1 # Copyright (C) 2019-2021 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
7 use PublicInbox::InboxWritable;
8 use PublicInbox::TestCommon;
10 require_git(2.6); # replace is v2 only, for now...
11 require_mods(qw(DBD::SQLite));
12 local $ENV{HOME} = abs_path('t');
14 sub test_replace ($$$) {
15 my ($v, $level, $opt) = @_;
16 diag "v$v $level replace";
17 my $this = "pi-$v-$level-replace";
18 my ($tmpdir, $for_destroy) = tmpdir($this);
19 my $ibx = PublicInbox::Inbox->new({
20 inboxdir => "$tmpdir/testbox",
24 -primary_address => 'test@example.com',
28 my $orig = PublicInbox::Eml->new(<<'EOF');
29 From: Barbra Streisand <effect@example.com>
32 Message-ID: <replace@example.com>
33 Date: Fri, 02 Oct 1993 00:00:00 +0000
35 Top secret info about my house in Malibu...
37 my $im = PublicInbox::InboxWritable->new($ibx, {nproc=>1})->importer(0);
38 # fake a bunch of epochs
39 $im->{rotate_bytes} = $opt->{rotate_bytes} if $opt->{rotate_bytes};
42 $opt->{pre}->($im, 1, 2);
43 $orig->header_set('References', '<1@example.com>');
45 ok($im->add($orig), 'add message to be replaced');
47 $opt->{post}->($im, 3, { 4 => 'replace@example.com' });
50 my $thread_a = $ibx->over->get_thread('replace@example.com');
52 my %before = map {; delete($_->{blob}) => $_ } @{$ibx->recent};
53 my $reject = PublicInbox::Eml->new($orig->as_string);
54 foreach my $mid (['<replace@example.com>', '<extra@example.com>'],
55 [], ['<replaced@example.com>']) {
56 $reject->header_set('Message-ID', @$mid);
57 my $ok = eval { $im->replace($orig, $reject) };
58 like($@, qr/Message-ID.*may not be changed/,
59 '->replace died on Message-ID change');
60 ok(!$ok, 'no replacement happened');
63 # prepare the replacement
64 my $expect = "Move along, nothing to see here\n";
65 my $repl = PublicInbox::Eml->new($orig->as_string);
66 $repl->header_set('From', '<redactor@example.com>');
67 $repl->header_set('Subject', 'redacted');
68 $repl->header_set('Date', 'Sat, 02 Oct 2010 00:00:00 +0000');
69 $repl->body_str_set($expect);
72 local $SIG{__WARN__} = sub { push @warn, @_ };
73 ok(my $cmts = $im->replace($orig, $repl), 'replaced message');
74 my $changed_epochs = 0;
75 for my $tip (@$cmts) {
76 next if !defined $tip;
78 like($tip, qr/\A[a-f0-9]{40,}\z/,
79 'replace returned current commit');
81 is($changed_epochs, 1, 'only one epoch changed');
84 my $m = PublicInbox::Eml->new($ibx->msg_by_mid('replace@example.com'));
85 is($m->body, $expect, 'replaced message');
86 is_deeply(\@warn, [], 'no warnings on noop');
88 my @cat = qw(cat-file --buffer --batch --batch-all-objects);
90 my @all = $git->qx(@cat);
91 is_deeply([grep(/confidential/, @all)], [], 'nothing confidential');
92 is_deeply([grep(/Streisand/, @all)], [], 'Streisand who?');
93 is_deeply([grep(/\bOct 1993\b/, @all)], [], 'nothing from Oct 1993');
94 my $t19931002 = qr/ 749520000 /;
95 is_deeply([grep(/$t19931002/, @all)], [], "nothing matches $t19931002");
97 for my $dir (glob("$ibx->{inboxdir}/git/*.git")) {
98 my ($bn) = ($dir =~ m!([^/]+)\z!);
99 is(xsys(qw(git --git-dir), $dir,
100 qw(fsck --strict --no-progress)),
101 0, "git fsck is clean in epoch $bn");
104 my $thread_b = $ibx->over->get_thread('replace@example.com');
105 is_deeply([sort map { $_->{mid} } @$thread_b],
106 [sort map { $_->{mid} } @$thread_a], 'threading preserved');
108 if (my $srch = $ibx->search) {
109 for my $q ('f:streisand', 's:confidential', 'malibu') {
110 my $mset = $srch->mset($q);
111 is($mset->size, 0, "no match for $q");
113 my @ok = ('f:redactor', 's:redacted', 'nothing to see');
115 push @ok, 'm:1@example.com', 'm:2@example.com',
116 's:message2', 's:message1';
119 push @ok, 'm:3@example.com', 'm:4@example.com',
120 's:message3', 's:message4';
123 my $mset = $srch->mset($q);
124 ok($mset->size, "got match for $q");
128 # check overview matches:
129 my %after = map {; delete($_->{blob}) => $_ } @{$ibx->recent};
130 my @before_blobs = keys %before;
131 foreach my $blob (@before_blobs) {
132 delete $before{$blob} if delete $after{$blob};
135 is(scalar keys %before, 1, 'one unique blob from before left');
136 is(scalar keys %after, 1, 'one unique blob from after left');
137 foreach my $blob (keys %before) {
138 is($git->check($blob), undef, 'old blob not found');
139 my $smsg = $before{$blob};
140 is($smsg->{subject}, 'confidential', 'before subject');
141 is($smsg->{mid}, 'replace@example.com', 'before MID');
143 foreach my $blob (keys %after) {
144 ok($git->check($blob), 'new blob found');
145 my $smsg = $after{$blob};
146 is($smsg->{subject}, 'redacted', 'after subject');
147 is($smsg->{mid}, 'replace@example.com', 'before MID');
149 # $git->cleanup; # needed if $im->{parallel};
151 is($im->replace($orig, $repl), undef, 'no-op replace returns undef');
152 is($im->purge($orig), undef, 'no-op purge returns undef');
153 is_deeply(\@warn, [], 'no warnings on noop');
154 # $im->done; # needed if $im->{parallel}
158 my ($im, @range) = @_;
161 if (ref($i) eq 'HASH') {
162 ($i, $irt) = each %$i;
164 my $sec = sprintf('%0d', $i);
165 my $mime = PublicInbox::Eml->new(<<EOF);
166 From: foo\@example.com
167 To: test\@example.com
168 Message-ID: <$i\@example.com>
169 Date: Fri, 02, Jan 1970 00:00:$sec +0000
176 $mime->header_set('References', "<$irt>");
183 my $opt = { pre => \&pad_msgs };
184 test_replace(2, 'basic', {});
185 test_replace(2, 'basic', $opt);
186 test_replace(2, 'basic', $opt = { %$opt, post => \&pad_msgs });
187 test_replace(2, 'basic', $opt = { %$opt, rotate_bytes => 1 });
190 require_mods(qw(Search::Xapian), 8);
191 for my $l (qw(medium)) {
192 test_replace(2, $l, {});
193 $opt = { pre => \&pad_msgs };
194 test_replace(2, $l, $opt);
195 test_replace(2, $l, $opt = { %$opt, post => \&pad_msgs });
196 test_replace(2, $l, $opt = { %$opt, rotate_bytes => 1 });