unarchive: use Python's tarfile module for tar listing (#3575)

* unarchive: use Python's tarfile module for tar listing

fixes https://github.com/ansible/ansible/issues/11348

Depending on the current active locale, `tar`'s file listing can end up
spitting backslash-escaped characters. Unfortunately, when that happens,
we end up with double-escaped backslashes, giving us a wrong path,
making our action fail.

We could try un-double-escaping our paths, but that would be complicated
and, I think, error-prone. The easiest way forward seemed to simply use
the `tarfile` module.

Why use it only for listing? Because the `unarchive` option also
supports the `extra_opts` option, and that supporting this would require
us to mimick `tar`'s interface.

For listing files, however, I don't think that the loss of `extra_opts`
support causes problems (well, I hope so).

* unarchive: re-add xz decompression support

Following previous change to use Python's `tarfile` module for tar file
listing, we lost `xz` decompression support. This commits re-add it by
adding a special case in `TarXzArchive` that pre-decompresses the source
file.

lib/ansible/modules/files/unarchive.py

@@ -5,6 +5,7 @@
 # (c) 2013, Dylan Martin <dmartin@seattlecentral.edu>
 # (c) 2015, Toshio Kuratomi <tkuratomi@ansible.com>
 # (c) 2016, Dag Wieers <dag@wieers.com>
+# (c) 2016, Virgil Dupras <hsoft@hardcoded.net>
 #
 # This file is part of Ansible
 #
@@ -114,6 +115,8 @@ import datetime
 import time
 import binascii
 from zipfile import ZipFile
+import tarfile
+import subprocess
 
 # String from tar that shows the tar contents are different from the
 # filesystem
@@ -492,22 +495,23 @@ class TgzArchive(object):
         self.zipflag = 'z'
         self._files_in_archive = []
 
+    def _get_tar_fileobj(self):
+        """Returns a file object that can be read by ``tarfile.open()``."""
+        return open(self.src, 'rb')
+
     @property
     def files_in_archive(self, force_refresh=False):
         if self._files_in_archive and not force_refresh:
             return self._files_in_archive
 
-        cmd = '%s -t%s' % (self.cmd_path, self.zipflag)
+        # The use of Python's tarfile module here allows us to easily avoid tricky file encoding
-        if self.opts:
+        # problems. Ref #11348
-            cmd += ' ' + ' '.join(self.opts)
+        try:
-        if self.excludes:
+            tf = tarfile.open(fileobj=self._get_tar_fileobj())
-            cmd += ' --exclude="' + '" --exclude="'.join(self.excludes) + '"'
+        except Exception:
-        cmd += ' -f "%s"' % self.src
-        rc, out, err = self.module.run_command(cmd)
-        if rc != 0:
             raise UnarchiveError('Unable to list files in the archive')
 
-        for filename in out.splitlines():
+        for filename in tf.getnames():
             if filename and filename not in self.excludes:
                 self._files_in_archive.append(filename)
         return self._files_in_archive
@@ -605,6 +609,19 @@ class TarXzArchive(TgzArchive):
         super(TarXzArchive, self).__init__(src, dest, file_args, module)
         self.zipflag = 'J'
 
+    def _get_tar_fileobj(self):
+        # Python's tarfile module doesn't support xz compression so we have to manually uncompress
+        # it first.
+        xz_bin_path = self.module.get_bin_path('xz')
+        xz_stdout = tempfile.TemporaryFile()
+        # we don't use self.module.run_command() to avoid loading the whole archive in memory.
+        cmd = subprocess.Popen([xz_bin_path, '-dc', self.src], stdout=xz_stdout)
+        rc = cmd.wait()
+        if rc != 0:
+            raise UnarchiveError("Could not uncompress with xz")
+        xz_stdout.seek(0)
+        return xz_stdout
+
 
 # try handlers in order and return the one that works or bail if none work
 def pick_handler(src, dest, file_args, module):