2 # Copyright (C) 2020-2021 all contributors <meta@public-inbox.org>
3 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
7 use PublicInbox::TestCommon;
9 use Fcntl qw(SEEK_SET);
10 use PublicInbox::Spawn qw(popen_rd which);
11 use List::Util qw(shuffle);
12 require_mods(qw(DBD::SQLite));
13 require PublicInbox::MboxReader;
14 require PublicInbox::LeiOverview;
15 require PublicInbox::LEI;
16 use_ok 'PublicInbox::LeiToMail';
17 my $from = "Content-Length: 10\nSubject: x\n\nFrom hell\n";
18 my $noeol = "Subject: x\n\nFrom hell";
21 my $kw = [qw(seen answered flagged)];
22 my $smsg = { kw => $kw, blob => '0'x40 };
23 my @MBOX = qw(mboxcl2 mboxrd mboxcl mboxo);
24 for my $mbox (@MBOX) {
26 my $cb = PublicInbox::LeiToMail->can($m);
27 my $s = $cb->(PublicInbox::Eml->new($from), $smsg);
28 is(substr($$s, -1, 1), "\n", "trailing LF in normal $mbox");
29 my $eml = PublicInbox::Eml->new($s);
30 is($eml->header('Status'), 'OR', "Status: set by $m");
31 is($eml->header('X-Status'), 'AF', "X-Status: set by $m");
32 if ($mbox eq 'mboxcl2') {
33 like($eml->body_raw, qr/^From /, "From not escaped $m");
35 like($eml->body_raw, qr/^>From /, "From escaped once by $m");
37 my @cl = $eml->header('Content-Length');
38 if ($mbox =~ /mboxcl/) {
39 is(scalar(@cl), 1, "$m only has one Content-Length header");
40 is($cl[0] + length("\n"),
41 length($eml->body_raw), "$m Content-Length matches");
43 is(scalar(@cl), 0, "$m clobbered Content-Length");
45 $s = $cb->(PublicInbox::Eml->new($noeol), $smsg);
46 is(substr($$s, -1, 1), "\n",
47 "trailing LF added by $m when original lacks EOL");
48 $eml = PublicInbox::Eml->new($s);
49 if ($mbox eq 'mboxcl2') {
50 is($eml->body_raw, "From hell\n", "From not escaped by $m");
52 is($eml->body_raw, ">From hell\n", "From escaped once by $m");
54 $s = $cb->(PublicInbox::Eml->new($crlf), $smsg);
55 is(substr($$s, -2, 2), "\r\n",
56 "trailing CRLF added $m by original lacks EOL");
57 $eml = PublicInbox::Eml->new($s);
58 if ($mbox eq 'mboxcl2') {
59 is($eml->body_raw, "From hell\r\n", "From not escaped by $m");
61 is($eml->body_raw, ">From hell\r\n", "From escaped once by $m");
63 if ($mbox =~ /mboxcl/) {
64 is($eml->header('Content-Length') + length("\r\n"),
65 length($eml->body_raw), "$m Content-Length matches");
66 } elsif ($mbox eq 'mboxrd') {
67 $s = $cb->($eml, $smsg);
68 $eml = PublicInbox::Eml->new($s);
70 ">>From hell\r\n\r\n", "From escaped again by $m");
74 my ($tmpdir, $for_destroy) = tmpdir();
75 local $ENV{TMPDIR} = $tmpdir;
76 open my $err, '>>', "$tmpdir/lei.err" or BAIL_OUT $!;
77 my $lei = bless { 2 => $err }, 'PublicInbox::LEI';
88 my $fn = "$tmpdir/x.mbox";
89 my ($mbox) = shuffle(@MBOX); # pick one, shouldn't matter
92 delete $lei->{dedupe};
