1 # Copyright (C) 2018-2020 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
6 use PublicInbox::TestCommon;
7 use File::Path qw(remove_tree);
10 local $ENV{HOME} = abs_path('t');
12 # Integration tests for HTTP cloning + mirroring
13 require_mods(qw(Plack::Util Plack::Builder
14 HTTP::Date HTTP::Status Search::Xapian DBD::SQLite));
17 use_ok 'PublicInbox::V2Writable';
18 use PublicInbox::InboxWritable;
20 use PublicInbox::Config;
21 # FIXME: too much setup
22 my ($tmpdir, $for_destroy) = tmpdir();
23 my $pi_config = "$tmpdir/config";
25 open my $fh, '>', $pi_config or die "open($pi_config): $!";
26 print $fh <<"" or die "print $pi_config: $!";
29 address = test\@example.com
31 close $fh or die "close($pi_config): $!";
33 local $ENV{PI_CONFIG} = $pi_config;
35 my $cfg = PublicInbox::Config->new($pi_config);
36 my $ibx = $cfg->lookup('test@example.com');
37 ok($ibx, 'inbox found');
39 my $v2w = PublicInbox::V2Writable->new($ibx, 1);
40 ok $v2w, 'v2w loaded';
42 my $mime = PublicInbox::Eml->new(<<'');
43 From: Me <me@example.com>
44 To: You <you@example.com>
46 Date: Thu, 01 Jan 1970 00:00:00 +0000
48 my $old_rotate_bytes = $v2w->{rotate_bytes};
49 $v2w->{rotate_bytes} = 500; # force rotating
51 $mime->header_set('Message-ID', "<$i\@example.com>");
52 $mime->header_set('Subject', "subject = $i");
53 ok($v2w->add($mime), "add msg $i OK");
56 my $epoch_max = $v2w->{epoch_max};
57 ok($epoch_max > 0, "multiple epochs");
60 my $smsg = $ibx->over->get_art(1);
61 like($smsg->{lines}, qr/\A[0-9]+\z/, 'lines is a digit');
62 like($smsg->{bytes}, qr/\A[0-9]+\z/, 'bytes is a digit');
66 my $sock = tcp_server();
67 ok($sock, 'sock created');
68 my $cmd = [ '-httpd', '-W0', "--stdout=$tmpdir/out", "--stderr=$tmpdir/err" ];
69 my $td = start_script($cmd, undef, { 3 => $sock });
70 my ($host, $port) = ($sock->sockhost, $sock->sockport);
74 foreach my $i (0..$epoch_max) {
75 my $sfx = $i == 0 ? '.git' : '';
76 @cmd = (qw(git clone --mirror -q),
77 "http://$host:$port/v2/$i$sfx",
78 "$tmpdir/m/git/$i.git");
80 is(xsys(@cmd), 0, "cloned $i.git");
81 ok(-d "$tmpdir/m/git/$i.git", "mirror $i OK");
84 @cmd = ("-init", '-j1', '-V2', 'm', "$tmpdir/m", 'http://example.com/m',
86 ok(run_script(\@cmd), 'initialized public-inbox -V2');
87 my @shards = glob("$tmpdir/m/xap*/?");
88 is(scalar(@shards), 1, 'got a single shard on init');
90 ok(run_script([qw(-index -j0), "$tmpdir/m"]), 'indexed');
92 my $mibx = { inboxdir => "$tmpdir/m", address => 'alt@example.com' };
93 $mibx = PublicInbox::Inbox->new($mibx);
94 is_deeply([$mibx->mm->minmax], [$ibx->mm->minmax], 'index synched minmax');
96 $v2w->{rotate_bytes} = $old_rotate_bytes;
98 $mime->header_set('Message-ID', "<$i\@example.com>");
99 $mime->header_set('Subject', "subject = $i");
100 ok($v2w->add($mime), "add msg $i OK");
105 my $fetch_each_epoch = sub {
106 foreach my $i (0..$epoch_max) {
107 my $dir = "$tmpdir/m/git/$i.git";
108 is(xsys('git', "--git-dir=$dir", 'fetch', '-q'), 0,
113 $fetch_each_epoch->();
115 my $mset = $mibx->search->reopen->mset('m:15@example.com');
116 is(scalar($mset->items), 0, 'new message not found in mirror, yet');
117 ok(run_script([qw(-index -j0), "$tmpdir/m"]), 'index updated');
118 is_deeply([$mibx->mm->minmax], [$ibx->mm->minmax], 'index synched minmax');
119 $mset = $mibx->search->reopen->mset('m:15@example.com');
120 is(scalar($mset->items), 1, 'found message in mirror');
123 $mime->header_set('Message-ID', '<10@example.com>');
124 $mime->header_set('Subject', 'subject = 10');
127 local $SIG{__WARN__} = sub { push @warn, @_ };
128 ok($v2w->purge($mime), 'purge a message');
129 my $warn = join('', @warn);
130 like($warn, qr/purge rewriting/);
131 my @subj = ($warn =~ m/^# subject .*$/mg);
