Fixed #6048: Distutils uses the tarfile module instead of the tar command now

Lib/distutils/archive_util.py

@@ -6,6 +6,9 @@ that sort of thing)."""
 __revision__ = "$Id$"
 
 import os
+from warnings import warn
+import sys
+
 from distutils.errors import DistutilsExecError
 from distutils.spawn import spawn
 from distutils.dir_util import mkpath
@@ -22,35 +25,44 @@ def make_tarball(base_name, base_dir, compress="gzip", verbose=0, dry_run=0):
     the appropriate compression extension (".gz", ".bz2" or ".Z").
     Returns the output filename.
     """
-    # XXX GNU tar 1.13 has a nifty option to add a prefix directory.
-    # It's pretty new, though, so we certainly can't require it --
-    # but it would be nice to take advantage of it to skip the
-    # "create a tree of hardlinks" step!  (Would also be nice to
-    # detect GNU tar to use its 'z' option and save a step.)
-
-    compress_ext = {'gzip': ".gz",
-                    'bzip2': '.bz2',
-                    'compress': ".Z" }
+    tar_compression = {'gzip': 'gz', 'bzip2': 'bz2', None: '', 'compress': ''}
+    compress_ext = {'gzip': '.gz', 'bzip2': '.bz2', 'compress': '.Z'}
 
     # flags for compression program, each element of list will be an argument
-    compress_flags = {'gzip': ["-f9"],
-                      'compress': ["-f"],
-                      'bzip2': ['-f9']}
-
     if compress is not None and compress not in compress_ext.keys():
         raise ValueError, \
-              "bad value for 'compress': must be None, 'gzip', or 'compress'"
+              ("bad value for 'compress': must be None, 'gzip', 'bzip2' "
+               "or 'compress'")
+
+    archive_name = base_name + '.tar'
+    if compress != 'compress':
+        archive_name += compress_ext.get(compress, '')
 
-    archive_name = base_name + ".tar"
     mkpath(os.path.dirname(archive_name), dry_run=dry_run)
-    cmd = ["tar", "-cf", archive_name, base_dir]
-    spawn(cmd, dry_run=dry_run)
 
-    if compress:
-        spawn([compress] + compress_flags[compress] + [archive_name],
-              dry_run=dry_run)
-        return archive_name + compress_ext[compress]
+    # creating the tarball
+    import tarfile  # late import so Python build itself doesn't break
+
+    log.info('Creating tar archive')
+    if not dry_run:
+        tar = tarfile.open(archive_name, 'w|%s' % tar_compression[compress])
+        try:
+            tar.add(base_dir)
+        finally:
+            tar.close()
+
+    # compression using `compress`
+    if compress == 'compress':
+        warn("'compress' will be deprecated.", PendingDeprecationWarning)
+        # the option varies depending on the platform
+        compressed_name = archive_name + compress_ext[compress]
+        if sys.platform == 'win32':
+            cmd = [compress, archive_name, compressed_name]
         else:
+            cmd = [compress, '-f', archive_name]
+        spawn(cmd, dry_run=dry_run)
+        return compressed_name
+
     return archive_name
 
 def make_zipfile(base_name, base_dir, verbose=0, dry_run=0):

Lib/distutils/tests/test_archive_util.py

@@ -3,12 +3,15 @@ __revision__ = "$Id$"
 
 import unittest
 import os
+import tarfile
 from os.path import splitdrive
+import warnings
 
 from distutils.archive_util import (check_archive_formats, make_tarball,
                                     make_zipfile, make_archive)
-from distutils.spawn import find_executable
+from distutils.spawn import find_executable, spawn
 from distutils.tests import support
+from test.test_support import check_warnings
 
 try:
     import zipfile
@@ -19,12 +22,13 @@ except ImportError:
 class ArchiveUtilTestCase(support.TempdirManager,
                           unittest.TestCase):
 
-    @unittest.skipUnless(find_executable('tar'), 'Need the tar command to run')
     def test_make_tarball(self):
         # creating something to tar
         tmpdir = self.mkdtemp()
         self.write_file([tmpdir, 'file1'], 'xxx')
         self.write_file([tmpdir, 'file2'], 'xxx')
+        os.mkdir(os.path.join(tmpdir, 'sub'))
+        self.write_file([tmpdir, 'sub', 'file3'], 'xxx')
 
         tmpdir2 = self.mkdtemp()
         unittest.skipUnless(splitdrive(tmpdir)[0] == splitdrive(tmpdir2)[0],
@@ -55,6 +59,111 @@ class ArchiveUtilTestCase(support.TempdirManager,
         tarball = base_name + '.tar'
         self.assert_(os.path.exists(tarball))
 
