Make some changes to the parser so that when multiple reasons are

logged not only are duplicates discarded, but also reason strings which
are subsets of one another.

So for example if a test logs X and then X,Y and then X,Y,Z then the
final reason will only include X,Y,Z.

Risk: Low
Visibility: Should reduce the amount of duplication in final reasons
for multi-reason failures.

Signed-off-by: John Admanski <jadmanski@google.com>



git-svn-id: svn://test.kernel.org/autotest/trunk@3346 592f7852-d20e-0410-864c-8624ca9c26a4

3 files changed, 48 insertions, 0 deletions

diff --git a/tko/parsers/version_1.py b/tko/parsers/version_1.py
index fb82a99d..ab3a62c4 100644
--- a/tko/parsers/version_1.py
+++ b/tko/parsers/version_1.py
@@ -292,6 +292,8 @@ class parser(base.parser):
                         # update the status of a currently running test
                         if running_test:
                             running_reasons.add(line.reason)
+                            running_reasons = tko_utils.drop_redundant_messages(
+                                    running_reasons)
                             sorted_reasons = sorted(running_reasons)
                             running_test.reason = ", ".join(sorted_reasons)
                             current_reason = running_test.reason
diff --git a/tko/utils.py b/tko/utils.py
index 8fc3711f..1bd1bf46 100644
--- a/tko/utils.py
+++ b/tko/utils.py
@@ -30,3 +30,22 @@ def find_toplevel_job_dir(start_dir):
             return None
         job_dir = os.path.dirname(job_dir)
     return job_dir
+
+
+def drop_redundant_messages(messages):
+    """ Given a set of message strings discard any 'redundant' messages which
+    are simple a substring of the existing ones.
+
+    @param messages - a set of message strings
+
+    @return - a subset of messages with unnecessary strings dropped
+    """
+    sorted_messages = sorted(messages, key=len, reverse=True)
+    filtered_messages = set()
+    for message in sorted_messages:
+        for filtered_message in filtered_messages:
+            if message in filtered_message:
+                break
+        else:
+            filtered_messages.add(message)
+    return filtered_messages
diff --git a/tko/utils_unittest.py b/tko/utils_unittest.py
index 2a04191d..7d184079 100644
--- a/tko/utils_unittest.py
+++ b/tko/utils_unittest.py
@@ -98,6 +98,33 @@ class find_toplevel_job_dir_test(unittest.TestCase):
         self.assertEqual(utils.find_toplevel_job_dir(jobdir), None)
 
 
+class drop_redundant_messages(unittest.TestCase):
+    def test_empty_set(self):
+        self.assertEqual(utils.drop_redundant_messages(set()), set())
+
+
+    def test_singleton(self):
+        self.assertEqual(utils.drop_redundant_messages(set(["abc"])),
+                         set(["abc"]))
+
+
+    def test_distinct_messages(self):
+        self.assertEqual(utils.drop_redundant_messages(set(["abc", "def"])),
+                         set(["abc", "def"]))
+
+
+    def test_one_unique_message(self):
+        self.assertEqual(
+                utils.drop_redundant_messages(set(["abc", "abcd", "abcde"])),
+                set(["abcde"]))
+
+
+    def test_some_unique_some_not(self):
+        self.assertEqual(
+                utils.drop_redundant_messages(set(["abc", "def", "abcdef",
+                                                   "defghi", "cd"])),
+                set(["abcdef", "defghi"]))
+
 
 if __name__ == "__main__":
     unittest.main()