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');
29 my $empty = PublicInbox::Eml->new("\n\n");
30 is($empty->as_string, "\n\n", 'empty message');
33 for my $cls (@classes) {
34 my $mime = $cls->new(my $orig = "From: x\n\nb");
35 is($mime->as_string, $orig, '->as_string works');
36 is($mime->header_obj->as_string, "From: x\n",
37 'header ->as_string works');
40 is($mime->header_raw('From'), 'x', 'header_raw scalar context');
41 $mime = $cls->new("R:\n\tx\nR:\n 1\n");
42 is_deeply([$mime->header_raw('r')], [ 'x', '1' ], 'multi-value');
43 $mime = $cls->new("R:x\nR: 1\n");
44 is_deeply([$mime->header_raw('r')], [ 'x', '1' ], 'multi-value header');
45 $mime = $cls->new("R:x\n R: 1\nR:\n f\n");
46 is_deeply([$mime->header_raw('r')], [ 'x R: 1', 'f' ],
47 'multi-line, multi-value header');
49 $mime->header_set('r');
50 is_deeply([$mime->header_raw('r')], [], 'header_set clears');
51 $mime->header_set('r');
52 is_deeply([$mime->header_raw('r')], [], 'header_set clears idempotent');
53 $mime->header_set('r', 'h');
54 is_deeply([$mime->header_raw('r')], ['h'], 'header_set');
55 $mime->header_set('r', 'h', 'i');
56 is_deeply([$mime->header_raw('r')], ['h', 'i'], 'header_set ary');
57 $mime->header_set('rr', 'b');
58 is_deeply([$mime->header_raw('r')], ['h', 'i'],
59 "header_set `rr' did not clobber `r'");
60 is($mime->header_raw('rr'), 'b', 'got set scalar');
61 $mime->header_set('rr', 'b'x100);
62 is($mime->header_raw('rr'), 'b'x100, 'got long set scalar');
63 if ($cls eq 'PublicInbox::Eml') {
64 like($mime->as_string, qr/^rr: b{100}\n(?:\n|\z)/sm,
65 'single token not wrapped');
67 $mime->header_set('rr', ('b'x100) . ' wrap me');
68 if ($cls eq 'PublicInbox::Eml') {
69 like($mime->as_string, qr/^rr: b{100}\n\twrap me\n/sm,
70 'wrapped after long token');
72 my $exp = "pre\tformatted\n with\n breaks";
73 $mime->header_set('r', $exp);
74 like($mime->as_string, qr/^r: \Q$exp\E/sm, 'preformatted preserved');
77 for my $cls (@classes) { # make sure we don't add quotes if not needed
78 my $eml = $cls->new("From: John Smith <j\@example.com>\n\n");
79 is($eml->header('From'), 'John Smith <j@example.com>',
80 "name not unnecessarily quoted $cls");
83 for my $cls (@classes) {
84 my $eml = $cls->new("Subject: foo\n\n");
85 $eml->header_str_set('Subject', "\x{100}");
86 like($eml->header_raw('Subject'), qr/utf-8\?B\?/i,
87 'MIME-B encoded UTF-8 Subject');
88 is_deeply([$eml->header('Subject')], [ "\x{100}" ],
89 'got wide character back');
92 # linux-mips apparently got some messages injected w/o Message-ID
93 # and long Subject: lines w/o leading whitespace.
94 # What appears in the blobs was generated by V2Writable.
