1 # Copyright (C) 2020 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # Lazy MIME parser, it still slurps the full message but keeps short
5 # lifetimes. Unlike Email::MIME, it doesn't pre-split multipart
6 # messages or do any up-front parsing of headers besides splitting
7 # the header string from the body.
9 # Contains ideas and code from Email::Simple and Email::MIME
10 # (Perl Artistic License, GPL-1+)
12 # This aims to replace Email::MIME for our purposes, similar API
13 # but internal field names are differ if they're not 100%-compatible.
15 # Includes some proposed fixes for Email::MIME:
16 # - header-less sub parts - https://github.com/rjbs/Email-MIME/issues/14
17 # - "0" as boundary - https://github.com/rjbs/Email-MIME/issues/63
20 # bdy => scalar ref for body (may be undef),
21 # hdr => scalar ref for header,
22 # crlf => "\n" or "\r\n" (scalar, not a ref),
24 # # filled in during ->each_part
25 # ct => hash ref returned by parse_content_type
27 package PublicInbox::Eml;
31 use Encode qw(find_encoding decode encode); # stdlib
32 use Text::Wrap qw(wrap); # stdlib, we need Perl 5.6+ for $huge
33 use MIME::Base64 3.05; # Perl 5.10.0 / 5.9.2
34 use MIME::QuotedPrint 3.05; # ditto
36 my $MIME_Header = find_encoding('MIME-Header');
38 use PublicInbox::EmlContentFoo qw(parse_content_type parse_content_disposition);
39 $PublicInbox::EmlContentFoo::STRICT_PARAMS = 0;
41 our $MAXPARTS = 1000; # same as SpamAssassin
42 our $MAXDEPTH = 20; # seems enough, Perl sucks, here
43 our $MAXBOUNDLEN = 2048; # same as postfix
45 my %MIME_ENC = (qp => \&enc_qp, base64 => \&encode_base64);
46 my %MIME_DEC = (qp => \&dec_qp, base64 => \&decode_base64);
47 $MIME_ENC{quotedprint} = $MIME_ENC{'quoted-printable'} = $MIME_ENC{qp};
48 $MIME_DEC{quotedprint} = $MIME_DEC{'quoted-printable'} = $MIME_DEC{qp};
49 $MIME_ENC{$_} = \&identity_codec for qw(7bit 8bit binary);
51 my %DECODE_ADDRESS = map { $_ => 1 } qw(From To Cc Sender Reply-To);
54 'Content-Description' => 1,
55 'Content-Type' => 1, # not correct, but needed, oh well
57 our %STR_TYPE = (text => 1);
58 our %STR_SUBTYPE = (plain => 1, html => 1);
63 # Do not normalize $k with lc/uc; instead strive to keep
64 # capitalization in our codebase consistent.
65 $re_memo{$k} ||= qr/^\Q$k\E:[ \t]*([^\n]*\r?\n # 1st line
67 (?:[^:\n]*?[ \t]+[^\n]*\r?\n)*)
71 # compatible with our uses of Email::MIME
73 my $ref = ref($_[1]) ? $_[1] : \(my $cpy = $_[1]);
74 if ($$ref =~ /\r?\n(\r?\n)/s) { # likely
75 # This can modify $$ref in-place and to avoid memcpy/memmove
76 # on a potentially large $$ref. It does need to make a
77 # copy for $hdr, though. Idea stolen from Email::Simple
78 my $hdr = substr($$ref, 0, $+[0], ''); # sv_chop on $$ref
79 substr($hdr, -(length($1))) = ''; # lower SvCUR
80 bless { hdr => \$hdr, crlf => $1, bdy => $ref }, __PACKAGE__;
81 } elsif ($$ref =~ /^[a-z0-9-]+[ \t]*:/ims && $$ref =~ /(\r?\n)\z/s) {
83 bless { hdr => \($$ref), crlf => $1 }, __PACKAGE__;
84 } else { # nothing useful
86 bless { hdr => \$hdr, crlf => "\n" }, __PACKAGE__;
91 my (undef, $ref) = @_;
92 # special case for messages like <85k5su9k59.fsf_-_@lola.goethe.zz>
93 $$ref =~ /\A(\r?\n)/s or goto &new;
94 my $hdr = substr($$ref, 0, $+[0], ''); # sv_chop on $$ref
95 bless { hdr => \$hdr, crlf => $1, bdy => $ref }, __PACKAGE__;
98 # same output as Email::Simple::Header::header_raw, but we extract
99 # headers on-demand instead of parsing them into a list which
100 # requires O(n) lookups anyways
102 my $re = re_memo($_[1]);
103 my @v = (${ $_[0]->{hdr} } =~ /$re/g);
105 # for compatibility w/ Email::Simple::Header,
110 wantarray ? @v : $v[0];
113 # pick the first Content-Type header to match Email::MIME behavior.