132 is_deeply(\@subj, ["# subject = 10"], "only rewrote one");
137 my $msgs = $mibx->over->get_thread('10@example.com');
138 my $to_purge = $msgs->[0]->{blob};
139 like($to_purge, qr/\A[a-f0-9]{40,}\z/, 'read blob to be purged');
140 $mset = $ibx->search->reopen->mset('m:10@example.com');
141 is(scalar($mset->items), 0, 'purged message gone from origin');
143 $fetch_each_epoch->();
146 PublicInbox::InboxWritable::cleanup($mibx);
148 my $cmd = [ qw(-index --prune -j0), "$tmpdir/m" ];
149 my ($out, $err) = ('', '');
150 my $opt = { 1 => \$out, 2 => \$err };
151 ok(run_script($cmd, undef, $opt), '-index --prune');
152 like($err, qr/discontiguous range/, 'warned about discontiguous range');
153 unlike($err, qr/fatal/, 'no scary fatal error shown');
156 $mset = $mibx->search->reopen->mset('m:10@example.com');
157 is(scalar($mset->items), 0, 'purged message not found in mirror');
158 is_deeply([$mibx->mm->minmax], [$ibx->mm->minmax], 'minmax still synced');
159 for my $i ((1..9),(11..15)) {
160 $mset = $mibx->search->mset("m:$i\@example.com");
161 is(scalar($mset->items), 1, "$i\@example.com remains visible");
163 is($mibx->git->check($to_purge), undef, 'unindex+prune successful in mirror');
167 local $SIG{__WARN__} = sub { push @warn, @_ };
169 is_deeply(\@warn, [], 'no warnings from index_sync after purge');
172 # deletes happen in a different fetch window
174 $mset = $mibx->search->reopen->mset('m:1@example.com');
175 is(scalar($mset->items), 1, '1@example.com visible in mirror');
176 $mime->header_set('Message-ID', '<1@example.com>');
177 $mime->header_set('Subject', 'subject = 1');
178 ok($v2w->remove($mime), 'removed <1@example.com> from source');
181 $fetch_each_epoch->();
182 PublicInbox::InboxWritable::cleanup($mibx);
184 my $cmd = [ qw(-index -j0), "$tmpdir/m" ];
185 my ($out, $err) = ('', '');
186 my $opt = { 1 => \$out, 2 => \$err };
187 ok(run_script($cmd, undef, $opt), 'index ran');
188 is($err, '', 'no errors reported by index');
189 $mset = $mibx->search->reopen->mset('m:1@example.com');
190 is(scalar($mset->items), 0, '1@example.com no longer visible in mirror');
193 if ('sequential-shard') {
194 $mset = $mibx->search->mset('m:15@example.com');
195 is(scalar($mset->items), 1, 'large message not indexed');
196 remove_tree(glob("$tmpdir/m/xap*"), glob("$tmpdir/m/msgmap.*"));
197 my $cmd = [ qw(-index -j9 --sequential-shard), "$tmpdir/m" ];
198 ok(run_script($cmd), '--sequential-shard works');
199 my @shards = glob("$tmpdir/m/xap*/?");
200 is(scalar(@shards), 8, 'got expected shard count');
201 PublicInbox::InboxWritable::cleanup($mibx);
202 $mset = $mibx->search->mset('m:15@example.com');
203 is(scalar($mset->items), 1, 'search works after --sequential-shard');
207 $mime->header_set('Message-ID', '<2big@a>');
209 $mime->body_str_set("z\n" x 1024);
210 ok($v2w->add($mime), "add big message");
213 $fetch_each_epoch->();
214 PublicInbox::InboxWritable::cleanup($mibx);
215 my $cmd = [qw(-index -j0), "$tmpdir/m", "--max-size=$max" ];
216 my $opt = { 2 => \(my $err) };
217 ok(run_script($cmd, undef, $opt), 'indexed with --max-size');
218 like($err, qr/skipping [a-f0-9]{40,}/, 'warned about skipping message');
219 $mset = $mibx->search->reopen->mset('m:2big@a');
220 is(scalar($mset->items), 0, 'large message not indexed');
223 open my $fh, '>>', $pi_config or die;
224 print $fh <<EOF or die;
230 $cmd = [ qw(-index -j0 --reindex), "$tmpdir/m" ];
231 ok(run_script($cmd, undef, $opt), 'reindexed w/ indexMaxSize in file');
232 like($err, qr/skipping [a-f0-9]{40,}/, 'warned about skipping message');
233 $mset = $mibx->search->reopen->mset('m:2big@a');
234 is(scalar($mset->items), 0, 'large message not re-indexed');
237 ok($td->kill, 'killed httpd');