/archives/ca-certificates-20210603-1-any.pkg.tar.zst swh/core/tests/data/archives/groff-1.02.tar.Z swh/core/tests/data/archives/hello.jar swh/core/tests/data/archives/hello.tar swh/core/tests/data/archives/hello.tar.bz2 swh/core/tests/data/archives/hello.tar.gz swh/core/tests/data/archives/hello.tar.lz swh/core/tests/data/archives/hello.tar.x swh/core/tests/data/archives/hello.tbz swh/core/tests/data/archives/hello.tbz2 swh/core/tests/data/archives/hello.zip swh/core/tests/data/archives/msk316src.zip swh/core/tests/data/archives/tokei-12.1.2.crate swh/core/tests/data/http_example.com/something.json swh/core/tests/data/https_example.com/file.json swh/core/tests/data/https_example.com/file.json,name=doe,firstname=jane swh/core/tests/data/https_example.com/file.json_visit1 swh/core/tests/data/https_example.com/other.json swh/core/tests/data/https_forge.s.o/api_diffusion,attachments[uris]=1 swh/core/tests/data/https_www.reference.com/web,q=What+Is+an+Example+of+a+URL?,qo=contentPageRelatedSearch,o=600605,l=dir,sga=1 swh/core/tests/fixture/__init__.py swh/core/tests/fixture/conftest.py swh/core/tests/fixture/test_pytest_plugin.py swh/core/tests/fixture/data/https_example.com/file.json \ No newline at end of file diff --git a/swh/core/tarball.py b/swh/core/tarball.py index e2d06b9..11dba55 100644 --- a/swh/core/tarball.py +++ b/swh/core/tarball.py @@ -1,226 +1,250 @@ # Copyright (C) 2015-2021 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import os import shutil import stat from subprocess import run import tarfile import zipfile import magic from . import utils def _unpack_tar(tarpath: str, extract_dir: str) -> str: """Unpack tarballs unsupported by the standard python library. Examples include tar.Z, tar.lz, tar.x, etc.... As this implementation relies on the `tar` command, this function supports the same compression the tar command supports. This expects the `extract_dir` to exist. Raises: shutil.ReadError in case of issue uncompressing the archive (tarpath does not exist, extract_dir does not exist, etc...) Returns: full path to the uncompressed directory. """ try: run(["tar", "xf", tarpath, "-C", extract_dir], check=True) return extract_dir except Exception as e: raise shutil.ReadError( f"Unable to uncompress {tarpath} to {extract_dir}. Reason: {e}" ) def _unpack_zip(zippath: str, extract_dir: str) -> str: """Unpack zip files unsupported by the standard python library, for instance those with legacy compression type 6 (implode). This expects the `extract_dir` to exist. Raises: shutil.ReadError in case of issue uncompressing the archive (zippath does not exist, extract_dir does not exist, etc...) Returns: full path to the uncompressed directory. """ try: run(["unzip", "-q", "-d", extract_dir, zippath], check=True) return extract_dir except Exception as e: raise shutil.ReadError( f"Unable to uncompress {zippath} to {extract_dir}. Reason: {e}" ) def _unpack_jar(jarpath: str, extract_dir: str) -> str: """Unpack jar files using standard Python module zipfile. This expects the `extract_dir` to exist. Raises: shutil.ReadError in case of issue uncompressing the archive (jarpath does not exist, extract_dir does not exist, etc...) Returns: full path to the uncompressed directory. """ try: with zipfile.ZipFile(jarpath) as jar: jar.extractall(path=extract_dir) return extract_dir except Exception as e: raise shutil.ReadError( f"Unable to uncompress {jarpath} to {extract_dir}. Reason: {e}" ) +def _unpack_zst(zstpath: str, extract_dir: str) -> str: + """Unpack zst files unsupported by the standard python library. Example + include tar.zst + + This expects the `extract_dir` to exist. + + Raises: + shutil.ReadError in case of issue uncompressing the archive (zstpath + """ + try: + run( + ["tar", "--force-local", "-I 'zstd'", "-xf", zstpath, "-C", extract_dir], + check=True, + ) + return extract_dir + except Exception as e: + raise shutil.ReadError( + f"Unable