+    def _tarinfo(self, path):
+        tar = tarfile.open(path)
+        try:
+            names = tar.getnames()
+            names.sort()
+            return tuple(names)
+        finally:
+            tar.close()
+
+    def _create_files(self):
+        # creating something to tar
+        tmpdir = self.mkdtemp()
+        dist = os.path.join(tmpdir, 'dist')
+        os.mkdir(dist)
+        self.write_file([dist, 'file1'], 'xxx')
+        self.write_file([dist, 'file2'], 'xxx')
+        os.mkdir(os.path.join(dist, 'sub'))
+        self.write_file([dist, 'sub', 'file3'], 'xxx')
+        os.mkdir(os.path.join(dist, 'sub2'))
+        tmpdir2 = self.mkdtemp()
+        base_name = os.path.join(tmpdir2, 'archive')
+        return tmpdir, tmpdir2, base_name
+
+    @unittest.skipUnless(find_executable('tar'), 'Need the tar command to run')
+    def test_tarfile_vs_tar(self):
+        tmpdir, tmpdir2, base_name =  self._create_files()
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        try:
+            make_tarball(base_name, 'dist')
+        finally:
+            os.chdir(old_dir)
+
+        # check if the compressed tarball was created
+        tarball = base_name + '.tar.gz'
+        self.assert_(os.path.exists(tarball))
+
+        # now create another tarball using `tar`
+        tarball2 = os.path.join(tmpdir, 'archive2.tar.gz')
+        cmd = ['tar', '-czf', 'archive2.tar.gz', 'dist']
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        try:
+            spawn(cmd)
+        finally:
+            os.chdir(old_dir)
+
+        self.assert_(os.path.exists(tarball2))
+        # let's compare both tarballs
+        self.assertEquals(self._tarinfo(tarball), self._tarinfo(tarball2))
+
+        # trying an uncompressed one
+        base_name = os.path.join(tmpdir2, 'archive')
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        try:
+            make_tarball(base_name, 'dist', compress=None)
+        finally:
+            os.chdir(old_dir)
+        tarball = base_name + '.tar'
+        self.assert_(os.path.exists(tarball))
+
+        # now for a dry_run
+        base_name = os.path.join(tmpdir2, 'archive')
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        try:
+            make_tarball(base_name, 'dist', compress=None, dry_run=True)
+        finally:
+            os.chdir(old_dir)
+        tarball = base_name + '.tar'
+        self.assert_(os.path.exists(tarball))
+
+    @unittest.skipUnless(find_executable('compress'),
+                         'The compress program is required')
+    def test_compress_deprecated(self):
+        tmpdir, tmpdir2, base_name =  self._create_files()
+
+        # using compress and testing the PendingDeprecationWarning
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        try:
+            with check_warnings() as w:
+                warnings.simplefilter("always")
+                make_tarball(base_name, 'dist', compress='compress')
+        finally:
+            os.chdir(old_dir)
+        tarball = base_name + '.tar.Z'
+        self.assert_(os.path.exists(tarball))
+        self.assertEquals(len(w.warnings), 1)
+
+        # same test with dry_run
+        os.remove(tarball)
+        old_dir = os.getcwd()
+        os.chdir(tmpdir)
+        try:
+            with check_warnings() as w:
+                warnings.simplefilter("always")
+                make_tarball(base_name, 'dist', compress='compress',
+                             dry_run=True)
+        finally:
+            os.chdir(old_dir)
+        self.assert_(not os.path.exists(tarball))
+        self.assertEquals(len(w.warnings), 1)
+
     @unittest.skipUnless(ZIP_SUPPORT, 'Need zip support to run')
     def test_make_zipfile(self):
         # creating something to tar

Misc/NEWS

@@ -307,6 +307,8 @@ Core and Builtins
 Library
 -------
 
+- Issue #6048: Now Distutils uses the tarfile module in archive_util.
+
 - Issue #6121: pydoc now ignores leading and trailing spaces in the
   argument to the 'help' function.