1#!/usr/bin/env python
2# Copyright 2014 the V8 project authors. All rights reserved.
3# Use of this source code is governed by a BSD-style license that can be
4# found in the LICENSE file.
5
6import argparse
7import subprocess
8import sys
9
10
11def GetArgs():
12  parser = argparse.ArgumentParser(
13      description="Finds a commit that a given patch can be applied to. "
14                  "Does not actually apply the patch or modify your checkout "
15                  "in any way.")
16  parser.add_argument("patch_file", help="Patch file to match")
17  parser.add_argument(
18      "--branch", "-b", default="origin/master", type=str,
19      help="Git tree-ish where to start searching for commits, "
20           "default: %(default)s")
21  parser.add_argument(
22      "--limit", "-l", default=500, type=int,
23      help="Maximum number of commits to search, default: %(default)s")
24  parser.add_argument(
25      "--verbose", "-v", default=False, action="store_true",
26      help="Print verbose output for your entertainment")
27  return parser.parse_args()
28
29
30def FindFilesInPatch(patch_file):
31  files = {}
32  next_file = ""
33  with open(patch_file) as patch:
34    for line in patch:
35      if line.startswith("diff --git "):
36        # diff --git a/src/objects.cc b/src/objects.cc
37        words = line.split()
38        assert words[2].startswith("a/") and len(words[2]) > 2
39        next_file = words[2][2:]
40      elif line.startswith("index "):
41        # index add3e61..d1bbf6a 100644
42        hashes = line.split()[1]
43        old_hash = hashes.split("..")[0]
44        if old_hash.startswith("0000000"): continue  # Ignore new files.
45        files[next_file] = old_hash
46  return files
47
48
49def GetGitCommitHash(treeish):
50  cmd = ["git", "log", "-1", "--format=%H", treeish]
51  return subprocess.check_output(cmd).strip()
52
53
54def CountMatchingFiles(commit, files):
55  matched_files = 0
56  # Calling out to git once and parsing the result Python-side is faster
57  # than calling 'git ls-tree' for every file.
58  cmd = ["git", "ls-tree", "-r", commit] + [f for f in files]
59  output = subprocess.check_output(cmd)
60  for line in output.splitlines():
61    # 100644 blob c6d5daaa7d42e49a653f9861224aad0a0244b944      src/objects.cc
62    _, _, actual_hash, filename = line.split()
63    expected_hash = files[filename]
64    if actual_hash.startswith(expected_hash): matched_files += 1
65  return matched_files
66
67
68def FindFirstMatchingCommit(start, files, limit, verbose):
69  commit = GetGitCommitHash(start)
70  num_files = len(files)
71  if verbose: print(">>> Found %d files modified by patch." % num_files)
72  for _ in range(limit):
73    matched_files = CountMatchingFiles(commit, files)
74    if verbose: print("Commit %s matched %d files" % (commit, matched_files))
75    if matched_files == num_files:
76      return commit
77    commit = GetGitCommitHash("%s^" % commit)
78  print("Sorry, no matching commit found. "
79        "Try running 'git fetch', specifying the correct --branch, "
80        "and/or setting a higher --limit.")
81  sys.exit(1)
82
83
84if __name__ == "__main__":
85  args = GetArgs()
86  files = FindFilesInPatch(args.patch_file)
87  commit = FindFirstMatchingCommit(args.branch, files, args.limit, args.verbose)
88  if args.verbose:
89    print(">>> Matching commit: %s" % commit)
90    print(subprocess.check_output(["git", "log", "-1", commit]))
91    print(">>> Kthxbai.")
92  else:
93    print(commit)
94