to uncompress {zstpath} to {extract_dir}. Reason: {e}" + ) + + def register_new_archive_formats(): """Register new archive formats to uncompress""" registered_formats = [f[0] for f in shutil.get_unpack_formats()] for name, extensions, function in ADDITIONAL_ARCHIVE_FORMATS: if name in registered_formats: continue shutil.register_unpack_format(name, extensions, function) _mime_to_archive_format = { "application/x-compress": "tar.Z|x", "application/x-tar": "tar", "application/x-bzip2": "bztar", "application/gzip": "gztar", "application/x-lzip": "tar.lz", "application/zip": "zip", "application/java-archive": "jar", + "application/zstd": "tar.zst", + "application/x-zstd": "tar.zst", } def uncompress(tarpath: str, dest: str): """Uncompress tarpath to dest folder if tarball is supported. Note that this fixes permissions after successfully uncompressing the archive. Args: tarpath: path to tarball to uncompress dest: the destination folder where to uncompress the tarball, it will be created if it does not exist Raises: ValueError when a problem occurs during unpacking """ try: os.makedirs(dest, exist_ok=True) format = None # try to get archive format from extension for format_, exts, _ in shutil.get_unpack_formats(): if any([tarpath.lower().endswith(ext.lower()) for ext in exts]): format = format_ break # try to get archive format from file mimetype if format is None: m = magic.Magic(mime=True) mime = m.from_file(tarpath) format = _mime_to_archive_format.get(mime) shutil.unpack_archive(tarpath, extract_dir=dest, format=format) except shutil.ReadError as e: raise ValueError(f"Problem during unpacking {tarpath}. Reason: {e}") except NotImplementedError: if tarpath.lower().endswith(".zip") or format == "zip": _unpack_zip(tarpath, dest) else: raise normalize_permissions(dest) def normalize_permissions(path: str): """Normalize the permissions of all files and directories under `path`. This makes all subdirectories and files with the user executable bit set mode 0o0755, and all other files mode 0o0644. Args: path: the path under which permissions should be normalized """ for dirpath, _, fnames in os.walk(path): os.chmod(dirpath, 0o0755) for fname in fnames: fpath = os.path.join(dirpath, fname) if not os.path.islink(fpath): is_executable = os.stat(fpath).st_mode & stat.S_IXUSR forced_mode = 0o0755 if is_executable else 0o0644 os.chmod(fpath, forced_mode) def _ls(rootdir): """Generator of filepath, filename from rootdir.""" for dirpath, dirnames, fnames in os.walk(rootdir): for fname in dirnames + fnames: fpath = os.path.join(dirpath, fname) fname = utils.commonname(rootdir, fpath) yield fpath, fname def _compress_zip(tarpath, files): """Compress dirpath's content as tarpath.""" with zipfile.ZipFile(tarpath, "w") as z: for fpath, fname in files: z.write(fpath, arcname=fname) def _compress_tar(tarpath, files): """Compress dirpath's content as tarpath.""" with tarfile.open(tarpath, "w:bz2") as t: for fpath, fname in files: t.add(fpath, arcname=fname, recursive=False) def compress(tarpath, nature, dirpath_or_files): """Create a tarball tarpath with nature nature. The content of the tarball is either dirpath's content (if representing a directory path) or dirpath's iterable contents. Compress the directory dirpath's content to a tarball. The tarball being dumped at tarpath. The nature of the tarball is determined by the nature argument. """ if isinstance(dirpath_or_files, str): files = _ls(dirpath_or_files) else: # iterable of 'filepath, filename' files = dirpath_or_files if nature == "zip": _compress_zip(tarpath, files) else: _compress_tar(tarpath, files) return tarpath # Additional uncompression archive format support ADDITIONAL_ARCHIVE_FORMATS = [ # name, extensions, function ("tar.Z|x", [".tar.Z", ".tar.x"], _unpack_tar), ("jar", [".jar"], _unpack_jar), ("tbz2", [".tbz", "tbz2"], _unpack_tar), # FIXME: make this optional depending on the