[libfuzzer] in-vcs corpus support.

testing/libfuzzer/archive_corpus.py

+#!/usr/bin/python2
+#
+# Copyright (c) 2015 The Chromium Authors. All rights reserved.

[mmoroz, 2016/03/07 14:08:55]

There is no "(c)" now, 2015 -> 2016.

Looks like we also need to update this header in gen_fuzzer_config.py.

[aizatsky, 2016/03/07 21:42:54]

Done.

+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+"""Archive corpus file into zip and generate .d depfile.
+
+Invoked by GN from fuzzer_test.gni.
+"""
+
+from __future__ import print_function
+import argparse
+import os
+import sys
+import zipfile
+
+
+def main():
+  parser = argparse.ArgumentParser(description="Generate fuzzer config.")
+  parser.add_argument('--depfile', required=True)
+  parser.add_argument('--corpus', required=True)
+  parser.add_argument('--output', required=True)
+  args = parser.parse_args()
+
+  corpus_files = []
+  # Generate .d file with dependency from corpus archive to individual files.
+  with open(args.depfile, 'w') as depfile:

[mmoroz, 2016/03/07 14:08:55]

Do we need this .d file? On ClusterFuzz side we can check existence of
zip-archive and use it together with corpus from GCS.

[aizatsky, 2016/03/07 21:42:54]

This file is needed for gn/ninja. I've also added a dependency to fuzzer binary,
so that corpus archive gets rebuilt.

+      print(args.output, ":", end="", file=depfile)
+      for (dirpath, _, filenames) in os.walk(args.corpus):
+          for filename in filenames:
+              full_filename = os.path.join(dirpath, filename)
+              print(" ", full_filename, end="", file=depfile)

[mmoroz, 2016/03/07 14:08:55]

Seems that this call will write strings like "<space><space><filepath>" into one
line of depfile. Is it expected format?

[aizatsky, 2016/03/07 21:42:54]

Number of spaces is irrelevant. The format is :

<file_name> : <dep1> <dep2> ....

+              corpus_files.append(full_filename)
+
+  with zipfile.ZipFile(args.output, 'w') as z:
+    for corpus_file in corpus_files:
+        z.write(corpus_file, os.path.basename(corpus_file))
+
+
+if __name__ == '__main__':
+  main()
+