[foozzie] Improve failure state deduplication

tools/foozzie/v8_foozzie.py

 #!/usr/bin/env python
 # Copyright 2016 the V8 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.
 
 """
 V8 correctness fuzzer launcher script.
 """
 
 import argparse
@@ skipping 19 matching lines @@
   ignition_turbo_opt=['--ignition-staging', '--turbo', '--always-opt'],
 )
 
 # Timeout in seconds for one d8 run.
 TIMEOUT = 3
 
 # Return codes.
 RETURN_PASS = 0
 RETURN_FAIL = 2
 
+# The number of hex digits used from the hash of the original source file path.
+# Keep the number small to avoid duplicate explosion.
+SOURCE_HASH_LENGTH = 3
+
 BASE_PATH = os.path.dirname(os.path.abspath(__file__))
 PREAMBLE = [
   os.path.join(BASE_PATH, 'v8_mock.js'),
   os.path.join(BASE_PATH, 'v8_suppressions.js'),
 ]
 
 FLAGS = ['--abort_on_stack_overflow', '--expose-gc', '--allow-natives-syntax',
          '--invoke-weak-callbacks', '--omit-quit', '--es-staging']
 
 SUPPORTED_ARCHS = ['ia32', 'x64', 'arm', 'arm64']
@@ skipping 145 matching lines @@
 def main():
   options = parse_args()
 
   # Suppressions are architecture and configuration specific.
   suppress = v8_suppressions.get_suppression(
       options.first_arch, options.first_config,
       options.second_arch, options.second_config,
   )
 
   # Get metadata.
+  # TODO(machenbach): We probably don't need the metadata file anymore
+  # now that the metadata is printed in the test cases.
   with open(options.meta_data_path) as f:
     metadata = json.load(f)
 
   if metadata_bailout(metadata, suppress.ignore_by_metadata):
     return RETURN_FAIL
 
   if test_pattern_bailout(options.testcase, suppress.ignore_by_content):
     return RETURN_FAIL
 
   common_flags = FLAGS + ['--random-seed', str(options.random_seed)]
@@ skipping 20 matching lines @@
     return RETURN_FAIL
 
   second_config_output = run_d8(options.second_d8, second_config_flags)
 
   # Bailout based on second run's output.
   if pass_bailout(second_config_output, 2):
     return RETURN_PASS
   if fail_bailout(second_config_output, suppress.ignore_by_output2):
     return RETURN_FAIL
 
-  difference = suppress.diff(
+  difference, source = suppress.diff(
       first_config_output.stdout, second_config_output.stdout)
   if difference:
     # The first three entries will be parsed by clusterfuzz. Format changes
     # will require changes on the clusterfuzz side.
     first_config_label = '%s,%s' % (options.first_arch, options.first_config)
     second_config_label = '%s,%s' % (options.second_arch, options.second_config)
-    hsh = lambda x: hashlib.sha1(x).hexdigest()[:8]
     print FAILURE_TEMPLATE % dict(
         configs='%s:%s' % (first_config_label, second_config_label),
-        sources=','.join(map(hsh, metadata['sources'])),
+        sources=hashlib.sha1(source).hexdigest()[:SOURCE_HASH_LENGTH],
         suppression='', # We can't tie bugs to differences.
         first_config_label=first_config_label,
         second_config_label=second_config_label,
         first_config_flags=' '.join(first_config_flags),
         second_config_flags=' '.join(second_config_flags),
         first_config_output=first_config_output.stdout,
         second_config_output=second_config_output.stdout,
         difference=difference,
     )
     return RETURN_FAIL
@@ skipping 16 matching lines @@
         configs='', sources='', suppression='wrong_usage')
     result = RETURN_FAIL
   except Exception as e:
     print FAILURE_HEADER_TEMPLATE % dict(
         configs='', sources='', suppression='internal_error')
     print '# Internal error: %s' % e
     traceback.print_exc(file=sys.stdout)
     result = RETURN_FAIL
 
   sys.exit(result)