Add presubmit check to prevent package boundary violations.

tools-webrtc/check_package_boundaries_test.py

+#!/usr/bin/env python
+
+# Copyright (c) 2017 The WebRTC project authors. All Rights Reserved.
+#
+# Use of this source code is governed by a BSD-style license
+# that can be found in the LICENSE file in the root of the source
+# tree. An additional intellectual property rights grant can be found
+# in the file PATENTS.  All contributing project authors may
+# be found in the AUTHORS file in the root of the source tree.
+
+import os
+import subprocess
+import unittest
+
+
+MSG_FORMAT = 'ERROR:check_package_boundaries.py: Unexpected %s.'
+
+
+def _RunCommand(command, cwd):
+  """Runs a command and returns the output from that command."""
+  p = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE,
+                       cwd=cwd)
+  stdout = p.stdout.read()
+  stderr = p.stderr.read()
+  p.wait()
+  p.stdout.close()
+  p.stderr.close()
+  return p.returncode, stdout, stderr
+
+
+class UnitTest(unittest.TestCase):
+  def test_check_package_boundaries(self):
+    os.chdir(os.path.dirname(os.path.abspath(__file__)))

[kjellander_webrtc, 2017/01/24 09:33:19]

It's generally bad practice to use os.chdir in scripts and IMO it's often a
source of bugs and hard-to-read code (although not here). 
To send the right signal to future readers, could you you create: 
SCRIPT_DIR = os.path.dirname(os.path.abspath(__file__))
and then join with that below instead? Otherwise I'm afraid if we add another
test case in the future, it will still have cwd set to what you change it to
here (it applies to the currently running interpreter executable IIRC).

+    with open(os.path.join('testdata', 'expected_stderr')) as f:
+      expected_stderr = f.read()
+    with open(os.path.join('testdata', 'expected_stdout')) as f:
+      expected_stdout = f.read()
+
+    test_command = ['python', 'check_package_boundaries.py', 'testdata']

[kjellander_webrtc, 2017/01/24 09:33:19]

Please refactor the check_package_boundaries module so you can invoke a function
in it here instead of firing up another Python process.
I'm afraid we loose several benefits with the unit test framework if we launch
another Python instance (like easy debugging).

+    return_code, stdout, stderr = _RunCommand(test_command, cwd=os.getcwd())
+
+    self.assertEqual(return_code, 1)
+    self.assertMultiLineEqual(expected_stdout, stdout)
+    self.assertMultiLineEqual(expected_stderr, stderr)
+
+
+if __name__ == '__main__':
+  unittest.main()