Merge lp:~mvo/apt-clone/lp1309447 into lp:apt-clone

[Michael Vogt]

This fixes the apt-clone part of bug #1309447.

[Iain Lane (laney)]

This looks good to me, will merge since Barry has been AWOL for months. :)

review: Approve

=== modified file 'apt-clone'
--- apt-clone	2014-03-19 07:27:54 +0000
+++ apt-clone	2014-06-06 12:27:01 +0000
@@ -22,6 +22,7 @@
 import argparse
 from apt_clone import AptClone
 
+
 if __name__ == "__main__":
 
     # command line parser

=== modified file 'apt_clone.py'
--- apt_clone.py	2014-03-12 15:34:43 +0000
+++ apt_clone.py	2014-06-06 12:27:01 +0000
@@ -247,8 +247,11 @@
 
     def _add_file_to_tar_with_password_check(self, tar, sources, scrub, arcname):
         if scrub:
-            with tempfile.NamedTemporaryFile(mode='wb') as source_copy, open(sources, 'r') as f:
+            with tempfile.NamedTemporaryFile(mode='wb') as source_copy, open(sources, 'rb') as f:
                 for line in f.readlines():
+                    # compat with both py2/py3
+                    if type(line) is bytes:
+                        line = line.decode("UTF-8")
                     if re.search('/[^/@:]*:[^/@:]*@', line):
                         line = re.sub('/[^/@:]*:[^/@:]*@',
                             '/USERNAME:PASSWORD@', line)
@@ -256,9 +259,7 @@
                     # open in Unicode mode in Python 2.  We can remove this
                     # once ubuntu-release-upgrader is run under Python 3
                     # (i.e. after Ubuntu 14.04).
-                    if line is not bytes:
-                        line = line.encode("UTF-8")
-                    source_copy.write(line)
+                    source_copy.write(line.encode("utf-8"))
                     source_copy.flush()
                 tar.add(source_copy.name, arcname=arcname)
         else:

=== modified file 'tests/test_clone.py'
--- tests/test_clone.py	2013-07-26 00:59:55 +0000
+++ tests/test_clone.py	2014-06-06 12:27:01 +0000
@@ -1,5 +1,5 @@
 #!/usr/bin/python3
-
+# -*- coding: utf-8 -*-
 from __future__ import print_function
 
 import apt
@@ -195,5 +195,6 @@
         self.assertFalse("/etc/issue" in unowned)
         #print("\n".join(sorted(unowned)))
 
+
 if __name__ == "__main__":
     unittest.main()

=== added file 'tests/test_lp1309447.py'
--- tests/test_lp1309447.py	1970-01-01 00:00:00 +0000
+++ tests/test_lp1309447.py	2014-06-06 12:27:01 +0000
@@ -0,0 +1,51 @@
+#!/usr/bin/python3
+# -*- coding: utf-8 -*-
+
+from __future__ import print_function
+
+import os
+import sys
+import unittest
+
+# this is important
+os.environ["LANG"] = "C"
+
+
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), ".."))
+from apt_clone import AptClone
+
+
+class MockTar(object):
+    def add(self, source, arcname):
+        with open(source, "rb") as f:
+            self.data = f.read().decode("utf-8")
+
+
+class TestClone(unittest.TestCase):
+
+    def setUp(self):
+        self.apt_clone = AptClone()
+        self.test_sources_fname = "test-sources.list"
+        with open(self.test_sources_fname, "wb") as f:
+            f.write(u"""# äüö
+deb http://mvo:secret@archive.u.c/ ubuntu main
+""".encode("utf-8"))
+
+    def tearDown(self):
+        os.unlink(self.test_sources_fname)
+
+    def test_scrub_file_from_passwords(self):
+        """Regression test for utf8 crash LP: #1309447"""
+        mock_tar = MockTar()
+        self.apt_clone._add_file_to_tar_with_password_check(
+            mock_tar, self.test_sources_fname, scrub=True, 
+            arcname="some-archname")
+        # see if we got the expected data
+        self.assertNotIn("mvo:secret", mock_tar.data)
+        self.assertEqual(mock_tar.data, u"""# äüö
+deb http://USERNAME:PASSWORD@archive.u.c/ ubuntu main
+""")
+
+
+if __name__ == "__main__":
+    unittest.main()