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 use_ok 'PublicInbox::LeiToMail';
16 my $from = "Content-Length: 10\nSubject: x\n\nFrom hell\n";
17 my $noeol = "Subject: x\n\nFrom hell";
20 my $kw = [qw(seen answered flagged)];
21 my @MBOX = qw(mboxcl2 mboxrd mboxcl mboxo);
22 for my $mbox (@MBOX) {
24 my $cb = PublicInbox::LeiToMail->can($m);
25 my $s = $cb->(PublicInbox::Eml->new($from), $kw);
26 is(substr($$s, -1, 1), "\n", "trailing LF in normal $mbox");
27 my $eml = PublicInbox::Eml->new($s);
28 is($eml->header('Status'), 'OR', "Status: set by $m");
29 is($eml->header('X-Status'), 'AF', "X-Status: set by $m");
30 if ($mbox eq 'mboxcl2') {
31 like($eml->body_raw, qr/^From /, "From not escaped $m");
33 like($eml->body_raw, qr/^>From /, "From escaped once by $m");
35 my @cl = $eml->header('Content-Length');
36 if ($mbox =~ /mboxcl/) {
37 is(scalar(@cl), 1, "$m only has one Content-Length header");
38 is($cl[0] + length("\n"),
39 length($eml->body_raw), "$m Content-Length matches");
41 is(scalar(@cl), 0, "$m clobbered Content-Length");
43 $s = $cb->(PublicInbox::Eml->new($noeol), $kw);
44 is(substr($$s, -1, 1), "\n",
45 "trailing LF added by $m when original lacks EOL");
46 $eml = PublicInbox::Eml->new($s);
47 if ($mbox eq 'mboxcl2') {
48 is($eml->body_raw, "From hell\n", "From not escaped by $m");
50 is($eml->body_raw, ">From hell\n", "From escaped once by $m");
52 $s = $cb->(PublicInbox::Eml->new($crlf), $kw);
53 is(substr($$s, -2, 2), "\r\n",
54 "trailing CRLF added $m by original lacks EOL");
55 $eml = PublicInbox::Eml->new($s);
56 if ($mbox eq 'mboxcl2') {
57 is($eml->body_raw, "From hell\r\n", "From not escaped by $m");
59 is($eml->body_raw, ">From hell\r\n", "From escaped once by $m");
61 if ($mbox =~ /mboxcl/) {
62 is($eml->header('Content-Length') + length("\r\n"),
63 length($eml->body_raw), "$m Content-Length matches");
64 } elsif ($mbox eq 'mboxrd') {
65 $s = $cb->($eml, $kw);
66 $eml = PublicInbox::Eml->new($s);
68 ">>From hell\r\n\r\n", "From escaped again by $m");
72 my ($tmpdir, $for_destroy) = tmpdir();
73 local $ENV{TMPDIR} = $tmpdir;
74 open my $err, '>>', "$tmpdir/lei.err" or BAIL_OUT $!;
75 my $lei = { 2 => $err };
82 my $fn = "$tmpdir/x.mbox";
83 my ($mbox) = shuffle(@MBOX); # pick one, shouldn't matter
86 delete $lei->{dedupe};
90 }, 'PublicInbox::LeiOverview';
91 my $l2m = PublicInbox::LeiToMail->new($lei);
93 require_mods('Storable', 1);
94 my $dup = Storable::thaw(Storable::freeze($l2m));
95 is_deeply($dup, $l2m, "$fmt round-trips through storable");
97 my $zpipe = $l2m->pre_augment($lei);
98 $l2m->do_augment($lei);
99 $l2m->post_augment($lei, $zpipe);
100 my $cb = $l2m->write_cb($lei);
106 my $wcb = $wcb_get->($mbox, $fn);
107 is(ref $wcb, 'CODE', 'write_cb returned callback');
108 ok(-f $fn && !-s _, 'empty file created');
109 $wcb->(\(my $dup = $buf), 'deadbeef', [ qw(seen) ]);
111 open my $fh, '<', $fn or BAIL_OUT $!;
112 my $raw = do { local $/; <$fh> };
113 like($raw, qr/^blah\n/sm, 'wrote content');
114 unlink $fn or BAIL_OUT $!;
116 local $lei->{opt} = { jobs => 2 };
117 $wcb = $wcb_get->($mbox, $fn);
118 ok(-f $fn && !-s _, 'truncated mbox destination');
119 $wcb->(\($dup = $buf), 'deadbeef', [ qw(seen) ]);
121 open $fh, '<', $fn or BAIL_OUT $!;
122 is(do { local $/; <$fh> }, $raw, 'jobs > 1');
125 for my $zsfx (qw(gz bz2 xz)) { # XXX should we support zst, zz, lzo, lzma?
