Intial checkin of layout test analyzer.

media/tools/layout_tests/test_expectations_history.py

Index: media/tools/layout_tests/test_expectations_history.py
diff --git a/media/tools/layout_tests/test_expectations_history.py b/media/tools/layout_tests/test_expectations_history.py
new file mode 100644
index 0000000000000000000000000000000000000000..3ec28386ea64cc2b2249e731f8389b9f4c17fb5e
--- /dev/null
+++ b/media/tools/layout_tests/test_expectations_history.py
@@ -0,0 +1,102 @@
+#!/usr/bin/python
+# Copyright (c) 2011 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+"""A Module for the history of the test expectation file."""
+
+import re
+import time
+import pysvn
+
+from datetime import datetime
+from datetime import timedelta
+
+# Default Webkit SVN location for chromium test expectation file.
+# TODO(imasaki): support multiple test expectation files.
+DEFAULT_TEST_EXPECTATION_LOCATION = (
+    'http://svn.webkit.org/repository/webkit/trunk/'
+    'LayoutTests/platform/chromium/test_expectations.txt')
+
+
+class TestExpectationsHistory:
+  """A class to represent history of the test expectation file.
+
+  The history is obtained by calling PySVN.log()/diff() APIs.
+
+  TODO(imasaki): Add more functionalities here like getting some statistics
+      about the test expectation file.
+  """
+
+  @staticmethod
+  def GetDiffBetweenTimes(start, end, testname_list,
+                          te_location=DEFAULT_TEST_EXPECTATION_LOCATION):
+    """Get difference between time period for the specified test names.
+
+    Given the time period, this method first gets the revision number. Then,
+    it gets the diff for each revision. Finally, it keeps the diff relating to
+    the test names and returns them along with other information about
+    revision.
+
+    Args:
+      start: A timestamp specifying start of the time period to be
+          looked at.
+      end: A timestamp object specifying end of the time period to be
+          looked at.
+      testname_list: A list of strings representing test names of interest.
+      te_location: A location of the test expectation file.
+
+    Returns:
+      A list of tuples (old_rev, new_rev, author, date, message, lines). The
+          |lines| contains the diff of the tests of interest.
+    """
+    # Get directory name which is necesary to call PySVN.checkout().
+    te_location_dir = te_location[0:te_location.rindex('/')]
+    client = pysvn.Client()
+    client.checkout(te_location_dir, 'tmp', recurse=False)
+    logs = client.log('tmp/test_expectations.txt',
+                      revision_start=pysvn.Revision(
+                          pysvn.opt_revision_kind.date, start),
+                      revision_end=pysvn.Revision(
+                          pysvn.opt_revision_kind.date, end))
+    result_list = []
+    # Find the last revision outside of time period and
+    # append it to preserve the last change before entering the time period.
+    # This is 1 day back as |start2| (called start2) from original |start|.
+    # The reason for using 1 day back is that there should be at least one
+    # revision during the 1 day period and so that at least 1 revision is
+    # obtained.

[dennis_jeffrey, 2011/08/26 19:01:26]

Will anything bad happen if at least 1 revision does not exist?  Suppose we set
up a worldwide "software engineer vacation day" such that some day occurs when
nobody checks in a revision.  If that situation would break this code, we should
probably have a check for that to handle this potential situation gracefully.

[imasaki1, 2011/08/26 22:28:44]

Modified the code to handle this case.

+    start2 = time.mktime(
+        (datetime.fromtimestamp(start) - timedelta(days=1)).timetuple())
+    logs2 = client.log('tmp/test_expectations.txt',
+                       revision_start=pysvn.Revision(
+                           pysvn.opt_revision_kind.date, start2),
+                       revision_end=pysvn.Revision(
+                           pysvn.opt_revision_kind.date, start))
+    if logs2:
+      logs.append(logs2[len(logs2) - 2])
+    for i in xrange(len(logs) - 1):
+      # PySVN.log() returns logs in reverse chronological order.
+      new_rev = logs[i].revision.number
+      old_rev = logs[i + 1].revision.number
+      # Parsing the actual diff.
+      text = client.diff('/tmp', 'tmp/test_expectations.txt',
+                         revision1=pysvn.Revision(
+                             pysvn.opt_revision_kind.number, old_rev),
+                         revision2=pysvn.Revision(
+                             pysvn.opt_revision_kind.number, new_rev))
+      lines = text.split('\n')
+      target_lines = []
+      for line in lines:
+        for testname in testname_list:
+          matches = re.findall(testname, line)
+          if matches:
+            if line[0] == '+' or line[0] == '-':
+              target_lines.append(line)
+      if target_lines:
+        # Needs to convert to normal date string for presentation.
+        result_list.append((
+            old_rev, new_rev, logs[i].author,
+            datetime.fromtimestamp(logs[i].date).strftime('%Y-%m-%d %H:%M:%S'),
+            logs[i].message, target_lines))
+    return result_list