scripts/mailmapper

   1 #!/usr/bin/env python2
   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     sys.exit('Please run in a git repository.')
  63
  64 # strip '\n'
  65 toplevel = toplevel.rstrip()
  66
  67 # Change the current working directory to the toplevel of the respository
  68 # for our easier life.
  69 os.chdir(toplevel)
  70
  71 # First, create 'auther name' vs 'number of commits' database.
  72 # We assume the name with the most commits as the canonical real name.
  73 shortlog = subprocess.check_output(['git', 'shortlog', '-s', '-n'])
  74
  75 commits_per_name = {}
  76
  77 for line in shortlog.splitlines():
  78     try:
  79         commits, name = line.split(None, 1)
  80     except ValueError:
  81         # ignore lines with an empty author name
  82         pass
  83     commits_per_name[name] = int(commits)
  84
  85 # Next, coalesce the auther names with the same email address
  86 shortlog = subprocess.check_output(['git', 'shortlog', '-s', '-n', '-e'])
  87
  88 mail_vs_name = {}
  89 output = {}
  90
  91 for line in shortlog.splitlines():
  92     # tmp, mail = line.rsplit(None, 1) is not safe
  93     # because weird email addresses might include whitespaces
  94     tmp, mail = line.split('<')
  95     mail = '<' + mail.rstrip()
  96     try:
  97         _, name = tmp.rstrip().split(None, 1)
  98     except ValueError:
  99         # author name is empty
 100         name = ''
 101     if mail in mail_vs_name:
 102         # another name for the same email address
 103         prev_name = mail_vs_name[mail]
 104         # Take the name with more commits
 105         major_name = sorted([prev_name, name],
 106                             key=lambda x: commits_per_name[x] if x else 0)[1]
 107         mail_vs_name[mail] = major_name
 108         if commits_per_name[major_name] > MIN_COMMITS:
 109             output[mail] = major_name
 110     else:
 111         mail_vs_name[mail] = name
 112
 113 # [1] If there exists a mailmap file at the location pointed to
 114 #     by the mailmap.file configuration option, update it.
 115 # [2] If the file .mailmap exists at the toplevel of the repository, update it.
 116 # [3] Otherwise, create a new mailmap file.
 117 mailmap_files = []
 118
 119 try:
 120     config_mailmap = subprocess.check_output(['git', 'config', 'mailmap.file'])
 121 except subprocess.CalledProcessError:
 122     config_mailmap = ''
 123
 124 config_mailmap = config_mailmap.rstrip()
 125 if config_mailmap:
 126     mailmap_files.append(config_mailmap)
 127
 128 mailmap_files.append('.mailmap')
 129
 130 infile = None
 131
 132 for map_file in mailmap_files:
 133     try:
 134         infile = open(map_file)
 135     except:
 136         # Failed to open. Try next.
 137         continue
 138     break
 139
 140 comment_block = []
 141 output_lines = []
 142
 143 if infile:
 144     for line in infile:
 145         if line[0] == '#' or line[0] == '\n':
 146             comment_block.append(line)
 147         else:
 148             output_lines.append(line)
 149             break
 150     for line in infile:
 151         output_lines.append(line)
 152     infile.close()
 153
 154 for mail, name in output.items():
 155     output_lines.append(name + ' ' + mail + '\n')
 156
 157 output_lines.sort()
 158
 159 sys.stdout.write(''.join(comment_block + output_lines))