95 for my $cls (@classes) {
96 my $eml = $cls->new(<<'EOF');
97 Message-ID: <20101130193431@z>
98 Subject: something really long
100 From: linux-mips archive injection
101 Object-Id: 8c56b7abdd551b1264e6522ededbbed9890cccd0
103 is_deeply([ $eml->header('Subject') ],
104 [ 'something really long and really wrong' ],
105 'continued long line w/o leading spaces '.$cls);
106 is_deeply([ $eml->header('From') ],
107 [ 'linux-mips archive injection' ],
108 'subsequent line not corrupted');
109 is_deeply([ $eml->header('Message-ID') ],
110 ['<20101130193431@z>'],
111 'preceding line readable');
115 my $eml = eml_load 't/msg_iter-order.eml';
117 my $orig = $eml->as_string;
118 $eml->each_part(sub {
119 my ($part, $level, @ex) = @{$_[0]};
120 my $s = $part->body_str;
122 push @parts, [ $s, $level, @ex ];
124 is_deeply(\@parts, [ [ qw(a 1 1) ], [ qw(b 1 2) ] ], 'order is fine');
125 is($eml->as_string, $orig, 'unchanged by ->each_part');
126 $eml->each_part(sub {}, undef, 1);
127 is(defined($eml) ? $eml->body_raw : '', # old msg_iter clobbers $eml
128 '', 'each_part can clobber body');
131 if ('descend into message/rfc822') {
132 my $eml = eml_load 't/data/message_embed.eml';
134 $eml->each_part(sub {
135 my ($part, $level, @ex) = @{$_[0]};
136 push @parts, [ $part, $level, @ex ];
138 is(scalar(@parts), 6, 'got all parts');
139 like($parts[0]->[0]->body, qr/^testing embedded message harder\n/sm,
141 is_deeply([ @{$parts[0]}[1..2] ], [ 1, '1' ],
142 'got expected depth and level for part #0');
143 is($parts[1]->[0]->filename, 'embed2x.eml',
144 'attachment filename found');
145 is_deeply([ @{$parts[1]}[1..2] ], [ 1, '2' ],
146 'got expected depth and level for part #1');
147 is_deeply([ @{$parts[2]}[1..2] ], [ 2, '2.1' ],
148 'got expected depth and level for part #2');
149 is_deeply([ @{$parts[3]}[1..2] ], [ 3, '2.1.1' ],
150 'got expected depth and level for part #3');
151 is_deeply([ @{$parts[4]}[1..2] ], [ 3, '2.1.2' ],
152 'got expected depth and level for part #4');
153 is($parts[4]->[0]->filename, 'test.eml',
154 'another attachment filename found');
155 is_deeply([ @{$parts[5]}[1..2] ], [ 4, '2.1.2.1' ],
156 'got expected depth and level for part #5');
159 # body-less, boundary-less
160 for my $cls (@classes) {
162 $cls->new(<<'EOF')->each_part(sub { $call++ }, 0, 1);
163 Content-Type: multipart/mixed; boundary="body-less"
166 is($call, 1, 'called on bodyless multipart');
169 $cls->new(<<'EOF')->each_part(sub { push @tmp, \@_; }, 0, 1);
170 Content-Type: multipart/mixed; boundary="boundary-less"
174 is(scalar(@tmp), 1, 'got one part even w/o boundary');
175 is($tmp[0]->[0]->[0]->body, "hello world\n", 'body preserved');
176 is($tmp[0]->[0]->[1], 0, '$depth is zero');
177 is($tmp[0]->[0]->[2], 1, '@idx is one');
180 # I guess the following only worked in PI::M because of a happy accident
181 # involving inheritance:
182 for my $cls (@classes) {
184 my $header_less = <<'EOF';
185 Archived-At: <85k5su9k59.fsf_-_@lola.goethe.zz>
186 Content-Type: multipart/mixed; boundary="header-less"
199 my $expect = "this is the body\n";
200 $cls->new($header_less)->each_part(sub { push @tmp, \@_ }, 0, 1);
201 my $body = $tmp[0]->[0]->[0]->body;
202 if ($cls eq 'PublicInbox::Eml') {
203 is($body, $expect, 'body-only subpart in '.$cls);
204 } elsif ($body ne $expect) {
205 diag "W: $cls `$body' != `$expect'";
207 is($tmp[1]->[0]->[0]->body, "something else\n");
208 is(scalar(@tmp), 2, 'two parts');
211 if ('one newline before headers') {
212 my $eml = PublicInbox::Eml->new("\nNewline: no Header \n");
213 my @v = $eml->header_raw('Newline');
214 is_deeply(\@v, ['no Header'], 'no header');
215 is($eml->crlf, "\n", 'got CRLF as "\n"');
219 for my $cls (@classes) { # XXX: matching E::M, but not sure about this
221 Content-Type: multipart/mixed; boundary="b"
227 my $eml = $cls->new(\$s);
230 $eml->each_part(sub {
231 @v = $_[0]->[0]->header_raw('Header');
234 is($nr, 1, 'only one part');
235 is_deeply(\@v, [], "nothing w/o body $cls");
238 for my $cls (@classes) {
239 my $s = <<EOF; # double epilogue, double the fun
240 Content-Type: multipart/mixed; boundary="b"
255 my $eml = $cls->new(\$s);
257 $eml->each_part(sub {
258 my $part = $_[0]->[0];
259 is_deeply([$part->header_raw('should')], ['appear'],
260 'only got one header');
261 is($part->body, "yes\n", 'got expected body');
264 is($nr, 1, 'only one part');
267 for my $cls (@classes) {
269 skip 'newer Email::MIME behavior inconsistent', 1 if
270 $cls eq 'PublicInbox::MIME';
271 my $s = <<EOF; # buggy git-send-email versions, again?