96 }, 'PublicInbox::LeiOverview';
97 my $l2m = PublicInbox::LeiToMail->new($lei);
99 require_mods('Storable', 1);
100 my $dup = Storable::thaw(Storable::freeze($l2m));
101 is_deeply($dup, $l2m, "$fmt round-trips through storable");
103 my $zpipe = $l2m->pre_augment($lei);
104 $l2m->do_augment($lei);
105 $l2m->post_augment($lei, $zpipe);
106 $l2m->write_cb($lei);
109 my $deadbeef = { blob => 'deadbeef', kw => [ qw(seen) ] };
111 my $wcb = $wcb_get->($mbox, $fn);
112 is(ref $wcb, 'CODE', 'write_cb returned callback');
113 ok(-f $fn && !-s _, 'empty file created');
114 $wcb->(\(my $dup = $buf), $deadbeef);
116 open my $fh, '<', $fn or BAIL_OUT $!;
117 my $raw = do { local $/; <$fh> };
118 like($raw, qr/^blah\n/sm, 'wrote content');
119 unlink $fn or BAIL_OUT $!;
121 local $lei->{opt} = { jobs => 2 };
122 $wcb = $wcb_get->($mbox, $fn);
123 ok(-f $fn && !-s _, 'truncated mbox destination');
124 $wcb->(\($dup = $buf), $deadbeef);
126 open $fh, '<', $fn or BAIL_OUT $!;
127 is(do { local $/; <$fh> }, $raw, 'jobs > 1');
130 for my $zsfx (qw(gz bz2 xz)) { # XXX should we support zst, zz, lzo, lzma?
131 my $zsfx2cmd = PublicInbox::LeiToMail->can('zsfx2cmd');
133 my $cmd = eval { $zsfx2cmd->($zsfx, 0, $lei) };
135 my $dc_cmd = eval { $zsfx2cmd->($zsfx, 1, $lei) };
136 ok($dc_cmd, "decompressor for .$zsfx");
138 my $wcb = $wcb_get->($mbox, $f);
139 $wcb->(\(my $dup = $buf), $deadbeef);
141 my $uncompressed = xqx([@$dc_cmd, $f]);
142 is($uncompressed, $orig, "$zsfx works unlocked");
144 local $lei->{opt} = { jobs => 2 }; # for atomic writes
145 unlink $f or BAIL_OUT "unlink $!";
146 $wcb = $wcb_get->($mbox, $f);
147 $wcb->(\($dup = $buf), $deadbeef);
149 is(xqx([@$dc_cmd, $f]), $orig, "$zsfx matches with lock");
151 local $lei->{opt} = { augment => 1 };
152 $wcb = $wcb_get->($mbox, $f);
153 $wcb->(\($dup = $buf . "\nx\n"), $deadbeef);
156 my $cat = popen_rd([@$dc_cmd, $f]);
158 PublicInbox::MboxReader->$mbox($cat,
159 sub { push @raw, shift->as_string });
160 like($raw[1], qr/\nblah\n\nx\n\z/s, "augmented $zsfx");
161 like($raw[0], qr/\nblah\n\z/s, "original preserved $zsfx");
163 local $lei->{opt} = { augment => 1, jobs => 2 };
164 $wcb = $wcb_get->($mbox, $f);
165 $wcb->(\($dup = $buf . "\ny\n"), $deadbeef);
169 $cat = popen_rd([@$dc_cmd, $f]);
170 PublicInbox::MboxReader->$mbox($cat,
171 sub { push @raw3, shift->as_string });
173 is_deeply(\@raw3, \@raw, 'previous messages preserved');
174 like($y, qr/\nblah\n\ny\n\z/s, "augmented $zsfx (atomic)");
180 $eml->header_set('Status');
184 unlink $fn or BAIL_OUT $!;
185 if ('default deduplication uses content_hash') {
186 my $wcb = $wcb_get->('mboxo', $fn);
187 $deadbeef->{kw} = [];
188 $wcb->(\(my $x = $buf), $deadbeef) for (1..2);