runtime lzip package install ("tar.lz", [".tar.lz"], _unpack_tar), ("crate", [".crate"], _unpack_tar), + ("tar.zst", [".tar.zst", ".tar.zstd"], _unpack_zst), ] register_new_archive_formats() diff --git a/swh/core/tests/data/archives/ca-certificates-20210603-1-any.pkg.tar.zst b/swh/core/tests/data/archives/ca-certificates-20210603-1-any.pkg.tar.zst new file mode 100644 index 0000000..f7373b7 Binary files /dev/null and b/swh/core/tests/data/archives/ca-certificates-20210603-1-any.pkg.tar.zst differ diff --git a/swh/core/tests/test_tarball.py b/swh/core/tests/test_tarball.py index a774477..366ca5a 100644 --- a/swh/core/tests/test_tarball.py +++ b/swh/core/tests/test_tarball.py @@ -1,240 +1,254 @@ # Copyright (C) 2019-2021 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import hashlib import os import shutil import pytest from swh.core import tarball @pytest.fixture def prepare_shutil_state(): """Reset any shutil modification in its current state""" import shutil registered_formats = [f[0] for f in shutil.get_unpack_formats()] for format_id in tarball.ADDITIONAL_ARCHIVE_FORMATS: name = format_id[0] if name in registered_formats: shutil.unregister_unpack_format(name) return shutil def test_compress_uncompress_zip(tmp_path): tocompress = tmp_path / "compressme" tocompress.mkdir() for i in range(10): fpath = tocompress / ("file%s.txt" % i) fpath.write_text("content of file %s" % i) zipfile = tmp_path / "archive.zip" tarball.compress(str(zipfile), "zip", str(tocompress)) destdir = tmp_path / "destdir" tarball.uncompress(str(zipfile), str(destdir)) lsdir = sorted(x.name for x in destdir.iterdir()) assert ["file%s.txt" % i for i in range(10)] == lsdir @pytest.mark.xfail( reason=( "Python's zipfile library doesn't support Info-ZIP's " "extension for file permissions." ) ) def test_compress_uncompress_zip_modes(tmp_path): tocompress = tmp_path / "compressme" tocompress.mkdir() fpath = tocompress / "text.txt" fpath.write_text("echo foo") fpath.chmod(0o644) fpath = tocompress / "executable.sh" fpath.write_text("echo foo") fpath.chmod(0o755) zipfile = tmp_path / "archive.zip" tarball.compress(str(zipfile), "zip", str(tocompress)) destdir = tmp_path / "destdir" tarball.uncompress(str(zipfile), str(destdir)) (executable_path, text_path) = sorted(destdir.iterdir()) assert text_path.stat().st_mode == 0o100644 # succeeds, it's the default assert executable_path.stat().st_mode == 0o100755 # fails def test_compress_uncompress_tar(tmp_path): tocompress = tmp_path / "compressme" tocompress.mkdir() for i in range(10): fpath = tocompress / ("file%s.txt" % i) fpath.write_text("content of file %s" % i) tarfile = tmp_path / "archive.tar" tarball.compress(str(tarfile), "tar", str(tocompress)) destdir = tmp_path / "destdir" tarball.uncompress(str(tarfile), str(destdir)) lsdir = sorted(x.name for x in destdir.iterdir()) assert ["file%s.txt" % i for i in range(10)] == lsdir def test_compress_uncompress_tar_modes(tmp_path): tocompress = tmp_path / "compressme" tocompress.mkdir() fpath = tocompress / "text.txt" fpath.write_text("echo foo") fpath.chmod(0o644) fpath = tocompress / "executable.sh" fpath.write_text("echo foo") fpath.chmod(0o755) tarfile = tmp_path / "archive.tar" tarball.compress(str(tarfile), "tar", str(tocompress)) destdir = tmp_path / "destdir" tarball.uncompress(str(tarfile), str(destdir)) (executable_path, text_path) = sorted(destdir.iterdir()) assert text_path.stat().st_mode == 0o100644 assert executable_path.stat().st_mode == 0o100755 def test_uncompress_tar_failure(tmp_path, datadir): """Unpack inexistent tarball should fail""" tarpath = os.path.join(datadir, "archives", "inexistent-archive.tar.Z") assert not os.path.exists(tarpath) with pytest.raises(ValueError, match="Problem during unpacking"): tarball.uncompress(tarpath, tmp_path) def test_uncompress_tar(tmp_path, datadir): """Unpack supported tarball into an existent folder should be ok""" filename = "groff-1.02.tar.Z" tarpath = os.path.join(datadir, "archives", filename) assert os.path.exists(tarpath) extract_dir = os.path.join(tmp_path, filename) tarball.uncompress(tarpath, extract_dir) assert len(os.listdir(extract_dir)) > 0 +def test_uncompress_tarzst(tmp_path, datadir): + """Unpack supported zst tarball into an existent folder should be ok""" + filename = "ca-certificates-20210603-1-any.pkg.tar.zst" + tarpath = os.path.join(datadir, "archives", filename) + + assert os.path.exists(tarpath) + + extract_dir = os.path.join(tmp_path, filename) + + tarball.uncompress(tarpath, extract_dir) + + assert len(os.listdir(extract_dir)) > 0 + + def test_register_new_archive_formats(prepare_shutil_state): """Registering new archive formats should be fine""" unpack_formats_v1 = [f[0] for f in shutil.get_unpack_formats()] for format_id in tarball.ADDITIONAL_ARCHIVE_FORMATS: assert format_id[0] not in unpack_formats_v1 # when tarball.register_new_archive_formats() # then unpack_formats_v2 = [f[0] for f in shutil.get_unpack_formats()] for format_id in tarball.ADDITIONAL_ARCHIVE_FORMATS: assert format_id[0] in unpack_formats_v2 def test_uncompress_archives(tmp_path, datadir): """High level call uncompression on supported archives""" archive_dir = os.path.join(datadir, "archives") archive_files = os.listdir(archive_dir) for archive_file in archive_files: archive_path = os.path.join(archive_dir, archive_file) extract_dir = os.path.join(tmp_path, archive_file) tarball.uncompress(archive_path, dest=extract_dir) assert len(os.listdir(extract_dir)) > 0 def test_normalize_permissions(tmp_path): for perms in range(0o1000): filename = str(perms) file_path = tmp_path / filename file_path.touch() file_path.chmod(perms) for file in tmp_path.iterdir(): assert file.stat().st_mode == 0o100000 | int(file.name) tarball.normalize_permissions(str(tmp_path)) for file in tmp_path.iterdir(): if int(file.name) & 0o100: # original file was executable for its owner assert file.stat().st_mode == 0o100755 else: assert file.stat().st_mode == 0o100644 def test_unpcompress_zip_imploded(tmp_path, datadir): """Unpack a zip archive with compression type 6 (implode), not supported by python zipfile module. """ filename = "msk316src.zip" zippath = os.path.join(datadir, "archives", filename) assert os.path.exists(zippath) extract_dir = os.path.join(tmp_path, filename) tarball.uncompress(zippath, extract_dir) assert len(os.listdir(extract_dir)) > 0 def test_uncompress_upper_archive_extension(tmp_path, datadir): """Copy test archives in a temporary directory but turn their names to uppercase, then check they can be successfully extracted. """ archives_path = os.path.join(datadir, "archives") archive_files = [ f for f in os.listdir(archives_path) if os.path.isfile(os.path.join(archives_path, f)) ] for archive_file in archive_files: archive_file_upper = os.path.join(tmp_path, archive_file.upper()) extract_dir = os.path.join(tmp_path, archive_file) shutil.copy(os.path.join(archives_path, archive_file), archive_file_upper) tarball.uncompress(archive_file_upper, extract_dir) assert len(os.listdir(extract_dir)) > 0 def test_uncompress_archive_no_extension(tmp_path, datadir): """Copy test archives in a temporary directory but turn their names to their md5 sums, then check they can be successfully extracted. """ archives_path = os.path.join(datadir, "archives") archive_files = [ f for f in os.listdir(archives_path) if os.path.isfile(os.path.join(archives_path, f)) ] for archive_file in archive_files: archive_file_path = os.path.join(archives_path, archive_file) with open(archive_file_path, "rb") as f: md5sum = hashlib.md5(f.read()).hexdigest() archive_file_md5sum = os.path.join(tmp_path, md5sum) extract_dir = os.path.join(tmp_path, archive_file) shutil.copy(archive_file_path, archive_file_md5sum) tarball.uncompress(archive_file_md5sum, extract_dir) assert len(os.listdir(extract_dir)) > 0