272 Content-Type: text/plain; =?ISO-8859-1?Q?=20charset=3D=1BOF?=
273 Content-Transfer-Encoding: 8bit
274 Object-Id: ab0440d8cd6d843bee9a27709a459ce3b2bdb94d (lore/kvm)
278 my $eml = $cls->new(\$s);
279 my ($str, $err) = msg_part_text($eml, $eml->content_type);
280 is($str, "\x{100}\n", "got wide character by assuming utf-8 ($cls)");
284 if ('we differ from Email::MIME with final "\n" on missing epilogue') {
286 Content-Type: multipart/mixed; boundary="b"
293 my $eml = PublicInbox::Eml->new(\$s);
294 is(($eml->subparts)[-1]->body, "no epilogue\n",
295 'final "\n" preserved on missing epilogue');
298 if ('header_size_limit stolen from postfix') {
299 local $PublicInbox::Eml::header_size_limit = 4;
301 local $SIG{__WARN__} = sub { push @w, @_ };
302 my $eml = PublicInbox::Eml->new("a:b\na:d\n\nzz");
303 is_deeply([$eml->header('a')], ['b'], 'no overrun header');
304 is($eml->body_raw, 'zz', 'body not damaged');
305 is($eml->header_obj->as_string, "a:b\n", 'header truncated');
306 is(grep(/truncated/, @w), 1, 'truncation warned');
308 $eml = PublicInbox::Eml->new("a:b\na:d\n");
309 is_deeply([$eml->header('a')], ['b'], 'no overrun header w/o body');
311 local $PublicInbox::Eml::header_size_limit = 5;
312 $eml = PublicInbox::Eml->new("a:b\r\na:d\r\n\nzz");
313 is_deeply([$eml->header('a')], ['b'], 'no overrun header on CRLF');
314 is($eml->body_raw, 'zz', 'body not damaged');
317 $eml = PublicInbox::Eml->new("too:long\n");
318 $eml = PublicInbox::Eml->new("too:long\n\n");
319 $eml = PublicInbox::Eml->new("too:long\r\n\r\n");
320 is(grep(/ignored/, @w), 3, 'ignored header warned');
323 if ('maxparts is a feature unique to us') {
324 my $eml = eml_load 't/psgi_attach.eml';
326 $eml->each_part(sub { push @orig, $_[0]->[0] });
328 local $PublicInbox::Eml::mime_parts_limit = scalar(@orig);
330 $eml->each_part(sub {
331 my $cur = $_[0]->[0];
332 my $prv = $orig[$i++];
333 is($cur->body_raw, $prv->body_raw, "part #$i matches");
335 is($i, scalar(@orig), 'maxparts honored');
336 $PublicInbox::Eml::mime_parts_limit--;
338 $eml->each_part(sub { push @ltd, $_[0]->[0] });
339 for ($i = 0; $i <= $#ltd; $i++) {
340 is($ltd[$i]->body_raw, $orig[$i]->body_raw,
343 is(scalar(@ltd), scalar(@orig) - 1, 'maxparts honored');
347 require_mods('PublicInbox::MIME', 1);
348 my $eml = eml_load 't/utf8.eml';
349 my $mime = mime_load 't/utf8.eml';
350 for my $h (qw(Subject From To)) {
351 my $v = $eml->header($h);
352 my $m = $mime->header($h);
353 is($v, $m, "decoded -8 $h matches Email::MIME");
354 ok(utf8::is_utf8($v), "$h is UTF-8");
355 ok(utf8::valid($v), "UTF-8 valid $h");
357 my $s = $eml->body_str;
358 ok(utf8::is_utf8($s), 'body_str is UTF-8');
