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>
6 use PublicInbox::TestCommon;
7 use PublicInbox::MsgIter qw(msg_part_text);
8 my @classes = qw(PublicInbox::Eml);
10 require_mods('Email::MIME', 1);
11 # TODO: Email::MIME behavior is not consistent in newer versions
12 # we need to evaluate and possibly adjust our behavior to decide
13 # between DWIM-ness with historical mail...
14 push @classes, 'PublicInbox::MIME';
16 use_ok $_ for @classes;
20 open(my $fh, '<', $path) or die "open $path: $!";
21 PublicInbox::MIME->new(\(do { local $/; <$fh> }));
25 my $eml = PublicInbox::Eml->new(\(my $str = "a: b\n\nhi\n"));
26 is($str, "hi\n", '->new modified body like Email::Simple');
27 is($eml->body, "hi\n", '->body works');
28 is($eml->as_string, "a: b\n\nhi\n", '->as_string');
31 for my $cls (@classes) {
32 my $mime = $cls->new(my $orig = "From: x\n\nb");
33 is($mime->as_string, $orig, '->as_string works');
34 is($mime->header_obj->as_string, "From: x\n",
35 'header ->as_string works');
38 is($mime->header_raw('From'), 'x', 'header_raw scalar context');
39 $mime = $cls->new("R:\n\tx\nR:\n 1\n");
40 is_deeply([$mime->header_raw('r')], [ 'x', '1' ], 'multi-value');
41 $mime = $cls->new("R:x\nR: 1\n");
42 is_deeply([$mime->header_raw('r')], [ 'x', '1' ], 'multi-value header');
43 $mime = $cls->new("R:x\n R: 1\nR:\n f\n");
44 is_deeply([$mime->header_raw('r')], [ 'x R: 1', 'f' ],
45 'multi-line, multi-value header');
47 $mime->header_set('r');
48 is_deeply([$mime->header_raw('r')], [], 'header_set clears');
49 $mime->header_set('r');
50 is_deeply([$mime->header_raw('r')], [], 'header_set clears idempotent');
51 $mime->header_set('r', 'h');
52 is_deeply([$mime->header_raw('r')], ['h'], 'header_set');
53 $mime->header_set('r', 'h', 'i');
54 is_deeply([$mime->header_raw('r')], ['h', 'i'], 'header_set ary');
55 $mime->header_set('rr', 'b');
56 is_deeply([$mime->header_raw('r')], ['h', 'i'],
57 "header_set `rr' did not clobber `r'");
58 is($mime->header_raw('rr'), 'b', 'got set scalar');
59 $mime->header_set('rr', 'b'x100);
60 is($mime->header_raw('rr'), 'b'x100, 'got long set scalar');
61 if ($cls eq 'PublicInbox::Eml') {
62 like($mime->as_string, qr/^rr: b{100}\n(?:\n|\z)/sm,
63 'single token not wrapped');
65 $mime->header_set('rr', ('b'x100) . ' wrap me');
66 if ($cls eq 'PublicInbox::Eml') {
67 like($mime->as_string, qr/^rr: b{100}\n\twrap me\n/sm,
68 'wrapped after long token');
70 my $exp = "pre\tformatted\n with\n breaks";
71 $mime->header_set('r', $exp);
72 like($mime->as_string, qr/^r: \Q$exp\E/sm, 'preformatted preserved');
75 for my $cls (@classes) { # make sure we don't add quotes if not needed
76 my $eml = $cls->new("From: John Smith <j\@example.com>\n\n");
77 is($eml->header('From'), 'John Smith <j@example.com>',
78 "name not unnecessarily quoted $cls");
81 for my $cls (@classes) {
82 my $eml = $cls->new("Subject: foo\n\n");
83 $eml->header_str_set('Subject', "\x{100}");
84 like($eml->header_raw('Subject'), qr/utf-8\?B\?/i,
85 'MIME-B encoded UTF-8 Subject');
86 is_deeply([$eml->header('Subject')], [ "\x{100}" ],
87 'got wide character back');
90 # linux-mips apparently got some messages injected w/o Message-ID
91 # and long Subject: lines w/o leading whitespace.
92 # What appears in the blobs was generated by V2Writable.