114 # It's usually the right one based on historical archives.
116 # PublicInbox::EmlContentFoo::content_type:
117 $_[0]->{ct} //= parse_content_type(header($_[0], 'Content-Type'));
120 # returns a queue of sub-parts iff it's worth descending into
121 # TODO: descend into message/rfc822 parts (Email::MIME didn't)
122 sub mp_descend ($$) {
123 my ($self, $nr) = @_; # or $once for top-level
124 my $bnd = ct($self)->{attributes}->{boundary} // return; # single-part
125 return if $bnd eq '' || length($bnd) >= $MAXBOUNDLEN;
126 $bnd = quotemeta($bnd);
128 # "multipart" messages can exist w/o a body
129 my $bdy = ($nr ? delete($self->{bdy}) : \(body_raw($self))) or return;
131 # Cut at the the first epilogue, not subsequent ones.
132 # *sigh* just the regexp match alone seems to bump RSS by
133 # length($$bdy) on a ~30M string:
134 my $epilogue_missing;
135 if ($$bdy =~ /(?:\r?\n)?^--$bnd--[ \t]*\r?$/sm) {
136 substr($$bdy, $-[0]) = '';
138 $epilogue_missing = 1;
141 # *Sigh* split() doesn't work in-place and return CoW strings
142 # because Perl wants to "\0"-terminate strings. So split()
143 # again bumps RSS by length($$bdy)
145 # Quiet warning for "Complex regular subexpression recursion limit"
146 # in case we get many empty parts, it's harmless in this case
147 no warnings 'regexp';
148 my ($pre, @parts) = split(/(?:\r?\n)?(?:^--$bnd[ \t]*\r?\n)+/ms,
150 # + 3 since we don't want the last part
151 # processed to include any other excluded
152 # parts ($nr starts at 1, and I suck at math)
153 $MAXPARTS + 3 - $nr);
155 if (@parts) { # the usual path if we got this far:
156 undef $bdy; # release memory ASAP if $nr > 0
158 # compatibility with Email::MIME
159 $parts[-1] =~ s/\n\r?\n\z/\n/s if $epilogue_missing;
161 @parts = grep /[^ \t\r\n]/s, @parts; # ignore empty parts
163 # Keep "From: someone..." from preamble in old,
164 # buggy versions of git-send-email, otherwise drop it
165 # There's also a case where quoted text showed up in the
167 # <20060515162817.65F0F1BBAE@citi.umich.edu>
168 unshift(@parts, $pre) if $pre =~ /:/s;
171 # "multipart", but no boundary found, treat as single part
172 $self->{bdy} //= $bdy;
176 # $p = [ \@parts, $depth, $idx ]
177 # $idx[0] grows as $depth grows, $idx[1] == $p->[-1] == current part
178 # (callers need to be updated)
179 # \@parts is a queue which empties when we're done with a parent part
181 # same usage as PublicInbox::MsgIter::msg_iter
182 # $cb - user-supplied callback sub
183 # $arg - user-supplied arg (think pthread_create)
184 # $once - unref body scalar during iteration
186 my ($self, $cb, $arg, $once) = @_;
187 my $p = mp_descend($self, $once // 0) or
188 return $cb->([$self, 0, 0], $arg);
190 my @s; # our virtual stack
192 while ((scalar(@{$p->[0]}) || ($p = pop @s)) && ++$nr <= $MAXPARTS) {
193 ++$p->[-1]; # bump index
194 my (undef, @idx) = @$p;