359 ok(utf8::valid($s), 'UTF-8 valid body_str');
360 my $ref = \(my $x = 'ref');
361 for my $msg ($eml, $mime) {
362 $msg->body_str_set($s .= "\nHI\n");
363 ok(!utf8::is_utf8($msg->body_raw),
364 'raw octets after body_str_set');
366 ok(utf8::is_utf8($s), 'body_str is UTF-8 after set');
367 ok(utf8::valid($s), 'UTF-8 valid body_str after set');
368 $msg->body_set($ref);
369 is($msg->body_raw, $$ref, 'body_set worked on scalar ref');
370 $msg->body_set($$ref);
371 is($msg->body_raw, $$ref, 'body_set worked on scalar');
373 $eml = eml_load 't/iso-2202-jp.eml';
374 $mime = mime_load 't/iso-2202-jp.eml';
376 is($s, $mime->body_str, 'ISO-2202-JP body_str');
377 ok(utf8::is_utf8($s), 'ISO-2202-JP => UTF-8 body_str');
378 ok(utf8::valid($s), 'UTF-8 valid body_str');
380 $eml = eml_load 't/psgi_attach.eml';
381 $mime = mime_load 't/psgi_attach.eml';
382 is_deeply([ map { $_->body_raw } $eml->subparts ],
383 [ map { $_->body_raw } $mime->subparts ],
384 'raw ->subparts match deeply');
385 is_deeply([ map { $_->body } $eml->subparts ],
386 [ map { $_->body } $mime->subparts ],
387 '->subparts match deeply');
388 for my $msg ($eml, $mime) {
389 my @old = $msg->subparts;
391 is_deeply([$msg->subparts], [], 'parts_set can clear');
392 $msg->parts_set([$old[-1]]);
393 is(scalar $msg->subparts, 1, 'only last remains');
396 # some versions of Email::MIME or Email::MIME::* will drop
397 # unnecessary ", while PublicInbox::Eml will preserve the original
398 my $exp = $mime->as_string;
399 $exp =~ s/; boundary=b\b/; boundary="b"/;
400 is($eml->as_string, $exp, 'as_string matches after parts_set');
403 for my $cls (@classes) {
405 Content-Type: text/x-patch; name="=?utf-8?q?vtpm-fakefile.patch?="
406 Content-Disposition: attachment; filename="=?utf-8?q?vtpm-makefile.patch?="
409 is($cls->new($s)->filename, 'vtpm-makefile.patch', 'filename decoded');
410 $s =~ s/^Content-Disposition:.*$//sm;
411 is($cls->new($s)->filename, 'vtpm-fakefile.patch',
412 "filename fallback ($cls)") if $cls ne 'PublicInbox::MIME';
413 is($cls->new($s)->content_type,
414 'text/x-patch; name="vtpm-fakefile.patch"',
415 'matches Email::MIME output, "correct" or not');
418 Content-Type: multipart/foo; boundary=b
421 Content-Disposition: attachment; filename="=?utf-8?q?vtpm-makefile.patch?="
425 Content-Type: text/x-patch; name="=?utf-8?q?vtpm-fakefile.patch?="
431 skip 'newer Email::MIME is inconsistent here', 1
432 if $cls eq 'PublicInbox::MIME';
434 $cls->new($s)->each_part(sub { push @x, $_[0]->[0]->filename });
435 is_deeply(['vtpm-makefile.patch', 'vtpm-fakefile.patch'], \@x,
436 "got filename for both attachments ($cls)");