93 for my $cls (@classes) {
94 my $eml = $cls->new(<<'EOF');
95 Message-ID: <20101130193431@z>
96 Subject: something really long
98 From: linux-mips archive injection
99 Object-Id: 8c56b7abdd551b1264e6522ededbbed9890cccd0
101 is_deeply([ $eml->header('Subject') ],
102 [ 'something really long and really wrong' ],
103 'continued long line w/o leading spaces '.$cls);
104 is_deeply([ $eml->header('From') ],
105 [ 'linux-mips archive injection' ],
106 'subsequent line not corrupted');
107 is_deeply([ $eml->header('Message-ID') ],
108 ['<20101130193431@z>'],
109 'preceding line readable');
113 my $eml = eml_load 't/msg_iter-order.eml';
115 my $orig = $eml->as_string;
116 $eml->each_part(sub {
117 my ($part, $level, @ex) = @{$_[0]};
118 my $s = $part->body_str;
120 push @parts, [ $s, $level, @ex ];
122 is_deeply(\@parts, [ [ qw(a 1 1) ], [ qw(b 1 2) ] ], 'order is fine');
123 is($eml->as_string, $orig, 'unchanged by ->each_part');
124 $eml->each_part(sub {}, undef, 1);
125 is(defined($eml) ? $eml->body_raw : '', # old msg_iter clobbers $eml
126 '', 'each_part can clobber body');
129 if ('descend into message/rfc822') {
130 my $eml = eml_load 't/data/message_embed.eml';
132 $eml->each_part(sub {
133 my ($part, $level, @ex) = @{$_[0]};
134 push @parts, [ $part, $level, @ex ];
136 is(scalar(@parts), 6, 'got all parts');
137 like($parts[0]->[0]->body, qr/^testing embedded message harder\n/sm,
139 is_deeply([ @{$parts[0]}[1..2] ], [ 1, '1' ],
140 'got expected depth and level for part #0');
141 is($parts[1]->[0]->filename, 'embed2x.eml',
142 'attachment filename found');
143 is_deeply([ @{$parts[1]}[1..2] ], [ 1, '2' ],
144 'got expected depth and level for part #1');
145 is_deeply([ @{$parts[2]}[1..2] ], [ 2, '2.1' ],
146 'got expected depth and level for part #2');
147 is_deeply([ @{$parts[3]}[1..2] ], [ 3, '2.1.1' ],
148 'got expected depth and level for part #3');
149 is_deeply([ @{$parts[4]}[1..2] ], [ 3, '2.1.2' ],
150 'got expected depth and level for part #4');
151 is($parts[4]->[0]->filename, 'test.eml',
152 'another attachment filename found');
153 is_deeply([ @{$parts[5]}[1..2] ], [ 4, '2.1.2.1' ],
154 'got expected depth and level for part #5');
157 # body-less, boundary-less
158 for my $cls (@classes) {
160 $cls->new(<<'EOF')->each_part(sub { $call++ }, 0, 1);
161 Content-Type: multipart/mixed; boundary="body-less"
164 is($call, 1, 'called on bodyless multipart');
167 $cls->new(<<'EOF')->each_part(sub { push @tmp, \@_; }, 0, 1);
168 Content-Type: multipart/mixed; boundary="boundary-less"
172 is(scalar(@tmp), 1, 'got one part even w/o boundary');
173 is($tmp[0]->[0]->[0]->body, "hello world\n", 'body preserved');
174 is($tmp[0]->[0]->[1], 0, '$depth is zero');
175 is($tmp[0]->[0]->[2], 1, '@idx is one');
178 # I guess the following only worked in PI::M because of a happy accident
179 # involving inheritance:
180 for my $cls (@classes) {
182 my $header_less = <<'EOF';
183 Archived-At: <85k5su9k59.fsf_-_@lola.goethe.zz>
184 Content-Type: multipart/mixed; boundary="header-less"
197 my $expect = "this is the body\n";
198 $cls->new($header_less)->each_part(sub { push @tmp, \@_ }, 0, 1);
199 my $body = $tmp[0]->[0]->[0]->body;
200 if ($cls eq 'PublicInbox::Eml') {
201 is($body, $expect, 'body-only subpart in '.$cls);