195 @idx = (join('.', @idx));
196 my $depth = ($idx[0] =~ tr/././) + 1;
197 my $sub = new_sub(undef, \(shift @{$p->[0]}));
198 if ($depth < $MAXDEPTH && (my $nxt = mp_descend($sub, $nr))) {
199 push(@s, $p) if scalar @{$p->[0]};
200 $p = [ $nxt, @idx, 0 ];
201 } else { # a leaf node
202 $cb->([$sub, $depth, @idx], $arg);
208 # prevent MIME::QuotedPrint from encoding CR as =0D since it's
209 # against RFCs and breaks MUAs
210 $_[0] =~ s/\r\n/\n/sg;
211 encode_qp($_[0], "\r\n");
215 # RFC 2822 requires all lines to end in CRLF, though... :<
216 $_[0] = decode_qp($_[0]);
217 $_[0] =~ s/\n/\r\n/sg;
221 sub identity_codec { $_[0] }
223 ########### compatibility section for existing Email::MIME uses #########
226 bless { hdr => $_[0]->{hdr}, crlf => $_[0]->{crlf} }, __PACKAGE__;
231 my $parts = mp_descend($self, 0) or return ();
232 my $bnd = ct($self)->{attributes}->{boundary} // die 'BUG: no boundary';
233 my $bdy = $self->{bdy};
234 if ($$bdy =~ /\A(.*?)(?:\r?\n)?^--\Q$bnd\E[ \t]*\r?$/sm) {
235 $self->{preamble} = $1;
237 if ($$bdy =~ /^--\Q$bnd\E--[ \t]*\r?\n(.+)\z/sm) {
238 $self->{epilogue} = $1;
240 map { new_sub(undef, \$_) } @$parts;
244 my ($self, $parts) = @_;
246 # we can't fully support what Email::MIME does,
247 # just what our filter code needs:
248 my $bnd = ct($self)->{attributes}->{boundary} // die <<EOF;
249 ->parts_set not supported for single-part messages
251 my $crlf = $self->{crlf};
252 my $fin_bnd = "$crlf--$bnd--$crlf";
253 $bnd = "$crlf--$bnd$crlf";
254 ${$self->{bdy}} = join($bnd,
255 delete($self->{preamble}) // '',
256 map { $_->as_string } @$parts
259 (delete($self->{epilogue}) // '');
264 my ($self, $body) = @_;
265 my $bdy = $self->{bdy} = ref($body) ? $body : \$body;
266 if (my $cte = header_raw($self, 'Content-Transfer-Encoding')) {
267 my $enc = $MIME_ENC{lc($cte)} or croak("can't encode `$cte'");
268 $$bdy = $enc->($$bdy); # in-place
274 my ($self, $body_str) = @_;
275 my $charset = ct($self)->{attributes}->{charset} or
276 Carp::confess('body_str was given, but no charset is defined');
277 body_set($self, \(encode($charset, $body_str, Encode::FB_CROAK)));
280 sub content_type { scalar header($_[0], 'Content-Type') }
282 # we only support raw header_set
284 my ($self, $pfx, @vals) = @_;
285 my $re = re_memo($pfx);
286 my $hdr = $self->{hdr};
287 return $$hdr =~ s!$re!!g if !@vals;
289 my $len = 78 - length($pfx);
291 # folding differs from Email::Simple::Header,
292 # we favor tabs for visibility (and space savings :P)
293 if (length($_) >= $len && (/\n[^ \t]/s || !/\n/s)) {
294 local $Text::Wrap::columns = $len;
295 local $Text::Wrap::huge = 'overflow';
296 $pfx . wrap('', "\t", $_) . $self->{crlf};
298 $pfx . $_ . $self->{crlf};
301 $$hdr =~ s!$re!shift(@vals) // ''!ge; # replace current headers, first
302 $$hdr .= join('', @vals); # append any leftovers not replaced
