1#!/usr/local/bin/perl
2
3# List people who might be interested in a patch.  Useful as the argument to
4# git-send-email --cc-cmd option, and in other situations.
5#
6# Usage: git contacts <file | rev-list option> ...
7
8use strict;
9use warnings;
10use IPC::Open2;
11
12my $since = '5-years-ago';
13my $min_percent = 10;
14my $labels_rx = qr/Signed-off-by|Reviewed-by|Acked-by|Cc|Reported-by/i;
15my %seen;
16
17sub format_contact {
18	my ($name, $email) = @_;
19	return "$name <$email>";
20}
21
22sub parse_commit {
23	my ($commit, $data) = @_;
24	my $contacts = $commit->{contacts};
25	my $inbody = 0;
26	for (split(/^/m, $data)) {
27		if (not $inbody) {
28			if (/^author ([^<>]+) <(\S+)> .+$/) {
29				$contacts->{format_contact($1, $2)} = 1;
30			} elsif (/^$/) {
31				$inbody = 1;
32			}
33		} elsif (/^$labels_rx:\s+([^<>]+)\s+<(\S+?)>$/o) {
34			$contacts->{format_contact($1, $2)} = 1;
35		}
36	}
37}
38
39sub import_commits {
40	my ($commits) = @_;
41	return unless %$commits;
42	my $pid = open2 my $reader, my $writer, qw(git cat-file --batch);
43	for my $id (keys(%$commits)) {
44		print $writer "$id\n";
45		my $line = <$reader>;
46		if ($line =~ /^([0-9a-f]{40}) commit (\d+)/) {
47			my ($cid, $len) = ($1, $2);
48			die "expected $id but got $cid\n" unless $id eq $cid;
49			my $data;
50			# cat-file emits newline after data, so read len+1
51			read $reader, $data, $len + 1;
52			parse_commit($commits->{$id}, $data);
53		}
54	}
55	close $reader;
56	close $writer;
57	waitpid($pid, 0);
58	die "git-cat-file error: $?\n" if $?;
59}
60
61sub get_blame {
62	my ($commits, $source, $from, $ranges) = @_;
63	return unless @$ranges;
64	open my $f, '-|',
65		qw(git blame --porcelain -C),
66		map({"-L$_->[0],+$_->[1]"} @$ranges),
67		'--since', $since, "$from^", '--', $source or die;
68	while (<$f>) {
69		if (/^([0-9a-f]{40}) \d+ \d+ \d+$/) {
70			my $id = $1;
71			$commits->{$id} = { id => $id, contacts => {} }
72				unless $seen{$id};
73			$seen{$id} = 1;
74		}
75	}
76	close $f;
77}
78
79sub blame_sources {
80	my ($sources, $commits) = @_;
81	for my $s (keys %$sources) {
82		for my $id (keys %{$sources->{$s}}) {
83			get_blame($commits, $s, $id, $sources->{$s}{$id});
84		}
85	}
86}
87
88sub scan_patches {
89	my ($sources, $id, $f) = @_;
90	my $source;
91	while (<$f>) {
92		if (/^From ([0-9a-f]{40}) Mon Sep 17 00:00:00 2001$/) {
93			$id = $1;
94			$seen{$id} = 1;
95		}
96		next unless $id;
97		if (m{^--- (?:a/(.+)|/dev/null)$}) {
98			$source = $1;
99		} elsif (/^@@ -(\d+)(?:,(\d+))?/ && $source) {
100			my $len = defined($2) ? $2 : 1;
101			push @{$sources->{$source}{$id}}, [$1, $len] if $len;
102		}
103	}
104}
105
106sub scan_patch_file {
107	my ($commits, $file) = @_;
108	open my $f, '<', $file or die "read failure: $file: $!\n";
109	scan_patches($commits, undef, $f);
110	close $f;
111}
112
113sub parse_rev_args {
114	my @args = @_;
115	open my $f, '-|',
116		qw(git rev-parse --revs-only --default HEAD --symbolic), @args
117		or die;
118	my @revs;
119	while (<$f>) {
120		chomp;
121		push @revs, $_;
122	}
123	close $f;
124	return @revs if scalar(@revs) != 1;
125	return "^$revs[0]", 'HEAD' unless $revs[0] =~ /^-/;
126	return $revs[0], 'HEAD';
127}
128
129sub scan_rev_args {
130	my ($commits, $args) = @_;
131	my @revs = parse_rev_args(@$args);
132	open my $f, '-|', qw(git rev-list --reverse), @revs or die;
133	while (<$f>) {
134		chomp;
135		my $id = $_;
136		$seen{$id} = 1;
137		open my $g, '-|', qw(git show -C --oneline), $id or die;
138		scan_patches($commits, $id, $g);
139		close $g;
140	}
141	close $f;
142}
143
144sub mailmap_contacts {
145	my ($contacts) = @_;
146	my %mapped;
147	my $pid = open2 my $reader, my $writer, qw(git check-mailmap --stdin);
148	for my $contact (keys(%$contacts)) {
149		print $writer "$contact\n";
150		my $canonical = <$reader>;
151		chomp $canonical;
152		$mapped{$canonical} += $contacts->{$contact};
153	}
154	close $reader;
155	close $writer;
156	waitpid($pid, 0);
157	die "git-check-mailmap error: $?\n" if $?;
158	return \%mapped;
159}
160
161if (!@ARGV) {
162	die "No input revisions or patch files\n";
163}
164
165my (@files, @rev_args);
166for (@ARGV) {
167	if (-e) {
168		push @files, $_;
169	} else {
170		push @rev_args, $_;
171	}
172}
173
174my %sources;
175for (@files) {
176	scan_patch_file(\%sources, $_);
177}
178if (@rev_args) {
179	scan_rev_args(\%sources, \@rev_args)
180}
181
182my $toplevel = `git rev-parse --show-toplevel`;
183chomp $toplevel;
184chdir($toplevel) or die "chdir failure: $toplevel: $!\n";
185
186my %commits;
187blame_sources(\%sources, \%commits);
188import_commits(\%commits);
189
190my $contacts = {};
191for my $commit (values %commits) {
192	for my $contact (keys %{$commit->{contacts}}) {
193		$contacts->{$contact}++;
194	}
195}
196$contacts = mailmap_contacts($contacts);
197
198my $ncommits = scalar(keys %commits);
199for my $contact (keys %$contacts) {
200	my $percent = $contacts->{$contact} * 100 / $ncommits;
201	next if $percent < $min_percent;
202	print "$contact\n";
203}
204