Reland of Size limit gtest collect

scripts/slave/recipe_modules/swarming/resources/collect_gtest_task.py

 #!/usr/bin/env python
 # Copyright 2014 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.
 
 import json
 import optparse
 import os
 import shutil
 import subprocess
@@ skipping 77 matching lines @@
         MISSING_SHARDS_MSG % as_str)
     # Not all tests run, combined JSON summary can not be trusted.
     merged['global_tags'].add('UNRELIABLE_RESULTS')
 
   # Convert to jsonish dict.
   for key in ('all_tests', 'disabled_tests', 'global_tags'):
     merged[key] = sorted(merged[key])
   return merged
 
 
+# 100 MB
+OUTPUT_JSON_SIZE_LIMIT = 100 * 1024 * 1024
+
+
 def load_shard_json(output_dir, index):
   """Reads JSON output of a single shard."""
   # 'output.json' is set in swarming/api.py, gtest_task method.
   path = os.path.join(output_dir, str(index), 'output.json')
   try:
+    filesize = os.stat(path).st_size
+    if filesize > OUTPUT_JSON_SIZE_LIMIT:

[Vadim Sh., 2016/09/27 00:10:56]

please log this to stderr

+      raise ValueError()
+
     with open(path) as f:
       return json.load(f)
   except (IOError, ValueError):

[Vadim Sh., 2016/09/27 00:10:56]

catch OSError here, since os.stat can raise it now if 'path' is missing

     print >> sys.stderr, 'Missing or invalid gtest JSON file: %s' % path
     return None
 
 
 def merge_list_of_dicts(left, right):
   """Merges dicts left[0] with right[0], left[1] with right[1], etc."""
   output = []
   for i in xrange(max(len(left), len(right))):
     left_dict = left[i] if i < len(left) else {}
     right_dict = right[i] if i < len(right) else {}
@@ skipping 74 matching lines @@
           'failed to process gtest output JSON', traceback.format_exc())
 
   finally:
     shutil.rmtree(task_output_dir, ignore_errors=True)
 
   return exit_code
 
 
 if __name__ == '__main__':
   sys.exit(main(sys.argv[1:]))