scripts/mailmapper

   1 #!/usr/bin/env python
   2 #
   3 # Copyright (C) 2014, Masahiro Yamada <yamada.m@jp.panasonic.com>
   4 #
   5 # SPDX-License-Identifier:      GPL-2.0+
   6 #
   7
   8 '''
   9 A tool to create/update the mailmap file
  10
  11 The command 'git shortlog' summarizes git log output in a format suitable
  12 for inclusion in release announcements. Each commit will be grouped by
  13 author and title.
  14
  15 One problem is that the authors' name and/or email address is sometimes
  16 spelled differently. The .mailmap feature can be used to coalesce together
  17 commits by the same persion.
  18 (See 'man git-shortlog' for furthur information of this feature.)
  19
  20 This tool helps to create/update the mailmap file.
  21
  22 It runs 'git shortlog' internally and searches differently spelled author
  23 names which share the same email address. The author name with the most
  24 commits is asuumed to be a canonical real name. If the number of commits
  25 from the cananonical name is equal to or greater than 'MIN_COMMITS',
  26 the entry for the cananical name will be output. ('MIN_COMMITS' is used
  27 here because we do not want to create a fat mailmap by adding every author
  28 with only a few commits.)
  29
  30 If there exists a mailmap file specified by the mailmap.file configuration
  31 options or '.mailmap' at the toplevel of the repository, it is used as
  32 a base file. (The mailmap.file configuration takes precedence over the
  33 '.mailmap' file if both exist.)
  34
  35 The base file and the newly added entries are merged together and sorted
  36 alphabetically (but the comment block is kept untouched), and then printed
  37 to standard output.
  38
  39 Usage
  40 -----
  41
  42   scripts/mailmapper
  43
  44 prints the mailmapping to standard output.
  45
  46   scripts/mailmapper > tmp; mv tmp .mailmap
  47
  48 will be useful for updating '.mailmap' file.
  49 '''
  50
  51 import sys
  52 import os
  53 import subprocess
  54
  55 # The entries only for the canonical names with MIN_COMMITS or more commits.
  56 # This limitation is used so as not to create a too big mailmap file.
  57 MIN_COMMITS = 50
  58
  59 try:
  60     toplevel = subprocess.check_output(['git', 'rev-parse', '--show-toplevel'])
  61 except subprocess.CalledProcessError:
  62     print >> sys.stderr, 'Please run in a git repository.'
  63     sys.exit(1)
  64
  65 # strip '\n'
  66 toplevel = toplevel.rstrip()
  67
  68 # Change the current working directory to the toplevel of the respository
  69 # for our easier life.
  70 os.chdir(toplevel)
  71
  72 # First, create 'auther name' vs 'number of commits' database.
  73 # We assume the name with the most commits as the canonical real name.
  74 shortlog = subprocess.check_output(['git', 'shortlog', '-s', '-n'])
  75
  76 commits_per_name = {}
  77
  78 for line in shortlog.splitlines():
  79     try:
  80         commits, name = line.split(None, 1)
  81     except ValueError:
  82         # ignore lines with an empty author name
  83         pass
  84     commits_per_name[name] = int(commits)
  85
  86 # Next, coalesce the auther names with the same email address
  87 shortlog = subprocess.check_output(['git', 'shortlog', '-s', '-n', '-e'])
  88
  89 mail_vs_name = {}
  90 output = {}
  91
  92 for line in shortlog.splitlines():
  93     # tmp, mail = line.rsplit(None, 1) is not safe
  94     # because weird email addresses might include whitespaces
  95     tmp, mail = line.split('<')
  96     mail = '<' + mail.rstrip()
  97     try:
  98         _, name = tmp.rstrip().split(None, 1)
  99     except ValueError:
 100         # author name is empty
 101         name = ''
 102     if mail in mail_vs_name:
 103         # another name for the same email address
 104         prev_name = mail_vs_name[mail]
 105         # Take the name with more commits
 106         major_name = sorted([prev_name, name],
 107                             key=lambda x: commits_per_name[x] if x else 0)[1]
 108         mail_vs_name[mail] = major_name
 109         if commits_per_name[major_name] > MIN_COMMITS:
 110             output[mail] = major_name
 111     else:
 112         mail_vs_name[mail] = name
 113
 114 # [1] If there exists a mailmap file at the location pointed to
 115 #     by the mailmap.file configuration option, update it.
 116 # [2] If the file .mailmap exists at the toplevel of the repository, update it.
 117 # [3] Otherwise, create a new mailmap file.
 118 mailmap_files = []
 119
 120 try:
 121     config_mailmap = subprocess.check_output(['git', 'config', 'mailmap.file'])
 122 except subprocess.CalledProcessError:
 123     config_mailmap = ''
 124
 125 config_mailmap = config_mailmap.rstrip()
 126 if config_mailmap:
 127     mailmap_files.append(config_mailmap)
 128
 129 mailmap_files.append('.mailmap')
 130
 131 infile = None
 132
 133 for map_file in mailmap_files:
 134     try:
 135         infile = open(map_file)
 136     except:
 137         # Failed to open. Try next.
 138         continue
 139     break
 140
 141 comment_block = []
 142 output_lines = []
 143
 144 if infile:
 145     for line in infile:
 146         if line[0] == '#' or line[0] == '\n':
 147             comment_block.append(line)
 148         else:
 149             output_lines.append(line)
 150             break
 151     for line in infile:
 152         output_lines.append(line)
 153     infile.close()
 154
 155 for mail, name in output.items():
 156     output_lines.append(name + ' ' + mail + '\n')
 157
 158 output_lines.sort()
 159
 160 sys.stdout.write(''.join(comment_block + output_lines))