303 # wantarray ? @_[2..$#_] : $_[2]; # Email::Simple::Header compat
304 undef; # we don't care for the return value
307 # note: we only call this method on Subject
309 my ($self, $name, @vals) = @_;
311 next unless /[^\x20-\x7e]/;
312 utf8::encode($_); # to octets
313 # 39: int((75 - length("Subject: =?UTF-8?B?".'?=') ) / 4) * 3;
314 s/(.{1,39})/'=?UTF-8?B?'.encode_base64($1, '').'?='/ges;
316 header_set($self, $name, @vals);
319 sub mhdr_decode ($) { eval { $MIME_Header->decode($_[0]) } // $_[0] }
322 my $dis = header_raw($_[0], 'Content-Disposition');
323 my $attrs = parse_content_disposition($dis)->{attributes};
324 my $fn = $attrs->{filename};
325 $fn = ct($_[0])->{attributes}->{name} if !defined($fn) || $fn eq '';
326 (defined($fn) && $fn =~ /=\?/) ? mhdr_decode($fn) : $fn;
329 sub xs_addr_str { # helper for ->header / ->header_str
330 for (@_) { # array from header_raw()
332 my @g = parse_email_groups($_); # [ foo => [ E::A::X, ... ]
333 for (my $i = 0; $i < @g; $i += 2) {
334 if (defined($g[$i]) && $g[$i] =~ /=\?/) {
335 $g[$i] = mhdr_decode($g[$i]);
337 my $addrs = $g[$i + 1];
338 for my $eax (@$addrs) {
339 for my $m (qw(phrase comment)) {
341 $eax->$m(mhdr_decode($v)) if
346 $_ = format_email_groups(@g);
351 require Email::Address::XS;
352 Email::Address::XS->import(qw(parse_email_groups format_email_groups));
355 # fallback to just decoding everything, because parsing
356 # email addresses correctly w/o C/XS is slow
357 %DECODE_FULL = (%DECODE_FULL, %DECODE_ADDRESS);
358 %DECODE_ADDRESS = ();
361 *header = \&header_str;
363 my ($self, $name) = @_;
364 my @v = header_raw($self, $name);
365 if ($DECODE_ADDRESS{$name}) {
367 } elsif ($DECODE_FULL{$name}) {
369 $_ = mhdr_decode($_) if /=\?/;
372 wantarray ? @v : $v[0];
375 sub body_raw { ${$_[0]->{bdy} // \''}; }
378 my $raw = body_raw($_[0]);
379 my $cte = header_raw($_[0], 'Content-Transfer-Encoding') or return $raw;
380 ($cte) = ($cte =~ /([a-zA-Z0-9\-]+)/) or return $raw; # For S/MIME, etc
381 my $dec = $MIME_DEC{lc($cte)} or return $raw;
388 my $charset = $ct->{attributes}->{charset};
390 if ($STR_TYPE{$ct->{type}} && $STR_SUBTYPE{$ct->{subtype}}) {
393 Carp::confess("can't get body as a string for ",
394 join("\n\t", header_raw($self, 'Content-Type')));
396 decode($charset, body($self), Encode::FB_CROAK);
401 my $ret = ${ $self->{hdr} };
402 return $ret unless defined($self->{bdy});
403 $ret .= $self->{crlf};
404 $ret .= ${$self->{bdy}};
407 # Unlike Email::MIME::charset_set, this only changes the parsed
408 # representation of charset used for search indexing and HTML display.
409 # This does NOT affect what ->as_string returns.
411 ct($_[0])->{attributes}->{charset} = $_[1];
414 sub crlf { $_[0]->{crlf} // "\n" }
416 sub willneed { re_memo($_) for @_ }
418 willneed(qw(From To Cc Date Subject Content-Type In-Reply-To References
419 Message-ID X-Alt-Message-ID));