202 } elsif ($body ne $expect) {
203 diag "W: $cls `$body' != `$expect'";
205 is($tmp[1]->[0]->[0]->body, "something else\n");
206 is(scalar(@tmp), 2, 'two parts');
209 if ('one newline before headers') {
210 my $eml = PublicInbox::Eml->new("\nNewline: no Header \n");
211 my @v = $eml->header_raw('Newline');
212 is_deeply(\@v, ['no Header'], 'no header');
213 is($eml->crlf, "\n", 'got CRLF as "\n"');
217 for my $cls (@classes) { # XXX: matching E::M, but not sure about this
219 Content-Type: multipart/mixed; boundary="b"
225 my $eml = $cls->new(\$s);
228 $eml->each_part(sub {
229 @v = $_[0]->[0]->header_raw('Header');
232 is($nr, 1, 'only one part');
233 is_deeply(\@v, [], "nothing w/o body $cls");
236 for my $cls (@classes) {
237 my $s = <<EOF; # double epilogue, double the fun
238 Content-Type: multipart/mixed; boundary="b"
253 my $eml = $cls->new(\$s);
255 $eml->each_part(sub {
256 my $part = $_[0]->[0];
257 is_deeply([$part->header_raw('should')], ['appear'],
258 'only got one header');
259 is($part->body, "yes\n", 'got expected body');
262 is($nr, 1, 'only one part');
265 for my $cls (@classes) {
267 skip 'newer Email::MIME behavior inconsistent', 1 if
268 $cls eq 'PublicInbox::MIME';
269 my $s = <<EOF; # buggy git-send-email versions, again?
270 Content-Type: text/plain; =?ISO-8859-1?Q?=20charset=3D=1BOF?=
271 Content-Transfer-Encoding: 8bit
272 Object-Id: ab0440d8cd6d843bee9a27709a459ce3b2bdb94d (lore/kvm)
276 my $eml = $cls->new(\$s);
277 my ($str, $err) = msg_part_text($eml, $eml->content_type);
278 is($str, "\x{100}\n", "got wide character by assuming utf-8 ($cls)");
282 if ('we differ from Email::MIME with final "\n" on missing epilogue') {
284 Content-Type: multipart/mixed; boundary="b"
291 my $eml = PublicInbox::Eml->new(\$s);
292 is(($eml->subparts)[-1]->body, "no epilogue\n",
293 'final "\n" preserved on missing epilogue');
296 if ('header_size_limit stolen from postfix') {
297 local $PublicInbox::Eml::header_size_limit = 4;
299 local $SIG{__WARN__} = sub { push @w, @_ };
300 my $eml = PublicInbox::Eml->new("a:b\na:d\n\nzz");
301 is_deeply([$eml->header('a')], ['b'], 'no overrun header');
302 is($eml->body_raw, 'zz', 'body not damaged');
303 is($eml->header_obj->as_string, "a:b\n", 'header truncated');
304 is(grep(/truncated/, @w), 1, 'truncation warned');
306 $eml = PublicInbox::Eml->new("a:b\na:d\n");
307 is_deeply([$eml->header('a')], ['b'], 'no overrun header w/o body');
309 local $PublicInbox::Eml::header_size_limit = 5;
310 $eml = PublicInbox::Eml->new("a:b\r\na:d\r\n\nzz");
311 is_deeply([$eml->header('a')], ['b'], 'no overrun header on CRLF');
312 is($eml->body_raw, 'zz', 'body not damaged');
315 $eml = PublicInbox::Eml->new("too:long\n");
316 $eml = PublicInbox::Eml->new("too:long\n\n");
317 $eml = PublicInbox::Eml->new("too:long\r\n\r\n");
318 is(grep(/ignored/, @w), 3, 'ignored header warned');
321 if ('maxparts is a feature unique to us') {
322 my $eml = eml_load 't/psgi_attach.eml';
324 $eml->each_part(sub { push @orig, $_[0]->[0] });
326 local $PublicInbox::Eml::mime_parts_limit = scalar(@orig);
328 $eml->each_part(sub {
329 my $cur = $_[0]->[0];
330 my $prv = $orig[$i++];
331 is($cur->body_raw, $prv->body_raw, "part #$i matches");
333 is($i, scalar(@orig), 'maxparts honored');