191 open my $fh, '<', $fn or BAIL_OUT $!;
192 PublicInbox::MboxReader->mboxo($fh, sub { $cmp .= $as_orig->(@_) });
193 is($cmp, $buf, 'only one message written');
195 local $lei->{opt} = { augment => 1 };
196 $wcb = $wcb_get->('mboxo', $fn);
197 $wcb->(\($x = $buf . "\nx\n"), $deadbeef) for (1..2);
199 open $fh, '<', $fn or BAIL_OUT $!;
201 PublicInbox::MboxReader->mboxo($fh, sub { push @x, $as_orig->(@_) });
202 is(scalar(@x), 2, 'augmented mboxo');
203 is($x[0], $cmp, 'original message preserved');
204 is($x[1], $buf . "\nx\n", 'new message appended');
208 open my $tmp, '+>', undef or BAIL_OUT $!;
209 local $lei->{1} = $tmp;
210 my $wcb = $wcb_get->('mboxrd', '/dev/stdout');
211 $wcb->(\(my $x = $buf), $deadbeef);
213 seek($tmp, 0, SEEK_SET) or BAIL_OUT $!;
215 PublicInbox::MboxReader->mboxrd($tmp, sub { $cmp .= $as_orig->(@_) });
216 is($cmp, $buf, 'message written to stdout');
219 SKIP: { # FIFO support
220 use POSIX qw(mkfifo);
221 my $fn = "$tmpdir/fifo";
222 mkfifo($fn, 0600) or skip("mkfifo not supported: $!", 1);
223 my $cat = popen_rd([which('cat'), $fn]);
224 my $wcb = $wcb_get->('mboxo', $fn);
225 $wcb->(\(my $x = $buf), $deadbeef);
228 PublicInbox::MboxReader->mboxo($cat, sub { $cmp .= $as_orig->(@_) });
229 is($cmp, $buf, 'message written to FIFO');
233 my $md = "$tmpdir/maildir/";
234 my $wcb = $wcb_get->('maildir', $md);
235 is(ref($wcb), 'CODE', 'got Maildir callback');
236 my $b4dc0ffee = { blob => 'badc0ffee', kw => [] };
237 $wcb->(\(my $x = $buf), $b4dc0ffee);
240 PublicInbox::LeiToMail::_maildir_each_file($md, sub { push @f, shift });
241 open my $fh, $f[0] or BAIL_OUT $!;
242 is(do { local $/; <$fh> }, $buf, 'wrote to Maildir');
244 $wcb = $wcb_get->('maildir', $md);
245 my $deadcafe = { blob => 'deadcafe', kw => [] };
246 $wcb->(\($x = $buf."\nx\n"), $deadcafe);
249 PublicInbox::LeiToMail::_maildir_each_file($md, sub { push @x, shift });
250 is(scalar(@x), 1, 'wrote one new file');
251 ok(!-f $f[0], 'old file clobbered');
252 open $fh, $x[0] or BAIL_OUT $!;
253 is(do { local $/; <$fh> }, $buf."\nx\n", 'wrote new file to Maildir');
255 local $lei->{opt}->{augment} = 1;
256 $wcb = $wcb_get->('maildir', $md);
257 $wcb->(\($x = $buf."\ny\n"), $deadcafe);
258 $wcb->(\($x = $buf."\ny\n"), $b4dc0ffee); # skipped by dedupe
260 PublicInbox::LeiToMail::_maildir_each_file($md, sub { push @f, shift });
261 is(scalar grep(/\A\Q$x[0]\E\z/, @f), 1, 'old file still there');
262 my @new = grep(!/\A\Q$x[0]\E\z/, @f);
263 is(scalar @new, 1, '1 new file written (b4dc0ffee skipped)');
264 open $fh, $x[0] or BAIL_OUT $!;
265 is(do { local $/; <$fh> }, $buf."\nx\n", 'old file untouched');
266 open $fh, $new[0] or BAIL_OUT $!;
267 is(do { local $/; <$fh> }, $buf."\ny\n", 'new file written');