126 my $zsfx2cmd = PublicInbox::LeiToMail->can('zsfx2cmd');
128 my $cmd = eval { $zsfx2cmd->($zsfx, 0, $lei) };
130 my $dc_cmd = eval { $zsfx2cmd->($zsfx, 1, $lei) };
131 ok($dc_cmd, "decompressor for .$zsfx");
133 my $wcb = $wcb_get->($mbox, $f);
134 $wcb->(\(my $dup = $buf), 'deadbeef', [ qw(seen) ]);
136 my $uncompressed = xqx([@$dc_cmd, $f]);
137 is($uncompressed, $orig, "$zsfx works unlocked");
139 local $lei->{opt} = { jobs => 2 }; # for atomic writes
140 unlink $f or BAIL_OUT "unlink $!";
141 $wcb = $wcb_get->($mbox, $f);
142 $wcb->(\($dup = $buf), 'deadbeef', [ qw(seen) ]);
144 is(xqx([@$dc_cmd, $f]), $orig, "$zsfx matches with lock");
146 local $lei->{opt} = { augment => 1 };
147 $wcb = $wcb_get->($mbox, $f);
148 $wcb->(\($dup = $buf . "\nx\n"), 'deadbeef', [ qw(seen) ]);
151 my $cat = popen_rd([@$dc_cmd, $f]);
153 PublicInbox::MboxReader->$mbox($cat,
154 sub { push @raw, shift->as_string });
155 like($raw[1], qr/\nblah\n\nx\n\z/s, "augmented $zsfx");
156 like($raw[0], qr/\nblah\n\z/s, "original preserved $zsfx");
158 local $lei->{opt} = { augment => 1, jobs => 2 };
159 $wcb = $wcb_get->($mbox, $f);
160 $wcb->(\($dup = $buf . "\ny\n"), 'deadbeef', [ qw(seen) ]);
164 $cat = popen_rd([@$dc_cmd, $f]);
165 PublicInbox::MboxReader->$mbox($cat,
166 sub { push @raw3, shift->as_string });
168 is_deeply(\@raw3, \@raw, 'previous messages preserved');
169 like($y, qr/\nblah\n\ny\n\z/s, "augmented $zsfx (atomic)");
175 $eml->header_set('Status');
179 unlink $fn or BAIL_OUT $!;
180 if ('default deduplication uses content_hash') {
181 my $wcb = $wcb_get->('mboxo', $fn);
182 $wcb->(\(my $x = $buf), 'deadbeef', []) for (1..2);
183 undef $wcb; # undef to commit changes
185 open my $fh, '<', $fn or BAIL_OUT $!;
186 PublicInbox::MboxReader->mboxo($fh, sub { $cmp .= $as_orig->(@_) });
187 is($cmp, $buf, 'only one message written');
189 local $lei->{opt} = { augment => 1 };
190 $wcb = $wcb_get->('mboxo', $fn);
191 $wcb->(\($x = $buf . "\nx\n"), 'deadbeef', []) for (1..2);
192 undef $wcb; # undef to commit changes
193 open $fh, '<', $fn or BAIL_OUT $!;
195 PublicInbox::MboxReader->mboxo($fh, sub { push @x, $as_orig->(@_) });
196 is(scalar(@x), 2, 'augmented mboxo');
197 is($x[0], $cmp, 'original message preserved');
198 is($x[1], $buf . "\nx\n", 'new message appended');
202 open my $tmp, '+>', undef or BAIL_OUT $!;
203 local $lei->{1} = $tmp;
204 my $wcb = $wcb_get->('mboxrd', '/dev/stdout');
205 $wcb->(\(my $x = $buf), 'deadbeef', []);
207 seek($tmp, 0, SEEK_SET) or BAIL_OUT $!;
209 PublicInbox::MboxReader->mboxrd($tmp, sub { $cmp .= $as_orig->(@_) });
210 is($cmp, $buf, 'message written to stdout');
213 SKIP: { # FIFO support
214 use POSIX qw(mkfifo);
215 my $fn = "$tmpdir/fifo";
216 mkfifo($fn, 0600) or skip("mkfifo not supported: $!", 1);
217 my $cat = popen_rd([which('cat'), $fn]);
218 my $wcb = $wcb_get->('mboxo', $fn);
219 $wcb->(\(my $x = $buf), 'deadbeef', []);
222 PublicInbox::MboxReader->mboxo($cat, sub { $cmp .= $as_orig->(@_) });
223 is($cmp, $buf, 'message written to FIFO');
227 my $md = "$tmpdir/maildir/";
228 my $wcb = $wcb_get->('maildir', $md);
229 is(ref($wcb), 'CODE', 'got Maildir callback');
230 $wcb->(\(my $x = $buf), 'badc0ffee', []);
233 PublicInbox::LeiToMail::_maildir_each_file($md, sub { push @f, shift });
234 open my $fh, $f[0] or BAIL_OUT $!;
235 is(do { local $/; <$fh> }, $buf, 'wrote to Maildir');
237 $wcb = $wcb_get->('maildir', $md);
238 $wcb->(\($x = $buf."\nx\n"), 'deadcafe', []);
241 PublicInbox::LeiToMail::_maildir_each_file($md, sub { push @x, shift });
242 is(scalar(@x), 1, 'wrote one new file');
243 ok(!-f $f[0], 'old file clobbered');
244 open $fh, $x[0] or BAIL_OUT $!;
245 is(do { local $/; <$fh> }, $buf."\nx\n", 'wrote new file to Maildir');
247 local $lei->{opt}->{augment} = 1;
248 $wcb = $wcb_get->('maildir', $md);
249 $wcb->(\($x = $buf."\ny\n"), 'deadcafe', []);
250 $wcb->(\($x = $buf."\ny\n"), 'b4dc0ffee', []); # skipped by dedupe
252 PublicInbox::LeiToMail::_maildir_each_file($md, sub { push @f, shift });
253 is(scalar grep(/\A\Q$x[0]\E\z/, @f), 1, 'old file still there');
254 my @new = grep(!/\A\Q$x[0]\E\z/, @f);
255 is(scalar @new, 1, '1 new file written (b4dc0ffee skipped)');
256 open $fh, $x[0] or BAIL_OUT $!;
257 is(do { local $/; <$fh> }, $buf."\nx\n", 'old file untouched');
258 open $fh, $new[0] or BAIL_OUT $!;
259 is(do { local $/; <$fh> }, $buf."\ny\n", 'new file written');