334 $PublicInbox::Eml::mime_parts_limit--;
336 $eml->each_part(sub { push @ltd, $_[0]->[0] });
337 for ($i = 0; $i <= $#ltd; $i++) {
338 is($ltd[$i]->body_raw, $orig[$i]->body_raw,
341 is(scalar(@ltd), scalar(@orig) - 1, 'maxparts honored');
345 require_mods('PublicInbox::MIME', 1);
346 my $eml = eml_load 't/utf8.eml';
347 my $mime = mime_load 't/utf8.eml';
348 for my $h (qw(Subject From To)) {
349 my $v = $eml->header($h);
350 my $m = $mime->header($h);
351 is($v, $m, "decoded -8 $h matches Email::MIME");
352 ok(utf8::is_utf8($v), "$h is UTF-8");
353 ok(utf8::valid($v), "UTF-8 valid $h");
355 my $s = $eml->body_str;
356 ok(utf8::is_utf8($s), 'body_str is UTF-8');
357 ok(utf8::valid($s), 'UTF-8 valid body_str');
358 my $ref = \(my $x = 'ref');
359 for my $msg ($eml, $mime) {
360 $msg->body_str_set($s .= "\nHI\n");
361 ok(!utf8::is_utf8($msg->body_raw),
362 'raw octets after body_str_set');
364 ok(utf8::is_utf8($s), 'body_str is UTF-8 after set');
365 ok(utf8::valid($s), 'UTF-8 valid body_str after set');
366 $msg->body_set($ref);
367 is($msg->body_raw, $$ref, 'body_set worked on scalar ref');
368 $msg->body_set($$ref);
369 is($msg->body_raw, $$ref, 'body_set worked on scalar');
371 $eml = eml_load 't/iso-2202-jp.eml';
372 $mime = mime_load 't/iso-2202-jp.eml';
374 is($s, $mime->body_str, 'ISO-2202-JP body_str');
375 ok(utf8::is_utf8($s), 'ISO-2202-JP => UTF-8 body_str');
376 ok(utf8::valid($s), 'UTF-8 valid body_str');
378 $eml = eml_load 't/psgi_attach.eml';
379 $mime = mime_load 't/psgi_attach.eml';
380 is_deeply([ map { $_->body_raw } $eml->subparts ],
381 [ map { $_->body_raw } $mime->subparts ],
382 'raw ->subparts match deeply');
383 is_deeply([ map { $_->body } $eml->subparts ],
384 [ map { $_->body } $mime->subparts ],
385 '->subparts match deeply');
386 for my $msg ($eml, $mime) {
387 my @old = $msg->subparts;
389 is_deeply([$msg->subparts], [], 'parts_set can clear');
390 $msg->parts_set([$old[-1]]);
391 is(scalar $msg->subparts, 1, 'only last remains');
394 # some versions of Email::MIME or Email::MIME::* will drop
395 # unnecessary ", while PublicInbox::Eml will preserve the original
396 my $exp = $mime->as_string;
397 $exp =~ s/; boundary=b\b/; boundary="b"/;
398 is($eml->as_string, $exp, 'as_string matches after parts_set');
401 for my $cls (@classes) {
403 Content-Type: text/x-patch; name="=?utf-8?q?vtpm-fakefile.patch?="
404 Content-Disposition: attachment; filename="=?utf-8?q?vtpm-makefile.patch?="
407 is($cls->new($s)->filename, 'vtpm-makefile.patch', 'filename decoded');
408 $s =~ s/^Content-Disposition:.*$//sm;
409 is($cls->new($s)->filename, 'vtpm-fakefile.patch',
410 "filename fallback ($cls)") if $cls ne 'PublicInbox::MIME';
411 is($cls->new($s)->content_type,
412 'text/x-patch; name="vtpm-fakefile.patch"',
413 'matches Email::MIME output, "correct" or not');
416 Content-Type: multipart/foo; boundary=b
419 Content-Disposition: attachment; filename="=?utf-8?q?vtpm-makefile.patch?="
423 Content-Type: text/x-patch; name="=?utf-8?q?vtpm-fakefile.patch?="
429 skip 'newer Email::MIME is inconsistent here', 1
430 if $cls eq 'PublicInbox::MIME';
432 $cls->new($s)->each_part(sub { push @x, $_[0]->[0]->filename });
433 is_deeply(['vtpm-makefile.patch', 'vtpm-fakefile.patch'], \@x,
434 "got filename for both attachments ($cls)");