Fix report symbolization on swarming bots.

tools/valgrind/asan/asan_symbolize.py

 #!/usr/bin/env python
 
 # Copyright (c) 2012 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.
 
 from third_party import asan_symbolize
 
 import argparse
 import base64
 import json
 import os
+import re
+import subprocess
 import sys
 
 class LineBuffered(object):
   """Disable buffering on a file object."""
   def __init__(self, stream):
     self.stream = stream
 
   def write(self, data):
     self.stream.write(data)
     if '\n' in data:
@@ skipping 18 matching lines @@
     script_dir = os.path.dirname(os.path.abspath(__file__))
     # Assume this script resides three levels below src/ (i.e.
     # src/tools/valgrind/asan/).
     src_root = os.path.join(script_dir, "..", "..", "..")
     symbolizer_path = os.path.join(src_root, 'third_party',
         'llvm-build', 'Release+Asserts', 'bin', 'llvm-symbolizer')
     assert(os.path.isfile(symbolizer_path))
     os.environ['LLVM_SYMBOLIZER_PATH'] = os.path.abspath(symbolizer_path)
 
 
+def is_hash_name(name):
+  match = re.match('[0-9a-f]+', name)

[earthdok, 2015/01/23 15:41:50]

Note that re.match() matches at the beginning of the string, but allows
unmatched characters at the end. You probably want '[0-9a-f]+$'.

[Alexander Potapenko, 2015/01/23 16:37:55]

Done.

+  return bool(match)
+
+
+def chrome_product_dir_path(exe_path):
+  if exe_path is None:
+    return None
+  path_parts = exe_path.split(os.path.sep)

[earthdok, 2015/01/23 15:41:50]

why not os.path.split()/os.path.join()?

[Alexander Potapenko, 2015/01/23 16:37:55]

Done

+  if len(path_parts) == 1:

[earthdok, 2015/01/23 15:41:50]

the logic is not entirely clear here. a comment would be helpful.

[Alexander Potapenko, 2015/01/23 16:37:55]

Done.

+    path_parts = ['.'] + path_parts
+  for index, part in enumerate(path_parts):
+    if part.endswith('.app'):
+      return os.path.sep.join(path_parts[:index])
+  return os.path.sep.join(path_parts[:-1])
+
+
+inode_path_cache = {}
+
+
+def find_inode_at_path(inode, path):
+  if inode in inode_path_cache:
+    return inode_path_cache[inode]
+  cmd = ['find', path, '-inum', str(inode)]
+  find_line = subprocess.check_output(cmd).rstrip()
+  lines = find_line.split('\n')
+  ret = None
+  if lines:
+    # `find` may give us several paths (e.g. 'Chromium Framework' in the
+    # product dir and 'Chromium Framework' inside 'Chromium.app',
+    # chrome_dsym_hints() will produce correct .dSYM path for any of them.
+    ret = lines[0]
+  inode_path_cache[inode] = ret
+  return ret
+
+
+# Create a binary name filter that works around https://crbug.com/444835.
+# When running tests on OSX swarming servers, ASan sometimes prints paths to
+# files in cache (ending with SHA1 filenames) instead of paths to hardlinks to
+# those files in the product dir.
+# For a given |binary_path| chrome_osx_binary_name_filter() returns one of the
+# hardlinks to the same inode in |product_dir_path|.
+def make_chrome_osx_binary_name_filter(product_dir_path=''):
+  def chrome_osx_binary_name_filter(binary_path):
+    basename = os.path.basename(binary_path)
+    if is_hash_name(basename) and product_dir_path:
+      inode = os.stat(binary_path).st_ino
+      new_binary_path = find_inode_at_path(inode, product_dir_path)
+      if new_binary_path:
+        binary_path = new_binary_path

[earthdok, 2015/01/23 15:41:50]

return new_binary_path ?

[Alexander Potapenko, 2015/01/23 16:37:55]

Done.

+    return binary_path
+  return chrome_osx_binary_name_filter
+
+
 # Construct a path to the .dSYM bundle for the given binary.
 # There are three possible cases for binary location in Chromium:
 # 1. The binary is a standalone executable or dynamic library in the product
 #    dir, the debug info is in "binary.dSYM" in the product dir.
 # 2. The binary is a standalone framework or .app bundle, the debug info is in
 #    "Framework.framework.dSYM" or "App.app.dSYM" in the product dir.
 # 3. The binary is a framework or an .app bundle within another .app bundle
 #    (e.g. Outer.app/Contents/Versions/1.2.3.4/Inner.app), and the debug info
 #    is in Inner.app.dSYM in the product dir.
 # The first case is handled by llvm-symbolizer, so we only need to construct
@@ skipping 101 matching lines @@
 def main():
   parser = argparse.ArgumentParser(description='Symbolize sanitizer reports.')
   parser.add_argument('--test-summary-json-file',
       help='Path to a JSON file produced by the test launcher. The script will '
            'ignore stdandard input and instead symbolize the output stnippets '
            'inside the JSON file. The result will be written back to the JSON '
            'file.')
   parser.add_argument('strip_path_prefix', nargs='*',
       help='When printing source file names, the longest prefix ending in one '
            'of these substrings will be stripped. E.g.: "Release/../../".')
+  parser.add_argument('--executable_path',

[earthdok, 2015/01/23 15:41:50]

dash, not underscore

[Alexander Potapenko, 2015/01/23 16:37:55]

Done.

+      help='Path to program executable. Will be used on OSX swarming bots to '

[earthdok, 2015/01/23 15:41:50]

Please drop the future tense.

[Alexander Potapenko, 2015/01/23 16:37:55]

Done.

+           'locate dSYM bundles for associated frameworks and bundles.')
   args = parser.parse_args()
 
   disable_buffering()
   set_symbolizer_path()
   asan_symbolize.demangle = True
   asan_symbolize.fix_filename_patterns = args.strip_path_prefix
-  loop = asan_symbolize.SymbolizationLoop(dsym_hint_producer=chrome_dsym_hints)
+  maybe_binary_name_filter = None

[earthdok, 2015/01/23 15:41:50]

"maybe_" is kind of unnecessary

[Alexander Potapenko, 2015/01/23 16:37:55]

Done.

+  if os.uname()[0] == 'Darwin':
+    maybe_binary_name_filter = make_chrome_osx_binary_name_filter(
+        chrome_product_dir_path(args.executable_path))
+  loop = asan_symbolize.SymbolizationLoop(
+      binary_name_filter=maybe_binary_name_filter,
+      dsym_hint_producer=chrome_dsym_hints)
 
   if args.test_summary_json_file:
     symbolize_snippets_in_json(args.test_summary_json_file, loop)
   else:
     # Process stdin.
     asan_symbolize.logfile = sys.stdin
     loop.process_logfile()
 
 if __name__ == '__main__':
   main()