feature: replace unzip usage with Python's zipfile

This removes the need to have `unzip` installed.

fixes pypa#258
closes pypa#312

README.rst

@@ -47,8 +47,7 @@ Installation
 
   $ pip3 install auditwheel
 
-It requires Python 3.6+, and runs on Linux. It requires that the shell command
-``unzip`` be available in the ``PATH``. Only systems that use `ELF
+It requires Python 3.6+, and runs on Linux. Only systems that use `ELF
 <https://en.wikipedia.org/wiki/Executable_and_Linkable_Format>`_-based linkage
 are supported (this should be essentially every Linux).
 

auditwheel/tools.py

@@ -1,8 +1,5 @@
 import argparse
 import os
-import shutil
-from glob import glob
-from os.path import join as pjoin
 from typing import Any, Iterable, List
 import zipfile
 import subprocess
@@ -38,19 +35,18 @@ def zip2dir(zip_fname: str, out_dir: str) -> None:
     out_dir : str
         Directory path containing files to go in the zip archive
     """
-    # Use unzip command rather than zipfile module to preserve permissions
-    # http://bugs.python.org/issue15795
-    subprocess.check_output(['unzip', '-o', '-d', out_dir, zip_fname])
-
-    try:
-        # but sometimes preserving permssions is really bad, and makes it
-        # we don't have the permissions to read any of the files
-        with open(glob(pjoin(out_dir, '*.dist-info/RECORD'))[0]):
-            pass
-    except PermissionError:
-        shutil.rmtree(out_dir)
-        with zipfile.ZipFile(zip_fname) as zf:
-            zf.extractall(out_dir)
+    with zipfile.ZipFile(zip_fname, "r") as z:
+        for name in z.namelist():
+            member = z.getinfo(name)
+            extracted_path = z.extract(member, out_dir)
+            attr = member.external_attr >> 16
+            if member.is_dir():
+                # this is always rebuilt as 755 by dir2zip
+                os.chmod(extracted_path, 0o755)
+            elif attr != 0:
+                attr &= 511  # only keep permission bits
+                attr |= 6 << 6  # at least read/write for current user
+                os.chmod(extracted_path, attr)
 
 
 def dir2zip(in_dir: str, zip_fname: str) -> None:

tests/unit/test_tools.py

@@ -1,6 +1,11 @@
 import argparse
+import lzma
+
+from pathlib import Path
+
 import pytest
-from auditwheel.tools import  EnvironmentDefault
+
+from auditwheel.tools import EnvironmentDefault, zip2dir, dir2zip
 
 
 @pytest.mark.parametrize(
@@ -43,3 +48,41 @@ def test_environment_action_invalid_env(monkeypatch):
             dest='PLAT',
             choices=choices,
             default='manylinux1')
+
+
+def _write_test_permissions_zip(path):
+    source_zip_xz = Path(__file__).parent / "test-permissions.zip.xz"
+    with lzma.open(source_zip_xz) as f:
+        path.write_bytes(f.read())
+
+
+def _check_permissions(path):
+    for i in range(8):
+        for j in range(8):
+            for k in range(8):
+                mode = (path / f"{i}{j}{k}.f").stat().st_mode
+                assert ((mode >> 6) & 7) == (i | 6)  # always read/write
+                assert ((mode >> 3) & 7) == j
+                assert ((mode >> 0) & 7) == k
+                mode = (path / f"{i}{j}{k}.d").stat().st_mode
+                assert ((mode >> 6) & 7) == 7  # always read/write/execute
+                assert ((mode >> 3) & 7) == 5  # always read/execute
+                assert ((mode >> 0) & 7) == 5  # always read/execute
+
+
+def test_zip2dir_permissions(tmp_path):
+    source_zip = tmp_path / "test-permissions.zip"
+    _write_test_permissions_zip(source_zip)
+    extract_path = tmp_path / "unzip"
+    zip2dir(str(source_zip), str(extract_path))
+    _check_permissions(extract_path)
+
+
+def test_zip2dir_round_trip_permissions(tmp_path):
+    source_zip = tmp_path / "test-permissions.zip"
+    _write_test_permissions_zip(source_zip)
+    extract_path = tmp_path / "unzip2"
+    zip2dir(str(source_zip), str(tmp_path / "unzip1"))
+    dir2zip(str(tmp_path / "unzip1"), str(tmp_path / "tmp.zip"))
+    zip2dir(str(tmp_path / "tmp.zip"), str(extract_path))
+    _check_permissions(extract_path)